CMOY starts just fine, then gets VERY clippy in 3 or 4 seconds

[Mizerable]

I built a Cmoy exactly like on the tangent website. However, I am using 470uF supply caps and .22uF input caps. I also have the 4.7k Ohm r5 resistors put in.
I am using two 9v batteries in series for 18 volts.


My problem is that when I first turn on the amp while music is playing, it's perfect. Beautiful. However after a few seconds, it suddenly starts distorting like crazy! clipping like mad! Why is that ? if i flick the switch off and then turn it on again, I get another few seconds of great performance and then, it goes bad again. why is that ?

 

[fault151]

It could be your batteries are nearly dead or the op amps are the wrong type. I experienced both in mine.

 

[Mizerable]

These are new batteries...haha

I am using a TL082CP dual op amp. The tangent website said that these are usuable though a bit bad. My headphones are only 24ohm crap ones though.

 

[FallenAngel]

1) 4.7K R5 is EXTREMELY large, skip it.
2) Strongly consider another opamp.

 

[Pars]

Agree with FallenAngel on the 4.75k resistors for R5. These will cause massive distortion. Jumper R5 instead. If you feel you have to use an R5 resistor, it must be <100ohms.

 

[Mizerable]

Ahhh wooops! I meant 47 ohms! sorry !

I noticed with my meter that one side of the powersupply was giving out only 1.3 volts while the other was giving slightly over 17 volts ! Damn, that must be the problem.

 

[Pars]

Check the two 470uf electrolytic caps for proper orientation (+/-) and that they are actually connected to the circuit. I've seen a number of people be 1 row off on these and the corresponding jumpers.

Otherwise, post detailed (macro) pics of the board top and bottom.

 

[Mizerable]

I got the distortion issue figured out. When I use lower voltage power supply, the clear-period lasts shorter and the distortion takes shorter to kick in. I think it's because I'm using some very low impedence headphones and that the weedy powersupply isn't able to handle it since i'm also using a very powerhungry TL082.

The other problem I have is the high DC offset.measured from ground to right and left outputs, they are 53 mOhms, and 64 mOhms. I dare not plug in my senn HD555's....

How can I fix that large DC offset ?I'm willing to sacrifice some sound quality if need be...

There's also a hissing background noise hah..

 

[tangent]

When the power supply is so greatly unbalanced, it means something's not connected correctly, causing a short or near short. This is why Pars recommended looking at the caps...not the only possible explanation, but one of them. Also likely are misconnected grounds, wrong resistor values, etc.

EDIT: We can diagnose most such problems with pics. Hint.

 

[Mizerable]

Quote:

Originally Posted by tangent /img/forum/go_quote.gif When the power supply is so greatly unbalanced, it means something's not connected correctly, causing a short or near short. This is why Pars recommended looking at the caps...not the only possible explanation, but one of them. Also likely are misconnected grounds, wrong resistor values, etc. EDIT: We can diagnose most such problems with pics. Hint.

Ahh okok. The voltage is equal now. However, there is still the problem of clipping/distortion within a few seconds after turning the amp on. I will provide some pictures.

Oh, and to cure the DC offset of around 50-60 mvolts, can I take the two C2 capacitors and move them in series with the output and leave the input without any capacitors so that it goes right into the opamp?

 

[Mizerable]

here are the pics :

 

[Pars]

Quote:

Originally Posted by Mizerable /img/forum/go_quote.gif Ahh okok. The voltage is equal now. However, there is still the problem of clipping/distortion within a few seconds after turning the amp on. I will provide some pictures. Oh, and to cure the DC offset of around 50-60 mvolts, can I take the two C2 capacitors and move them in series with the output and leave the input without any capacitors so that it goes right into the opamp?

No. I believe if you were going to cap couple the output that you would need much larger caps for most headphones; certainly for low impedance phones.

50-60mV, while not good, won't hurt most phones, particularly higher impedance phones like you Senns. Grados or something low impedance I might worry a bit about. That said, the goal is to get the offset down to <5mV or less levels.

Looking at your pics, it would appear that you have the electrolytic PS caps in correctly from what I can tell. At any rate, verify those. If you are looking at the board with the components up, opamp to the bottom of the board and PSU caps to top, the left cap + should ohm out to the positive bat. lead. The - should ohm out to virtual ground. The right cap + should ohm out to virtual ground and the - should ohm out to the negative bat. lead.

All ratshack parts it appears. You are sure that the R5 resistors are 47ohm? What are the R4? Could just be the camera colors, but what should be black stripes look green... but I'll assume they're black. Otherwise, I don't see anything obviously wrong... you're missing some insulation here and there on the wires, which you should clean up. Solder joints look ok... I'd probably reflow many of them just to make sure.

 

[Mizerable]

hmm. so for 120 ohm headphones, 0.22 uF metal film caps would not be sufficient eh ?

I also noticed something else: the clipping/distortion is MUCH worse on a pair of 16 ohm headphones than 24 ohm headphones. Is this what people refer to when they say that a certain op amp " can't handle low impedence headphones well " ?

Why is that ? Why can't an omp handle low resistence headphones ? I was thinking since V=IR, and if R is very low, then the current must be very high and some op amps can't supply that much current. is that line of thought correct ?

EDIT: I mean for the caps to be put in series with the output with the 120 ohms headphones to eliminate dc offset.

 

[Pars]

Quote:

Originally Posted by Mizerable /img/forum/go_quote.gif hmm. so for 120 ohm headphones, 0.22 uF metal film caps would not be sufficient eh ? I also noticed something else: the clipping/distortion is MUCH worse on a pair of 16 ohm headphones than 24 ohm headphones. Is this what people refer to when they say that a certain op amp " can't handle low impedence headphones well " ? Why is that ? Why can't an omp handle low resistence headphones ? I was thinking since V=IR, and if R is very low, then the current must be very high and some op amps can't supply that much current. is that line of thought correct ? EDIT: I mean for the caps to be put in series with the output with the 120 ohms headphones to eliminate dc offset.

Regarding the current sourcing for low impedance phones, you got it. What phones (impedance) are you testing this with BTW?

Regarding your 0.22uf caps as output coupling caps, a series cap in combination with the impedance of the phones will form a high pass filter, whose -3dB point is given by 1/2 * Pi * R * C, where C is in farads. So, for your 120ohm phones, you have a -3dB point of ~6 KHz... a little high

It will be even worse for your low impedance phones. A 1000uf coupling cap would give you a -3dB point of 1.32Hz, much more reasonable. Personally, I would fix the amp rather than toss band aids on it, but that may mean (well probably will mean) replacing the opamp.

 

[Mizerable]

For testing this buggy amp I'm using the sony MDR-g57 "street style" headphones. They're pretty junk ones that are meant for jogging and such. 24 ohms.

There's also another pair of sony sporty-headphones , MDR-w014. this one is only 16 ohms and when used with my amp the clipping is many times worse than the g57. well... i'm not sure its many times worse, but it is much much worse.

Oh also, what does the -3dB point refer to? Is that the point where the sound becomes in audible ?