An Acoustic Simulator For Headphone Amplifiers.

by Chu Moy

This article describes modifications to an acoustic simulation circuit for headphones that appeared in a magazine article called “Improved Headphone Listening” by Siegfried Linkwitz (Audio, December 1971). It is a simple RC-type filter that creates a more realistic sound image in headphones by electronically mimicking the “shaded” interchannel crossfeed of normal hearing. The circuit was based on a design published by Benjamin Bauer ten years earlier, but the Linkwitz version does not use inductors and is less sensitive to load impedance. (Crossfeed filters should not be confused with virtualizers, which use digital signal processing to simulate binaural or externalized 3D sound.) These modifications improve the sound of Linkwitz filter and optimize the circuit to work with headphone amplifiers.

I tested the modifications with my DIY pocket headphone amplifier (shown above), the Musical Fidelity X-Cans (v.1) headphone amplifier, Sennheiser 465 and Sennheiser HD600 headphones and the Stax SRS-3030 electrostatic headphone system. The source was a Panasonic SL-SX500 CD player. The filter can drive either a headphone amplifier or headphones directly with the headphone amplifier as the source. I tested the original filter directly connected to the HD465 phones and fed by the X-Cans. The modified filter was tested both ways: as the source to headphone amplifiers and directly connected to both headphones. Putting the filter before the amplifier eliminates any impedance interaction between the filter and headphones.

For more information about the original Linkwitz circuit, please refer to the Audio magazine article cited above. For more information about the Bauer circuit and acoustic simulators generally, see Technologies For Presentation of Surround Sound in Headphones and The Psychoacoustics of Headphone Listening. For information on commercial crossfeed filters, see A Quick Guide to Headphone Accessories.

Background

At work, I listen to music with headphones several hours each day. My system is a Musical Fidelity X-Cans amplifier and a pair of Sennheiser HD465 headphones (discounted 50% from a liquidation house!). The source material is either CD or FM radio. Although the audio reproduction is excellent, I cannot listen to this system for more than four hours without suffering headaches, caused by the super-stereo effect that is characteristic of headphone listening.

Stereo recordings are meant to be heard through loudspeakers. Headphones create a soundfield that is unnaturally spacious, in which some sounds seem to be crowded around each ear. A few months ago, I did some research into the problem. Headphone amplifiers from HeadRoom Corporation had a built-in crossfeed circuit to mitigate this unpleasant effect. However, I did not want to replace the X-Cans amplifier. Then I ran across a stereo crossfeed circuit in an old issue of Audio Magazine, which seemed to do something very similar to HeadRoom’s audio image processor.

I upgraded the design with higher quality parts, which I ordered from Digi-key Electronics and Mouser Electronics. The circuit was assembled on a 2″ x 1.75″ piece of printed circuit protoboard and put in a PacTec case from Radio Shack (RS 270-211). The front panel had two mini-stereo jacks (for the input and output), and a bypass toggle switch. (See A Pocket Headphone Amplifier for more information about purchasing parts.) The output of the X-Cans went to the input of the filter, and the headphones plugged into the output. After a month of listening with the Linkwitz circuit, I became dissatisfied with the sound quality. It did eliminate the super-stereo effect, but introduced other sonic problems.

First, the high frequencies were severely attenuated compared to the original signal, despite the built-in treble boost in each channel. The imbalance imparted a muffled effect to all kinds of music. With some vocalists (such as Barbra Streisand – yes, I am a fan), voices took on a “thick” quality. I suspected that the effect was caused by phase anomalies at the crossover point interacting with phasey artifacts in recordings.

Second, the level of crossfeed was a bit excessive for my tastes. The soundstage was pulled inward, away from the ears, but I felt as though I was sitting in the back of an auditorium with heavily padded walls. Flipping the bypass switch restored the spaciousness and reverb at the expense of the benefits of the crossfeed.

Third, the filter did not drive headphones well with a headphone amp input. The original circuit was designed to be connected to the output of a power amplifier. Running it from the speaker outputs of a 15W receiver produced ample volume. Headphone amps, on the other hand, huff and puff along at a relatively measly 100mW maximum output. A few Thevenin computations also revealed that the output impedance of the filter was slightly frequency dependent and varied from 61 to 73 ohms. My Sennheiser HD465s were rated at 60 ohms. These factors explained why I had to keep the X-Cans volume control at 75% rotation to achieve acceptable listening levels.

The Modifications

Figure 1

The original circuit (figure 1) crossfeed frequencies below 700 Hz (figure 2). Linkwitz noted that the low frequency blending raised the bass response in each channel by about 3dB. He designed a 2dB treble boost to compensate partially for the increase, reasoning that full compensation was not needed because the low frequencies in each channel weren’t always in phase. On some recordings, the filter appeared to reduce the amount of deep bass due to phase cancellation. However, it also overemphasized the lower midrange and imparted a heaviness to the sound. Even though the separation between channels increased above 700 Hz, there was some high frequency attenuation – probably due both to phase effects at the crossover point and the fluctuating output impedance.

Figure 2

The most obvious solution was to increase the treble boost, but that would not have affected the circuit’s high drive requirements or the apparent width of the soundfield. A voltage divider at the input (R1/R2) set the original attenuation factor at about 1:6. The crossfeed signal was summed into the output resistor (R2) of the divider. Increasing the value of the output resistor would increase the voltage output, but would also increase the output impedance and vary the level of crossfeed. Not a good idea.

Figure 3

Lowering the input resistor (R1) would increase the output voltage, while keeping the crossfeed level constant (but having the effect of reducing percent of crossfeed in each channel, because a smaller R1 increases the level of the main signal). If crossfeed level remained constant, decreasing the value of R1 would “widen” the soundstage and create a smoother response with the existing treble boost. Also, a smaller R1 value would minimize output impedance fluctutation. After experimenting with several R1 values, I set R1a to 200 ohms (40% of the original value).

Figure 4

Since all recordings are not the same, I added a “PERSPECTIVE” switch (S1) to customize the processing with an alternate R1 value. At R1b = 150 ohms, the low frequency separation between channels goes up to about 10dB and the overall output increases by about 2dB. Toggling from R1b to R1a, the soundstage appears to move further away (lower output, more narrow soundstage, slightly softened highs). After a period of listening, I still heard a slight thickness in the lower midrange, which I suspected was due to the threshold frequency of the treble boost being a bit low. Decreasing R3 to 910 ohms moved the threshold to about 800Hz and cleaned up the midrange emphasis. It also gave a touch more treble boost for a more balanced, clearer sound overall.

I settled on the R1b setting (low crossfeed) as the default, and use R1a when the sonic presentation would otherwise be too wide or if the recording is too bright. The PERSPECTIVE switch has a distinct effect with these R1 values, with the high crossfeed setting effectively simulating greater distance from the soundstage. On good acoustic recordings, the high crossfeed produces a palpable sense of depth in the headphone image.

Figure 3 shows the final schematic. The “Low-Z” version (where “Z” is short for impedance) is the most versatile. It is the version that I built and can drive headphones directly, because it has a low output impedance. The rest of this article mostly refers to the Low-Z version. If the simulator will be used ONLY as an input stage to a headphone amplifier, consider building the “High Z” version, which scales the resistor and capacitor values by a factor of 10 (x10 resistors and ÷10 capacitors) to get a higher input impedance (about 2000 ohms), which is a better match for the line outputs of preamps and other audio sources. The headphone amp itself should have an input impedance of 5K ohms or higher. Several DIYers have built the high-Z filter. See the addendum for details. Do NOT scale the parts if the simulator will ever drive headphones directly.

Figure 5

A comparison of the original to modified signal levels (figure 4) shows that the modified output is about 3dB louder in each channel because of the new attenuation factor of 1:3. The low frequency separation is 6 dB, 3 dB wider than before. The overall output impedance is lower and is flatter over the audio range: 51 to 60 ohms. My calculations indicate that the high and low frequencies are 25% closer in level than the original. The improvement in sound is so great that I suspect the reduction of high frequency phase effects due to the wider separation also contributes significantly. The modification increases the crossfeed threshold frequency by about 8% and treble boost threshold frequency by 10%.

Construction

Assembling this acoustic simulator is fairly straightforward, and would be a good project for the intermediate DIY beginner. I used a printed-circuit protoboard from Radio Shack, which since has been discontinued. Instead, I recommend the Vector Circbord from Mouser Electronics (Stock No. 574-3677-6), which is tin-plated. If possible, layout the circuit as compactly as possible on a breadboard first to get an idea of where the components will go. The circuit is fairly simple, but the 16 capacitors and resistor can be a challenge to place.

Then cut a small square (about 2″ x 1.75″) of the protoboard with a section of the foil pattern that best suits the breadboard layout. To cut out the board, a coping saw will work fine. If the DIYer is building my pocket amp project (which uses a similar-sized board), another good method is to score the board with a utility knife and break off the section needed. This method can result in some waste, unless the DIYer is building more than one project that will be installed in the same type of enclosure.

1. With a ruler, draw, on the non-foil side, a line parallel to the long side of the Circbord, about 1.75 to 2 inches from the edge. Be sure that the selected section contains a useful foil pattern.
2. Score the Circbord several times with a utility knife and ruler along the drawn line.
3. Position the scored line of the Circbord (foil-side down) over a table with a sharp edge – the marked section should hang over the edge of the table.
4. With one hand, press down on the Circbord against the table to anchor it.
5. With the other hand, apply several sharp blows to the area of the Circbord overhanging the table. The section should snap, but still remain hinged because of the copper foil. Be careful not to pull off the foil.
6. To separate the section, cut through the foil by scoring with the utility knife.
7. Repeat this procedure on the Circbord section itself to get a board about 2 inches long.

The 2.75″ x 4.6″ x 1″ enclosure is Pac Tec model K-RC24-9VB and comes in the colors Bone and Black. I purchased it at Radio Shack, and it is the same case I used for the pocket headphone amp project. It has a 9V battery compartment and both opaque and red plastic front panels. Radio Shack stores no longer stock the case, but it can be ordered from RadioShack.com (RS 910-1096, black only). Digikey and Mouser also sell these cases.

The headphone jacks are enclosed units for 1/8″ stereo plugs. Radio Shack sells a version of these jacks (RS 274-249). I ordered higher quality units that have spring-loaded contacts from Mouser Electronics (Stock No. 161-3502). The DPDT switches are micro-mini toggles from Radio Shack (RS 275-626). It may be more convenient to mount the switches and jacks before wiring them, because some of the resistors may have to be wired behind the front panel and not on the protoboard, as space allows. I chose the red plastic panel because the opaque panel was too thick to mount the headphone jacks. Resistors that are mounted behind the front panel should have their leads insulated to prevent short circuits. I cut plastic tubes to length from the insulation on my hookup wire (24 ga.). Keep the wiring as short and neat as possible (which is not easy to do) to minimize RF pickup.

In Linkwitz’s schematic, the original C2 is a 1.3uF non-polar electrolytic, but this value is difficult to find in a film capacitor. I used a 1.2uF film capacitor. Many DIYers like the sound of the simulator with the parts values shown, but others have written to tell me about their modifications to the circuit. I encourage anyone interested in building this simulator to experiment with parts values to get the best sound for that person’s hearing preferences and characteristics. The easiest way to customize the simulator is to build it first on an experimenter’s breadboard and then permanently solder the connections after the best sound is achieved.

I strongly recommend auditioning the filter on a breadboard first, before finalizing the component values. The most popular modification is to decrease or increase the values of R1a/R1b in the range from 50 ohms to 330 ohms. Again, the lower the value of R1, the wider the soundstage will be. Another popular modification is to decrease C2 to 1uF, which can give a bit more depth to the soundstage due to the higher threshold frequency. The crossfeed threshold frequency is given by: F_crossfeed = 1/[2*pi*C2*[R4 || (R5 + {R2 || R1})]]. The treble boost threshold frequency is approximately 1/[2*pi*C1*R3]. The amount of treble boost can be increased/decreased by decreasing/increasing R3 (to keep the threshold frequency constant, choose C1 so that R3*C1 = 0.00018).

Download Gus Wanner’s Excel Application

For DIYers who want to customize the simulator but don’t have the time or patience to build many versions, Gus Wanner has prepared the above Microsoft Excel 2000 application, which can instantly plot changes in frequency response, time delays and other circuit characteristics as component values change. To use this spreadsheet, Excel must have installed the engineering functions in the analysis toolpak (which comes with the Excel package but is NOT automatically installed by the standard MS Excel install program). To install this toolpak, use the add-ins submenu of Excel. Verify that the correct toolpak is installed by clicking on Insert|Functions options and look for the engineering functions menu. Excel 97 should work the same way.

Mr. Wanner describes the application as follows:

The analysis is straight-forward, using Thevenin equivalent impedances and voltage sources as explained on the “analysis” tab. Note that Excel does NOT have any formatting capability for complex numbers (they display with maximum precision all the time, taking a huge amount of space). I have “cut them off” (for display only) by putting blanks or other values into adjacent columns.

The spreadsheet is parameterized for a medium and high crossfeed case designed for use with a low power amplifier (20 watts/ch or so) and a load of 70 ohms (the Sennheiser HD-25 impedance). The network sheets are protected (to prevent my accidentally wiping out formulas and values); there is no password, so simply unprotect the sheets (Tools|Unprotect) to change the values.

See the addendum for more details about this application.

The Results

The best sound quality was obtained with the Linkwitz filter driving the input of my pocket headphone amplifier or the Stax electrostatic amplifier. To optimize the 3D effect with supra-aural headphones, wear the earpieces as forward as possible, to enhance in-front imaging (slightly down and forward on the ears). This positioning helps direct the sound waves to enter the ears at an angle as happens with normal hearing, instead of going straight to the eardrums. The technique does work with circumaural headphones, but the effect is not as pronounced.

The modified Linkwitz filter sounds much more open and clearer than the original, especially when the crossfeed level was set to low (R1a). The highs were back! The bass was stronger and better defined. Instruments and vocals were focused and had “air” around them again (Barbra’s voice sparkled). Although the soundstage was wider than with the original filter on both low and high crossfeed settings, the crossfeed was still effective and subtlely pulled the image forward. Reverb no longer “bounced” off my ears the way that unprocessed headphone reverb does, and the bass was more centered like loudspeaker bass.

With the PERSPECTIVE switch set to high crossfeed, the soundstage narrowed and the top treble softened, yet there was more depth, more dimensionality – as though it had been moved further back. Recordings that polarized the stereo presentation with instruments or vocals to the extreme left and right had an substantially improved sense of aural continuity.

Sound Quality with Simulator Before Headphone Amp

The first set of extended listening tests were with the simulator as the source to the headphone amplifier and set to low crossfeed. In the HD465, there was just a trace of the emphasis in the lower midrange and a softening of the high end, which still made for a natural presentation. The HD465 is a very fine supra-aural headphone, but the looser ear coupling thins out the sound somewhat and highlights treble anomalies from the crossfeed. Regardless of these slight tonal shifts, the simulator’s effect still generated a very pleasant, forward soundstage.

The HD600 has a more even frequency response than the HD465. The extended low end lends the sound a lushness. The effect of the simulator in the HD600 was so smooth that I could call it “ethereal.” Switching the filter on and off, there was only a hint of tonal impact in treble. It was almost seamless, like a video “morphing” transition. The HD600’s larger, more enveloping soundfield (due to the different ear coupling) created a vibrant, dimensional acoustic space that portrayed instruments and vocals with greater accuracy and tonal fidelity than the HD465.

