"WHAMMY" Pass DIY headphone amp guide


Joined 2010
Paid Member
EDIT DEC 2021 - There are a huge amount of lost photos in this thread, please go to this link to see the new guide:


Yes, it's a silly name. But it's cute. And it explains a lot about it. So it stuck.

Video of Wayne talking about the Whammy at BAF 2017 -- Wayne Colburn at BAF '17

PCB available here - WHAMMY Headphone Amplifier – diyAudio Store

Given the explosion in the popularity of headphones, as well as the unbelievable amount of choice and variety in the market, it seemed that a simple and great-sounding headphone amp would be a great addition to the DIY community.

This one will drive any headphone you want to throw at it. Wayne wanted a headphone amp for his desk. This is the fruits of that idea. It’s made to be made in an afternoon or an evening and it has no adjustments, it’s going to have a high likleyhood of success for the builders. :)

It's an all in one PCB, just wire the AC & fuse, input and output jacks. Add a selector switch if you like.

Class-A output stage with enough current to drive anything. It also makes a good linestage with about 14db of gain.


Transformer 15VA or 25VA 15+15 to 22+22

25VA 22V+22V is best and used in this guide.

Currently the transformers that fit the board are available at Digikey
Amgis 6663, 6664 / Amveco 70053

Here's a link to a factory surplus transformer that fit the PCB will work beautifully - You will want to make R16, R22, R29, R32 15ohm if using this transformer.

70054K PC Mount Transformer 110/230V-18/36V 110V-36V 110V-18V 230V-18V 230V-36V

There are pads on the PCB for a non-PC mount transformer if you have room in your chassis. Something like an Antek AN-0220 or AN-0222 would work well. AN-0220 - 25VA 20V Transformer - AnTek Products Corp

If you use smaller transformer you may need to adjust the bias down a bit. It will still be pure class-A for 99.999999999% of all headphone listening.

The bridge is made from 1N4004 or use high speed diode if you like. Snubber capacitor C20 0.22uF 250v X-rated.

AC filtering is done in a big and effective way, utilizing a CRCRC filter with 3300uF capacitors and 5.1ohm resistors. You can use smaller resistors and caps if you like, it's a very effective filter and will work well with even 1/2 the values.

The regulators using 7815/7915 can be elevated a bit, using a red or green LED as the reference if you wish. Don’t use blue, they are noisy and they will set the regulators to too high of a voltage.


On the input there is a dual opamp used for voltage gain. We’ve tried these with great success LM833, Muses 8820, RC4580, OPA2604, AD823, TL072.

If you want to try a different opamp, try something made for audio. Feel free to try some surface-mount opamps in a DIP adaptor if you like, there are lots of neat opamps to try.

Gain is set by R8/R12. Lower gain, make R12 bigger, unity gain, R12=10K

Potentiometer - Alps RK27 fits the PCB, feel free to use what you like. If you have room in your chassis, this is a fine place for a stepped attenuator.

After the opamp there is a Mosfet source follower for current capibility, and the feedback loop includeds the opamp. This keeps the DC offset stable as well as lowers the output impedance to a very low level, less than 1/10 of an ohm.

Output stage

The output stage is a Mosfet NP pair in source-follower configuration. Being a follower it can add no voltage to the signal but can contribute lots of current. Since the opamp os being used for gain this is not problem. It also has the advantage of adding very little sonic flavor to the signal, it's esentally transparent. The output stage is simple, powerful, will drive anything, and is self adjusting due to optocoupler and the opamp controls DC offset because the output stage is in the feedback loop. No potentiometers to adjust or voltages to read when biasing.

Output impedance is less than 1/10 ohm

The following Mosfets work well in the circuit.

Toshiba Mosfet 2SK2013 / 2SJ313
Fairchild Mosfet On Semiconductor Fairchild FQP3N30 / FQP3P20
IRF Mosfet Vishay IRF610PBF / IRF9610PBF

No matching is required.

Bias arrangement (low offset due to opamp precision)

(4) 10K resistors make a voltage divider to give lots of bias voltage to the gates; this bias will be 1/2 the rail voltage. Assuming a standard build with 17v rails this will give the Mosfets a maximum of 8.5v of bias. With no other controls his would make the output devices conduct like there’s no tomorrow, and probably let the smoke out, but the 4N35 optocoupler helps control and set the necessary bias voltage. With this it happily operates in Class-A all the time.

