Harmonic Percolator (George Giblet's schem) build report

[Rodgre]

I built up George Giblet's version of the later Harmonic Percolator.

So far so good. It sounds great through my test bench amp, but the proof is in the pudding (or probably my AC30). I put in a socket for the Ge tranny and went through a few different trannies until something sounded special. All that I tried sounded pretty good and had a good range of cleanish to overdrive. Some were random unmarked ones I inherited somewhere. Some were ECG 158s that I bought years ago. So far the winner is some one-off purchase of an unknown transistor that I got from Electronic Goldmine a year or two ago. This one had way more gain and when the input control was maxed, it went into a fantastic sounding fuzz that would compress and swell up in a very cool and splatty way. It still had a nice range of clean with the input control dialed down (or my guitar's volume control).

Anyone else have any reports about this version?

Roger

[foxfire]

i played with that circuit while back. i liked it but ended up making some changes. i don't have my folder with me though. i do remember it was a bit thin for me. i'll check my notes when i get a chance. rylan

[Rodgre]

Since I haven't tried it through a "real" amp, I can't tell if it's thin or not, so I'll look out for that.

Do you recall if you tried transistor substitutions, or was it more or less tailoring capacitors and resistors until you liked it better?

What interests me in this circuit is the non-Albini video clip which shows how the harmonic character of it lets more complex chords cut through. I'm not looking for something completely clean and clear (I still haven't found a use for my Roland GR-300's hex-fuzz!) but I would like to have one pedal that has a unique tone among the pedals I already have (which, of course is already too many).

Roger

[alex frias]

I just passed by a compulsive Percolator seeker phase...

I tested several component values and trannies.

But I must admit, until now the best sound I achieved with Harmonic Percolator circuits was the Doug hammond's "The Brick" basic scheme, but using a diferent clipping/diode section.

Thereís something very special about that simple circuit. It's diferent, it gets good results with a good clean amp set, but with a good tube crunched amp it really delivers!!!

[George Giblet]

Just for the record, this circuits isn't my version, it's a trace of the actual circuit.  There's a few versions of the actual circuit and that appeared to be the one people like.   Because of  transistor differences you might need to play around with the values.  Look in the archives for some tips/ideas.

[Gus]

Has anyone measured the DC voltages at the
Emitters node VDC?
Q1 collector VDC?
Q2 collector VDC?

This will help understand what is going on with a nice sounding version of this circuit

Interesting circuit both transistors are forced to run at the same collector currents being in series and having no other DC current paths because of C2, C3 and C6.

It is a 2 stage circuit
  First stage gain, Q1s gain with the feedback network of the 220K C to B / 100K input pot and source impedance (guitar)

  2nd stage gain, Q2s gain and the 750K feedback C to B /  R4 value

The size of the C4 cap(I think of it as an AC ground) will have an effect on how much the emitter node moves with signal

Also the voltage at the emitter node has an effect on the max gain from each stage by how it divides the 9VDC supply.

It is a two stage amp Q1 then Q2 at a lower than 9VDC for each stage

Looks like it would be sensitive to the transistors used and R1,R2,R3 and R4

[George Giblet]

Some of that behaviour was addressed in the original thread,

http://www.diystompboxes.com/smfforum/index.php?topic=55987.0

The circuit as is had the maximum gain and I seemed to recall the Vce voltages of the transistors were quite low.  The key could be how near saturation the transistors operate, the problem is large changes in the external components make small changes to VCE in that region.  Variations in the GE versions are likely to be related to leakage.

[Solidhex]

Yo

  Yeah I built one using that schem and a layout I did in Eagle. Used all the recommend values and types of components. Socketed the transistors and tried a bunch in both positions. All that seems to come out of mine is a pretty mediocre fuzzy overdrive. Maybe I'll rip out the circuit and build "The brick" in there.

--Brad

[Gus]

George

      I did not notice the 50ua total current draw before.  Good post 2nd page from your link.

  If I wanted to build a close version of this circuit I would make two circuit fragments to select transistors and C to B resistors for each stage.

Using voltages worked out by George in one of the posts.

Q1
A 10K and 50nf in series with the base, 20K collector resistor and fixed resistor and a pot wired C to B
AND a Power Supply set to 3.48VDC (from George's post).
  The 10K is a crude guitar sim with this type of biasing you are also affecting the feedback ratio and it also works with the open loop gain of the transistor.   So drop in a transistor turn the pot and look for 1V from ground to the collector if the feedback resistor is much greater or less than 220K try a different transistor. The input resistor and cap is for testing using a low output resistance signal generator.  2.48VDC  E to C


Q2
  A 20K (collector resistor value of Q1) in series with a 100nF to the base, a 91K collector resistor, a fixed resistor and pot for the 750K collector to base feedback resistor.
  A Power Supply set to 5.52VDC.
  Adjust the feedback resistor for 0.97VDC(4.45 -3.48).  The input resistor and cap is for testing using a low output resistance signal generator.  Q2s bias point is closer to saturation and set for more gain(.97VDC C to E)

Pick your transistors and feedback resistor and drop them in a board.  You don't want the feedback/bias resistors to be much different than the stock values if you want to build it close to the original because they affect the stages gain.

If you understand more about transistor you could pick other collector current and hfe transistors

Again has anyone measured a real good sounding HP operating points?

[Gus]

Did a DC test of the fragments and then with them connected together.

