Differential overdrive

TejfolvonDanone · March 12, 2016, 05:15:04 AM

Hi everybody!

I designed this overdrive circuit: https://drive.google.com/file/d/0BzGRfrmHSjb4Q2FNcGJBYUlxckk/view?usp=sharing
It has got 2 OP-amps which create a buffered and an inverted signal. Those overdrive the differential amp. The last transistor creates a buffer and a simple bass cut and high cut tone control.
My problem is that the outputs of the differential amp (+Output and -Output on the schematic) according to my knowledge should be the same. I don't have a scope at home so i can't really show the exact waveforms but when i looked at them they were like this:
https://drive.google.com/file/d/0BzGRfrmHSjb4aDBfWHV2MzF3QzA/view?usp=sharing The two aren't in phase which is as i expected but the top of them are really different.
The differential amp is almost symmetrical the trans have their Hfe in 10% difference the other components are int 5% tolerance.

Does anybody have any thought on how can the outputs differ from each other?

R.G. · March 12, 2016, 09:19:50 AM

They are the same shape, but inverted, excepting in those funny regions at the top and bottom of the waveform. And at the tops and bottoms, something funny happens.

"Something funny" is when the two opamps drive the diffamp into distortion. I suggest you look at the outputs of the opamps driving the differential amp and see what they're doing when the diffamp starts being "funny".

TejfolvonDanone · March 12, 2016, 11:07:13 AM

That was my thoughts too. The op-amps act as they should: same amplitude perfect sine wave but one is inverted.
I once disconnected the load and just watched how the diffamp works without driving anything. Everything stayed in this "funny" state.

(I forgot to mention that the input signal is a 100mVp-p sine wave from a generator.)

R.G. · March 12, 2016, 11:51:14 AM

There are some things to look at. It might be instructive for you to probe the joined emitters of the two diffamp transistors, and also to just disconnect the signal to one of the bases, leaving the bias connected.

A theoretical diffamp with a current source on the joined emitters has a linear range of only about 25mV before the higher Vbe on one base causes that device to have all the current from the emitters. In fact the diffamp, as cascaded amplifying sections, is used as an amplifier and limiter in the RF world for exactly this purpose. RF is fed from the antenna matching circuit to a series of diffamps, with up to 120db of gain in the cascade, each one both amplifying and limiting the signal.

But that's working on truly tiny signals to start with.

Bipolar transistors do odd things when their base-collector is driven positive. If the drive is from a high impedance source, it just clamps the device into saturation. But if the drive is from a low impedance source, the source itself can pull the collector up to follow the source if it can supply enough current to the base-collector and base-emitter junctions to raise the emitter voltage. The shapes of the traces hint that this may be what is going on here.

TejfolvonDanone · March 12, 2016, 01:47:40 PM

I'm not saying that my circuit is perfect and there isn't any rookie mistakes but i triple checked it. I checked without signal all the bias voltages then I checked the connections and didn't find any mistakes. I check with a scope all the inputs and outputs of the op-amps and the input of the diffamp. Everything is as should be.

The base-collector is well in the negative range even for a 10 times greater input signal then my test signal. (The op-amp stage amplifies maximum about 3 times and the base-collector voltage is about -4 V.)

But my main question is: why aren't the outputs look the same (except the inversion)? If i understand correctly this effect should occur to the other BJT thus giving me symmetrical output. Did i misunderstood you?

PRR · March 12, 2016, 01:58:24 PM

It would help to know the scale of those waveforms. 0.1V? 10V?

I suspect you are driving a collector all the way down to the emitters. Alternatively exceeding the 7V breakdown of emitter-base.

TejfolvonDanone · March 12, 2016, 03:08:17 PM

The waves are 5Vp-p if i remember correctly.
Unfortunately i can't provide pictures of the scope for about 5 days.

TejfolvonDanone · March 16, 2016, 02:47:05 PM

My name is TejfolvonDanone and welcome to Facepalm Time.
The main issue was that somehow my generator put out 10 times bigger signal than intended and my scope's probe was a 10:1 which i also didn't know. So the input op-amp got a 1Vp-p sine which it could handle easily and amplified by 3. But the diffamp didn't like that idea so much. Also i soldered it in a prefboard and turned out that on my breadboard i also messed up something. So you were right the collector was pulled down to the emitter which caused the "funny things".
All in all it works fine now. Here is the output of an op-amp:
https://drive.google.com/file/d/0BzGRfrmHSjb4aDk0REZ5NV96NGM/view?usp=sharing (the other has the same output.)
And the output of the diff amp:
https://drive.google.com/file/d/0BzGRfrmHSjb4YnR3aHk1bXRqVXc/view?usp=sharing
The output stage needs a redesign but it's a whole different story...

PRR · March 16, 2016, 07:14:14 PM

> the output of the diff amp:

Mostly what you expect from a HIGH!-gain amplifier. (As R.G. says, a few dozen mV will drive a BJT from one end to the other.) The round top and flat bottom is not expected for FULLY differential, but I think you have the output single-sided and also the collector loads are large enough to slam collectors to emitters (why R.G. said to look at the emitters node).

But there's a billion ways to do heavy clipping. What does it look like when "slightly distorted"? Musically the interesting part is as you pass from "clean" to "flavor" to "shout" and back on the decay.

TejfolvonDanone · March 17, 2016, 03:53:51 AM

> Mostly what you expect from a HIGH!-gain amplifier.
I don't know if we are talking about the same thing but i really expected this. The collector is slightly closer to the Vcc than the emitter. So it bumps its head to the Vcc (falt top) and not fully goes down to the emitter (round bottom).

> Musically the interesting part is as you pass from "clean" to "flavor" to "shout" and back on the decay.
I not quite sure that this circuit is viable as a musical distortion and i don't know if anybody can tell from only the schematics. I'm really hopeful that it will serve its purpose. But only with a guitar and an amp i can truly test and it this will only begin in the next couple of days. I know I know I know i should have tested that when it was on the breadboard. But it wasn't working well back then.

After all thanks for the help I really learned a lot along the way.

News:

Differential overdrive