Curiously Strong CMOY Headphone Amp Review

[amirm]

This is a review and measurements of the "biosciencegeek" CMOY headphone amplifier. The CMOY is named after its designer, Pow Chu Moy. It is built out of myriad of IC (dual) opamps with the circuit so simple that just about anyone can put one together. I probably built mine 20 years ago and helped my son build one about 8 years ago. I remember being shocked how much better it sounded compared to the native output of my laptop.

The specific version loaned kindly to me by a member was prebuilt in a beautiful Altoids box (trademark of the Cmoy builds):

Went to look for this Altoids enclosure and only found one on ebay for $65! Apparently it is a collector's item now due to its rarity.

As noted, the circuit is quite simple:

It powers on when you plug in the headphone which is nice. A red LED indicates it is on. There is no volume control. You would use the one in the source for that.

CMOY Audio Measurements
Let's start with our usual dashboard:

That's quite a bit of gain at 15 dB. Anything more than 0.35 volts input would cause it to severely clip.

Signal to noise ratio running on batteries was very good:

Frequency response is nice and flat:

With just 2 volts output, there is not a lot of power there into 300 ohm load:

As you see, I also measured it using 12 volt power supply as the owner indicated the unit is rated to work better. Indeed it does. Distortion is markedly reduced and power increased.

Strangely, the reverse happened with 33 ohm load:

Conclusions
It is trivial to beat this specific CMOY with just about any headphone amplifier. Heck, even a phone dongle may produce more power. But that is not the point of CMOY. It is great "first electronic project" that should fit right in to the needs of a young person wanted to build something and use it immediately. I suggest building one using a breadboard as you don't even need a soldering iron.

So despite horrible objective results, I am going to recommend the CMOY portable headphone amplifier. It should started my journey into headphone amps.

------------
As always, questions, comments, recommendations, etc. are welcome.
Still without a working septic tank. Technician came, took a look and told me what I already knew: "your tank is full -- needs to be pumped first." Then asked for $200. Pumping will happen tomorrow. Meanwhile my wife is all unhappy and we are having to keep running outside to use the toilet in our RV.

The worse part is that I can't process any fruit. Here is a small sampling of harvest waiting pretty:

Please wish me success in getting our septic tank fixed by donating as much as you can using:
https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

 

[ROOSKIE]

Awesome.

I think I am going to build this with my GF's kid.
He already did a set of headphones via Kiwi Co
https://www.kiwico.com/headphones-project-kit.html

They sound bad but they work. He deff noticed a SQ difference as he is used to using my Sennheiser and AKG sets.

 

[Vini darko]

Cute little amp. Looking forward to the inevitable scematics posting and conversation.

 

[JohnYang1997]

O2 got me going. I was lucky that I got a fairly good audio interface focusrite forte so I was doing measurements since day 1.

 

[AudioTodd]

I’m enthusiastic about ANY way to get kids - or people generally - interested and hopefully excited about electronics and particularly analog electronics!!!

 

[ruinevil]

Pow Chu Moy made the first Internet forum specific to the headphone hobby (Head-Wize) back in the 1990s, and filled it with designs and reviews and other content.

 

[AnalogSteph]

Looks like we've got a TLE2426 rail splitter in this one. It may be running out of current into 33 ohms, the worse behavior at 12 V suggests some sort of current limiter kicking in for sure. Otherwise it's the opamp. We're talking 10 mA rms or 14 mA peak at this point. cMoys were always intended for at least medium-impedance (100-150 ohms up) cans, and they normally had a 47 ohm output impedance as well. If you wanted to drive Grados, you'd go for an apheared-47 circuit or similar instead.

27 mW into 300 ohms is 8.0 Vpp, which is realistic to expect out of a 12 V supply. (14 mW would mean 5.7 Vpp, again a fairly typical value out of 9 V.) You could probably get 9-9.6 Vpp out of the thing with an OPA2132/34 or NJM4556AD. Those serious about output power into higher-impedance loads (250+ ohms) would have been building a dual 9 V (split supply) cMoy, which may have provided up to 4-5 Vrms out and as such close to 80 mW into 300 ohms..