My new (actually used) Stax SRS-3030 headphone system arrived just in time to be included in the latest revision of this article. The system consists of the Lambda SR-303 electrostatic headphones and the SRM-313 amplifier. I quickly auditioned the Stax with the modified Linkwitz filter installed between the Panasonic CD player and the SRM-313 amplifier. With the simulator switched off, the SR-303 headphones presented a spatial, open soundfield without the “in-each-ear” sensation of other headphones, which I attribute to the oversized Stax transducers that sit at a slight angle on the ears. The sound was brighter than with the HD600s, not as lush. The Stax imaging seemed larger, but still two dimensional – essentially forming in a straight line between my ears.

With the modified Linkwitz filter switched in (low crossfeed setting), the image instantly took on depth and dimensionality. On recordings with a front vocal, instruments seemed to localize around and behind the voice. The treble softening was more noticeable than in the HD600 (though less noticeable than with the HD465) but did not appreciably alter the frequency balance of the SR-303s. The Staxes had tighter bass than the HD600. There was a touch fullness in the bass frequencies that conveyed the energy of the performances well. After a sustained period of auditioning with a variety of music from classical to pop, in every case I preferred leaving the filter on. The filter greatly enhanced the realism of the Stax presentation. Overall, the Stax/modified Linkwitz was a very successful pairing.

Sound Quality with Simulator After Headphone Amp

The sound quality from headphones directly connected to the Linkwitz filter varies, depending on the impedance response of the headphones. In theory, the filter should drive high impedance headphones more consistently than low impedance headphones, but that was not the result of my listening tests. Headphones are not the same as loudspeakers, and tend to have flatter impedance curves. So long as headphone impedance remains flat over the audible range or never falls below a threshold (say 10 times the filter’s output impedance), the headphone amplifier’s current output is the dominant factor in sound quality. Where headphone impedance falls below that threshold in the audible range, the Linkwitz circuit may introduce some coloration due to impedance fluctuation.

The modified Linkwitz is much more efficient for directly driving headphones than the original filter. My X-Cans headphone amp could now comfortably drive the filter and the HD465 with the volume control at around 30% rotation. Substituting my pocket amp for the X-Cans was similarly gratifying. Since the HD600 has a higher impedance which should have made it better load for the Linkwitz filter, I expected it to sound more tonally consistent with the filter than the HD465. In fact, the HD465 had the better interfacing experience. Directly connected the Linkwitz, the HD600 had a distinctly dark sound. By comparison, the sound of the HD465 was similar to that when the filter was put before headphone amp. The lesson here is that there is no easy way to predict the success of the filter directly driving headphones. The best way to determine the sound quality in this configuration is to try the headphones with the filter.

Conclusion

The modified Linkwitz circuit sounds natural and spacious with a forward-projecting image. The filter’s effect is more subtle compared with the original design, and is, therefore, less intrusive into the musical experience. On many recordings, it adds an almost three-dimensional quality to the presentation. Headphone listening is now definitely less fatiguing. I can listen to my headphone systems for hours every day without headaches. If I do feel the need to take a headphone off, it is usually because my ears hurt from having the earpieces physically pressed against them for a long time.

The filter works best as the input to a headphone amplifier and pairs well with my pocket headphone amplifier in the matching enclosure (shown above) for a very nice portable listening system. Both units are lightweight and will fit into many portable stereo carrying cases. The unused battery compartment in the acoustic simulator is good for storage of small accessories such as a headphone plug adapter. Finally, the price for the modified Linkwitz circuit is hard to beat, especially if one already has a headphone amplifier. The upgraded parts, case, switches and jacks came to about $20. A headphone amplifier with the modified Linkwitz circuit could be built for less than $40. If you build the Linkwitz filter with or without my modifications or have other mods you’d like to share, please don’t hesitate to e-mail me.

Thanks to Tyll Hertsens for his helpful comments.

c. 1998, 1999, 2000, 2001 Chu Moy.

Addendum

8/15/98: For better in-front localization with the Linkwitz filter, try wearing the headphones slightly forward and lower on the ears (supra-aural phones are the easiest to position this way). Experiment with the positioning to obtain the best localization. The goal is to get the sound to enter the ears at an angle, which is closer to the way normal hearing works. With the right recordings, this technique can produce a stunning sense of depth. It also works without crossfeed, but does not sound as natural.

10/6/98: Updated comparison of HeadRoom circuit to Linkwitz circuit. I also want to report that depending on the recording, the R1a setting of the perspective switch (“further from the soundstage”) can render a more 3D sound image, although the apparent width of the presentation would be fine without the additional crossfeed.

10/25/98: Added discussion about placing filter in front of headphone amp to eliminate any impedance interaction between the filter and headphones. Also clarified a few points throughout article.

11/16/98: Added image and description of portable headphone system. Also received report from user that the filter can drive Grado headphones directly with good results.

6/22/99: Added graph of time delays for the modified Linkwitz filter.

8/20/99: Mika V鳵r鄚iemi built the acoustic simulator and pocket amp in a single aluminum enclosure. He experimented with various values of C1 and R1 and found that C1 = 1uF and R1a = 50 ohms, R1b = 100 ohms had the widest soundstage and least effect on the high frequencies (Mika used the original R3 = 1000 ohms). “[B]efore I was positioned in the middle of band playing music. Now I’m in the front row as close as you can be…. Music just sounds realistic and that’s what I was looking for.” A more complete description of his work can be found in the DIY Workshop Forum.

8/26/99: Here is the parts layout and wiring diagram for Mika V鳵r鄚iemi’s simulator/amplifier project. Pictures of the finished amplifier can be see in A Pocket Headphone Amplifier.

11/23/99: Added more guidelines for customizing the simulator. Also, Chester Simpson has created a version of the modified Linkwitz with scaled parts values for headphone amps with high input impedances (greater than 250K ohms). See his article A Soundfield Simulator for Stereo Headphones.

12/9/99: Siegfried Linkwitz (himself!) e-mailed me the equation for calculating the crossfeed threshold frequency, which I have added to the article. Check out his new website: Linkwitz Lab.

1/12/00: scrazy@gcn.net.tw built this version of the pocket amp, which has a 10K ohm volume control and an acoustic simulator front-end that is based on the circuit by Chester Simpson (see design by Fred Peng below). Full details can be found at DIY Zone (in chinese only). His system consists of a Rega Planet CD Player and Audio Technica ATH-f15 headphones.

1/13/00: Fred Peng based his headphone amplifier on the acoustic simulator by Chester Simpson (which is based on the Linkwitz design), except that he replaced the R4,R6 combination in Simpson’s circuit with a 100K ohm resistor and added a unity gain input buffer stage made from an OPA134 and a high current output stage made from a PMI BUF-03 buffer. When compared with a McCormack Micro Headphone Drive, the BUF-03 driving his Grado HP-1 headphones with the simulator bypassed sounded better in the high and low frequencies than the McCormack, but the McCormack was better in the mid frequencies. With the simulator switched in, the sound was more relaxed, the low frequencies more centered, and the soundfield moved from inside his head to outside. He is very satisfied with the result and is planning to make another simulator for his Stax Lambda headphones. Full details and schematics (in chinese only) can be found at DIY Zone.

1/28/00: Added figure 1a. Thanks to Siegfried Linkwitz for sending me the graph!

5/1/00: Gus Wanner has sent in a Microsoft Excel Spreadsheet application that analyzes changes to component values of the modified Linkwitz circuit (see the text above for instructions to download). He writes:

I enjoyed your article on your modification to the Linkwitz crossfeed network. Since I have a low power (20 watt/channel) high quality amplifier integral to my McIntosh C-40 audio control center, I wanted to develop a version of this network to work with the C-40 and my Sennheiser HD-25 phones (and other headphones with greater than 60 ohm impedance). The HD25s have a maximum input level of 200mW. With the HD25s at 70 ohms, this will require a voltage of approximately 3.7 volts across each channel or approximately 11.8 volts into the crossfeed network. This voltage level corresponds to an amplifier output of 17.5 watts into 8 ohms.

To aid in doing the design, I developed a complete analysis of the Linkwitz network using the complex number analysis capability built into the MS Excel spreadsheet (and I think also available in newer versions of Quattro Pro). The spreadsheet allows you to enter values for the various components, and immediately computes the resulting levels, channel separation, and delay times for frequencies from 20 – 20,000 Hertz. Graphical plots for these parameters versus frequency are included as well.

The component values on the spreadsheet are the final values for a crossfeed circuit I designed for use with my McIntosh C-40 and my Sennheiser HD-25 headphones. The circuit is relatively insensitive to load impedances 70 ohms or greater, so it would work with other headphones as well. The modified Linkwitz crossfeed filter works great with my Sennheiser HD25s.

3/14/01: Major rewrite of article. Added detailed comparison of sound quality of filter placed before and after the headphone amplifier and review of Stax SRS-3030 headphones with filter. Added new high resolution pictures.

3/14/01: Coffin Lin put his version of the pocket headphone amplifier (with a Linkwitz crossfeed front-end) in an old TV remote control case. He modified the filter by making R2 and R3 adjustable, instead of R1. The component values in his version of the filter are:

R1: 30K ohms
R2a, R2b: 15K, 10K
R3a, R3b: 50K, 100K
R5: 33K
C1: 3,300pF
R4: 33K
C2: 10,000pF

The resistors are Dale RN55D. About making R2 and R3 adjustable, he says:

I mistook R2 for R1, but on the Excel worksheet simulator, R2 can still alter some balance. I think that adjusting R3 is more effective than adjusting R2 (I forget which switch is for what resistor.) One has more stereo (good for dance and rock) and the other is more natural (good for jazz).

3/25/01: Changed the value of R3 in figure 3 to 910 ohms (originally 1K-ohms) to remove emphasis in lower midrange and to increase the treble boost. This update results in a more balanced, clearer sound. I STRONGLY recommend it.

3/27/01: R2 and R3 incorrectly drawn in figure 3 from 3/25 update. Fixed.

11/24/2001: Mark D. Johnson writes:

I just finished building your Acoustic Simulator and have been auditioning it over the last several days on my Sennheiser HD600s (even now I’m listening to Miles Davis as I type this). I just want to tell you how much more I’m enjoying my music and how much less fatiguing it is to listen for long periods of time.

The thing I like best of all is the three dimensionality I hear in recordings that was never present before. As a drummer, the most amazing thing to me is that I can actually hear (whether true or not) the location of individual drums/cymbals being played – and not just sound coming from a point source called “drums.” On the latest Dianna Krall CD I could hear even elevation changes taking place as different cymbals were struck. Again, whether or not this is just a byproduct of the design I cannot be sure, but it sure makes listening more enjoyable.

5/21/2002: J. Ian Ramsey (from the forums) built a pocket amp and the high impedance version of the Linkwitz acoustic simulator in separate enclosures. He obtained most of the parts from RS Components Ltd. For the 120nF cap (C2), he paralleled two caps: 100nF and 22nF. He writes:

I made the Simulator in the high impedance version as recommended by Chu in the project notes. I have increased the gain of the amp from x11 to x17 by putting a 1.5K resistor in parallel with R3 (1K) to compensate for the insertion loss of the simulator which is only used between the source and the amp.

I laid out the simple circuit for the acoustic simulator in stripboard. The simulator effect is very subtle and at first I was unsure if I had made mistakes in the layout, which were preventing the correct circuit action. Listening to this year’s Grammophone magazine Award winner – Vaughan Williams ‘A London Symphony’ confirms the following:

• With the simulator switched to B the whole soundfield is very gently more centre focused and there is a slight loss of ambience.
• Switched to A, the field narrows and the volume level drops slightly thus making the sound image appear to be heard from a greater distance – just as Chu remarked in the article.

Overall I could easily live without the simulator as my HD600 headphones with the cmoy amp are very, very good. As an intellectual exercise, the concept of the simulator is satisfying in the way it addresses some of the headphone effects against speakers and this design does not degrade the sound – so I will most likely continue to use it.

Chu was spot on when he said to spend time with it in the design values as one would soon tire if the values were changed to produce a more dramatic effect. The subtlty soon gives way to a distinct change with it in and out and between mode A and B. Congratulations and thanks to Chu Moy for these brilliant designs.

5/21/2002: Phidaeaux (from the forums) built a pocket amp and the high impedance version of the Linkwitz acoustic simulator in a single enclosure. The white LED power indicator is mounted on the circuit board INSIDE the transparent Serpac enclosure. He writes:

Ok, I just finished this bad boy, and man oh man, am I pleased! This thing sounds so good. I’m only driving 32-ohm phones (Sennheiser HD497s) but the difference in sound quality is VERY obvious over my SlimX MP3/CD player’s built-in amp. It feels good to have power to spare, instead of driving the stock amp to its limits. Also, the crossfeed is very nice. I didn’t really notice at first, but now that I can sit down and play with it, I really like how it sounds. 

First off, thanks much to PRR, cmoy, tekir, tangent, and everyone who had either helpful suggestions, or had problems in the past, that I could read about and avoid doing myself. Cmoy and tangent’s info were both very useful too, for both the obvious reason (the schematics) but also the piles of helpful tweaks and pictures they had.

Technical notes! I’m using all the ‘usual suspect’ parts: Panasonic pot, Digikey jacks, metal film resistors, polypro caps (except for the power supply caps, which are electrolytic). I’m using a Burr-Brown OPA2134 dual op-amp (in a nice machined socket).

The crossfeed circuit is a pre-amplifier modified Linkwitz (the same circuit that cmoy uses as an independent item) and has had its resistor values all multiplied by 10, and its capacitor values divided by 10. This was done to raise input impedence a bunch, because its not actually driving headphones, it just connects directly to the amplification stage. The two ‘perspective’ settings are 2000 ohms, and 1500 ohms.

I thought about replacing one of those with a pair of mini pcb-mount multiturn pots, so I would have a ‘default’ setting and then a ‘custom’ setting I could adjust by opening the case. Or maybe put a ‘stereo’ pot controling those values, and cram it somewhere else on this case. Who knows. Anyway, I’m satisfied with those two values for now, I’ll wait before I do any ‘tinkering’. 

Anyway, to those who said it could not be done, I managed to fit the following panel items in the case: input and output jacks, one on each side of the case. I’ll get some right angle adapters I think to clean it up a bit, so it doesn’t have the plugs sticking off to the side like cowlicks. Power switch, crossfeed bypass switch, crossfeed perspective switch and Panasonic EVJ volume control (with pretty aluminum knob) are all on the front panel. The volume control, and the crossfeed switches are touching each other. They are quite literally pressed up against each other inside that panel!

White LED. Looks good shining out from inside the case. Bright too. I was going to use a blue, but the blues were less efficient than this white (oddly enough). I’m running it with a 2.2k resistor. Current through it seems to be about 3mA. That could stand to be lower, and the LED is very bright right now. If I make another modification, it’ll be to dim that LED a bit, and save some power. You can see neat little shadows inside the case from the components, light glinting off of things inside there.  Very cool.

I also really like this Serpac H-65 transparent blue case. I’m all about form AND function. I know that people usually get on one side of the fence or the other. They either LOVE pretty little gadgets, regardless of how they actually work. And then there are people who say “Screw it, does it work right? Then it can be ugly, I don’t care.”

For me, I want it to look good, work well, and feel ‘right’. If I’m going to be using something every day, it needs to be ergonomically designed. Engineers are notorious for totally forgetting the fact that real people have to USE the things they build. I work with equipment that was NOT inuitively designed so often; its like a breath of fresh air to find equipment where someone actually sat down to use it for a while, and thought ‘huh, this knob should really go down here on this side, instead of up here… that would make a lot more sense.’