I find this circuit interesting because of the interactions.  Took two random, common, Si transistors 2n2222a (hfe 210) and 2n2907 (hfe 190).  Changed the Ic and collector resistors for more drive in this case around .1ma (100ua).  Selected the C to B resistors for the voltage at the collectors and therefore the Ic I wanted

Q2 fragment 47K collector, 470K C to B, 5.5VDC supply  Collector at .957

Q1 fragment 10K collector, 1.2 meg C to B, 3.53VDC supply, + to emitter, Collector at 1.057VDC ref to ground

Connected emitters and used a supply of 9VDC using Georges schematic as a starting point
Q1 collector .986VDC
emitters 3.42VDC
Q2 collector 4.37VDC
+ 9.02VDC

NOTE gain of Q1 and Q2 stage will be different: However the division of voltage is close to what George calculated and posted.

I think I will try Si diodes for more output when I box it up to hear what it does.

[Gus]

Are there no replys because people maybe don't understand what I have posted?  If so ask.

  To me this circuit should have both gain sections tested before building because of the series connection between the gain stages.
Ic collector current sound be close for both fragments and the gain in each stage is affected by both the transistor gain and the feedback resistors (C to B resistor and the source resistance feeding the base,  This also reduces the transistor distortion being feedback).
Note the signal inversion and in then at Q1 collector then at Q2 collector.

[DougH]

I went back through the old thread that George linked above. I didn't understand his "bypassed emitter" comments completely until I realized that each transistor is the emitter circuit for the other transistor. Almost like an "active load" in a cascode or mu-amp circuit- except it's in the emitter not the collector circuit.

I never measured Vc in either stage when I built the Brick. I just biased it by ear. DC-wise it is similar to George's schem except for the 100k in place of the 20k Rc in the first stage. AC-wise it is tuned differently. Also, the diode "offset resistor" in George's circuit is 4.7k instead of 10k in mine. I would like to breadboard George's circuit, check voltages and then compare with the Brick sound.

[dangerboy]

I've built three of these now to this schematic, and they have all sounded so good that people wanted to buy them. I make sure to get the Hfe's in the range discussed on the other thread, and they work great. I tune all 4 resistors, then basically end up back with the values George has given. This is using some East German Ge PNP Hfe 40-50, and a 2N2222A 180-200 for the Si.

[Rodgre]

I've built up three versions recently, and I've come up with a variant that I like a lot. I've changed a few things that might make it worth posting my schematic when I get a chance.

I built up 90% of the Brick version, but I was left with a question about the 1uf caps: whether they are polarized or not. I'd probably have to use 2 .47uf caps if they're not polarized. I don't think I've ever had such a large cap in my parts bin that wasn't electrolytic or tantalum.

Roger

[DougH]

Just use film caps for the 1uf's and it's a done deal. They are not polarized.

FWIW I had a little time this morning and built up George's schem on the breadboard. I used a 2n404a hfe 30 for Q1 and a 2n3565 hfe 220 for Q2. My voltages came out a little different than Gus came up with- I had 2.7v on the Q2 collector for example. Still sounded fine though. With R4=22k it is pretty wild - loud, lots of pinch harmonic emphasis, slightly "mis-biased" sounding even when gain and/or guitar is turned down. With R4 = 100k it is not as loud, sounds more "together" (not mis-biased) , but in either case just as fuzzy with the gain cranked. I had to run to the dentist before I had a chance to compare to the Brick, which is tuned differently (1u's instead of .1u's, etc).

I have another 2n404a as well as other assorted Ge's and Si's. I'd like to experiment with different transistors and see how they affect voltage levels and sound. I think what sounds "good" may be subjective, and maybe the Albini vid could be used to compare. I suspect the "wilder" it sounds the more it will sound like the vid.

[birt]

i made quite a few and like them best with both transistors in the 80-100 range.

[danielwarner]

I've been looking at various incarnations of this circuit and I'm curious as to why George Giblet's schem only has two diodes. In the photos posted on Harmony Central, there are three diodes on the PCB. Is it just a different version of the Percolator?

[mac]

I went back through the old thread that George linked above. I didn't understand his "bypassed emitter" comments completely until I realized that each transistor is the emitter circuit for the other transistor.

I do not remember if I mentioned in the above link a mod I did to mine: a 100k pot between emiters --> kind of octave up fx. When pot is not at min, Ge emiter is fully bypassed, but Si emiter is not. (Mine has two Ge, 2sa102 and 2sd352, sounds a lot better)
Also, a fixed resistor between emiters of the order of the collector resistors, and no bypass cap turn the circuit into a nice clean even harmonics booster 

mac

[DougH]

I've been looking at various incarnations of this circuit and I'm curious as to why George Giblet's schem only has two diodes. In the photos posted on Harmony Central, there are three diodes on the PCB. Is it just a different version of the Percolator?

That's D5 in Alf's version 2 schematic. Looks like input overvoltage protection. Those photos also show the 1k-2u2 from Q1-C to ground.

[DougH]

I compared George's circuit on the breadboard to my Brick. I much prefer the voicing of George's circuit. With the .047u, .1u, .1u caps it sounds fuller- the Brick is just too thin with the .01u input cap. However, the Brick is *much* louder, due to the 1n46 clipping diodes I used.

So far I'm drawing much more than the 50uA current in the spec sheet. I'm going to experiment with lower hfe npn si transistors and see if i can get the current down to see how that will affect the sound.