Essential Harmonic Percolator Q1 PnP/ Q2 NpN Transistor gains?

[Bandwagonesque]

Hi everyone. I realize this has been beaten to death but ive been spending the last couple nights spending hours studying everything I can about the harmonic percolator before a build this weekend and will be going to my parts depot tomorrow with my meter/transistor tester in hand in hopes of taking home a bunch of Q1 Germanium PNP's and as many Q2 2365's/2222 silicon NPN's I can. If i'm studying things right, it seems that the school of thought is to use whateeever low gain (between 40-100 ufe) leaky PNP's I can find for Q1 and then sticking reallll high gain 2365's in the Q2 position...anything above 250 hFE if higher.

But I am reading other build reports of people doing the opposite with the high gain in the pnp position and then low gain silicon's in Q2. So I am a bit confused even though this approach seems to be in the minority.

Currently I have a russian made (I think a 2n527) PNP germ measuring at 45 hFE in Q1 and a plastic package 2365 at around 350 hFE in Q2 on the breadboard and the results are quite pleasing. Not as scrambled as I hear on demos and the octave effect I hear I hear on demo's is more subdued on mine. Cleans up amazingly...but I do want to get the most out of this as it seems this circuit is reliant more on the transistor combo than anything else including the diodes. And ive never played an original or a clone or even a Harmonic Percolator to test my ears personally. All very much virgin territory.

So while I apologize for the waste in space here and not sticking this in another thread, I do hope that someone as green as I can come across this in the future if they face the same confusion.

But could someone who has some real experience with this circuit lay down the law for me before I spend any money tomorrow on parts? Would love to hear anything I can about the 'right' way to do this. So far my ears doing the walking are leading me in the right direction so far I believe, but you lot are far smarter than I and appreciate any input I can get.

Have a great weekend everyone!

[iainpunk]

i have a lot of experience experimenting with HP's and have even come up with my own unique version, the Harmonic Instant Coffee, if you are interested in a unique version that has more BassFuzz type bias diodes:
https://www.diystompboxes.com/smfforum/index.php?topic=126336.msg1207629#msg1207629

a friend of mine has an original first version (he's a collector) i tried it and it had no octave what so ever, really smooth and overdrive-ish, but that differs a lot to the later unit, i have been told.
with fuzz pedals, i think you should rely on you ears, not on Hfe's and other numbers.
a good transistor is a reverse bias 2n2222A without the bias resistor. it works on my breadboard, with all 3 i tried, but your mileage may vary.

i recommend just getting a few transistors of a bunch of different types, so you can mix and match to your own ears.

cheers, Iain

EDIT: i totally forgot to answer your question;
a leaky low gain Ge PNP and a high gain Si are in the first version of the original, without bias resistor for the PNP,
later, a less leaky higher gain Ge PNP and a lower gain Si was used, which sounds raspier and uses an extra bias resistor

[mac]

I did the maths in one post here to understand DC bias, but I can't find it.
Last time I experimented with it was 15 years ago.
If you want octave, a resistor between emiters do it.

The 2N2222 reverse beta sounds interesting. IIRC it has negative resistance this way so it's used for oscillators.

See in my gallery the all Germ version with octave.

mac

[garcho]

a friend of mine has an original first version (he's a collector) i tried it and it had no octave what so ever, really smooth and overdrive-ish, but that differs a lot to the later unit, i have been told

I built the George Giblet schematic with Small Bear transistors (2N404 and 2N3565). I didn't measure or mod a single thing and it's been the only overdrive I've played with for years now, I love it. No octave, very smooth and overdrive-ish indeed. Cleans up nice and plays well with the guitar's volume knob. Put it 1st in chain/before buffer.

[digi2t]

There's a pretty big thread on the HP here; https://www.diystompboxes.com/smfforum/index.php?topic=99596.0

My experimentation with this circuit can be narrowed down to these posts;
https://www.diystompboxes.com/smfforum/index.php?topic=99596.msg874830#msg874830
https://www.diystompboxes.com/smfforum/index.php?topic=99596.msg874870#msg874870

The entire thread is chock full of useful info, and at the end of the day, it's simply a matter of rolling transistors through it until you hit on the magic combo that tickles your ears the best. This is not a "right or wrong" affair, simply "try and try again". Even in cases where voltages match up, different transistors can change the soundscape.

All in all, it's a circuit that allows for a ton of combinations and permutations, even where resistors and caps are concerned. Yeah... I'm not helping your cause, but that's the straight bottom line. In my book, there's no "standard" HP build schematic. Only "my HP is the best sounding HP" schematics.

Video of my breadboard testing from the big thread;

If this peaks your curiosity, then you may even wish to try a full germanium HP as well. It's a different flavor, for sure, but just as interesting.

[Mark Hammer]

Once upon a time, there would be a benchmark unit of a fuzz, and someone like RG would take all the measurements and identify what voltages needed to be met.  The components in the benchmark unit would determine what the voltages and currents needed to be, and smart minds could/would provide guidance regarding how to meet them with different transistors, by adjusting this or that component value.

The Percolator, however, doesn't really seem to have a benchmark unit. Most successful builds sound pretty decent.