It’s so easy to figure out where controls should go. Everything flows left to right on mine. Left is the input, right is the output. Starting from the left on the panel, you can turn it on or off, then you can choose crossfeed on or off, then you can raise or lower the crossfeed, then you can change the final volume. I don’t need labels, cause it makes sense. And then, I love this case. Its curvy and sexy, but still has quite a bit of internal volume. The transparent blue doesn’t scream “I bought this at Radio Shack like a nerd!” but rather, “This is a modern piece of technological wizardry.” But at the same time, you can peer inside and see the parts, making a muted statement that it was a DIY project. The LED inside is nice too. Don’t have to use up a panel spot, and the inside of the device lights up all pretty. Form AND function. Together at last!

Anyway, I don’t mean to lecture you guys; you all build very good things.  I got the ideas for this from seeing various other projects people have built. I just urge you all to take note of the stuff you deal with every day, and while you are thinking about all the technological aspects of it, to give a thought to intuitive design and control placement, the delicate art of making something ‘easy to use’ regardless of what its actually doing, etc.

Also, read the book “The Design of Everyday Things” by Don Norman. It talks about doors, VCRs, ovens, and all sorts of things you use every day, but don’t think much about. Ever push on a door when you are supposed to pull? Everyone has! But why? It’s a simple matter to make the operation of a door obvious; you don’t even need ‘push’ and ‘pull’ signs that people have to look at and read. A well designed door gets used right each time, without anyone even noticing. It’s all about intuitive engineering and human-centered design.

Oh, more technical jibba-jabba. Gain on the amp is set to 11. Current measured while in use, at a moderate volume, is 12mA. Not too shabby. Voltage between ground and the rails is 4.40V and -4.41V while in operation. Not bad, if I do say so myself! This single 9V is plenty to drive the 32-ohm HD497 to utterly insane levels. No need to give it more voltage.

Advertisement

A Pocket Headphone Amplifier.

by Chu Moy

“Thank you for your amplifier design. I built it and can’t believe how wonderful it makes my AKG K340 headphones sound as well as my Sennheiser 600.”
– A DIYer.

While doing research for the article Designing an Opamp Headphone Amplifier, I built a portable headphone amplifier for testing purposes. Each channel uses a single Burr-Brown OPA134 opamp in a non-inverting configuration. It has adequate current capability to drive most headphones without an output stage. I have used it with Sennheiser 465s (94dB SPL) and achieved ear-splitting volume. The amplifier is ideal as a booster for power-conserving stereo sources such as portable CD players and for interfacing with passive EQ networks such as tone controls or a headphone acoustic simulator.

The Amplifier Design

Figure 1

The schematic for one channel of the amplifier is shown in figure 1. All of the parts, except for the opamps, are available from Radio Shack. In several instances though, higher quality parts are available from other sources for about the same price that Radio Shack charges. The parts are commonly available, so look around for good buys. I do recommend Radio Shack’s 1/4W Metal Film Resistor Assortment (RS 271-309). It contains 50 resistors in popular values and nearly all of the values needed for this project. The total cost for this project should be no more than $20 – $25 US, assuming you already have general purpose items such as wire (I used solid 22 ga.).

The original opamp for this design, the OPA132, has been discontinued. The OPA134 is the audio-specific version of the OPA132 and will work identically in this circuit. It was selected for its excellent specs: FET inputs for high input impedance and low offset current, 8 MHz bandwidth, 20V/uS slew rate, ultra low noise, ultra low distortion, etc. It has fine PSRR (power supply rejection) numbers, can run on as little as ±2.5V (very important in a portable design) and includes built-in current limiting. The OPA134 costs less than $3.00 per unit from Digi-Key Electronics. It comes in a popular dual version: the OPA2134, which contains two opamps in a single package. Be sure to get the “DIP” package opamps; SOIC opamps are miniatures that are very difficult to handle.

Other opamps can be substituted, but make sure they will work with battery voltages (as little as ±3V) and are stable without external compensation. Also check the opamp’s current capability and current draw. The OPA134 has a quiescent current of about 4mA and will not drain the battery excessively. It can output almost 40mA into a short circuit at room temperature. Modern dynamic headphones need about 10mW to reach full volume. For more information, see Understanding Headphone Power Requirements.

Amplifier Frequency Response

The OPA134 is wired as a non-inverting amp with a gain of 11. At this gain, the output impedance of the amplifier is less than 0.2 ohms throughout the audio range. The high-pass filter C1-R2 at the input blocks DC current and has a corner frequency of about 15Hz. Substituting a 1uF capacitor will lower the corner frequency to 1.5Hz. However, 1uF capacitors tend to be too large for the recommended enclosure. Instead, if a lower corner frequency is mandatory, try increasing R2 to 1M (and scale R1 accordingly). You could omit C1 entirely, if DC input protection is not important. I recommend leaving C1 in the circuit.

If the amp will be driving low impedance headphones (32 ohms or less) such as the Grados, see appendix 1 for ways to optimize the amp for low impedance loads. R5 is an optional load resistor, which is explained in appendix 1. It can help reduce residual hiss and keep the power supply balanced.

The original pocket amp did not have a volume control, due to insufficient space in the enclosure (but see the next section for information on adding mini-pot volume control). Nor was a volume control necessary since the intended audio sources such as portable CD players and FM stereos already had volume controls. I did want the ability to reduce the input level as required to avoid overloading the amplifier (for example, some portable stereos have very high output voltage levels even when the volume control is set near 0). With R1 = 100K ohms, the LEVEL switch (SW1) drops the input voltage by 50% (6dB). At R1 = 470K ohms (the value I used), the switch attenuates the input by 15dB.

Figure 2

Several DIYers have written me to ask about adding a true volume control to the amplifier. In figure 2, R1 and SW1 are replaced with a dual, audio-taper mini potentiometer. The suggested pot values are 10K to 50K ohms. The enclosure in the prototype is barely 1″ tall, and the front panel is already crowded with and LED, switch and jacks. Mini dual pots are hard to find. Currently, Tangent’s Parts Shop is selling the ALPS RK097, a dual 10K audio mini pot, for a reasonable $3.25. Digikey sells the Panasonic EVJY10 series pots in 10K and 50K versions (part nos. P2G1103-ND for 10K, P2G1503-ND for 50K) for less than $3 each. The excellent dual 10K Clarostat 585 conductive plastic pots can be ordered from Newark Electronics (part no. 585DX4Q25F103ZP) for less than $3 each. Radio Shack sells a physically larger, dual 100K pot (RS 271-1732), which will work if the value of R2 is increased to between 200K and 1M. (C1 can remain at 0.1uF, and the threshold frequency of the high pass filter will decrease with larger values of R2.)

The diagram above shows how to wire the Clarostat and Panasonic pots. The ALPS pot has the same wiring as the Clarostat. Use an ohmmeter to confirm the wiring diagram. First, choose one section of a dual pot to check. Connect an ohmmeter to measure the pot resistance from the middle terminal (wiper) to one of the end terminals. Then monitor the meter as the pot shaft is turned clockwise from minimum to maximum. If the resistance increases as the pot shaft is turned clockwise, then the end terminal being measured goes to the amplifier ground. If the resistance decreases as the pot is turned clockwise, then the other end terminal should be grounded.

The Power Supply

Figure 3

The power supply circuit (figure 3) converts the 9V battery into a ±4.5V dual supply. Although the OPA134 could run from a single supply, it (and other opamps) are designed for dual supplies, and a dual supply is required for direct-coupling the output. This virtual ground sits at 4.5V, but works because opamps only care about relative power supply voltages. At idle, the opamp output is still 0V (minus a millivolt or two of offset) without capacitor coupling. However, if the headphone amp will also double as a preamp, add a capacitor to the opamp output to block DC, if the input stage of the power amplifier is direct coupled.

The left and right channels are connected in parallel to the power supply. Choose the largest filter caps (C1 and C2) that will fit in the enclosure. I used 220uF caps, but would gladly have replaced with 330uF or higher caps if my enclosure had been bigger. Appendix 3 below discusses power supply options in depth: adding dual 9V supply, making a battery pack, recharging 9V NiCad/NiMH batteries, choosing an AC adapter, etc.

Putting It Together

I assembled the circuit on a printed circuit, 3-hole pad protoboard. I used a Vector Circbord board from Mouser Electronics (Stock No. 574-3677-6). This Circbord has an excellent circuit pattern (featuring numerous bus strips throughout) for this project. Radio Shack sells non-solder-plated boards, which are an acceptable substitute, but the copper will oxidize in time. I cut a small square (about 2″ x 1.75″) of the protoboard with a utility knife to fit the case (mark a section on the board, score it several times with the utility knife and straight-edge, and then break off the section). When cutting the board, make sure to include at least 3 foil “buses” for the power supply and ground. I socketed the ICs using gold-plated machined-contact sockets which work with low insertion force.

The case is a PacTec HML-9VB (Mouser 616-62582-510-039 or 616-62578-510-000). It measures 2.75″ x 4.6″ x 1″ with a built-in 9V battery compartment and both opaque and transparent red plastic front panels. (Note: PacTec may discontinued the red panels). I chose the red plastic panel because it’s thinner and easier to mount the headphone jacks. Many DIYers have been using colorful candy mint tins as enclosures. If the tin’s interior is conductive, it must be insulated with electrical tape or it could cause short circuits. The headphone jacks are enclosed units for 1/8″ stereo plugs. Radio Shack sells a version of these jacks (RS 274-249). I ordered higher quality units that have spring-loaded contacts from Mouser Electronics (Stock No. 161-3502).

Figure 4

The layout of the switches, jacks and the power LED on the front panel is shown in figure 4. The placements are a little tight, but I think it turned out well. By the way, the LED can be either a low current type or an ultra-bright type. It is biased at less than 1mA to conserve battery power and still produces a very strong light. I used a 5mm LED placed in a LED bezel (RS 276-079) before being mounted on the front panel.

Note: If the amplifier is housed in a plastic enclosure, the LEVEL switch must be grounded or the amplifier will hum when the switch is touched. To ground the switch, strip about 1.5″ of insulation from a 5″ length of 22 ga. solid wire, tin the exposed end if necessary, and tightly wrap the exposed end around the groove at the rear of the metal mounting flange of the switch, twisting the end to form a secure, closed loop. Trim the other end of the wire to a suitable length and solder it to the circuit ground. The same is true if a volume control replaces the level switch. If the pot has a metal shaft and the amplifier will be mounted in a plastic case, the pot housing may have to be grounded to prevent hum. Follow the same directions for grounding the level switch housing.

The project came together very quickly – about two evenings – and without incident. I attribute the quick assembly to the simple design of the circuit and the neat layout provided by the Vectorbord. The circuit was first built on a standard breadboard and then transferred to the Vectorbord. The amp worked immediately when the power was applied. I did tweak the power supply for improved stability. My amplifier does not have a belt clip, but add-on belt clips are available at Radio Shack.

The Results

The sound of the amplifier is excellent, with solid bass and a sizzle-free, detailed high end. It powered my Sennheiser 465 headphones effortlessly. A 9V alkaline battery can power the amp for several days of continuous play (high-capacity NiCad and NiMH rechargeable batteries will also work). When paired with my modified Linkwitz acoustic simulator, which is housed in an identical enclosure, the set make for a truly “dynamic duo”. I pack them and a CD player for travel in a Case Logic KSDM-1 case. Since the amp and acoustic simulator are lightweight, they are well-suited for people on the go who like to take with them a complete listening system (of course, you could build both projects into a single enclosure for even greater convenience). Given the low overall cost and the high quality parts used, this project “amply” rewards for the modest expenditure.

Appendix 1: Tweaking the Amp for Low Impedance Headphones

The OPA134 opamp produces a small DC offset voltage, which does not affect the amp’s performance when driving medium to high impedance headphones (over 100 ohms). Low impedance headphones (32 ohms or less) can cause the power supply to become unbalanced, because a small current flows though the load, even when the amp is at idle. This table compares the power supply voltages with the Sennheiser HD600 (300 ohms) and Sony MDR-G52LP (24 ohms) headphones connected to the amp.

Amplifier Load	V+	V-
No headphone	3.9V	-3.9V
HD600 (300 ohms)	3.9V	-3.9V
MDR-G52LP (24 ohms)	4.2V	-3.7V

Note: the battery by itself measured 8VDC.

There is disagreement about whether this almost negligible offset is worth the trouble to fix. With opamps other than the OPA134 series, the offset might be higher and the power supply imbalance could be greater. The offset has not damaged any of my headphones, but it might impact performance slightly by reducing the amp’s power output, injecting noise and/or draining the battery. To determine whether a certain headphone unbalances the power supply, measure the V+ and V- values with and without the headphones plugged in (and no music playing).

For those who want to reduce or block the offset current, figure A1 shows two ways to modify the amp for optimal performance with low impedance headphones: a) add a load resistor or b) AC-couple the amp’s output. A third way is to rebuild the power supply with an active virtual ground device like the TLE2426 or an opamp-based equivalent. Active virtual ground circuits are described in the addendum.

Figure A1

Solution A is the simplest and allows the output to remain DC coupled. The load resistor (figure A1a) will help stabilize the virtual ground and reduce any hiss or noise in the system. The load resistor does create a voltage divider effect with low impedance headphones, and so may lower the amp’s gain and maximum output power and possibly alter the frequency response. Some say that the pocket amp’s gain of 11 is too high for low impedance headphones, so the small drop in gain due to R5 might be desirable anyway. Choose a R5 value just large enough to stabilize the power supply without too much volume loss. I recommend a 1/4 watt, metal film resistor in the 20-50 ohm range.

Solution B avoids a voltage divider effect because although the capacitor blocks DC current, it is largely invisible to audio frequencies. The circuit in figure A1b shows how to switch between AC-coupled and DC-coupled outputs for the highest fidelity with medium and high impedance headphones (the load resistor in solution A could be switched too). Choose the largest value electrolytic capacitor that will fit in the enclosure. A 220uF capacitor will give a flat response down to about 22Hz in 32-ohm headphones.

Use a high quality, low impedance electrolytic capacitor to minimize any sonic coloration. High quality electrolytic caps don’t have to be expensive. The Nichicon Muse KZ series 470uF, 25V sells for less than $1.00 at the time of this writing. The Panasonic FC and FM series caps are also less than $1.00 each. The exotic Elna Silmic II series (which feature a silk fiber dielectric instead of paper) has a 470uF, 25V unit for less than $2.00 each. By comparison, an ultra high-end type like the Black Gate 470uF, 16V typically sells for around $12.00 each and is not recommended for this amp.

Appendix 2: Ideas for Troubleshooting Noise

When built as recommended above, this amplifier is a quiet performer with virtually no background noise. It is more immune to EM and RF interference than some other amplifiers I have heard. The pocket amplifier remained quiet when tested near an old elevator facility that was known for generating loud crackles in another, more susceptible design. Nor did I hear any RF despite that the building had an internal RF communications system.

Nevertheless, there have been a few reports of problems with noise. The first step in troubleshooting noise is to make sure it is coming from the amplifier itself, and not from the audio source. Disconnect the audio source and listen to the pocket amp for any background hiss, static, RF (radio frequency) or EM (electromagnetic) interference. If the amp is driving low impedance headphones (32 ohms or less), try installing R5 (see figure 1) and/or AC coupling the amp’s output as described in appendix 1.

