bass amp compressor distorts

[Alex R]

Please stay with me on this one - a bit complicated but of interest I hope, and I could do with a bit of help.

Here is the compressor section of a very expensive SS boutique bass amp made in Germany. The customer complains that a subtle distortion creeps in when the compressor is working. It is indeed both audible as a higher harmonic and visible on the output as a sort of soft clip to the top of the wave. Turn down the compression and the gain, and it goes away.

The compressor is an optical type. OC1 is the LDR controlled by the LED labelled OptoLED.

The TL071 that drives the LED puts out about 4vAC max, and averages about 1.5vAC under normal signals. The bridge rectifies this AC signal and thus the compressor is activated. It uses a green LED in a hand-built opto-whatsit held together with heatshrink. The LED sees from 0 to 0.8vDC, but at 0.8vDC there's also about 1.4vAC at signal frequency driving the LED.

I observed that at such low voltages, the diodes were leaving significant portions of the waveform intact because of their thresholds, so the LED was being driven by a rather dirty mixed DC/AC input that echoed the signal frequency. So I took the LED out of the circuit and ran it from the bench supply with clean DC. The distortion disappeared.

OK, what's wrong here? Am I missing something in this circuit that should be smoothing out the LED's supply? What does C19 do? Or is is it (I hate to say this) a bad design? Any solutions? I was thinking, germanium diodes... (shows I'm getting desperate).

The diodes are good; the bridge is connected right.

Thanks in advance as usual.

[Alex R]

...maybe this schemo is easier to read...

[Rick Erickson]

I had a similar complaint with a Hartke Mdl 2000. iirc - the factory told me this was "normal" and there was no cure for it at the time.
I didn't feel like re-designing their compressor so I told the customer the "good news".

RE

[Alex R]

Shit could that really be it? They didn't realise the bridge wouldn't really rectify such a low voltage? I'm starting to think 4 germaniums might actually work...

But what about C19? Would that smooth things out? Might do some experimenting tomorrow. I did try poking some electros in there after the bridge to see if I could smooth it. Just made it worse.

[captntasty]

No other reason than past experience... might it be the opamp? I've had bad ones and that's what they do - a small amount of really annoying distortion.

[Alex R]

Well... I tried:

checking every component value in the compressor
opamp swaps
swapping and increasing values of caps
adding smoothing caps
a replacement vactrol
and, yes, germanium diodes in the bridge...

and none of them worked.

I fed clean DC to the LED again, and the distortion disappeared.

Looking at other optical compressor designs, which have much more complex drivers for the optoisolator's LED, I'd say that I have to conclude that the supply has too few stages and too little smoothing to drive the LED properly. Am I right, do you think? ...or missing something obvious?

Oh boy. Looks like I'll have to drop 3 hours of pointless labour on this.

[52 Bill]

I assume that when you say that you "fed clean DC to the LED", that the signal would vary as you varied the DC voltage without distortion?

If this is so, have you scoped the DC supply rails at the LED Driver chip? Maybe the LED is drawing too much current from the opamp and it is causing it to screw up the rails supply? Maybe try a red LED or add a current limit resisitor?

Just a thought.

[captntasty]

R40

Wouldn't R40 (220R) be the current limiting resistor? If so, perhaps replace or maybe sub in temporarily a 1k pot and try to dial in the "brightness" of the LED?
Just thoughts...

[52 Bill]

Yeah, you're right...it's that short attention span thing again.

Also, maybe a 1K pot sub for R41 might be thing to try. Wouldn't a higher value there reduce the loading on the opamp as well?

[Alex R]

Thanks guys, those are interesting ideas.

Yes when I fed DC to the led I could vary the 'compression' up and down by changing the voltage. Even when I overloaded the led there was no distortion.

When the distortion happens it's a kind of higher harmonic, seems to slightly push down the early part of the top of the curve.

No I didn't check the opamp rails. I'll give that a go.

[Alex R]

R40/41. Put a brake on the current. Could try that - but I don't think it's drawing that much because, well, the compressor compresses with the right sort of range as you twiddle the control up; I'd say the led isn't glowing too much. but still, maybe it's too much for the opamp. Mm. I'll check the rails (tomorrow)

When the distortion is full on you can still get plenty more gain out of the preamp, though.

By the way it's a rack bass head by Tecamp (used to be called Tech, name change to avoid confusion with another company I think).

[Steve Conner]

I agree that the distortion in this compressor circuit is "by design". If you want it sounding cleaner, you'll probably have to redesign it with a proper rectifier that has defined attack and decay characteristics, so that it is mostly feeding DC to the gain control element.

Of course, this will completely change the sound of the compressor! What that amp has is more of a soft clipping limiter kind of thing than an actual compressor, since there don't seem to be any time constants built into the circuit.

What fool tried to make a boutique amp out of three-legged fuses anyway?!

[Alex R]

Thanks for looking at it Steve. I'm slow to go for the 'blame the designer' diagnosis so beloved of a certain sort of repairman, but this design just looked a bit too simple to me. The customer is I think going to complain to the manufacturer... stand by for fireworks... my position is simply that it seems to be functioning as it was intended to.

These amps do sound good - very 'natural' - but they cost about a grand. The customer has two, a 350 and a 600 watt. I've already put a new power board in the 350, the old one was a bit too charred to repair; don't know what went wrong with that but it seems to have settled down now. Just about every 3-legged fuse on it had blown...