Harmonic Percolator - make one!

[pinkjimiphoton]

no worries, dino, just wanted to make sure somebody didn't think it was right and get dissappointed.

'sides, now i got YOUR layout to go with. works for me!!

[LP Hovercraft]

@ Mark-I will have to play with it a bit to see if I can get the Octave up sound.  It seemed to have the best sound (to my ears, anyway) with a leaky ge transistor in q1.  I remember trying a tightly spec'd 2N404 from Small Bear as q1 and it did not sound very good.  Yes, it must be bandwidth limited just at the right place in upper register somehow-one of the few 2 knobbers that truly do not need a tone control. 

[DougH]

Part of the bandwidth limiting I believe is due to the low input impedance and the way it loads your guitar pickup. The Rangemaster and Fuzz Face both behave the same way. The low impedance load will roll highs off of your guitar signal. And at the same time, when you turn the guitar down it will get clear and bright. I suspect there's not much more to it than that, other than the passive electronics in your guitar which affect it as well But in the end, I believe it's due to the interaction between the guitar circuit and the input to the fuzz circuit.

[Mark Hammer]

Part of the bandwidth limiting I believe is due to the low input impedance and the way it loads your guitar pickup. The Rangemaster and Fuzz Face both behave the same way. The low impedance load will roll highs off of your guitar signal. And at the same time, when you turn the guitar down it will get clear and bright. I suspect there's not much more to it than that, other than the passive electronics in your guitar which affect it as well But in the end, I believe it's due to the interaction between the guitar circuit and the input to the fuzz circuit.

Which would also imply that it would sound "best" when placed in an uninterrupted path path from the guitar, rather than preceded by anythng between guitar and Percolator that might offer a low-enough output impedance.  Does that strike you as sensible?

[pinkjimiphoton]

i run it first, before my fuzzface even.

[tca]

So, I dind't have the parts for the Percolator, so I've breadboard Escobedo's Jerkulator with two pairs of silicon diodes at the output and with the transistors 2N4401/2N4403. I couldn't believe my ears!! Here is the sound clip from a portable mp3 recorder with a 1W solid state amp with a 3'' speaker. Need to fix the frequency response, but it took me about 15m to breadboard it!


Cheers.

[DougH]

Part of the bandwidth limiting I believe is due to the low input impedance and the way it loads your guitar pickup. The Rangemaster and Fuzz Face both behave the same way. The low impedance load will roll highs off of your guitar signal. And at the same time, when you turn the guitar down it will get clear and bright. I suspect there's not much more to it than that, other than the passive electronics in your guitar which affect it as well But in the end, I believe it's due to the interaction between the guitar circuit and the input to the fuzz circuit.

Which would also imply that it would sound "best" when placed in an uninterrupted path path from the guitar, rather than preceded by anythng between guitar and Percolator that might offer a low-enough output impedance.  Does that strike you as sensible?

Yes.

[digi2t]

Something struck a chord (pardon the pun) with me last night. I was reading Gus' analysis of the circuit, when this jumped out at me;

If you look close it is just two simple gain stages in series: HOWEVER the 47uf sets up how much the current drain in each stage interacts. This is most likely the cause of the musical feedback in the video. Work out your gain phase and current draw.

The 47uf is most likely one of the big tuning points in the circuit make it bigger and smaller and hear what happens. The emitter node voltage is also important. If it is 9V supply and 4.5V both sections get 4.5V as a supply voltage. If it is 5V the NPN gets 4VDC and the PNP 5VDC. If it is 3VDC the NPN gets 6VDC and the PNP 3VDC etc.

Before I go any further, I'll just mention that I added a 100K pot to the input. What DouH said about loading bothered me. So 100k pot in, and no pot on the out, direct to amp. Pot is maxed during my tests.