If the noise is primarily RF or EM interference and is not coming from the audio source, it is probably due to long interconnects and headphone cords, which can act as antennas that channel RF signals into the headphone amplifier. The easiest way to block RF noise is to place one or more clip-on ferrite noise suppressors on the audio cables. They should be located on the end of a cable as close as possible to the input or output of the headphone amplifier. The clip-ons can be removed if the interference is temporary and subsides. See A Quick Guide to Headphone Accessories for more information on ferrite clip-ons.

Another way to deal with RF/EMI interference is to shield the circuit either by putting the it in a steel or mu-metal enclosure (connect the circuit ground to the metal case) or by lining the interior of the plastic enclosure with a shielding foil (such as copper). The bottom of the case where the circuit board rests must be insulated with electrical tape to avoid shorting out the amp. If foil is used, it must be connected to the circuit ground. Copper foil shielding tape could also be used (stain glass supply retailers sell inexpensive copper tape).

DIYers have told me that the high gain of the pocket amplifier can emphasize hiss from noisy portable CD players or other audio sources, especially when driving low impedance, high efficiency headphones. If CD player hiss is a problem, try taking the CD output from the Line Out instead of the Headphone Out – in which case, the amplifier must be constructed with a true volume control instead of the LEVEL switch as discussed above.

Figure A2

Another option is to reduce the gain of the amplifier to minimize hiss. Try a gain between 2 and 6 (R3 = 10K ohms to 4.7K ohms). If the amplifier will also be used with higher impedance headphones that can benefit from higher gain, make the gain adjustable with a switch to select between different value feedback resistors (figure A2). Again, make sure to ground the metal housing of this feedback resistor switch to prevent hum and noise from the switch itself (see instructions for grounding the level switch above).

Appendix 3: Power Supply Options

Figure A3

There are several situations, where the pocket amp could benefit from a higher voltage power supply – when driving high impedance headphones, when the amplifier is being fed from a high gain equalizer or when the listener just wants more volume. With very high impedance headphones (600 ohms or more), the amp may not be able to develop sufficient voltage across the load for maximum power transfer. If the amp is fed from an equalizer or tone control with a high boost, the output of the pocket amp could be driven into clipping.

Figure A4

In such cases, I recommend using a ±9V dual battery supply, which is nothing more than two 9V batteries in series (figure A3) or an external power source such as an AC adapter or battery pack (figure A4). R1 can remain 10K ohms, but any value between 10K and 15K ohms will work fine. Unfortunately, two 9V batteries will not fit in the specified enclosure for this project. The Pac-Tec model K-HML-ET-9VB measures 4.6″ x 2.75″ x 1.5″ and has a compartment for two 9V batteries (Newark Electronics part. no. 93F9946).

Figure A5

Figure A5 shows a simple 15VDC external battery pack consisting of 10 AA batteries in a battery holder. The battery holder is Caltronics model BH107 and has snap terminals which fit standard 9V battery snap clips. Radio Shack sells an 8 cell version (RS 270-387) which will output 12VDC. The cable can be any thin 2-conductor cable. I made my own cable by braiding 3 lengths of 24 ga. stranded hookup wire (2 black and 1 red). Only 1 red and 1 black wire carry voltage; the second black wire functions as a shield.

One end of the cable is terminated with a 9V battery clip (RS 270-324). The red wire from the battery clip will carry the (+) voltage when connected to the battery holder and is connected to the red wire of the cable. Only one of the black wires is connected to the (-) wire of the battery clip; the other black wire is not connected on this side. The other end of the cable is terminated with a submini (2.5mm) 2-conductor phone plug, such as the Switchcraft 850X (Mouser 502-850X). Wire the plug so that the tip carries the (+) voltage. The two black wires connect to the ground of the plug. Insulate any exposed connections with a thin layer of electrical tape.

Figure A6

The power jack is the matching submini (2.5mm) 2-conductor phone jack, closed circuit type, such as the Switchcraft TR2A (Mouser 502-TR-2A). The jack is wired so that when the plug is inserted, the internal 9V battery is automatically cut off (figure A5). If the 9V battery were not cut off, the higher external voltage would flow into the battery and possibly cause it to explode. Therefore, the wiring of this jack must be done very carefully. Use a voltmeter to test the jack:

With the jack unplugged and the 9V internal battery installed, the V+ output terminal should read about 9VDC.

Insert the plug (do not connect the battery holder) into the jack. The voltage at the V+ terminal should read 0V (meaning that the internal battery has been cut off).

Remove the internal 9V battery and connect the battery holder (with batteries) to the cable. The voltage at the V+ terminal should be about 15V (or 12V with the 8-cell holder). The voltage across the internal 9V battery clip should be 0V (meaning that there is no backflow of voltage into the battery).

The jack should be mounted in the upper right-hand corner at the rear of the enclosure’s cover. Enlarge the mounting hole of the jack, as necessary, so that mounting nut will be installed flush with the top of the insertion tube (see figure A5). Note: the mounting nut MUST be flush with the top of the jack’s insertion tube or the power plug will not seat properly – a dangerous situation that could short the battery pack. If either the internal 9V battery or external battery pack gets hot during use, there is short circuit somewhere. Disconnect the battery pack immediately and resolve the problem.

The battery pack also could short if the plug were to come partially loose in the jack. For this reason, I do NOT recommend using this battery pack while traveling. Safer alternatives to the phono plug and jack are coaxial DC connectors, which will not short if the plug is unseated. When the amp was being constructed, I could not find DC coaxial jacks small enough to fit on the side of the case. The Switchcraft 712A (Mouser 502-712A, Jameco 281842) fits in a 0.313 inch hole. The mating plug must accept a 2.5mm (0.1″) pin, such as Switchcraft 760 (Mouser 502-760, Jameco 281877).

The mounting threads of the power jack are in electrical contact with the power jack’s ground. If the amplifer is put in a metal enclosure, the virtual ground and the power jack ground must NOT be connected together or the virtual ground will be shorted out. To prevent this occurrence, insulate the power jack’s mounting threads from the metal enclosure with nylon washers or electrical tape on both sides of and within the jack’s mounting hole. Use an ohmmeter to confirm that the power jack ground is not in electrical contact with the enclosure.

An AC adapter could replace the external battery pack. Most AC adapters are poorly filtered and will introduce noise into the amplifier. The best AC adapter for this project is a wall-wart with a regulated, non-switching supply. The adapter shown above (RS 273-1662) can output up to 12VDC at 300mA regulated. It also comes with a set of interchangeable power plugs, including a 2.5mm phono plug that should be compatible with the power jack in figure A5, so long as the voltage polarity is correct.

Figure A7

The circuits in figure A7 turn the AC adapter into a NiCad/NiMH trickle charger with a 20mA charging current. Trickle charging takes longer but is gentler on the battery. The circuits are identical except for the value of the resistor that sets the charging current. Figure A7a is for the specified NiCad battery, and A7b is for the specified NiMH battery. The 9V NiCad from Radio Shack (RS 23-448) has a capacity of 120mAh and should achieve a full charge (8.2V) in about 5 hours. The 9V NiMH (PowerEx MH-96V230 by Maha) has a higher voltage and almost double the capacity of the NiCad. It will take almost 10 hours to fully charge (9.6V). NiMH batteries are very sensitive to overcharging. The charger must be turned off when the battery is fully charged to avoid shortening the battery’s lifespan. The 1N4001 diode prevents the battery from discharging backwards if the 12V adapter is not turned on but is still plugged into the amp.

Appendix 4: Turning the Pocket Amp into a Personal Monitor

Figure A8

Commercial personal monitors for musicians can be expensive, yet are essentially nothing more than headphone amplifiers with a limiter and/or a balanced input option. Figure A8 shows the pocket amplifier with both balanced and unbalanced inputs. This simple wiring trick for converting balanced signals to single-ended signals isn’t free: the signal amplitude is cut in half, but the loss can be compensated by turning up the volume. A true balanced converter that preserves the signal amplitude and noise rejection can be found in Designing an Opamp Headphone Amplifier.

Figure A9

Figure A9 shows an adjustable clipper, which can limit headphone volumes to safe levels. The maximum voltage that the headphone can see is 0.7Vp (the forward bias voltage of the diodes), so the clipper is most effective with high efficiency headphones of low to medium impedance (less than 200 ohms). High impedance headphones may not achieve enough volume even at the maximum setting. In that case, try replacing each diode with two diodes in series to raise the clipping voltage to 1.4Vp. The clipping effect is a little harsh because of the hard cutoff by the diodes. P1 is a trimmer pot or an inline stereo volume control, such as those made by Koss or Radio Shack.

Figure A10

The limiter can be set for only one headphone at a time. Different models of headphones have different sensitivity ratings, so the limiter must be readjusted if the headphones are changed. The more accurate and safest way to set the limiter is with an audio level meter and a headphone coupler (or artificial ear) sold by audiometric suppliers. If such equipment is not available, the limiter can be set by ear, but with less reliable results.

For initial testing, it is a good idea to use a pair of disposable headphones with the same impedance and the same or higher sensitivity as the intended headphones. Begin by turning the amp’s volume control to minimum. Do not connect the headphones yet. Feed an audio signal into the amp and turn up the volume until the diodes are forward-biased and clipping the signal. Use a voltmeter set on AC to confirm that there is about 0.7V across each of the diodes. The voltage should stay at about 0.7V even if the volume is turned up higher, indicating that the diodes are clamping the signal.

Set P1 in both channels for maximum resistance or set the inline volume control to minimum volume. With the trimmer pots, only one channel can be set at a time. With the inline control, both channels are set simultaneously, but if the channels don’t track precisely, always set the limiter based on the channel that is louder.

Connect the disposable headphones. Adjust P1 or the inline volume control slowly until the headphone volume reaches the desired level. Confirm that the limiter is working by turning up the amp’s volume control. If the volume increases, reduce the volume by readjusting P1 or the inline control. Repeat until the circuit clips at a consistent volume level. Then turn the amp’s volume control down to minimum and plug in the intended headphones. Slowly increase the volume and confirm that the clipping level is set correctly.

Once the pots are set, the settings must be protected against accidental change. While trimmer pots on a circuit board would be protected by the amp’s enclosure, it’s best to fix the thumbwheels in place with a dab of white glue. If an inline volume control is used, wrap the thumbwheel with electrical tape. For tips on setting maximum headphone volume, see Preventing Hearing Damage When Listening With Headphones. For more information on limiters, see Designing a Limiter for Headphone Amplifiers.

Addendum

12/4/98: Adding wiring diagram for headphone jack in figure 1.

11/25/98: Rewired SW1 in figure 1 to eliminate hum. Corrected R1 in figure 2.

11/20/98: Revised R1 in figure 1 to range from 100K ohms to 470K ohms, depending on desired input attenuation.

5/22/99:

Jason Portman built the above version of the pocket headphone amplifier with an anodized aluminum case by Context Engineering, Inc. (available at Fry’s Electronics), volume control (10K) and blue LED. The larger size of the case allowed the use of 1uF WIMA polypropylene capacitors to couple the input. Very nice!

7/7/99: I have just been told that Digi-Key is backordered on the Burr-Brown opamps used in this project for the next 15-23 weeks! Here are some other sources: Insight Electronics and Sager Electronics. I have never order from these companies, but they are listed as Burr-Brown distributors.

7/12/99: Corrected polarity of LED in figure 2.

7/14/99: Added section on converting the pocket amp into a personal monitor.

8/24/99: Mika Vääräniemi built the modified Linkwitz acoustic simulator and pocket amp in a single aluminum enclosure. The power supply is an AC adapter that outputs 9VDC regulated. Here is the parts placement and wiring diagram that he used:

He added a switch (S3) to turn off the treble boost and changed the values of C1 and R1 to C1 = 1uF and R1a = 50 ohms, R1b = 100 ohms. These values seem to give the widest soundstage with the least effect on the high frequencies. “[B]efore I was positioned in the middle of band playing music. Now I’m in the front row as close as you can be…. Music just sounds realistic and that’s what I was looking for.” A more complete description of his work can be found in the DIY Workshop Forum.

DIYers who would like to built both the simulator and amplifier together may want to scale the resistors and capacitors of the simulator section to increase the input impedance to about 2K ohms (x10 for resistor values, ÷10 for capacitor values – and use a volume pot between 10K and 50K ohms). Increasing the input impedance is not absolutely necessary, but it may then work better with some preamps which have a high output impedance.

8/25/99: Here are pictures of Mika Vääräniemi’s completed headphone amplifier with acoustic simulator:

9/2/99: Jim Burruss built a “micro mixer” based on the pocket amp design. He used a metal candy box for a compact enclosure that also provides excellent shielding:

Attached are some digital photos of the micro mixer I built based on your design. I’m an electrical engineer and musician. I play a MIDI horn and needed a way to mix the signal from an electronic metronome with the output of the synthesizer for quiet practice. Your design ideas were great. I had an old Altoids box that looked just big enough to house it.

It has one mono input for the metronome with on-off and volume control on the pot with the short shaft. The other channel is stereo with its own ganged volume control. [Editor: The pots are available from Radio Shack.] The output is to drive headphones. I built it with an LM358 dual opamp just to verify the wiring and have an OPA (same pinout) on order to improve the sound.

The Altoids box provides great shielding. The board is insulated from the box with a fold-up plastic box made out of the packaging material from the metronome.

9/7/99: This version of the pocket amp by Tomohiko Ishigami uses the acoustic simulator circuit by Jan Meier (see A DIY Headphone Amplifier With Natural Crossfeed). He reduced the gain of the amp to unity to minimize problems with noise, which he later traced to the CD player itself. The larger case is from Radio Shack (RS 270-213).

I feel it is very good idea to use modular approach. I used separate board for crossover and the buffer itself. This way, I did not have to go crazy load all the parts on one board which will result in a hay wire. Also, this approach is useful when I was trying to achieve smaller size.
I was able to use 1uF polymer capacitor for input…. These are so tiny. It is made by Phillips and you should be able to find it in Digikey [Digikey part nos. shown below]. I used this same type for my crossover circuit allowing me to conserve a lot of space:

3019PH-HD 1uF Metal Film Box ( 10mm (H) by 7mm by 6mm (L) )
3015PH-ND .22uF Metal Film Box
3011PH-ND .047uF Metal Film Box

11/21/99: Added section on replacing level switch with a volume control.

11/21/99: Stephen Jenkins wrote: Wow, I just finished building the headphone amplifer that you designed. I am in awe at the sound quality while using my little (but fabulous) Koss Porta Pro Jr’s and my Pansonic SL-S360 portable CD player. The only change I made was that I included an AC jack on the side so that I that I could plug into the wall while at home, this was really easy and I highly recommend it. Thank you for the plans, you’ve made my day!

12/18/99: Added section on implementing a dual 9V power supply for driving very high impedance headphones.

1/7/00: Several DIYers have installed Jan Meier’s natural crossfeed filter as a front-end to the pocket amp. Jan offers these tips re: selection and placement of a volume control for this combination: It all depends on the specific circuitry. Generally it might be better to place the pot after the filter instead in front of it. The influence of impedance changes might be less pronounced. A 10 kOhm pot will certainly be too small. 50 kOhm will be a kind of minimum I think. However, note that with certain opamps this will result in changing offset voltages, since the DC impedance changes with volume.