Given a HD600 or similar, this amp should still make it to 103 dB SPL at max -80 dB of distortion (likely less than the cans' at this point), or about 108 dB SPL max, even on battery. cMoys never were about excess, just about getting the job done - running some proper hi-fi headphones from what might have been a 5 mW @ 32 ohm Discman output, back in the early-mid 2000s. And that they did. You couldn't just go out and buy a $30 LiIon powered portable headphone amplifier back in those days. Headphones were a pretty niche subject in general. Things didn't truly go mass market until the first FiiO amplifiers came out in around 2010-ish.

 

[infinitesymphony]

Wow, this is actually better than I expected. I built a number of these years ago, the first using the classic Radio Shack 276-150 prototyping board, the rest using a custom PCB from a Head-fi member. It was trivial to add a volume potentiometer and an 8-pin DIP socket for op-amp swapping. I decked it out with a Panasonic FM cap, Vishay Dale resistors, an ALPS pot, the works.

My favorite op-amps were LM4562, OPA2228, and OPA2107. Also had OPA2132 and AD823.

I was using low-impedance headphones at the time and had to add 0.1µF decoupling capacitors to the op-amp power supply pins to stop oscillation (increased distortion with increased volume) from the pickier op-amps, LM4562 and OPA2228 in particular.

Went to look for this Altoids enclosure and only found one on ebay for $65! Apparently it is a collector's item now due to its rarity.

That's hilarious. I used to love this flavor and am almost certain I have this kicking around in my CMoy parts box along with some other tins.

 

[YSC]

it actuallly performed much better than expected considering there's almost no circuitry to speak of and the case is just a reused food packing! power should be fine for the age, when you get practically sh_t power from all sort of phone outputs and computer on board stuffs

 

[JohnYang1997]

it actuallly performed much better than expected considering there's almost no circuitry to speak of and the case is just a reused food packing! power should be fine for the age, when you get practically sh_t power from all sort of phone outputs and computer on board stuffs

Same parts on O2 performs way better tho.

 

[infinitesymphony]

Here's some commentary about OPA2227 in the CMoy circuit:

"I don't understand why you would use the OPA2227. Distortion is quite high above a few KHz. It's really designed to be used in DC or low frequency applications. The 8066 is...better I guess, but I still wouldn't recommend it.

...

The 8066 and 2227 are far inferior in load driving capability and you're pretty much inviting much higher distortion by using them. I'm not trying to be flippant about it (I'm not upset or anything ), but the choices people make for their CMoys to go with chips like the OPA2227, I think, contributes greatly to the later impression people come to that they'll benefit tremendously from buying a commercial portable amp."

Maybe better results with a different op-amp and gain setup?

 

[JohnYang1997]

Here's some commentary about OPA2227 in the CMoy circuit:

"I don't understand why you would use the OPA2227. Distortion is quite high above a few KHz. It's really designed to be used in DC or low frequency applications. The 8066 is...better I guess, but I still wouldn't recommend it.

...

The 8066 and 2227 are far inferior in load driving capability and you're pretty much inviting much higher distortion by using them. I'm not trying to be flippant about it (I'm not upset or anything ), but the choices people make for their CMoys to go with chips like the OPA2227, I think, contributes greatly to the later impression people come to that they'll benefit tremendously from buying a commercial portable amp."

Maybe better results with a different op-amp and gain setup?

OPA2227 is very good when within its capability. The dashboard is not loaded(or 100k), there's no reason for it to perform poorly rather than the design itself. It's easily in the <0.0002% territory.

 

[infinitesymphony]

OPA2227 is very good when within its capability. The dashboard is not loaded(or 100k), there's no reason for it to perform poorly rather than the design itself. It's easily in the <0.0002% territory.

That's good to know. If you were building a CMoy which op-amp would you pick?

 

[JohnYang1997]

That's good to know. If you were building a CMoy which op-amp would you pick?

I would firstly make a pcb. For opamps, that really depends on what I want. High impedance load? Low impedance load? Best battery life? Do I change the circuit etc. I'm currently trying out a battery friendly design using opa2189 in parallel.