So, back to what Gus said. I started rummaging through my capacitors, and found two things;
1) If I lower the 47uF to 4.7uF, raise the cap between the two trannies to 0.1u, and work the trimmer on Q1C, then yes, I do get sub-octave. But, in my case, it's not stable, ugly, and I get oscillation as well. If I turn down the 100K pot, it goes away, but the circuit sounds lifeless.
2) If I go with a 6.8uF tantalum, in place of the 47uF, and return my cap between the two trannies to my original value of 0.01uF, a huge grin appears on my face. Now, the octave up bloom appears slightly more pronounced, and my root notes/chords go into feedback at a much more sane amp volume level. The octave bloom just continues to bloom into sweet feedback. OH HAPPY MAN!!
3) THE sweet spot with a 6.8uF tant is now 1.0v on Q1C. Plucking a note while working the trimmer, as soon as I get close to 1.0v, it's... it's... well, let me put it this way; it finds Jesus.

At the moment my voltages are;

Q1
E - 1.146
B - 1.049
C - 1.015
Q2
E - 1.146
B - 1.675
C - 2.256

Everything else is unchanged on the board, it's as I drew the vero, except for the 6.8uF replacing the 47uF.

Freakin' crazy!!

[wavley]

Part of the bandwidth limiting I believe is due to the low input impedance and the way it loads your guitar pickup. The Rangemaster and Fuzz Face both behave the same way. The low impedance load will roll highs off of your guitar signal. And at the same time, when you turn the guitar down it will get clear and bright. I suspect there's not much more to it than that, other than the passive electronics in your guitar which affect it as well But in the end, I believe it's due to the interaction between the guitar circuit and the input to the fuzz circuit.

Which would also imply that it would sound "best" when placed in an uninterrupted path path from the guitar, rather than preceded by anythng between guitar and Percolator that might offer a low-enough output impedance.  Does that strike you as sensible?

Yes.

So, it might be a good idea to try an AMZ pickup sim in this thing if you don't want it first in your rig?  I put one in my fuzz face and it's certainly close enough for rock and roll.

[DougH]

Part of the bandwidth limiting I believe is due to the low input impedance and the way it loads your guitar pickup. The Rangemaster and Fuzz Face both behave the same way. The low impedance load will roll highs off of your guitar signal. And at the same time, when you turn the guitar down it will get clear and bright. I suspect there's not much more to it than that, other than the passive electronics in your guitar which affect it as well But in the end, I believe it's due to the interaction between the guitar circuit and the input to the fuzz circuit.

Which would also imply that it would sound "best" when placed in an uninterrupted path path from the guitar, rather than preceded by anythng between guitar and Percolator that might offer a low-enough output impedance.  Does that strike you as sensible?

Yes.

So, it might be a good idea to try an AMZ pickup sim in this thing if you don't want it first in your rig?  I put one in my fuzz face and it's certainly close enough for rock and roll.

Yes, you need to treat this like any vintage fuzz with a low input impedance.

And for digi2t: The 100k pot on the input won't affect the impedance much. The input impedance will basically be constrained by the input impedance of the transistor, which is pretty low, typical for a BJT like this. That's why vintage fuzz circuits using BJT transistors typically have low input impedance which loads the guitar pickup and is part of the "sound". Unless you bootstrap the first stage, a la Gus's NPN Boost, a typical BJT gain stage like this will be low Z.

[digi2t]

And for digi2t: The 100k pot on the input won't affect the impedance much. The input impedance will basically be constrained by the input impedance of the transistor, which is pretty low, typical for a BJT like this. That's why vintage fuzz circuits using BJT transistors typically have low input impedance which loads the guitar pickup and is part of the "sound". Unless you bootstrap the first stage, a la Gus's NPN Boost, a typical BJT gain stage like this will be low Z.

Got it Doug. I just wanted to eliminate any and all possible variables here. Bit of a pain when something sounds one way on the board, and then different boxed, especially if it's due to cutting corners when I had it on the board.

[digi2t]

CRAP!!!

