Troubleshooting a cMoy Headphone Amplifier

[Sp4wi]

Hello!

I am really new to electronics and my first real project is the cMoy Amp, the Crystal cMoy to be excact.

Here is the schematic:



I have built it as seen on the schematic. The only change i have done to the circuit was using 1000uF instead of 470uF caps in the voltage divider.
According to http://tangentsoft.net/audio/cmoy-tutorial/tweaks.html#caps this should significantly improve the bass handling of the amp.

Now my cMoy looks like this:


First, my measurements:
I am getting -9.5V on pin 4 and +9,5V on pin 8, and that is perfectly fine since i am using a 19V power supply.

Now my problem:
The sound coming out of the amp is really crappy, but only if i am using headphones. While i was writing this post, i plugged the amp´s output in to the aux input on my stereo and the sound was just absolutely great.

When i am using headphones, the voltage on pin 4 goes to about 1.5-2.5 volts, depending on the input signal and the voltage on pin 8 goes up to about 17-18 volts.
As i said, i am a beginner but as far as i can tell this may be an issue with the voltage divider.

I would be really grateful if somebody could help me fix this.

Some additional information that may helps:

the Op-Amp i am using: Burr & Brown OPA2107AP
headphones i have tried: 18R and 38R impedance
wall power supply i am using: 19V at 3.5A
1/4W metal film resistors

Thank you!

[Paul Price]

You are asking too much of the op-amp. Your headphones are a low resistance device and the op amp cannot supply enough output current without distortion and clipping of the signal  to your headphones without the op amp going into  current limiting to protect the op-amp.

[Paul Price]

If you want a simple circuit that can deliver enough power to headphones try a simple power amp, like the LM386, for instance.

The op amp you are using is great for voltage amplification, not great for current amplification.

A headphone works with low voltages at higher currents, sending your circuit's output to another amplifier doesn't load the output of your amplifier and demand anything but a tiny current for this circuit to work well.

[Rufus]

Check the output voltage of both channels with no audio input and no load. It should be less than 10mV. Something wrong with your build if it isn't.

[Sp4wi]

If you want a simple circuit that can deliver enough power to headphones try a simple power amp, like the LM386, for instance.

The op amp you are using is great for voltage amplification, not great for current amplification.

A headphone works with low voltages at higher currents, sending your circuit's output to another amplifier doesn't load the output of your amplifier and demand anything but a tiny current for this circuit to work well.

Thank you for your reply.

I am using the excact same Op-Amp as the person who designed the circuit and it worked in his case.
Here is the link to the original project: http://runawaybrainz.blogspot.co.at/2011/12/mystery-project-update-3-time-vampire.html

[Sp4wi]

Check the output voltage of both channels with no audio input and no load. It should be less than 10mV. Something wrong with your build if it isn't.

Hello and thank you for your reply!

I am getting -150mV, any ideas?

[Rufus]

Check the output voltage of both channels with no audio input and no load. It should be less than 10mV. Something wrong with your build if it isn't.

Hello and thank you for your reply!

I am getting -150mV, any ideas?

Not really. The voltage on the opamp +ve input should be zero, the voltage on the -ve input not more than 1mV away from zero. Check your feedback resistors R4 and R3 are the right value. Are you sure the opamp is good, not something from china via ebay which might not even be an OPA2107AP.

[vlf3]

You need to make your voltage meter measurement from the GND with the negative lead, and the positive lead tests each output pin. There should be next to no voltage difference reading +/-... if you do have more than 10 mV then the feed-back resistors R 4/8 & 3/7 may be miss-matched as a pair ? also, make sure the power supply from the GND reference, shows a reading equal for a Positive and the Negative readings... this also takes us to look at the half-rail resistors RP 1 & 2, these values must be the same 4k7 resistance.

Otherwise, the only way you would get this headphone amplifier to work on low impedance headphones, is to place the positive end of two 47 or 100 Mfd electrolytic caps, to each of the op-amp output pins, the negative end to your headphones L/R... the headphones he used are AKG 62 ohm, even so those series R 5 & 9 would be required.

However, R 5 & 9 looks like there 120 ohm ! not sure colour-wise, and if you have-not fitted them ? your placing a low ohm short to the Gnd of the supply, via your headphones... anyway, R 5 & 9 are in the wrong position relative to the feed-back resistor network R 4/8 & 3/7, each end of the resistor R 4 & 8 should be directly connected to the output pins.

So I can't see how it would work undistorted without R 5 & 9 with a value around 560 ohm in series with your headphones, either 8 or 32 ohm; the design is poor, with minimal parts used. The specification for that op-amp under test conditions, rate the output at (RL = 600?, VO = –12.8V to +12.5V)... at the end of the day, this is just a voltage gain pre-amp, not a low impedance headphone driver.