1/12/00: scrazy@gcn.net.tw built this version pocket amp, which has a 10K ohm volume control and an acoustic simulator front-end by Chester Simpson (see design by Fred Peng below). He used OPA134 opamps and set the gain to unity because his CD player’s line out supplies more than adequate drive voltage. Full details can be found at DIY Zone (in chinese only). His system consists of a Rega Planet CD Player and Audio Technica ATH-f15 headphones.

1/13/00: Fred Peng’s headphone amplifier incorporates the acoustic simulator by Chester Simpson, except that he replaced the R4,R6 combination in Simpson’s circuit with a 100K ohm resistor and added a unity gain input buffer stage made from an OPA134 and a high current output stage made from a PMI BUF-03 buffer. The opamp power supply is double regulated for the cleanest output. The first stage of the power supply outputs ±34VDC, which is regulated to ±22VDC and again to ±15VDC. TWhen compared with a McCormack Micro Headphone Drive, the BUF-03 driving his Grado HP-1 headphones with the simulator bypassed sounded better in the high and low frequencies than the McCormack, but the McCormack was better in the mid frequencies. With the simulator switched in, the sound was more relaxed, the low frequencies were slightly “nasal”, and the soundfield moved from inside his head to outside. He is very satisfied with the result and is planning to make another simulator for his Stax Lambda headphones. Full details and schematics (in chinese only) can be found at DIY Zone.

2/7/00: Eric Lee‘s pocket amp has a modified Linkwitz acoustic simulator front-end. He says “it works great…nice design…and I can hear almost no audible noise from it.” He used a slider pot for the volume control and installed dual headphone jacks for 1/4″ and 1/8″ headphone plugs. The enclosure is from Radio Shack.

5/1/00: Forest Chang built a pocket amplifier with a modified Linkwitz simulator front-end and the component value changes suggested by Mika Vääräniemi (see above). He writes: The circuit that I built is the same as Mika’s, but the OPA that I used is the OPA134, and I put an OPA2134 as a buffer in the front-end of the acoustic simulator. The grounding method that I use is to tie the output ground, power supply virtual ground and switch housing together. Then I connect this common ground to touch the metal watch box (the enclosure that I used) with a spring. The amp has no hiss, even when I put it right beside the monitor. And I cut a beautiful picture from a metal candy box and put it into the watch box. My girlfriend uses the amp with a Panasonic SL-280 and Sennheiser HD-320 headphones. She is very happy with the sound improvement, and the cute headphone amp.

5/1/00: Revised figure 2a and section on using dual 9V supply. Added section on constructing pocket amp with adjustable gain. Expanded description of how to cut protoboard with a utility knife. Added figure 6a – pocket amp with balanced input.

5/1/00: Jeff Medin‘s pocket amplifier has 3 sections: a gain stage, the crossfeed filter by Jan Meier and an output buffer stage. The power supply creates a virtual ground with a Texas Instruments TLE2426 voltage reference instead of a resistor divider network. The 1uF (or less) capacitors are Philips box-type metal film; capacitors larger than 1uF are Panasonic FC/Z series. All resistors are 1/4W Yaego metal film. Medin writes: This is the FIRST amp I built after discovering HeadWize. It is a “basic” pocket amp with the natural crossfeed circuit by Jan Meier. ALL parts are from Digikey. It has very good decoupling with 3 capacitors per opamp and 3.9uH chokes (the 4 green things that look like resistors – they are connected in series with each V+ and V- lead). The first stage (on the left side of the first picture) is an OPA2132 with a gain of 10.

This then feeds a Meier crossfeed circuit (4 caps in a row) and you can see the crossfeed resistor on TOP of the board (2.2k) with long leads. The output from the filter feeds a voltage follower (OPA2132) stage. The switches are for low and high crossfeed, power, and bypass for binaural recordings. I used Philips Box style metal poly caps. The two large caps on top & bottom of board are 1uF input caps. The output is taken from the OPA2132… with a 100 ohm resistor… which is included in the feedback loop so it will drive very low z phones and to prevent oscillation due to capacitance from long cables. I used 100 ohm resistors in BOTH stages.

If the resistor is OUTSIDE the loop, the impedance WILL have an effect on the sound of the phones, sometimes more bass, sometimes MUCH less signal based on the efficiency of the phones, etc. etc. Some phones as you know are spec’d to be run from an impedance of 100-150 ohms or so. I have a 15 year old APT/HOLMAN preamp (designed by same guy that invented THX-Tom Holman) and it’s Headphone Jack is driven by a 5532 with a 120 ohm resistor OUTSIDE loop right to the jack. I would suggest people can try both (like Jan did) and see what sounds better to them. I would DEFINITELY recommend that you include this resistor in at least the last stage.

Note that I did not have any problems, I always “over-build” opamp circuits so I don’t have to worry about problems later on. It’s just habit.

5/4/00: Jasmin Levallois‘s amplifier is similar to Jeff Medin’s, except that he uses the Meier enhanced-bass natural crossfeed filter (and the original resistor-based virtually-grounded power supply). He writes: Finally I got some free time to complete my project…. I got a lot of work to do for school during the last few weeks and I didn’t have time to work on my amp. This weekend I decided to take one day to transfer the amp from the breadboard to the pc board. I used about the same circuit as Jeff Medin. The input stage has a gain of 10, the output stage is a voltage follower, and in the middle I put the Meier bass-enhanced crossfeed circuit.

I used 2 OPA2132 opamps, but if I had to do it again I would use 2 OPA2134. An OPA2132 costs $6.99 while an OPA2134 costs $2.67. Since there is almost no audible difference between both opamps, I would go with the OPA2134 to save money. Since the second stage has no voltage gain, I decided to omit the capacitor in front of the output stage. I also removed the resistor in front of the output stage, and I don’t hear any noise from the output stage. The only noise I can hear, sometimes, is coming from my CD player.

As you’ll see on the photos, the inside of my amp is very messy, but, hey, its my first electronic project. Fortunately, even if it’s messy inside, the outside looks pretty good. I really like this Serpac Enclosure (Digikey part no. SRH65-9VB-ND); it looks ways better than the PacTec case.

The photo of the battery compartment is to show that the Serpac enclosure has a 9v Battery compartment with battery contacts. It’s easier to remove the battery with that kind of battery compartment than the PacTec Enclosure. Also the Serpac enclosure is just about the same size as the Pactec enclosure except that it’s a bit longer, and the height is a little bit less. This might be a problem for the electrolytic capacitors. I would recommend the Philips ones with this enclosure rather than the Panasonic Z series because the Philips electrolytic caps are much smaller.

Download parts list for Levallois Amplifier (MS Excel format)

6/16/00: Jasmin Levallois writes: This weekend I finished to build another pocket headphone amplifier for a project in my physic class. I used an Altoid box like Jim Burruss did, but I used the cinnamon kind to not be accused of plagiarism ;). This enclosure has the advantage that I could show to other students the circuit, and it is also very small and provides great shielding.

I took my original circuit and I improved it a little bit. First, I replaced all my Phillips polymer capacitors by some Polyester made by Panasonic. I followed Jeff Medin’s recommendations and added a 100 ohms resistor in the feedback loop of the last stage. I used a .12uF capacitor to decouple each power supply pins. I also added a 100k resistor connected to the ground in front of the output stage. On my last circuit I had omitted this resistor, but many people in the forum convinced me to put it back.

I built the complete circuit on a very small board (4cm by 5cm) and I don’t think it would have been possible to make it much smaller than this. To save some space on the board, but also because Digikey was out of OPA2132, I used a single OPA4134. It is pretty cheap, $2.30, I think, and I really recommend it. I had a hard time to find some good electrolytic capacitors that would fit in the small enclosure. Finally I used some mini alum electrolytic capacitors made by Panasonic. You can find the Digikey part # of these capacitors in my part list.

This amp sounds great and looks great; I love it.

Download parts list for Levallois Amplifier (MS Excel format)

10/11/00: Added sections on AC adapters, troubleshooting noise. Revised section on volume controls.

11/23/00: Bob Scott put his pocket amp into an Altoids candy box and uses it between his Sony MD player and his Sennheiser HD495 headphones. He writes: Attached are photos of my amp. I built it into an Altoids tin, partly for shielding, partly for the entertainment value. The only changes I made from your schematic was a slightly larger resistor for the LED to reduce current draw and using a “pigtail” for the input to save some panel space and reduce bulk when “cabled up”.

I got the short-handled switches from Digikey. They kept the unit compact and reduced the likelihood of the amplifier turning on accidently. I may build a second copy using “dead bug” construction to see if I can make it REALLY small.

11/23/00: Carl Hansen has designed PC boards for the Levallois version of the pocket amp with Jan Meier’s enhanced-bass crossfeed. He writes: I have been spending the past nine months following your forums and building a number of variations of the pocket amp. I have decided that I have more than a few friends that would like to have one for Christmas in either kit or a variety of assembled form…. Because I have the resources available to me through my work I have gone ahead and laid out a nice little double sided board using Tango PCB, which I have sent to one of the commercial board houses in the Seattle area for a small “prototype” run. My boards arrived last week and I have assembled three of them and they work great!

The board house that fabricated the boards is fully automated meaning that no human hands were involved in the manufacturing process including a complete optical inspection using a robotic vision system…. I would like to sell off some of my excess boards. The price to sideliners in the forum like myself will be $6.50 each (or 3 for $17.00) plus $3.00 S&H; which is about the same as the cost for using Vectorboard. To those that have posted contributions to the forum that have furthered the dialogue, particularly regarding the pocket amp, I would like to offer two boards each for free except the cost for S&H.;

The specifications for the board are:

Dimensions: 1.80″ X 2.45″ with routed notches and corners to precision fit Pac-Tec case HML-9VB, leaving a 1.25″ space behind the panel for components such as switches and jacks. The amplifier section is designed for dual OPA2132/4s with the crossfeed filters between the amplifier sections. There are provisions for two levels of enhanced-bass crossfeed filters plus flat. A 3 pole, 3 position rotary switch or some equivalent would be required to use all three settings. The filter capacitor component locations have multi-holes each to allow the use of different size capacitors. There is a provision for volume control or high-pass filter resistor. Gain of course is a matter of component selection. Personally I have found a gain of 5 to be the most versatile. There is also a provision for power indicator LED.

Shown below are the Levallois schematic and pictures of the Hansen PC board. For more information about the circuit, see Levallois’ entry in the addendum update (p. 1) for May 4, 2000.

Update: C.E. Hansen is no longer selling the PC boards or the Noble XVB93 mini-pots described in the article. Instead, Jon M. Tsukiji (JMT in the forums) is now selling the PC boards for the same price, although he is NOT selling the Noble pots. JMT is also selling completed amps in the Penguin Mints boxes first shown by “Apheared.” Contact JMT for pricing on the completed amps and to order the Hansen PC boards.

Jon M. Tsukiji
3142 Spruce Hill Ct.
Antelope, CA 95843
Email: JMT@surewest.net

3/14/2001: Major rewrite of article, including new appendix section on power supply options. Added new high resolution pictures.

3/14/2001: Coffin Lin put this amplifier (with a modified Linkwitz crossfeed front-end) in an old TV remote control case. He used an OPA627 opamp and made R2 and R3 in the Linkwitz filter adjustable instead of R1. The volume control is an Aiko pot in a shunted configuration with a 50K resistor (Dale RN55D), so that the audio signal passes through a single high quality resistor. Regarding his selection of the opamp, he writes:

I found that the OPA637 oscillated, even though the gain was set to greater than 5. The power supply voltage was not symmetric (2V/10V using a 12VDC AC supply). Then I changed the opamp to an OPA627, which was quite good for both my Sennheiser and Pro2 headphones, but the supply voltage was still not symmetric enough (6.4V/6.8V). The OPA134 got best result in stability (6.5V/6.7V), but the sound is too fat for me. So the final version is OPA627 – great detail, sound balance, clear, dynamic.

Lin put the Linkwitz filter at the input to the amplifier. The component values in his version of the filter are:

R1: 30K ohms
R2a, R2b: 15K, 10K
R3a, R3b: 50K, 100K
R4: 33K
R5: 33K
C1: 3,300pF
C2: 10,000pF

The resistors are Dale RN55D. About making R2 and R3 adjustable, he says: I mistook R2 for R1, but on the Excel worksheet simulator, R2 can still alter some balance. I think that adjusting R3 is more effective than adjusting R2 (I forget which switch is for what resistor.) One has more stereo (good for dance and rock) and the other is more natural (good for jazz).

12/26/2001: Revised value of the current limiting resistor in figure A7. I reviewed Stephen Lafferty’s circuit for charging a single 9V NiMH battery. The value of current-limiting resistor in Lafferty’s circuit assumes that the specified unregulated 12VDC adapter will output 14VDC, because the amp is a very light load for the adapter. The recommended adapter in my project has a regulated output, so the output should be 12V exactly (or fairly close). Therefore, I changed the value of the resistor from 330 ohms to 220 ohms to get a charging current of about 20mA.

12/26/2001: Here are three candy box amps from forum members Doh, Droche and LivingPlasma. Doh put his Hansen-board amp in a Penguin Mints box (first shown by Michael Shelton – a.k.a. “Apheared”). He writes:

It looks like Apheared beat me to posting a Penguin Mints amp, but I swear I didn’t steal the idea! Penguins rock! I’m afraid my amp isn’t nearly as DIY cool as Apheared’s creation, but it’s only my second amp and I just learned how to solder a few weeks ago!

As you can see, the Hansen board is mounted upside-down in the tin with the power and crossfeed switches sitting right underneath the op-amps. It has dual pigtails and one position of crossfeed plus flat. There’s no LED and volume is adjusted via an inline volume control from radio shack (soon to be replaced by a DIY version that uses the panasonic pot once those parts get in). I don’t have any of that fancy tape, so I actually just stick a metrocard under the lid before I close it. (Haven’t gotten around to glueing it in with some artist’s spray mount quite yet).

I think that there is still enough space in the box to wire the pot inside if anyone feels like giving it a try. Personally, I like the flexibility that a modular volume control gives me. On the other hand, I’m still trying to think up a way to get rid of the pigtails to improve portability.

Just a note on drilling the holes in an altoids tin or other metal candy container. What I found to work really well are the black and decker “bullet” tip drill bits. They have a small extension at the point that bites into whatever you’re drilling into so that the drill bit doesn’t slip. The tip works its way through the metal fairly quickly, so after it’s through you have a pilot hole that holds the bit steady while the rest of the bit does the work. Using these bits, I found drilling holes up to 1/4-inch to be no problem. The bits are available through amazon.com, but should be widely available.

Droche put his amp in the popular Altoids tin. He writes:

For any of you who think that building an amp is too difficult for a beginner- I am proof that it isn’t. I started browsing the forums a month ago with no electronics experience whatsoever. After browsing for a while, I put in a few orders and before I knew it, I had a headphone amp. It took a few tries to get it into the box, but I finally got it in after removing the headphone jacks and adding pigtails and removing the pot. It was well worth the effort. I was amazed at how much better the sound out of my portable MD player got. Thanks to everyone here for all the helpful info.

Livingplasma put his amp in a round candy tin. He writes:

Not to take the attention away from Apheared, but I just couldn’t help it after seeing all thse proud people post their version of the CMoy pocket amp. Those who have been here a while will know I had a string of bad luck making my first cmoy, this is what I came up with the leftover parts. It’s the basic CMoy amp made with an OPA2134 and modified for the Meier crossfeed (changed some values so I could use a 50k pot and smaller input capacitor; yes, it’s unbuffered). Input is through the pigtail, has a LED power indicator and uses one submini toggle for power and one for the crossfeed (on or off). Measures about 3 inches in diameter (not counting the controls), and just under an inch in height. Schematically, I think it’s very similar to Tomo’s version.

