Controfuzz bias (loss?)

[John Lyons]

With this circuit I get a delayed start up which I think is loss of bias...



When I first first turn it on after not playing overnight it will pass clean signal and after a couple
strong chords the fuzz will quickly fade in and all will be well.
Should pin 6 have some sort of voltage reference? Ideas?
Thanks.

John

[PRR]

The circuit needs a full 2 seconds after power-on to bias-up, but I gather you are not that impatient.

C3 may be leaking. Replace it with something fresh.

> Should pin 6 have some sort of voltage reference?

It does. For DC 6 follows 7 which follows 5, which is tied to the 100K+100K V/2 node at pin 3.

Oh wait. TL072 is not good about inputs at the rails, and pin 2 starts from zero. Can you try 4558?

[John Lyons]

I'll try a 4558. Thanks Paul.

[Mark Hammer]

Hmmm, curious.  Why would it need a full 2 seconds to bias up?

And John, I've made them with a 4558.  Sounds fine.

The original schematic uses compensation caps with the 748 op-amp that will likely limit the top end.  If using an internally compensated op-amp, it may well be appropriate to stick a small feedback cap in parallel with the 220k resistor in U1a.  That stage has a whole whack of gain, and you don't really want it to be applied to the entire spectrum.

[PRR]

> Why would it need a full 2 seconds to bias up?

Why are there *two* "U1a" on that drawing? I'll call the top one U1a, the bottom one can be called "Fred" for all I care.

C3 10uFd starts at zero but must come up to very-nearly half-voltage before U1a "works".

C3 is charged through 220K+100r.

At turn-on: R1 R2 tend toward half-supply 4.5V, and C1 hardly slows it down. C3 is low while current trickles from the 220K. In an ideal opamp, the + input high and - input low will slam the output high. So there's 9V across 220K and C3 does tend to rise. But 10uFd*220K is 2.2 seconds to rise to 63%.

The amp stays slammed high until the - input comes up to 4.5V.

In a low-gain circuit the C3 cap only has to get near 4.5V for the - input to cross 4.5V, un-slam the output, and signal will pass.

Here the gain is 2,200, so the cap has to charge to 1/2200 of 4.5V to bring the opamp under control. That actually doesn't take much longer than the low-gain case, because the charging source is the slammed-high (9V, 8V) output.

I wasn't sure if the time-constant multiplied so I asked the computer. The output comes into range in 1.9 seconds. So it works out: 2.2 sec for 63% and 1.9 sec for 50%.

[blackieNYC]

Mark Hammer just gave me a good learnin'. - the "delay" is what this pedal is all about. The clean is heard first, then the fuzz comes in and outlasts the clean in sustain of course.  So it evolves from clean to dirty, if you have the pots set just right. . And the fuzz continues to evolve as the notes rings on.   I later found that he addresses the cfuzz in a thread I found while googling. I've lost it.
Everyone seems to want to kill the clean and have just the fuzz, but the transition, achieved with the two signals being summed out of phase apparently, is the deal.  I don't want the clean, but I want this transition, so I'm researching replacing the clean with another distortion that i love but drops out rather abruptly.  The Escobedo PWM schmidt trigger. Ever built one?  Unique sound but that cutoff is rude. Ya didn't happen to find a way around that did you? With the above method I don't expect to get the envelopes to match up perfectly.  Thats fine. And I only hope to mask the cutoffs a bit. If this doesn't work maybe an autoswell should bring in the sustaining "2nd half" of the note.   If you have any thoughts on that, please let rip.

[Mark Hammer]

Just to clarify, and separate some things, the slow bias onset that was he subject of the start of the thread, only pertains to when when first plugs into a Controfuzz.  Once it has biased up, there is no delay in going back and forth from effect to bypass.

What Alan/blackie is attempting to do is something that popped into my head when he described his desire to have some sort of decay/tail to the PWM sound, instead of having it sputter and die.  I was reminded of the manner in which the Controfuzz exploits two sounds, one lasting much longer than the other because of the gain applied to it.

In theory, a PWM circuit and either the Controfuzz distortion side, or something very much like it, could be blended at a mixing stage, with the fuzz side taking over from the PWM side and tacking on a decay/tail.

As discussed with blackie, the other alternative is using some sort of envelope control of the PWM such that the guitar dynamics controlled the output level of the circuit.  The PWM would continue to trigger as the guitar input signal goes from peak to almost quiet, but the last dying gasps of the PWM output would be truncated by an imposed fadeout.  That would certainly get one an audible decay, rather than a sudden choke-off, but the total duration of the note would be shorter than the PWM->fuzz transition noted above.

Thoughts?

[blackieNYC]

Controfuzz parts on the way.  Another idea I had was to blend in a distortion that passes thru something like this
http://www.aronnelson.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=45221
Autoswell. Looks quick and dirty.  Might have to tweak attack time.
Sorry if I'm hijacking the thread. 

[Mark Hammer]

Not at all.  I'm just sorry that the general topic of "morphing" across signals is not conveyed well by the thread header, because there could be something of use for folks here.