The 4N35 optocoupler does a few things -

The optocoupler has two sides when looking at the schematic, the diode and the transistor. They are linked optically, not electrically, so the two sides of the optocoupler can share different voltages that don't effect the other side. as the current change in the LED side of the opto it will glow brighter or dimmer, which controls how much the transistor side conducts - in this circuit the current through the LED is directly equal to the mosfet current, and as it gets brighter it controls the BJT, whereby the BJT "burns up" the excess dc bias voltage.

The optocoupler appears to be a variable resistor in parallel with the inside 10k resistors - it changes the gate bias with the collector-emitter junction as the opto coupler looks at the current through the mosfet sources. The LED part of the opto has a 1.2v constant drop, this is used in conjunction with R18 to set bias current across the source resistors. If the current is too high it will make the LED brighter, that modulates the base of the transistor, and the collector-emitter junction will decrease its apparent resistance in parallel with the inside 10k resistors, changing the ratio of rail voltage to ground, decreasing the amount of voltage on the gates, and keeping the bias stable as the load swings.

Is a simple solution - its a single part and it automatically adjusts. If there is any drift the optocoupler will compensate immediately No resistors to measure across and potentiometers to adjust

Output bias

The diode in the 4N35 gives voltage across source resistors to set current, a 1.2V reference.
Total source R is added, so 10R resistors is 1.2V/20R=60mA
Want more bias? make the resistors smaller. 4.7R = 120mA
Last edited:
  • Like
Reactions: 1 user


Joined 2010
Paid Member
The WHAMMY headphone amp


WHAMMY shown here in the Hammond enclosure and with my very favorite headphones, the Denon AH-D2000

How to use this guide - READ THE CAPTIONS BELOW EACH PHOTO! That’s where the good info is.:D :D :D

Also you can click on any photo to view in full size.


The PCB. This is a rev.3, the latest version, identifiable by the pads near the pot for the 2.2k resistors. Production versions in the diyAudio store may be very slightly different, with slightly different silkscreen and probably in a different color.




Wiring the PCB


The transformer have dual primaries and can be used for 120VAC or 240VAC mains. There is room for jumpers on the PCB and they must be connected properly depending on your mains voltage. Instructions for the jumpers are written plainly on the PCB.


120VAC jumpers


240VAC wiring


With the Hammond enclosure you can see that things get pretty tight. I used a small IEC inlet with integral fuse. Note that the AC Live should be the lead that is fused, and some IEC sockets have fuses on the Live and Neutral, check with your country’s codes to see if having Neutral fused is acceptable. Live should go to the lead winding of the primary which is the AC pad towards the middle of the PCB, and the Neutral to the trailing lead, the AC pad closer to the corner. Safety earth is attached to the chassis via a screw and a star washer. If you add a power switch, switch the Live lead.

Using a slightly longer enclosure greatly simplifies this. Also, my layout isn’t ideal — I’d suggest placing the IEC closer to the center of the back panel to it’s not smashed against the transformer.


You can (sort of) see the screw and star washer used for safety earth in this photo. The orange capacitor is attached to safety earth on one side and the ground tabs of the RCA jacks on the other. The RCA jacks are isolated from the chassis with their shoulder washers.

Building and testing the PSU

Making sure the PSU is behaving properly first is an integral part of a successful build. So it will be populated first.


Print out the Schematic and have it in front of you at all times when soldering.

This PSU requires some choices to be made, so DO NOT populate everything on the schematic. It will not work properly if you do. More about this in a few steps.


When working on any part of this project, print out the schematic and have it in front of you at all times.


Start with the passive filtering - the rectifier diodes, the filter resistors, the grey snubber capacitor, and the tall filter capacitors. Note that the big caps are polarized, and need to be on the PCB in the right direction or they will blow up. The + is noted on the PCB, and the + also has a square pad. + is indicated on the capacitor by the long lead, and also - is shown on the cap’s insulation.


At this time also stuff D7 and D8, they are protection diodes for the regulators and will be used in all configurations.

Regulator configuration PLEASE READ

The PCB has the ability to be wired to use the regulators in their most simple mode, to have a LED voltage reference added, or to have the voltage set by using a resistor voltage divider. This is to give builders a chance to use many different configurations to suit what parts and transformers that may have.

All configurations work beautifully and are very quiet.

I suggest that you build it with the LED voltage reference, and the bulk of the photos in the guide will show this.



Configuration 1 - “Naked Regulators”

This is the simplest configuration, it uses the 78xx and 79xx regulators in their most simple configuration. R9 and R13 are jumpered in order for the regulators to see ground and nothing else is added.

