dangerous hum with opa637 in a Cmoy style amp

[Garbz]

Hi all.

I've built a mutant version of a Cmoy amp, linked here

Basically it's exactly the same as in CMoy's design expect that it has a 50k Audio Taper volume control and R2 (from noninverting input to gnd) is 1Mohm. THe power supply is also modified to use a descrete rail splitter made of trasistors diodes and a few resistors and it is doing wonderfully keeping the voltages at +-6.2v with a deviation of +-0.1v. The psu is also decoupled with a 220uf, 10uf and 0.1uf caps (over kill probaby not good but normally sorts a lot of problems out)

The amp worked like a charm with the OPA134 driving a set of 300ohm sennheisers quite nicely. However in the search for better quality this chip didn't do it for me. THe opa637 however does.

THe problem is that if i turn the volume control down below about 10-15% of it's swing (so that the resistance between input to output is about 52kohm (yes it's a 50k pot but it's max resistance is about 55.6k)) the amp starts to hum. It briefly thumps at the point where the hum stops/starts.
The hum is not a grounding problem. I've eliminated every grounding posibility, and the hum is also pitched slightly higher than the usual grounding frequency. THe hum also changes pitch while I swing the pot between 0% ~15%.
Furthermore while the hum is present so is DC on the output. Normally it sits quite comfotably between 0.1mV and 0.6mV but while it hums it puts out an easy 500mV (same as the thump when i switch the amp off, but in this case i know the problem)

Does anyone have any idea how to fix this? I'm at the end of the rope. I don't want to have to switch back to the OPA134.

[tlmadsen]

Garbz

Are you aware that the OPA637 is only stable at gains >= 5 ??? I think you have a gain of 10, but please check (your "high pitchs" humming could be some kind of oscillation).. You can use a the OPA627 instead. I use that with a BUF634 on the output to drive my Sennheiser HD580. It sounds VERY good.



THomas

[borges]

THomas,

you just beat me to it. My experience is to start off with the OPA627 and then switch in the 637 if I need the extra gain and rise time.

Speculation: Might there be a problem with too much resistance driving the input capacitance? Is that forming a pole which does unfavorable things to your feedback?

--
Børge

[Garbz]

i doubt gain is the problem it's set to a gain of 11 and since it's stable at >=5 according to the datasheet it should work just fine.

Please tell me more about the input resistance problem? I'm not sure what you mean but it may make sense if the hum only happens when the volume is low (i -> o resistance high, o -> gnd resistance low). I think this is a plausable solution.

However to eliminate gain problems i'll give the 627 a try providing i can find my chips.

[Garbz]

Hey guys thanks a lot the hum is gone, and so are any nasty voltages. Although there appears to be a rise in output DC to 1.1mV but who cares. My body puts out more than that .

I do however have a question. (reads 2).

1. The comparison between 637, and 627 is the difference purely benficial to really sensitive equipment (thinks sonar / HF dependant devices) ? I.e. is there a measureable difference to audio? I can't really pick any problems, but then my source isn't that good.

2. Why exactly is this chip sooo unstable? I mean the gain of 11 meets the minimum requiremnts. So why is wasn't it working ? Any idea why only at low volumes too?

Anyway thanks again guys. It sounds simply beautiful. Never thought a cheap onboard soundcard could sound this good. (New card is in transit atm. damn customs making me wait)

[needtubes]

I tried the 637 in a CMOY amp to drive my 580s and ended up with anywhere from .1-.7V on the output no matter what I did to it. I suppose this is the same problem you were having.

I am going to guess that this had something to do with the opamp's speed. It is significantly faster than the 627 and needs special treatment, I suppose. I guess solutions might be a feedback cap and small value decoupling caps on each power supply rail (close to the opamp's pins). I didn't try anything else with the amp and just scrapped it. I am assuming the chip was oscillating, otherwise it would have no excuse to put out .7V on the output. Maybe sometime I'll try to fix the problem in another build.

[peranders]

Originally posted by Garbz
2. Why exactly is this chip sooo unstable? I mean the gain of 11 meets the minimum requiremnts. So why is wasn't it working ? Any idea why only at low volumes too?

The OPA637 is designed for gains > 5. That's it!

OPA637 is rather fast. This means that the pcb layout must be done proberly.

If you can supply us with a schematics and/or picture we maybe can see what's wrong. I supect two things: 1 unsuitable component values. 2 Not so good decoupling of the supply voltages. 3 Maybe some error in your design.

[Garbz]

I'd rather not supply pictures. I'm rather messy regarding Soldering and now my amp is closed and finished with the OPA627. Just for future reference all components on the amp are located within 1cm of the amp itself and the Decoupling caps (0.1uf, 10uf, 220uf) are no more than 2cm from the amp. I suppose the problem is bad values or something. It couldn't be a gain problem. And considering my ouput problems occure only when the volume is turned down i assume again that it's the parts.

Either way thanks again for the replies. I would have never thought of trying the opa627 because like i said the gain is the high enough. But that just proved me wrong.
Sounds great now.