some fresh gut shots of an original interfax harmonic percolator

[pinkjimiphoton]

someone on facebook posted these pics today

[digi2t]

This looks to be the dual NPN version. Here's my "Connect-the-Dots".

[pinkjimiphoton]

nice, thanks brother dino!!

didn't know how useful it may be all these years out, but...

in for a dollar...

[Mark Hammer]

Is it me or are the two transistors individually encased in epoxy?

And what are the two clear things?  They look like wirewound resistors.

[Rob Strand]

is it me or are the two transistors individually encased in epoxy?

Perhaps they are different transistors being made to look like the same transistor.

Is there really and dual NPN version at all?   Is dual NPN a myth?

If those transistors were PNP and NPN then that circuit matches up closely with the common schematic.

And what are the two clear things?  They look like wirewound resistors.

They are very likely to just be carbon films.    Common garden carbon film resistors look
like that under the coating.

Go 1/4 the way down the page,
https://learn.sparkfun.com/tutorials/resistors

[bushidov]

I think Rob's correct. Looks like a carbon resistor with the coating gone.

[pinkjimiphoton]

Is there really and dual NPN version at all?   Is dual NPN a myth?

i don't think it's a myth, i've built them in the past, rob! this circuit will run on darned near anything, if the gain is right. even reverse beta'd stuff will work sometimes. its a weird circuit. i'm gonna be building a run of juergulators, i'll try n remember to make an all npn so you can hear it.

They are very likely to just be carbon films.    Common garden carbon film resistors look
like that under the coating.

Go 1/4 the way down the page,
https://learn.sparkfun.com/tutorials/resistors

nice trick!

[mozz]

Is it me or are the two transistors individually encased in epoxy?

And what are the two clear things?  They look like wirewound resistors.

Ive got Western Electric resistors that look close to that style.

[Rob Strand]

i don't think it's a myth, i've built them in the past, rob! this circuit will run on darned near anything, if the gain is right. even reverse beta'd stuff will work sometimes. its a weird circuit. i'm gonna be building a run of juergulators, i'll try n remember to make an all npn so you can hear it.

I don't doubt the circuit will produce an output.   I'm thinking more along the lines of what the (vintage) production units were.   I have vague recollection some clones used 2xNPN's but I was thinking they stuffed up by just following web info and web schematics.   Plenty of clone HP's sounded pretty bad.  The deciding factor would be someone measuring NPN/PNP'ness of those red transistor parts in a genuine vintage unit.   If it were a clone I'd still be skeptical  .

I rememeber Alphonso Hermida put up two schematics at one point.   It looked like it he traced those circuits.   Then there was a lot of DIY versions trying to get it to sound good, some deviating from the real unit but still sounding good.  Then some time later the George Giblet schematic came out, which at least represented one of the production versions.   However people still needed to tinker around with it because it's a touchy circuit.

Here's Alphonso's schematic(s),  both have one of each polarity transistor,  very early on 1998 to 2001,

https://www.diystompboxes.com/pedals/Percolator.gif

Then a big gap before the George Giblet schematic 2007,

https://www.fredric.co.uk/misc/HarmonicPercolator.sch.png

[digi2t]

Is there really and dual NPN version at all?   Is dual NPN a myth?

If those transistors were PNP and NPN then that circuit matches up closely with the common schematic.

I agree, but just for shits and grins, let's say it's a dual NPN. Say this was drawn up, based on a dual NPN unit...



Note: There was an error with this schemo, it showed the 91K resistor going to ground, rather than to positive. I took the liberty of correcting that.

Given that the schemo states that both transistors are 2N3565's (TO-106), then the two encapsulations being identical would lend some credence to the claim. I can't see one being a germanium since there would be a definite size difference between the two. Besides, the Si/Ge version is pretty much documented to death. Did the schemo's author cut them open? Seems a bit arbitrary to select 2N3565 as the transistor, so benefit of the doubt there. But... Rob's point still stands, maybe one is a PNP. Perhaps a PNP TO-106 package? Who knows. One thing for sure though, except for the ground/positive error, the rest seems to match up.

My first thought....

Breadboard. And que sera sera.

[Rob Strand]

two encapsulations being identical would lend some credence to the claim. I can't see one being a germanium since there would be a definite size difference between the two. Besides, the Si/Ge version is pretty much documented to death. Did the schemo's author cut them open? Seems a bit arbitrary to select 2N3565 as the transistor, so benefit of the doubt there. But... Rob's point still stands, maybe one is a PNP. Perhaps a PNP TO-106 package? Who knows. One thing for sure though, except for the ground/positive error, the rest seems to match up.

That matches up with my thoughts.    The TO-106 seems a good candidate it's one of the few common packages that will fit in something that size.   Another theory is the two encapsulations is design protection and perhaps some deception thrown in - in that two different things look identical.    A germanium TO-106 puts a spanner int the works though.   I can't imagine cutting tops of germanium transistor cases and sticking on new ones!   

It's not hard to work out PNP vs NPN and Si vs Ge from external measurements.

EDIT:
One thing I forgot to mention is notice the gold leads.   Those were quite common on TO-106 parts and some TO-18's.  I don't recall any AC128 style germaniums with gold leads.  I do recall a few germanium TO-18's with gold leads.

[digi2t]

EDIT:
One thing I forgot to mention is notice the gold leads.   Those were quite common on TO-106 parts and some TO-18's.  I don't recall any AC128 style germaniums with gold leads.  I do recall a few germanium TO-18's with gold leads.