Also note that the flat of the regulator is facing the line on the silkscreen. they are shown here for illustration. Heatsinks will be added to the regulators, do not run these without heatsinks.



( Yes, in this photo D5 is installed backwards…:( )

Configuration 2 - “LED reference”

If you don’t have a specific other reason to do so, build this configuration.

This is the quietist configuration of the three, though not by all that much. The LED is used to elevate the regulator off ground by it’s V fwd, and the cap is used to control the impedance of the diode at high frequency, making the whole circuit quieter. Use a red LED to raise the voltage by about 1.8v more than the regulator, use orange, green or yellow for a few more tenths of a volt. Do not use blue or white, they are 3.3v and very noisy when used as a voltage reference.

If you want to use one of these as a light for the front panel you may, but personally I wouldn’t. Instead just attach a LED and resistor to the V+ and ground output of the raw PSU. (Use a blue one for power… this is a Pass design after all. :) )



Configuration 3 - “Resistor reference”

Using resistors at positions R9,10 and R13, 14 and the capacitor, will set up a voltage divider across the reg setting the output voltage at 21.5 when using the suggested values.



Attach the transformer and the regulators + heatsinks. See below for notes on thermal grease, heatsinks and screws.


Testing the PSU

Plug in the transformer and power it on. The LEDs should light.


Place your meter to DC volts and measure the V+ rail between the ground pad near the pot and the square pad of capacitor C9. (Or both pads of C9, red on square and black on the round) This should read approximately +17V give or take a volt.



The square pad of C9 is a good place to test V+

Measure V- with the black probe on R13 pad and red on the R14 pads as shown. Read -17V plus or minus a volt or two.



Probe position for measuring V-

Why are the voltages not precise? Because there is no load on the regulators at all, and they have no current flowing through them to regulate. This is normal with the rest of the board unstuffed.

If the voltages measure good and there’s no smoke or shorts, move on to the next step! :)
Last edited:


Joined 2010
Paid Member
Stuffing the PCB


Print out the Schematic and have it in front of you at all times when soldering.


Although somewhat invalidated by the PSU being complete, the old advice of stuffing the smallest to the biggest components holds true. Start with the resistors.

Resistors - All resistors have 0.5” lead spacing.


Please align the resistors in the same way - if banded put the brown band on the bottom or the right, and if the value is printed make sure you bend the leads so that the value is on top, and align to they read left to right or bottom to top. This will help troubleshooting later


Capacitors -


Remember that when using polarized capacitors to put the long lead (+) into the square pad.



Optocoupler alignment


The dot on the body of the Optocoupler is placed on the same side as the notch in the drawing on the silkscreen.

Opamp Alignment

The dot on the body of the opamp is placed on the same side as the notch in the drawing on the silkscreen. The dot should be closer to the pair of 220uF capacitors.


Mosfet and regulator -


The heatsinks are not attached electrically to the PCB, you do not need to use insulators if your TO-220 devices have metal tabs. A bit of thermal grease is extremely helpful.

Align the device to he heatsink and snug on the screw, bit do not make it tight yet.



Align all the pins and tabs before soldering. The heatsinks solder in. Then solder the pins. Now tighten the screws. (Not too tight! You can distort the device or break it.)


The procedure is the same for the regulators.


Input and output wiring -

I used shielded mic cable in this build. Other wire will work, use what you have on hand.


Input at RCA jacks.


Input at PCB. In this photo the ground wire is mostly hidden by the jacketed cable.


Output at PCB



Output at headphone jack.





Use this photo as an example of what NOT to do -- please move the IEC inlet to the middle of the back panel to get it away from the transformer.
Last edited:


Joined 2010
Paid Member
WHAMMY measurements

PSU ripple


Ripple after 1st caps.


Ripple after 2nd caps.


Ripple after 3rd caps, before the regulators. The lesson here? The “big and simple” approach of noise filtering using multiple CRC is extremely effective. :)


THD + noise vs. frequency


THD + noise vs. power output


Distortion residual from AP, this is mainly 2nd harmonic, and overall very low.


Frequency response


Noise floor.


Square wave response. Clean, no ringing, very, very small shoulder at the leading edge.
Last edited:


Joined 2010
Paid Member

Here's a rough (but useful) list with some important annotations.