Drilling the holes on the side of the tin is annoying, to say the least. After a certain hole size, it’s really hard to drill a hole, the bit catches on the metal and goes ripping the case apart (lesson learned trying it with a Altoids tin). I just made the hole as big as possible with the bit, then reamed it with either a screwdriver or a knife. The opening for the volume pot (it’s those panasonic ones) is a square, I think I used some old diagonal cutters I didn’t mind messing up and some pliers to bend and break the tin.

12/28/2001: Kenji Rikitake (a.k.a. “bdx” in the forums) built two versions of the pocket amp with an opamp-based virtual ground. He says:

The amp on the left in the picture is the OPA2134 version; the one in the middle is a single-amp version (OPA134). The breadboard on the right is for the OPA2134 version. The basic amplifier circuitry is same, but I changed the value of the feedback resistors to 1k/4.7k ohms pair. The 1k ohm resistor at the + input of the opamp protects it from accidental overcurrent or overvoltage (though the probability is very low). This is generally recommended when you make a non-inverted amplifier.

I tested with several different opamps. OPA2134 showed its excellence (only 3mV maximum output offset). NJM4580DD worked OK though it had 70mV maximum offset. The NJM082D (TL082 compatible) also worked but it could not fully drive my Sehnheiser HD414 Classic. Note that the NJM2043DD didn’t work (caused self oscillation).

I also added an opamp voltage follower for providing the virtual ground, to stabilize the voltage exactly to the 1/2 of the unipolar power supply (namely a 9V dry battery). I tried the OPA134 as a unity gain buffer for the virtual ground driver and then the BUF634 (as ppl suggested). The quiscent current of the chip lowered from 4mA to 1.5mA or so, and the amplifier sounded the same. Note that OPA134 and BUF634 are virtually pin-compatible if you use the OPA134 for a unity-gain buffer. The 1N4002 diode protects the circuit from accidental inverse voltage connection.

The circuit is put into an aluminum case (T-SIN Denki TM-1) which can hold a 9V dry battery inside and can mount a 47mm x 72mm breadboard widely available here (Sanhayato ICB-88 compatible) in Japan. The FatBrain.com stickers are some of which I’ve got from the bookstore. The size of each amp is 87mm(W) * 31mm(H) * 103mm (D). I built the circuit on a glass-epoxy DIP breadboard. Since I proved the amp works fine with my Sony D-E880, Diamond Rio500, and Sehnheiser HD580 as well as with HD414Classic, I think I’ve got to build another one for my wife sooner or later. 

More details about Bdx’s amp and power supply can be found here.

2/21/2002: Added note about insulating the power jack ground when the power jack is mounted in a metal enclosure.

2/22/2002: Forum member Tangent has created a tutorial for electronics newbies who are interested in building the pocket amp. Many DIYers have found the tutorial very helpful. Please note that Tangent’s opinions are not necessarily the same as this author’s.

5/20/2008: Added a new appendix 1 about how to optimize the amp for use with low impedance headphones. Updated sections on choosing a volume control and using rechargeable batteries. Revised figures 1 and A7.

6/30/2008: Revised Appendix 4 and figures A8, A9 and A10.

Designing an Opamp Headphone Amplifier.

Because a few milliwatts will drive headphones to full volume, a great headphone amplifier design can be relatively simple. Yet, there are any number of reasons for experimenting with more complex topologies, such as improved performance and the ability to incorporate custom options. Of course, some DIYers like to try different circuits just for the fun of it. The major disadvantage of complex circuits is that they are complex. It can take a DIYer months to locate and purchase the parts, not to mention the time to assemble and troubleshoot the project.

Integrated circuit opamps are both complex and simple. They may contain hundreds of components on a chip, but are relatively easy to configure. For DIYers short on time and patience (and few people have the luxury of both), opamps are a convenient entre into the world of complex design. The audio cognescenti have attacked opamps as being one of the major causes of “mid-fi” sound, but if the truth be known, they are lurking everywhere – even under the covers of prestigious high-end gear. Opamps are not all the same. Building a headphone amplifier with good sound is a matter of careful selection and design.

This article discusses several opamp-based headphone amplifier circuits, including suggestions for selecting opamps, input coupling and filtering, high current output stages and power supply options. There are no recommendations for specific opamp brands or models. For tube devotees, there is also an introduction to designing with tube amp-blocks. Tube amp-blocks (AC feedback amplifiers and tube opamps) are not as compact as their silicon brethren, nor do they measure as well, but they do offer smooth tube sound with the ease of feedback configuration.

SELECTING SOLID STATE OPAMPS

Entire books cover the subject of interpreting opamp specifications. Here are a few guidelines for choosing opamps when designing headphone amplifiers. Opamps inch closer to the “ideal” with every succeeding generation. Modern devices are internally compensated for stability, have slew rates going through the roof and noise and distortion numbers at threshold of measurement. There are even opamps that will run off a 1-volt supply. For portable devices, the power supply requirements should be the first consideration. The majority of modern opamps will run with as little as ±4V, but low voltages may degrade performance. Check the manufacturer’s V_CC specs to confirm that low voltage operation is, in fact, recommended. The most common battery supply voltages are ±1.5V, ±3V, ±4.5V and ±9V. Single supplies are another possibility. Keep in mind also that the idling current for the entire amplifier must also be low – around 10mA or less for good battery life. For more information, see the section on battery power options below.

Opamp performance specifications are an unreliable indicator of sound quality. So long as the numbers are below audibility thresholds, specs that are magnitudes better than the averages will not necessarily translate into better sound. Regardless of type (bipolar or FET), modern opamps do very well on the test bench. Total harmonic distortion figures are so low (typically less than 0.1%) that datasheets have stopped listing them. Look for noise specifications, listed as “noise density” in units of nV/Ö(Hz), of 25 or less, slew rates of 5uV/sec or more and “wide” unity gain-bandwidths of 3 MHz and higher.

Figure 1a

When reviewing the gain-bandwidth specification of a bipolar-input opamp, also examine the open-loop bandwidth. The gain-bandwidth defines the amount of small-signal gain at any frequency and is the product of the open-loop bandwidth and the open-loop gain. Most opamps have a high open-loop gain (100dB or more) and a relatively narrow open-loop bandwidth (100Hz or less). In a multi-stage system with overall feedback, if the opamp has a bipolar input stage and narrow open-loop bandwidth, it can manifest dynamic phase shifts and other response non-linearities with high level, high frequency input signals.

To reduce this type of distortion, choose a bipolar-input opamp with a wide open-loop bandwidth (into the kHz range) or use a FET-input opamp. FET input stages are more linear and so less susceptible to this type of distortion. Finally, the open-loop bandwidth of the voltage-gain input stage can be effectively extended with local feedback (see the section on output stages below).

Also look for unity gain stability and low offset voltage. Opamps that are internally compensated are less likely to oscillate at high frequencies, and save the builder the hassle of adding external compensation (however, it never hurts to check the amplifier output on an oscilloscope anyway.) The ideal opamp has zero DC output at idle, so that DC coupling can be accomplished without trimming. Real-world opamps have a small output voltage at idle. If the opamp is not followed by a gain stage, a 15mV or less offset at idle should be acceptable. FET-input opamps are known for their low offset voltages.

Are there audible differences between opamps with similar or identical specs? Some listeners can distinguish between products, but not all. Because modern opamps are internally compensated and are usually plug-in replacements for each other, building circuits with IC sockets or on protoboard first allows the DIYer to audition a variety of opamps at will. Theories abound as to why opamps may have sonic signatures in spite of stellar test results that suggest neutral sound. Years ago, IM, DIM, TIM, etc. distortion were held to be the culprits. Two of the most recent courses of research on this topic have been the effects of the harmonic structure of opamp noise and opamp input errors.

Figure 1b

The first course of research posits that much distortion in audio signals is actually noise, which may be too low to measure but is still audible. Human hearing is very sensitive to high order harmonics produced by high negative feedback ratios. Noise structures with a predominance of even order harmonics seem to sound less harsh. Opamp systems with poor harmonic structures can have improved performance if the systems are designed so that the harmonics cancel or harmonize with the products of other stages in the system. Figure 1b is a plot of the noise spectrum of a common bipolar-input opamp. Manufacturers generally do not include such analyses in datasheets. Since these tests must be done with sophisticated equipment that can measure noise 140dB or more below the signal, most DIYers will have to rely on other published sources for this type of data.

There are three types of opamp input errors that potentially affect sound quality: source-impedance, power-supply and thermal errors due to the output loading of an opamp. Source-impedance errors arise when there are unequal source impedances at each of the two inputs to an opamp, which interact with the opamp’s internal capacitances to create even-order harmonic distortion. It is a common-mode type error, and so applies only when the opamp operates in a non-inverting configuration. JFET-input opamps have an internal capacitor at each of the inputs, and are likely to show higher levels of source-impedance distortion than bipolar-input types.

Source-impedance errors can measured by comparing distortion levels when the feedback network impedance (R_f||R) differs from the input source impedance R_s and when they are the same. Selecting opamps with low internal capacitance or balancing the source impedances will minimize this form of distortion. The latter technique is discussed in the section on configuring opamps for voltage gain below.

Figure 1c

Power-supply errors occur when noise from the power supply mixes with the input signal. The PSR (power supply rejection) specification is a measure of how well an opamp is able to block power supply noise and values of 100dB or more are common. PSR will vary with frequency, but the spec usually refers only to DC behavior. Instead, search for a graph in the datasheets of the PSR over the audio frequency range. In addition to choosing opamps with high PSR over a wide audio range, power-supply errors can be reduced by using power supplies that are highly regulated and bypassed.

Figure 1d

Opamp-based headphone amplifiers can be prone to thermal errors due to output loading of the opamps. The power dissipation in an opamp when driving low impedance loads can raise the temperature of the device and cause changes in the input offset voltage, thus compromising the linearity. In dual and quad IC devices, the thermal conditions of one opamp can affect all others in the package because the opamp circuits share a common substrate (“power-dissipation-related crosstalk”).

Thermal errors can be measured by comparing the output distortion of an opamp under load and no-load conditions. Figure 1d shows the thermal loading effect of one buffer on the other in a dual buffer IC (the system circuit is similar to figure 5d). Channel A is being fed a frequency sweep signal; channel B is idle. When buffer A is driving a 25 ohm load (as opposed to driving no load), it induces a stronger thermal-related error signal at the input of buffer B.

When the dual (or quad) IC buffers are used in circuits with voltage gain front-end as seen in figure 5d, these thermal errors can be corrected via global feedback. Conversely, buffering an input-stage opamp can reduce thermal errors in the input stage by isolating the power dissipation to the output stages. (For more information about buffering, see the section on output stages below.) In general, using single opamps instead of duals or quads will prevent power-dissipation crosstalk distortion. Precision low-noise opamps appear to have the lowest thermal errors.

TUBE-BASED OPAMPS

Unlike transistor amplifier design, tube amplifier design is more dependent on the electrical characteristics of the tubes themselves. Tube opamps attempt to bring the simplicity and higher preformance of amp-block design to tube audio. In audio applications, they can aspire to the same high performance as their solid state cousins and have the additional benefit of even-order distortion harmonics. There has been a revival of interest in these devices with the publication in recent years of several amp-block circuits, ranging from basic AC feedback amplifiers to tube-MOSFET hybrids – all configurable with the familiar opamp feedback scheme.

Not widely available back in the heyday of glass audio, tube opamps are very hard to find today. The following circuits develop the tube amp-block concept with increasing complexity. They all have limited current output and may need an output buffer stage to comfortably drive headphones (see the section on output stages below). To adjust the closed-loop gain of any of these amp-blocks, simply add a feeback resistor from the output to the inverting input and an input resistor – just as with solid state opamps. Some of these amp-blocks may prefer higher magnitudes of resistance than are typical of solid state opamps, so sample gain resistances are included. It’s a good idea to construct several of these at a time to have a handy supply for experimentation.

Figure 2

Eric Barbour’s “1+1 Cascade” amp-block is an AC feedback amplifier (figure 2), consisting of a common cathode gain stage and a cathode follower output stage. This amp-block has a single inverting input and limited open loop gain, but is entirely suitable as a front-end of a headphone amplifier. The open loop figures are less than spectacular: G ~ -50, F_h ~ 30 kHz, THD > 2%, R_out ~ 2K ohms. When configured for a closed loop gain of -10 (R_f = 100K ohms, R_in = 10K ohms), the situation changes dramatically: F_h > 100 kHz, Rout ~ 500 ohms and THD drops below 0.4%. Since the 12AX7 is a dual triode, this design uses only 1 tube per channel. 12AU7s can also be subsituted, but the open loop gain will be lower. Purists may want to add an inverting output stage for correct output phase.

Figure 3

Fred Forssell’s circuit (figure 3) has the differential input stage of a true opamp with a high common-mode rejection ratio (CMRR) and a mu follower output stage (biased at 12mA). The open loop gain is about 510 (30 from the first stage, 17 from the second). When configured for a closed loop gain of 18, the performance approaches that of solid state opamps: THD < 0.1%, R_out= 8 ohms, F_h > 400 kHz and s/n ratio = -86dB. Despite the low output impedance, the load impedance should be 3K ohms or greater to avoid increased distortion. When configuring this opamp for gains of less than 18, Forssell recommends using a lower mu input tube such as the 12AU7A for a lower open loop gain, so that less feedback is required.

Figure 4

Both of the Barbour and Forssell amp-blocks use high voltage supplies, and neither is DC-coupled. Erno Borbely’s hybrid design (figure 4) is both low voltage and DC-coupled. The differential input stage uses a single ECC86/6GM8 dual triode, which has a maximum anode voltage of 25V (a good substitute is the 6DJ8/ECC88). The current mirror Q1 and the constant current diodes (D1A and D1B) increase the CMRR and improve linearity. The output stage is a P-channel MOSFET configured as a common source amplifier with Q3 as its current source (the bias current is 10mA and can be adjusted by varying R_s). R_p is adjusted for 0 output voltage.

C2 provides phase compensation and if the opamp is configured for less than 6dB of gain, the R15-C5 low pass network must be added for stability (for G = 6dB, C5 = 100pF; for G = unity, C5 = 330pF). The open loop characteristics of the Borbely hybrid are excellent, especially for a tube opamp: G ~ 53dB, F_h ~ 90kHz and THD < 1%. When set to a gain of 10 (R_f = 10K ohms, R_in= 1.1K ohms), the specs once again are excellent: F_h > 700 kHz, THD < 0.1% and the output impedance is 50 ohms. A high load impedance (10K ohms) is recommended for maximum voltage output (15V).

CONFIGURING OPAMPS FOR VOLTAGE GAIN

Figure 5

Opamps are most commonly used as voltage gain stages. The basic voltage-gain configurations are shown in figures 5a, 5b. The input impedance is the value of the input resistor. The output impedance Z_o depends on the particular opamp, but generally decreases with decreasing gain (see the opamp datasheet for output impedance specs). If the opamp will be driving headphones directly, the output impedance should be less than 1/10th the headphone impedance across the audio spectrum. When choosing between inverting or non-inverting stage, the goal to keep in mind is that the opamp’s contribution should result in correct phase at the amplifier output. As a rule of thumb, non-inverting configurations tend to have lower noise, higher input impedance and wider bandwidth, but may be subject to certain design constraints (see manufacturer specifications).