Yup. I noticed that too, but neglected to mention it before. I've got a stash of TO-106 transistors, and all of them have those beefy gold leads. As for Ge transistors with gold pins, I seem to remember seeing some of those as well, but I believe the leads were skinnier. The TO-106 leads always tend to be thicker, sorta like the ones in Jimi's pictures.

[Rob Strand]

The TO-106 leads always tend to be thicker, sorta like the ones in Jimi's pictures.

Agreed, they are both similar as well.    I don't know about you guys but I don't remember germaniums in TO-106.  So maybe this version is all Silicon?

Jimi's pics show one collector resistor to + and the other to gnd like the PNP/NPN circuit.  A "more correct" dual NPN version would be like this one but now the collector resistor of the lower transistor needs to more from ground to virtual ground point ie. the big cap.

https://farm7.static.flickr.com/6010/5927031813_01ab0bb6ef.jpg

[pinkjimiphoton]

i'm telling you, just about anything will percolate... including reverse beta, npn pairs, pnp... its a weird circuit! lol

[Rob Strand]

its a weird circuit! lol

Indeed it is!   And there's 20+ years of internet tinkering and posts to prove it!

[garcho]

The Steward of Interfax himself has gone on record saying there has never been a dual NPN version, it's only ever been NPN/PNP Si/Ge. Not saying it matters, outside of historicity

[Mark Hammer]

Let's be frank, here.  The HP is yet one more circuit that degrades signal fidelity, albeit in a pleasing way.  There are MILLIONS of ways for an audio signal to be output with much less fidelity than it came in with.  So it should not surprise us in the least that all these variations in the basic HP circuit could and would result in distortion. 

The question for me is whether it results in the same kind/form/quality of "signal degradation" that the original NPN/PNP circuit does.  Not trying to be a cork-sniffer.  Rather, it's more of an if-you-do-this-you-get-to-keep-that-but-forfeit-a-bit-of-this approach that wants to know what such changes to the basic circuit alter in the waveform, and by how much.  The dual-NPN version might get someone 87% of what they like about an HP.

[Gus]

Is there really and dual NPN version at all?   Is dual NPN a myth?

If those transistors were PNP and NPN then that circuit matches up closely with the common schematic.

I agree, but just for shits and grins, let's say it's a dual NPN. Say this was drawn up, based on a dual NPN unit...



Note: There was an error with this schemo, it showed the 91K resistor going to ground, rather than to positive. I took the liberty of correcting that.

Given that the schemo states that both transistors are 2N3565's (TO-106), then the two encapsulations being identical would lend some credence to the claim. I can't see one being a germanium since there would be a definite size difference between the two. Besides, the Si/Ge version is pretty much documented to death. Did the schemo's author cut them open? Seems a bit arbitrary to select 2N3565 as the transistor, so benefit of the doubt there. But... Rob's point still stands, maybe one is a PNP. Perhaps a PNP TO-106 package? Who knows. One thing for sure though, except for the ground/positive error, the rest seems to match up.

My first thought....

Breadboard. And que sera sera.

Schematic looks wrong I would move the 20K at the first emitter to between the first transistor collector and the 47uf cap (remove the existing connection) move the 1uf? coupling cap to the 1st collector to 2nd base
Is the output volume wired like in the picture  or is that a mistake as well. It will work as shown but in an odd way.

I built a dual NPN Silicon it works at higher current.
https://www.diystompboxes.com/smfforum/index.php?topic=91764.0

Simmed the circuit digit2t linked with the corrections using 2n5089s Q2 C 2.7VDC, Q1 C .67VDC

[digi2t]

Let's be frank, here.  The HP is yet one more circuit that degrades signal fidelity, albeit in a pleasing way.  There are MILLIONS of ways for an audio signal to be output with much less fidelity than it came in with.  So it should not surprise us in the least that all these variations in the basic HP circuit could and would result in distortion. 

Let's be franker still... signal fidelity degradation doesn't necessarily need a circuit. Me plugging a guitar straight into an amp and playing can do a fine job as well, albeit in a less pleasing way. Sometimes... the circuit just helps. 

[pinkjimiphoton]

Let's be frank, here.  The HP is yet one more circuit that degrades signal fidelity, albeit in a pleasing way.  There are MILLIONS of ways for an audio signal to be output with much less fidelity than it came in with.  So it should not surprise us in the least that all these variations in the basic HP circuit could and would result in distortion. 

Let's be franker still... signal fidelity degradation doesn't necessarily need a circuit. Me plugging a guitar straight into an amp and playing can do a fine job as well, albeit in a less pleasing way. Sometimes... the circuit just helps. 

and in the end, this is the only actual frank assesment lol

@gus, if ya move those components, is it still a percolator? those components are actually key to the sound of the unit...less the values than their place in the signal path.

i've built a bunch of variants of this circuit, including both npn. my pref is the npn/pnp but truth be told, harmonically, most folks couldn't even tell the diff in the first place.

its funny, watching the debate about this on facebook right now...

but anyways, i suppose maybe if i get a chance i'll have to build one of these suckers and post some video if i ever get the time to cobble it together.

like dino said... time for someone to take the breadboard challenge.

i would be surprised to see 5089's work well with this circuit. gain is far too high with most of them.

use low gain si, keep it under 200 and you'll find a nite and day difference tonally. above 400 or so it sounds more like escobedo's jerkulator to my ear.

but everyone's different. what pleases one may not work so well for another.