Most part numbers are from Mouser.com (unless they end with "-ND", and then it's Digikey.com)

WHAMMY Bill of Materials


(2) .1UF C13, C23 .1uF 505-MKP2D031001FKA00

(1) .22/250V C20 .22/250V Snubber KEMET R46KN322000M1M 22.5mm leads, radial, 26mmx6mmx15mm, X2 rated polypropylene

(6) 1N4004 D1,D2, D3, D4, D7, D8 1N4004 583-1N4004-T

(2) 1UF_PP C1 C5 1UF_PP Input coupling capacitor 594-2222-416-71005

(2) P Mosfet Q2 Q4 2SJ313 May also use IRF9610, FQP3P20 Fairchild - 512-FQP3P20 Vishay - 844-IRF9610PBF

(2) N Mosfet Q1 Q3 2SK2013 May also use IRF610, FQP3N30 Fairchild - 512-FQP3N30 Vishay - 844-IRF610PBF

No matching required for Mosfets

(2) 4N35 OPTO 1 2 4N35 78-4N35

(6) 22UF/25V C12 C17 C22 C25 C26 C27 22UF/25V Nichicon UES 647-UES1E220MEM Silmic - 555-RFS25V220MF3#5 Panasonic FC - 667-EEA-FC1E220H

(1) U3 7815 7815 positive regulator 863-MC7815CTG Digikey - NJM7815FA-ND

(1) U2 7915 7915 negative regulator 863-MC7915CTG Digikey - NJM7915FA-ND

(2) 100PF C2 C7 100PF Feedback compenstion cap, leave empty unless needed

(6)220UF/50V C3 C4 C8 C9 C10 C28 220UF/50V 647-UKZ1E221MHM 50v parts were specified as there are a million of them at Pass Labs. You can use 25v capacitors with no disadvantages. A lower-voltage capacitor might also be physically smaller, and that would be good, as there's not much room around the opamp socket. Also using a smaller value capacitor (around 10-100uF) in these positions will be fine as the PSU is so incredibly clean by the time it makes it to these caps... :)

(6) 3300UF/35V C6 C11 C14 C15 C18 C19 3300UF/35V 7.5mm lead spacing 18mm max diameter 35.5mm max length 647-UPW1V332MHD

(1) ALPS_POT P1 RK27 688-RK27112A0A16

(1) AMVECO_25VA TR1 Amveco 70053 Amgis L01-6365 Talema 70065K Digi-Key - 1295-1079-ND Digi-Key - TE2261-ND

(2) LED D5 D6 {RED} These are not lights, but voltage references use red 604-WP710A10LID

(1) Socket 8 pin U1 Not needed, unless you want to swap opamps. 575-199308

(1) Opamp 863-LM833NG 595-RC4580IP 584-AD823ANZ

(6) Heatsinks 532-531102B25G

1/4W (or bigger) resistors

Note - The part numbers shown are for the Vishay/Dale RN55D series, a mil-spec resistor with non-magnetic leads, non-RoHS, and 100% thermal derating. Which is to say it's a 1/4W resistor that says 1/8W on the box. And has tinned copper leads. Because they are non-RoHS (I.E., not lead-free) they are cheaper. (And honestly better. Look up "tin whiskers")
They are of fantastic precision, sound great, and the value is written clearly on the body of the resistor. They are far and wide my favorite resistor.

(6) RESQ R1 R2 R7 R10 R12 R14 1K 71-RN55D-F-1.0K

(2) RESQ R39 R40 2.21K 71-RN55D-F-2.21K

(2) RESQ R4 R8 4.75K 71-RN55D-F-4.75K

(4) RESQ R16 R22 R29 R32 10ohm 71-RN55D-F-10

(8) RESQ R6 R15 R23 R24 R25 R28 R33 R34 10K 71-RN55D-F-10k

(4) RESQ R17 R27 R35 R26 47.5 ohm 71-RN55D-F-47.5

(2) RESQ R18 R30 100 ohm 71-RN55D-F-100

(2) RESQ R3 R5 100K 71-RN55D-F-100k

(2) RESQ R9 R13 330 ohm 71-RN55D3300FTR

(4) RESQ R11 R19 R26 R31 499 ohm 71-RN55D-F-499

1/2w (or bigger) resistors

(4) RES_HALF R20 R21 R37 R38 5.1ohm 1.0ohm - 5.1ohm is fine 71-CCF02-J-5.1

1/4" Headphone jack ("D" style mounting, similar to an XLR jack) 568-NJ3FP6C-B / or Panel mount, as shown in above photos 568-NYS212
Last edited:
Great build Guide as always!! I would suggest that you insert an information regarding grounding the potentiometer.
In some cases you can get hum noise on the output and grounding the potentiometer with a piece of wire running from one of the pot chassis screws to signal ground.