Headphone amplifiers are usually fed from the outputs of a preamp or portable stereos which have plenty of voltage gain (instead, they lack the current capability to drive headphones cleanly). If a headphone amplifier has a voltage gain stage, the gain is typically set between 2 and 10. Some opamps sound cleaner at lower gains. The feedback resistor R_f probably should be less than 1M for optimal stability (check manufacturer specs for other feedback network design issues), and lower feedback network impedances (R_f||R_in) result in lower noise.

Figure 5c

Modern opamps do just fine with the basic configurations, but there are many design tweaks that can improve performance. One such optimization reduces source-impedance input errors in JFET-input opamps, which were discussed in the section on selecting opamps above. Recall that source-impedance input errors affect non-inverting gain configurations only and are caused by unequal source impedances at the + and – opamps inputs. The non-inverting amplifier in figure 5c balances the source impedances by choosing R_s = R_f||R. In a headphone amplifier, R_s is likely to be variable, in the form of a volume control, and so the 2K to 3K value is an approximation.

Figure 5d

In a multi-stage opamp system (such as a voltage gain stage followed by a current buffer – see the section on output stages for more information), if the input stage opamp has a bipolar input stage and narrow open-loop bandwidth, it may exhibit nonlinearities when fed high level, high frequency signals. The system in figure 5d has an input stage opamp, which has had its open-loop bandwidth effectively extended under local feedback. The overall gain of the system is 5, but the local gain of the input stage is about 100 for an effective open-loop bandwidth of 100kHz. The bandwidth extension should go well beyond the audio range.

Figure 6

If the opamp is configured for a gain of 1 (R = R_f), it becomes a voltage follower. Most solid state opamps will also function as non-inverting followers with a straight wire in place of the feedback resistor (figure 6). Non-inverting voltage followers have the input impedance and a low output impedance. The input impedance of an inverting follower is the resistance of the input resistor. Voltage followers are often used as buffers which could drive headphones, but voltage gain opamps have modest current capability. A high current buffer opamp is specially designed to provide large amounts of current – perfect for driving headphones. For more information, see the section on output stages below

Handling Balanced Inputs

Figure 7

Pro audio equipment may have balanced inputs and outputs – where the ground is separate from the signal ground for more effective noise shielding. Thus, each channel has a total of 3 connections: signal, signal ground and ground. The circuit in figure 7 converts a balanced input into a single-ended signal with unity gain (the input resistors are split to implement a RF filter – see below). The resistors must be matched to within 0.1% or the CMRR will degrade (e.g., an 80dB CMRR can drop to 60dB due to input resistor mismatch). The converter can also be configured with gain determined by the ratio of R_f / R, but keeping all Rs that same value makes matching the resistor array easier.

AC Coupling and RF Input Filters

Bandwidth-limiting the signal input can block DC voltages or filter out RF noise. DC protection is not necessary if the audio source already has 0 DC output, but some designers prefer extra insurance. With values of 1uF and 100K, the high-pass input filter in figure 5a has a corner frequency of about 1.6Hz and will minimally affect bass response or overall sound quality – if a high quality parts are used (e.g., film capacitors and metal film resistors). Instead of the resistor, an audio taper potentiometer could be substituted to serve as a volume control.

If the signal has RF noise, it can be cleaned up with a low-pass filter at the inputs. The low-pass network in figure 7a has a corner frequency of about 200kHz. An alternative RF network is shown in figure 7b. Frequencies above the corner frequency are mixed together, so that they are canceled out by the opamp’s CMRR. As with the resistor array, the RF capacitors should also be matched as closely as possible. Also use shielded cable when wiring the inputs to further reduce noise pickup.

Figure 8

Low and high pass filters can be cascaded at the input, so long as the resistor values of each filter are different by at least a factor of 10. Also, the impedance of input network will affect the overall impedance of the input stage, so must be accounted for in selecting filter resistance values. These bandpass filters can alternatively be incorporated into the feedback loop. The circuit in figure 8 has an approximate bandpass from 2Hz to 150kHz. Some audiophiles may be able to hear distortion from capacitors in the signal path. As with opamp distortion, capacitor distortion is not audible to everyone. Before discarding the benefits, audition the amp on a protoboard with and without the capacitors.

THE OUTPUT STAGE

Voltage-gain opamps may output enough current to drive some headphones directly (check the manufacturer specs). For example, the author built a pocket headphone amp (shown above) with Burr-Brown OPA132 opamps in a non-inverting configuration as shown in figure 5. The amp has no trouble reaching ear-splitting volumes with most headphones. For more information about this project, see A Pocket Headphone Amplifier. Modern dynamic headphones will play loudly with just a few milliwatts (see Understanding Headphone Power Requirements).

However, when an opamp does not have enough current capability or if it is susceptible to output loading errors (for information on output loading errors, see the section on selecting opamps above), it must then be augmented by an output stage. This section reviews solid state and tube class A followers, class AB symmetric emitter followers and buffer opamps (which are nothing more than elaborate emitter followers) as output stages. For tube amp-block front-ends, there is a discussion on interfacing tubes to solid state output stages. Note: the class A followers and high current buffers described below also make excellent standalone headphone amplifiers, where voltage gain is not necessary.

Class A MOSFET Follower

Figure 9

Purists prize class A amplifiers as capable of reproducing audio signals with the ultimate fidelity, because the output voltage swing is under the control of a single transistor or tube. Class A amplifiers are inefficient, consuming up to 400% more power than they output, but are enjoying revived popularity for their simple topologies (especially single-ended class A amps). Whereas class A loudspeaker amps run hot enough to heat a room, headphone fans can indulge without guilt, since headphones require very little power.

The MOSFET source follower in figure 9a is a single-ended class A output stage. MOSFET followers (and their bipolar cousins, emitter followers) are current amplifiers, which have non-inverting unity (or slightly less than unity) gain. A voltage divider biases the MOSFET. The bias pot Rp adjusts the output voltage to 0V for DC coupling (see Greg Szekeres’ Class A Headphone Driver for an AC coupled design). An input coupling capacitor blocks incoming DC and isolates the MOSFET’s bias network. Rs sets the MOSFET’s bias current I_D. Then Rs = V/I_D.

The MOSFET can be any power MOSFET, so long as the voltage and current ratings are adequate. MOSFETs have the “soft” overload characteristics of vacuum tubes and are preferred in this type of application over bipolars. The gate resistor helps to stabilize the MOSFET. The V_DS spec should be at least twice the idle voltage. If the MOSFET idles at 1/2 the total supply voltage, then V_DS should be at least the value of the total supply or higher. Rs is a power resistor. Starting with an idle current Id of about 100mA and -V = -12V, then R_S = 12/.1 = 120 ohms. The resistor’s power rating should be much greater than 12 * 0.1 = 1.8W (at least 3.6W to be safe). Also make sure that the MOSFET is heatsinked to dissipate a similar amount of power.

The amp in figure 9b (by PRR) adds an opamp gain stage. The design exploits the ability of an opamp to serve as a voltage source. Once Id is set via Rs, the MOSFET gate draws the bias voltage it needs directly from the opamp’s output without a biasing network. (The MOSFET’s device specs will have a graph of gate voltage V_G vs. idling current I_D.) The feedback network (5K and 1K resistors) sets the overall gain to 5 and automatically nulls the output to 0V for DC coupling. That 20pF capacitor rolls off the frequency response above 100kHz to prevent oscillations. With Rs = 100 ohms, the MOSFET idling current I_D is 120mA.

Instead of a resistor for R_S, a precision current source would improve linearity. Current sources are usually made with a transistor, but the version above employs a LM117/317 floating regulator, which needs only one resistor to adjust current output from 10mA to 1.5A. The voltage differential between V_in and V_out (which is the 1.25V internal reference voltage) should be between 7 and 15V. At higher differentials, the current output starts dropping due to internal safe-area protection, in which case more than one current source can be paralleled for higher output. While not required, the output capacitor helps eliminate any instability. Again, heatsinking is recommended.

AC-Coupled Cathode Follower

Figure 10

From a design point of view, tubes are less favored as output stages, because the output impedance is higher than can usually be achieved with transistors. Yet, there are many excellent headphone amps with tube outputs. The AC-coupled cathode follower in figure 10 (from Andrea Ciuffoli’s headphone amp project) achieves a relatively low output resistance of about 33 ohms by paralleling two sections of a dual triode. The cathode resistor is tapped to provide self-bias. Each section is biased at 26mA or 52mA total.

The output impedance of a single-tube cathode follower is calculated as: Z_out = R_k / (1 + G_mR_k), where G_m is the tube’s transconductance and R_k is the total resistance of the cathode resistors. Therefore, when building a cathode follower, select tubes with a high transconductance to get the lowest output impedance.

Class B and AB Symmetric Emitter Followers

Figure 11

The high power consumption of class A amplifiers makes them impractical in battery-powered headphone amplifiers. The current booster circuits in figure 11 have complementary output devices that each reproduce one half of the audio signal. These schemes are more efficient because the idle current can be very low or even 0mA. The circuit in figure 11a is a class B amplifier with Q1 and Q2 off at turned off at idle. When the audio signal is positive, Q1 conducts; when it is negative, Q2 conducts. However, both transistors conduct only when the signal exceeds the forward bias voltages which is around 0.7V. Therefore, both transistors remain off when the audio signal is between ±0.7V, resulting in crossover distortion at the output. Since headphones are driven at low output voltages, this type of distortion is particularly noticeable in a headphone amplifier.

The circuit in figure 11b improves performance by allow the opamp stage to supply current until the voltage drop across R is large enough to forward bias both transistors. However, this design suffers from fluctuating output impedance. The output stage in figure 11c solves both problems by having both transistors conducting at very low idle currents. The voltage drop across the two diodes forward biases Q1 and Q2; the emitter resistors determine the idle current – about 0.6mA with these values. The output stage operates in class A at low levels – until the load draws more current or voltage swing than one of the transistors can provide. For battery operation, the output stage is often biased from 1-10mA, trading off between sound quality and battery life. The minimum idle current is best determined by monitoring a sine wave output on an oscilloscpe while adjusting the bias until the crossover distortion just disappears. AC powered amplifiers can take advantage of extended class A operation by increasing the bias current. Earle Eaton’s headphone amplifier uses a variation of this design. Sheldon Stokes’ headphone amplifier has a class AB MOSFET output stage.

High Current Buffers

Figure 12

High current buffers are basically output stages on a chip. Because they are specialty products and are meant for use in specific applications, buffers are optimized to be particularly good at one job. In general, these chips have fantastic specs: slew rates in the hundreds, low distortion and of course, high current capability. For a headphone amplifier, a buffer that can output 100mA is probably more than sufficient, but additional current drive doesn’t hurt, so long as the power supply requirements meet the builder’s goals. Figure 12a shows a voltage gain opamp with its current capability doubled with the help of an identical opamp configured as a voltage follower. The load balancing resistors ( R_c ) are about 50 ohms. The output impedance of this circuit would be R_c || R_c || ( R + R_f ), but the impedance seen by the headphones is much less – reduced by the effect of the feedback taken at the outputs of the combined R_cs: Z_out = R_out / amount of feedback.

Figures 12b and 12c show voltage gain opamps augmented with current buffers – 12b has a buffer outside the feedback loop and 12c has buffers inside. The overall gain for both versions is the same, but the version with global feedback might function with greater linearity. However, some designers argue that these buffers are already very linear, and global feedback can introduce instability into a system. Both configurations work. If the circuit of figure 12c is wired for local feedback only, such as in figure 12b, then the load balancing resistors can be as little as 1 ohm for a lower output impedance. When using dual or quad buffer ICs, global feedback can help correct for output loading errors (see the sections on selecting opamps and configuring opamp voltage gain stages for more information about output loading errors).

Note: Class A output stages can similarly be excluded from the feedback loop, but class AB stages should be included, since they are more prone to nonlinear operation.

In the case where a single buffer does not supply enough current or has an output impedance that is too high, it is possible to parallel output buffers. Figure 12c doubles output current capability and cuts output impedance in half by paralleling 2 output buffers. The current-summing output resistors R_c (typically 50 ohms) ensure that all of the buffers contribute equally to the output. Again, because the feedback is taken after the R_cs, the output impedance seen by the headphones is less than 1 ohm. Ben Duncan’s PHONES-01 headphone amplifier substitutes ferrite beads and incandescent lamps (see below) for the output resistors, reasoning that any unequal sharing is likely to be in the RF range. The beads also help block RF. Again, the feedback loop can be placed either before or after the parallel buffers.

Interfacing Tubes To Solid State Output Stages

When interfacing tubes with solid state output stages, the higher operating voltages of tubes pose two potential problems. First, the power supply may have to be “stepped down” and second, tube circuits can send out high voltage transients that could damage solid state components. The solutions: use high voltage opamps and buffers and/or limit the voltage going into solid state inputs. With a high-voltage MOSFET, the class A source follower described above would interface well with tube gain stages, as tubes and MOSFETs have similar sonic characteristics. The MOSFET amp has zener protection against overvoltage damage. There are also high-voltage bipolar devices, but they are less common. Apex Microtechnologies and Burr Brown are two manufacturers of high voltage opamps and buffers. Many of these are well-suited for audio applications, and a few chips are able thrive on power supplies of up to ±600V.

Figure 13

High voltage output stages may also have input voltage limitations that tubes could breach. The following are two overvoltage protection schemes that can be used with any solid state output stage. Figure 13a is a suggestion by Eric Barbour. When fed high voltage transients, the zeners clamp the input to a maximum of ±15V. Figure 13b is the protection scheme that Greg Szekeres uses in his MOSFET headphone driver. Here, transients in excess of the power supply voltages will forward bias the silicon diodes and be conducted out of the system. The input resistor sets the minimum load impedance seen by the tube output.

Output Current Limiting

Figure 14

When a headphone plug is inserted or removed from the jack, the possibility arises that the amplifier outputs will be shorted, if only briefly. Without current limiting, such a short could burn out opamps and/or output stage transistors. Rather than resort to complex current sensing schemes, figure 14 shows two common limiting mechanisms that protect against short-circuit damage: current limiting resistors and incandescent bulbs. Current limiting resistors set the minimum load that the amplifier can see – typically 100 ohms, 1/2W. Output resistors will reduce the output power and increase the amplifier’s output resistance, but most headphones will be unaffected. Another option is to locate the current limiting resistors inside the feedback loop (figure 12) so that the effective output impedance of the amplifier is minimized from the feedback. See Headphone FAQs for more information about the impact of amplifier output impedance on headphone sound.

In place of a current limiting resistor, an incandescent lamp has the advantage of very low resistance when the filament is cold. Lamp filaments have a positive temperature coefficient. As increasing current heats the filament, the resistance also goes up, thereby reducing the output current. Choose lamps with voltage and current characteristics similar to that of the output stage. Incandescent lamps were once popularly deployed to protect loudspeakers from overdrive. The idea resurfaced as output limiting for headphone amplifiers in Ben Duncan’s PHONES-01 headphone amplifier project.

EQUALIZATION

Figure 15

Designing an equalization stage is an entire subject by itself (see Designing a Pocket Equalizer for Headphones). Equalization can be implemented in separate circuit blocks – either as active stages or passive networks – to ensure that they can be switched out completely without compromising the quality of the main gain stage. But there are instances where equalization is so important and basic to the use of the amplifier that the EQ filter network is incorporated in the feedback loop of the main gain stage for convenience and economy. For example, headphone amplifiers for guitar practice almost always require a bass boost.