It was too good to be true. Nix on the 6.8uF. I finally got the chance to crank it up more to, ahem, "stage" levels. Problem is the there is a short oscillation "after howl" that occurs whenever you hit chords hard. So, 6.8uF is out.

10uF tant sounds pretty good though. Better than a high-end ELNA cap that I tried. But, 10uF has a drawback; doesn't clean up as well as a 47uF when you roll of the volume. I have a 33uF Philips axial in right now, and I think it's the best compromise to date. Good harmonic content, feedback attainable (though higher volume is needed), and cleans up pretty decent.

I get the feeling this thing is telling me to stay between 22 and 47uF.

I also reduced the cap on Q1C to ground to 0.0047uF. I saw that the Collins clone uses a 0.001 here, but I couldn't hear a difference between 0.001, and nothing. 0.0047 is a good compromise between the 102, and the 103 I was using. It's just enough to kill the noise from the tranny, while being less buzzy than the 0.01.

Ah... the joys of breadboarding! The saga continues...

[pinkjimiphoton]

yah, them's the breaks bro...

when i first built this, i was amazed...thru my ruby.
when i played it thru my princeton, i realized there was a humongous difference.. a little tweaking tho, and now it sounds pretty awesome!!

[Gus]

I built two Si versions
One a PNP 2n2907 and NPN 2n2222
And an all NPN 2n2222
I did change the circuit a little I increased the collector currents and adjusted the feedback bias resistors.


pinkjimiphoton

yah, them's the breaks bro...

when i first built this, i was amazed...thru my ruby.
when i played it thru my princeton, i realized there was a humongous difference.. a little tweaking tho, and now it sounds pretty awesome!!

If I adjusted effects for friends of mine I would ask them to bring there stage amp, guitar(s), and signal cords that they used live.  I would then ask the play at stage volume and explain what they wanted to change.  I would adjust the circuit(EQ, gain, bias etc) and then we would test again.  One thing I noticed is often you want a thin sounding effect when tested it at low volumes in a house for distortions and fuzzes at stage volume.  A few things are going on how the players share the frequency range, if the guitar has open back speakers thing can sound bad when the volume is turned up if you have to much bass, and there are other things.

digi2t
Thanks for doing the cap value and type test at the two emitters.  I never got around to testing this tuning point.

[Paul Marossy]

Well I finally got mine to work. It wouldn't work because I did something so shockingly stupid that I can't even mention it. 

Anyway, I did the Giblet version with a few tweaks:

Q1 is a 2N1108 with an Hfe of about 62
Q2 is a 2N2222 with Hfe of about 125
R2 is a 510K
C3 & C6 are 0.15uF
R4 is a 50K trimpot. Sounds best to my ears at about 6.8K
50K balance pot
100uF power supply cap added
0.001uF cap from Q1 collector to ground
0.0022uF cap from output of balance pot to ground

[pinkjimiphoton]

MUST...HAVE...CLIPS!!!!!!!!!!!!!!!!!!!!!!![/color]



PROPS ALL AROUND, I HOIST THIS HERE BREWSKY TO ALL YA'LL!

i found the jerkulator i built, and listened to it a little... i think it has the flavor of the perc, but doesn't have the same sound.
pretty much the only difference i think is gonna be the diode clipper.

[digi2t]

Well I finally got mine to work. It wouldn't work because I did something so shockingly stupid that I can't even mention it. 

Way to keep us up at night, wondering...

[Paul Marossy]

Well I finally got mine to work. It wouldn't work because I did something so shockingly stupid that I can't even mention it. 

Way to keep us up at night, wondering...

Ha ha, right. I think this was the dumbest mistake I ever made. My days are just too long sometimes... 

[DougH]

How does it sound, Paul? How do you like it?

[pinkjimiphoton]

c'mon, paul fess up...

i bet i've done  WAYYYYYYYY dumber. probably on every build!!

the important thing is it's alive!!

we need clips!!!!