If this will work at all with the current build you have, then placing electrolytic's in series with each output pin, will isolate the op-amp output from the DC volts half supply; and only allow the AC signal to energise the headphones; having said that, it will be a low audio output, as already stated... in reality, this is only a pre-amp and would be suitable for driving into a main amplifier via an auxiliary input.

Op-amp data sheet here:

http://pdf1.alldatasheet.com/datasheet-pdf/view/82142/BURR-BROWN/OPA2132P.html

[Sp4wi]

Check the output voltage of both channels with no audio input and no load. It should be less than 10mV. Something wrong with your build if it isn't.

Hello and thank you for your reply!

I am getting -150mV, any ideas?

Not really. The voltage on the opamp +ve input should be zero, the voltage on the -ve input not more than 1mV away from zero. Check your feedback resistors R4 and R3 are the right value. Are you sure the opamp is good, not something from china via ebay which might not even be an OPA2107AP.

Arrgh you got me, this is the one i used: http://www.ebay.com/itm/330632044444?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649

I always had good experience with parts from china, these chinese sellers always seem really dedicated to what they do.
It´s hard to believe that it is fake because i put my cMoy in between my audio source and Pro-Ject Headbox S amplifier and it sounded as good as always.
If my Op-Amp was something like a relabelled LM358 or LM386 it just wouldn´t sound that good.

I ordered an OPA2132PA from a german retailer now though to test this.

[Sp4wi]

You need to make your voltage meter measurement from the GND with the negative lead, and the positive lead tests each output pin. There should be next to no voltage difference reading +/-... if you do have more than 10 mV then the feed-back resistors R 4/8 & 3/7 may be miss-matched as a pair ? also, make sure the power supply from the GND reference, shows a reading equal for a Positive and the Negative readings... this also takes us to look at the half-rail resistors RP 1 & 2, these values must be the same 4k7 resistance.

Otherwise, the only way you would get this headphone amplifier to work on low impedance headphones, is to place the positive end of two 47 or 100 Mfd electrolytic caps, to each of the op-amp output pins, the negative end to your headphones L/R... the headphones he used are AKG 62 ohm, even so those series R 5 & 9 would be required.

However, R 5 & 9 looks like there 120 ohm ! not sure colour-wise, and if you have-not fitted them ? your placing a low ohm short to the Gnd of the supply, via your headphones... anyway, R 5 & 9 are in the wrong position relative to the feed-back resistor network R 4/8 & 3/7, each end of the resistor R 4 & 8 should be directly connected to the output pins.

So I can't see how it would work undistorted without R 5 & 9 with a value around 560 ohm in series with your headphones, either 8 or 32 ohm; the design is poor, with minimal parts used. The specification for that op-amp under test conditions, rate the output at (RL = 600?, VO = –12.8V to +12.5V)... at the end of the day, this is just a voltage gain pre-amp, not a low impedance headphone driver.

If this will work at all with the current build you have, then placing electrolytic's in series with each output pin, will isolate the op-amp output from the DC volts half supply; and only allow the AC signal to energise the headphones; having said that, it will be a low audio output, as already stated... in reality, this is only a pre-amp and would be suitable for driving into a main amplifier via an auxiliary input.

Op-amp data sheet here:

http://pdf1.alldatasheet.com/datasheet-pdf/view/82142/BURR-BROWN/OPA2132P.html

Thank you a lot, this could be really helpful.
Regardless, i will wait until i get my breadboard (yes, i do not have a breadboard yet) to test all this.

[Andy Watson]

When i am using headphones, the voltage on pin 4 goes to about 1.5-2.5 volts, depending on the input signal and the voltage on pin 8 goes up to about 17-18 volts.
As i said, i am a beginner but as far as i can tell this may be an issue with the voltage divider.

Yes, it's the voltage divider, because:

RP1 and RP2 together form the half-way potential at an impedance of 2.35k. When you plug your stereo in it probably has an input impedance of the order of 10k, and it's probably A.C. coupled as well, so the amp works as expected. When you plug the headphones in, they have a low impedance and are D.C. coupled, so you have the output of the op-amps driving through a low impedance, D.C. path into the GND. The GND is also the same point from which you have biased your op-amps, so you have created positive feedback loop at D.C. - the op-amps ramp up or down to a rail and stay there.

You need to derive an op-amp bias that is independent of the output signal.
Edit: and/or A.C. couple the output, or define a significantly lower impedance GND reference at D.C.