Figure 15 shows a bass boost feedback network by T. Giesberts that gives a 10dB boost at 50 Hz when turned on. The network is a shelving EQ. With the boost deactivated, R1-C1 and R2-R3-C2 form a bandpass with threshold frequencies of about 20Hz and 30kHz. The gain of the amplifier is determined by (R2 || R3)/R1 and is approximately 4 with the values shown. With the boost switched in, R3-C3 create a bass shelf, with a threshold frequency of about 500Hz. The downturn in the low frequency response below 50Hz is caused by the attenuation from the input high pass filter.

ACOUSTIC SIMULATION

Figure 16

Headphone sound suffers from a “super-stereo” effect caused by the isolation of each audio channel to one ear. Acoustic simulators electronically alter the stereo signal to create a more natural soundfield in headphones. They may be implemented with digital or analog filters (also called crossfeed filters). While digital and active analog simulators have amplification for headphones built into the design, passive simulators are RC networks that shape and time delay the crossfeed. Passive networks are sensitive to the source and load impedances that can affect the frequency response of the networks. (For examples of passive acoustic simulators, see the HeadWize Projects Library. For more information about digital and active network simulators, see Technologies for Surround Sound Presentation in Headphones.)

Depending on the input and output impedances of a passive simulator, it can appear at the input or output of a headphone amplifier (figures 16a, 16b), but isolating the network between two amplifier stages will often result in the best performance (figure 16c). With two isolating stages, the network can be assured of seeing a low input source impedance and a high output load impedance, such that the frequency response of the network remains constant. Both stages can be voltage gain blocks and/or unity-gain buffers, as the application may require.

However, with battery-powered amplifiers, which may operate the opamps at lower voltages, the preferred way to construct a headphone amplifier with an acoustic simulation is to make the second stage a voltage gain block to compensate for any insertion loss through the network as well as provide for overall voltage gain. If the voltage-gain block does not output sufficient current to drive headphones, add a high current, unity-gain buffer after the voltage-gain block.

When using multiple opamp gain stages, be sure to check the idle voltage at the output of the last stage. If it is more than a few millivolts, the DC-offset voltages of the opamps must be adjusted – either by trimming the DC offsets, by adding capacitors between stages and at the output to block offsets or by selecting feedback resistors to minimize offsets (see next section).

ADJUSTING DC-OFFSET VOLTAGES IN MULTI-STAGE AMPLIFIERS

Figure 17

In single stage amplifiers, the opamp’s DC offset voltage is only a few milliamps and is rarely a problem. In multistage amplifiers, DC offsets may be amplified by successive stages until the idle voltage at the output of the final stage reaches several volts, although the overall gain the system may not be very high. Jan Meier experienced this situation while building and testing a headphone amplifier:

Referring to figure 17a, the non-inverted input of an opamp wired as a voltage follower requires a small input bias current (i+) that, since it flows through the resistor R1, generates a non-zero voltage V+ = (i+)*R1 at the input. Typical values for i+ are 1uA to 2uA (LM6171/LM6181/LTC1206) for bipolar-input, or 1 to 50 pA (OPA627/OPA604) for FET-input amplifiers. With a R1 of 100K, V+ (and thus Vout) can have values up to 200 millivolts!

In figure 17b, a feedback loop is added that amplifies V+ by a factor (R3+R2)/R2. It is not unusual for a headphone amplifier to have a gain factor of around 5. This will, however, also amplify V+ for a Vout of up to 1000mV, which can damage headphones – especially low impedance headphones. Fortunately the inverted input also generates a bias current (i-) that generates a DC-voltage (V-) at the inverted terminal and thus counteracts the effects of V+. The effective resistance to ground seen by the inverted input is the value of R2 and R3 in parallel which equals (R2R3)/(R2+R3). To eliminate the output voltage offset generated by i+, the input voltage V- should be equal to V+:

(i+)R1 = (i-)*(R2R3)/(R2+R3)To select values for R2 and R3, first take a look at the specifications of the opamp for i+ and i-. Note that they do not have to have the same value. For instance, the LTC1206 has an i+ value of 2uA whereas i- goes up to 10uA! By a proper selection of the resistor values, the offset can be strongly reduced. With a headphone amplifier made from a LM6171 opamp and having R1 = 47 kOhm, R2 = 56 kOhm, R3 = 300 kOhm, one channel shows a very good offset of only 20 mV. The other channel came down to a hardly measurable 0.2 mV! The fact that the channels were not equal simply has to do with manufacturing variations in opamps of the same type.

A problem remains with the input stage. If the input potentiometer is directly coupled to the opamp, the value of R1 now changes with the volume control, and a perfect fit of the resistances can not be made. A possible solution is shown in figure 17c. The resistance of the potentiometer no longer has an influence on the DC-resistance of the opamp. Alternatively, if the headphone amplifier has a second stage, the input stage can be decoupled from the second stage as shown in figure 17d. If the output of a first stage is directly coupled to the input of a second stage, the effective value of R1 is zero and a match can not be made. However, you simply can put a resistor between output and input.

Last warning: If the headphone amplifier will also be a preamplifier, any DC-offset at the output will be amplified by the power amp and will be fed into a low resistance loudspeaker. In this situation a few millivolts offset can damage the loudspeaker. To prevent any damage to loudspeakers or to the power amp, always use (decent quality) capacitors at the output of the preamp.

HEADPHONE DISTRIBUTION AMPLIFIERS

Figure 18a

Multitrack recording allows musicians to record songs in layers. Tracks can be added or overdubbed. Musicians may be positioned far apart from each other or play at different times to isolate their performances for the greatest flexibility in editing. Headphone monitoring is the most common way for musicians to hear each other under these circumstances, and a headphone distribution amplifier is central to this function.

Headphone distribution amplifiers can drive several pairs of headphones from a single set of inputs. While it is fairly easy to build one from a power amplifier with a ladder of output resistors (see the headphone FAQs for instructions), there are advantages to driving each headphone from its own amplifier, such as greater control over gain. The first stage of the basic distribution amplifier shown in figure 18a is a voltage follower that provides impedance buffering and signal inversion for correct phase at the headphone output. The buffer feeds any number of headphone amplifier blocks with their own volume controls.

Figure 18b

As more musicians demand custom mixes, so commercial distribution amplifiers have begun adding mixer features. Figure 18b shows how to convert the input buffer stage of the basic distribution amplifier into a mixer stage. The input buffers (A1 and A2 for the left and right channels) now have a series of 100K summing resistors, one resistor for each stereo or mono input. The level controls for the stereo and mono inputs are balance-volume and pan-volume sets of pots. To move the balance or panning characteristic closer to the ends of the pot rotation, decrease the value of Ri.

Figure 18c

A full-featured headphone distribution amplifier will have limiters and possibly equalization stages for each headphone output. An acoustic simulation network, equalizer and/or limiter can be placed between the buffer and headphone amplifier blocks (see Designing a Limiter for Headphone Amplifiers for information about limiters and Designing a Pocket Equalizer for Headphones for equalization schematics). To increase the drive capability of an amp block, add a current buffer output stage (after any active EQ stages).

POWER SUPPLIES

AC Power Supplies

Figure 19

To regulate or not to regulate, the answer depends on the circuit. Modern opamps have excellent power supply rejection ratios (PSRR) and are less affected by voltage fluctuations than older products, but discrete output stages may be more vulnerable. Since headphone amplifiers draw so little power and 3-pin regulators are cheap, it cannot hurt to have a regulated supply. Figure 19 shows a dual supply with the LM150/LM133 floating regulators configured for slow-start to minimize turn-on thumps. The delay is R*C = 8900 * 1000 E-6 = 9 seconds. Also check for power supply schematics in HeadWize Projects articles or in the datasheets for regulators. In any case, each opamp should be decoupled from the power supply with a 0.1uF ceramic capacitor and possibly a 10uF electrolytic, connected from the power supply pins to ground (see figure 14e below).

Figure 20

Tube circuits often do not use regulated supplies, but where recommended, it is usually a single high voltage regulated supply for a gain stage and/or a low voltage regulated supply for tube heaters. For example, the Forssell tube opamp requires a regulated +350V supply for the output stage (at least 30mA for two opamps) and a 6V regulated supply for the heaters to minimize hum. Floating regulators, such as the LM150, can output hundreds of volts, so long as the input/output differential voltage remains within spec (and don’t forget to diode protect the regulator). Figure 20 shows a simple zener-based high voltage regulator. If the output voltage goes up, the potential across V_GS decreases and the MOSFET reduces output current. If the output voltage goes down, then V_GS increases, and output current increases. The IRF420 specified has a V_DS of 450V and an I_D of 2A continuous and must be mounted on a heatsink. The zener can be any series of 5W zeners (for the Forssell circuit) that total about 350V.

Battery Supplies

Figure 21

There are opamps that will operate on a single 1.5V cell. Such micro-power opamps can drive very efficient, low-impedance headphones. With other headphones, the inability of micro-power opamps to develop higher voltages across the load will limit the volume. One method of getting higher voltages from batteries is to stack the batteries in series. Another is to raise the battery voltage with a DC-to-DC converter (to several volts or even several hundred volts in the case of portable electrostatic headphones). DC-to-DC converters, also called switching regulators, do their magic by changing the DC voltage to an AC voltage via an oscillator, which feeds a step-up transformer or capacitive/inductive reservoir to build the voltage, and then is converted back to DC at the higher voltage. As with any AC-based supply, a DC-to-DC supply must have a good filter network at the output to minimize power supply noise.

Opamps can run off single supplies, but are designed for dual supplies. Headphone amps with direct-coupled outputs must be powered from dual supplies. If there is room in the amplifier enclosure, separate batteries for the positive and negative supplies is the suggested implementation (figures 21b). If the opamp can run on a supply of ±3V or less, a single 9V battery can be converted into a dual supply as shown in figure 21a. A voltage divider creates a virtual ground at the center junction and draws less than 1 ma. at idle. The electrolytic ouput capacitors both reduce the supply impedance at high frequencies and function as a power reservoir to simulate two separate battery sources.

This version of a virtually grounded supply works best with amplifiers that draw lower idling current, as the capacitors must be able to “recharge” quickly. Start with 100uF capacitors. With an oscilloscope or a multimeter, monitor the supply for any ringing or fluctuation with the amplifier driving headphones loudly. Increase the capacitance or decrease the resistor values of the voltage divider to compensate (decreasing the voltage divider resistance values will increase idle current). The most important test of all is the listening test. Despite supply fluctations, the amplifier may function without audible detriment. For a simple AC power supply, apply this circuit to an adapter with 12V regulated output (Radio Shack sells one) to get regulated ±6V (figure 21c). Regardless of the supply option used, decoupling the opamps from the supply (figure 21d) will improve stability.

Figure 22

There may be times, though, when a virtually-grounded dual supply has a tendency to “rail” when the resistor-type voltage divider cannot maintain the virtual ground at 1/2 Vcc. Such cases may occur when the opamp draws too much current or input signal (for example, a high boost equalizer) pushes the opamp into heavy clipping and power supply is unable to recover. There are several inexpensive commercial voltage references that can output a stable 1/2 Vcc regardless of the load. Figure 22 shows a virtually-grounded dual supply implemented with the Texas Instruments TLE2426 voltage reference and one 9V battery (the TLE2426 is excellent also with two 9V batteries for a stable, dual 9V supply). Before selecting a voltage reference, check the specifications to see if it has adequate current capability for the load.

Addendum

11/11/98: Expanded discussion of using opamps to drive headphones directly without an output stage. Also updated figures 21c, 21d – lowered voltage divider resistor values to 5K, based on experimentation.

11/13/98: Added image and description of pocket headphone amp.

5/10/99: Added the following new sections: Equalization, Acoustic Simulation, Adjusting Opamp DC-Offsets. Also, minor revisions in other sections.

7/19/99: Added section on headphone distribution amplifiers.

7/20/99: Updated section on headphone distribution amplifiers.

7/24/99: Updated section on headphone distribution amplifiers.

1/3/00: Updated the following sections: Selecting Solid State Opamps, Configuring Opamps for Voltage Gain, The Output Stage.

1/24/00: Corrected reversed opamp input connections in figure 18a and 18b. Also corrected calculation of Rs for MOSFET driver in figure 9.

4/10/00: Revised figure 16c and corresponding description.

12/10/07: Revised figure 9a. Added figure 9b and corresponding description. Thanks to PRR for the circuit.

References:
___, Electronic Circuit Design Handbook, 1971.
___, National Semiconductor Audio/Radio Handbook (1986).
___, “Applications For An Adjustable IC Power Regulator,” National Semiconductor, January 1997.
___, “BUF634 Application Notes,” Burr-Brown, 1993.
___, “LF353 Application Notes,” Texas Instruments, 1993.
___, “NE5532 Application Notes,” Philips Semiconductor, April 1996.
___, “MLM82 Mixer Schematic,” Rane Corporation, May 1998.
___, “LM150 Application Notes,” National Semiconductor, May 1998.
___, “LM358 Datasheet,” National Semiconductor, Sept. 1999.
___, “LM6171 Datasheet,” National Semiconductor, May 1998.
Banks, Morgan, “AB-051 Double The Output Current To A Load With The OPA2604 Audio Opamp”, Burr-Brown, 1993.
Barbour, Eric, “Vacuum Tube AC Feedback Amplifiers,” Glass Audio, 2/92.
Barbour, Eric, “1 + 1 Cascade Amplifier,” Glass Audio, 6/92.
Borbely, Erno, “Low-Voltage Tube-MOSFET Line Amp,” 1/98, p. 1.
Danner, Denzil G., “Tube Circuit Primer: A Reference,” Glass Audio, 3/92, p. 20.
Duncan, Ben, “How Clean Is Your Opamp,” Electronics World + Wireless World, January 1993, p. 42.
Duncan, Ben, “Phone Home,” Hi-Fi News and Record Review, May 1997, p. 68.
Duval, Bernard, “IRF150 Hybrid Amp,” LED, Winter 1996.
Fox, Harry, Master OpAmp Application Handbook, 1978.
Forssell, Fred, “A Vacuum Tube Op Amp,” Glass Audio, 1/95, p. 1.
Giesberts, T., “Headphones Amplifier for Guitarists,” Elektor Electronics, 11/96, p. 32.
Horowitz, P. and Hill, W. The Art of Electronics (1989).
Jung, W., Audio IC OpAmp Applications (1988).
Jung, W., “Composite Line Driver With Low Distortion,” Electronic Design, 6/24/96, p. 78.
Jung, W., Minimizing Input Errors, Electronic Design, 12/14/98.
Jung, W., Realising High Performance: Bandwidth Limitations, Electronic Design, 12/1/98.
Jung, W., Realising High Performance: Buffers (Part 1), Electronic Design, 9/1/98.
Jung, W., Realising High Performance: Buffers (Part 2), Electronic Design, 10/1/98.
Pass, Nelson, “DIY Opamps,” Audio Electronics, 6/98.
Pass, Nelson, “The Bride of Zen,” Audio Amateur, 4/94.
Pass, Nelson, “The Pass Zen Amplifier,” Audio Amateur, 2/94.
Pass, Nelson, “Zen Revisited,” Audio Amateur, 3/94.
Van Alstine, Frank, “The AVA Versa-Kit Headphone Amplifier Do-it yourself Project,” Audio By Van Alstine Website, 9/97.

Also thanks to Richard Steven Walz for his insight on virtually-grounded dual power supplies.

c. 2001 Chu Moy.