Breadboard Distortion

[newfish]

Last night's little project was to fool around with the phase splitting of a BJT / JFET to see what would happen (based on the role the Bobtavia transformer works...).

Circuit is a very simple buffer (Jack Orman / Beavis Audio style) using a J201.
Outputs are taken from both the Drain and Source.

Drain output goes through a 22nF, then 330K resistor in series into the inverting input of a TL071.
Source output goes through a 1uF, then 330K resistor in series into the non-inverting input of a TL071.

Output from the TL071 goes through a 100nF cap, with a pair of Silicon diodes to give symmetrical clipping.

It sounds extremely aggressive - which is a huge bonus.

Am just wondering why it works.  Is it that as the Drain output goes 'high', this sends the inverting input higher than the 'low' source / non-inverting side of things - and vice-versa?

I can post soundclips and schematics later this evening.  This was a throw-together in the early hours.
One of those, "I can get two outputs from the J201, and the TL071 has two inputs.  Wonder what would happen if I put them together..." moments.

[earthtonesaudio]

Hey Ian,
Last night I was actually working on a very similar idea.  I used a BJT phase splitter, one output going into each of the inputs of a 741.  I put the diodes in the feedback loop, however, and a pot across the inputs (wiper to Vref) to control the type of distortion.  It worked fairly well, but as I was removing it from the breadboard I realized I had connected one end of the offset null to the "no connection" terminal.   
So it probably could have worked better if I'd hooked everything up right.

What you have stumbled upon is a very common use of an op-amp, converting differential signals (such as the phase splitter outputs or a balanced microphone) to a single ended output.  This gives you the full differential gain (limited by the devices used and how symmetrical the inputs are), which is more than the gain from using only one input.

For more info, look up the "difference amplifier" circuit, or "instrumentation amplifier" or "op-amp mic preamp."

[edvard]

Oh dear.
Now I have to clear my breadboard again. 

I've often wondered this myself, but never got around to actually testing it. Thanks for coming up with a working example.

P.S. How would you control the gain? With single ended designs, you just control the feedback to the other input, but this design eliminates that altogether.
Hmmm... 500k pot on the input to the FET to do double duty of gain control and gate bias?

[earthtonesaudio]

For a single op-amp circuit, you could do the gain control in many (unconventional) ways.  If you put the gain control on the op-amp side, you could use a variable feedback resistor, variable resistor across the inputs, or separate (voltage divider) attenuation controls for each polarity.  Or, you could put the gain control before the op-amp, such as using an emitter bypass cap, or panning from collector to emitter (this is what I was working on), or varying the bias to the transistor.

[newfish]

I was planning on either having the circuit at maximum gain, and rolling off the Guitar's Volume, or using a feedback resistor on the op-amp.

Alternatively, a dual-ganged pot wired as a volume pot (to ground) across both Collector / Emitter (Drain / Source) would work also.

I've added a buffered blend to mine - simply because I was asked to do this on my Bro's Bass Fuzz I'm building him and I fancied having a play.  This is also another reason for leaving the circuit at maximum 'dirt'.

*EDIT* Blend removed. 

Now using two 100K resistor into inputs of TL071.  1Meg Linear pot now between output of Op-Amp (pin 6) and input on pin 2 to change the amount of gain / dirt / filth / grit / call-it-what-you-will-but-it's-a-bit-louche.

Gets trés dirty. 

100K Linear as volume control and it's good to go.

The story so far is here...

http://www.iansnape.co.uk/schematics/DiffDrive_300709.PDF

[earthtonesaudio]

No bias for pin 3?

[earthtonesaudio]

Here's what I was working on:




I don't even know what values I was using, but it's not ready for primetime yet.  Sounds bad at the extremes of the "mix" pot rotation, but pretty decent otherwise.  Just needs tweaking really.

One thing is fixed: Transistor base is biased at 1/3 supply.

In theory it should be blending from "op-amp Muff" type distortion with the mix wiper at the bottom, to Tube Screamer type distortion with the wiper at the top, but in practice it requires little tweaks here and there to play nice and not howl with feedback or make radio noises. 

[newfish]

Hi,

Guess not.

I've tried with and without - as per many searches for the proper way to do this.

Sounds fine without, so have left it off.

Will hopefully post soundclips later this week.

With the gain pot turned all the way down, it's more of  FET booster, going through the buffer that the TL071 then becomes.
Lots of gnarly gain available with the pot turned up though...

[panterafanatic]

earthtone, maybe set some resistance on one side of the blend so you at least eliminate one extreme, but you culd use a smaler pot t eliminate the other, depending on your pot of course

newfish, if you dont mind, maybe a scheme to me so i can test it out for myself at one point? i'd love that gnarly gain idea.

[earthtonesaudio]

Yeah, fine tuning would probably be pretty straightforward.  I think there's good potential there.

Newfish posted a .pdf of his schem in reply #4 above.  He says it works without biasing the op-amp, but I think this might make his results difficult for others to repeat.  My guess is the leakage current through the coupling caps is providing the bias... I can imagine it working, if you have caps that are just right.

[snap]

earthtone, maybe set some resistance on one side of the blend so you at least eliminate one extreme, but you culd use a smaler pot t eliminate the other, depending on your pot of course

newfish, if you dont mind, maybe a scheme to me so i can test it out for myself at one point? i'd love that gnarly gain idea.

Instead of placing limiting resistors in series with the outer pot lugs, I`d insert them in series with the collector and emitter caps.
In both cases they will prohibit the collector or emitter resistor of the transistor effectively being AC-shorted (which leads to instability or exaggerated gain factors), but in the first case they will not allow the signal to either of the opamp inputs to be completely shut off.
The described second case allows for this. Take a look at passive pan-pot circuits for comparison.

[earthtonesaudio]

Thanks, that's what I was planning to try next.
I was also thinking a buffer between the collector and the rest of the circuit might be advantageous, but maybe not worth the extra complexity if a simple resistor would suffice.