 

[preload]

If I can add a couple of things - I don't think the (original) CMoy circuit was designed for low impedance headphones. Chu Moy's original article on Head-Wize contains an entire appendix that describes the additional modifications that are necessary to drive low-Z phones. So, my opinion is that it isn't exactly fair to test a device for an application it wasn't designed for and then claim it measures poorly.

Secondly, as @AnalogSteph points out, 1mW is plenty for many high impedance headphones. For instance, the popular Sennheiser HD650 and HD800 (300 ohms) each have a power sensitivity of 100dB at 1mW. I HOPE people aren't listening at volumes of 100dB! So I'm also not sure I would agree with the statement "max power = 14mW (poor)." In fact, 14mW would be in the top tier if we were comparing to portable DAC+HPA'S:

My personal interpretation of the data provided is that the CMoy amp actually measures quite well for its intended purpose (high impedance headphones with reasonable sensitivity). Yes, THD rises when driven beyond its design range, but so does every single other amplifier in existence.

I am therefore puzzled by the headless panther rating.

 

[JohnYang1997]

If I can add a couple of things - I don't think the (original) CMoy circuit was designed for low impedance headphones. Chu Moy's original article on Head-Wize contains an entire appendix that describes the additional modifications that are necessary to drive low-Z phones. So, my opinion is that it isn't exactly fair to test a device for an application it wasn't designed for and then claim it measures poorly.

Secondly, as @AnalogSteph points out, 1mW is plenty for many high impedance headphones. For instance, the popular Sennheiser HD650 and HD800 (300 ohms) each have a power sensitivity of 100dB at 1mW. I HOPE people aren't listening at volumes of 100dB! So I'm also not sure I would agree with the statement "max power = 14mW (poor)." In fact, 14mW would be in the top tier if we were comparing to portable DAC+HPA'S:
View attachment 86731

My personal interpretation of the data provided is that the CMoy amp actually measures quite well for its intended purpose (high impedance headphones with reasonable sensitivity). Yes, THD rises when driven beyond its design range, but so does every single other amplifier in existence.

I am therefore puzzled by the headless panther rating.

Unless it's intended to measure poorly, no device should be treated differently. It does not matter if it's designed for high impedance.
In fact, there's literally no design almost at all. No sign of optimizing for high impedance load.

The chart you gave is very old. Fiio, topping etc have several portable devices measured with way higher power than 14mW.

 

[jasonhanjk]

My recent cmoy built, this will be my last cmoy.
Start my first one around Y2k. Ahh, it brings back memories.


This board only active component is a single 4556, configure gain 3x.
Base on 12V supply:
Cross talk <-70dB for 32 ohm, best THD+N is 0.004% for 300 ohm at 3V output.
32 ohm have increase distortion at low frequencies due to it's output cap.


In future, my design will only have 2 IC. The target is 3Vrms achieving 0.0005%.
Powered by battery (hopefully).

 

[amirm]

Secondly, as @AnalogSteph points out, 1mW is plenty for many high impedance headphones. For instance, the popular Sennheiser HD650 and HD800 (300 ohms) each have a power sensitivity of 100dB at 1mW.

I have little use for such sensitivity numbers and the math around them. My standard for desktop products is 100 milliwatts with 300 ohm load. Anything less and I can get the amp to distort before the headphone. No way 1mw is remotely sufficient.

 

[amirm]

So I'm also not sure I would agree with the statement "max power = 14mW (poor)." In fact, 14mW would be in the top tier if we were comparing to portable DAC+HPA'S:

Distortion-less performance is only 0.1 mwatts:

Here is google v1 dongle with similar power:

Notice how it has little rise in distortion.

 

[preload]

I have little use for such sensitivity numbers and the math around them. My standard for desktop products is 100 milliwatts with 300 ohm load. Anything less and I can get the amp to distort before the headphone. No way 1mw is remotely sufficient.

Can I ask how you came up with the 100mW figure?
Also, you mentioned that 100mW of output was your standard for desktop products. I think most people would agree that the CMoy is a portable product?

Distortion-less performance is only 0.1 mwatts:

Can you clarify what you meant by this? Unless I'm interpreting your THD+N vs. Power chart incorrectly, the CMoy's THD+N is the same at 0.1 mW (100 microW) as it is at 1mW.