Dino has it right, as does Iain.  Install sockets for the transistors and plug in different units until you hit something you like.  The great part of the circuit is that "sounds great", much like enlightenment, can be arrived at via many routes.

[Gus]

Is there a way to collect all the HP threads?
This would make it easier for the newer people to find the information.

You can have fun with this circuit. I have built a 2N2222, 2N2907 and a Dual NPN 2N2222, 2N2222 version.

Others have posted different takes on the circuit

The interesting things on this circuit is how the two gain stages are powered in series so both operate at lower than 9VDC.
They interact at the middle so resistance and cap size are interesting to adjust at that node.
The current(transistor and resistor currents added together) in both stages is the same due to the biasing used.

[PRR]

Is there a way to collect all the HP threads? ....

Harmonic Percolator site:diystompboxes.com
https://www.google.com/search?q=Harmonic+Percolator+site%3Adiystompboxes.com
About 355 results

[iainpunk]

the fun of the HP is that almost every bias point can sound awesome, if the transistors are right with each other.

Dual NPN 2N2222, 2N2222 version.

do you have a schematic?

They interact at the middle so resistance and cap size are interesting to adjust at that node.

well, if they interact at that node, you have oscillations, so the cap value needs to keep the voltage fluctuation at that point really close to zero.
if you have oscillation issues, a single 1k resistor on one of the gain stage's emitter to the capacitor helps a lot to tame it.

The Percolator, however, doesn't really seem to have a benchmark unit.

i could ask Daniel if i can take measurements on his HP, he's probably hesitant about opening up the case (he is extremely autistic, and has a hard time dealing with others handling his stuff), but asking won't hurt.
that way, we might have an official benchmark.
my experience is that it sounds most Punk when the middle cap has a voltage between 6v and 7.5v, and i suspect the original* would be around 2v to 3v. i think the higher the voltage, the more prominent the octave, this makes me want to add a 250k dual gang pot, each side being a collector resistor, having 2k2 stopper resistors.

cheers, Iain

* with the original, i mean version one on this schematic:

[Gus]

Dual NPN in the following link. It not a HP with the Si NPN Ge PNP however, it is something fun to build.
I noted the reading from the Harbor Freight DMM with a transistor checker socket on the screen shot.
I guessed a number of people used this DMM(or one like it) at the time this was posted.
Different testers will give different numbers due to how they are set up.

I also reduced the collector transistors to operate the transistor at higher current.
One can increase the collector resistor values and adjust the C to B resistors for biasing

https://www.diystompboxes.com/smfforum/index.php?topic=91764.msg784041#msg784041

[iainpunk]

https://www.diystompboxes.com/smfforum/index.php?topic=91764.msg784041#msg784041

the image in your original post doesn't load, and the layout makes no sense, there is no bias for the first transistor, and the input signal gets sent to the 2nd transistors B as well as the first transistors B.

cheers, Iain

[Gus]

https://www.diystompboxes.com/smfforum/index.php?topic=91764.msg784041#msg784041

the image in your original post doesn't load, and the layout makes no sense, there is no bias for the first transistor, and the input signal gets sent to the 2nd transistors B as well as the first transistors B.

cheers, Iain

The image loads on my PC

The .1uf goes from Q1 collector to Q2 base I don't understand what you are posting.
1meg C to B for Q1
470K C to B for Q2

[iainpunk]

https://www.diystompboxes.com/smfforum/index.php?topic=91764.msg784041#msg784041

the image in your original post doesn't load, and the layout makes no sense, there is no bias for the first transistor, and the input signal gets sent to the 2nd transistors B as well as the first transistors B.

cheers, Iain

The image loads on my PC

The .1uf goes from Q1 collector to Q2 base I don't understand what you are posting.
1meg C to B for Q1
470K C to B for Q2

it loads now, but not before, i guess the site hosting it was down for a few minutes

i was trying to decypher this image:
but there seems to be a trace missing,

that schematic looks quite nice, ill try it Tuesday

cheers, Iain

[Bandwagonesque]

you guys are just fantastic and thank you so much for all the replies and taking your time to give me some reallll good feedback. really can't say thanks to the lot of you enough, can't wait to dive into this stuff in practice after i finally grab some transistors this morning <3

[bmsiddall]

Greetings from Australia.

As it happens, I'm in the middle of another round of HP Q1 transistor swapping/tuning.  I don't think the 2N404 are an absolute must have as I've had great tones from a variety of types, hfe and leakage.  I've tuned my circuit to give the octave down from about the 12th fret (lower E) down to about G# same string- put a 5M trimpot in series with the 750K on Q2 and crank it (currently about 3.4M i think?).  Q2 is a 2SC536 (179 hfe). 

ASY25 sound best around 50-60 hfe and 50uA or so leakage.  Go higher (75 or so) and fuzz/gain is reduced- also loses the octave down.
2N404 sound good all the way down to 12 hfe (!).  Leakage varies but the one I was leaning towards is about 105hfe, 300ua leakage.
P416 are very low leakage but sound ok up to 70hfe, if a little burbly.  i think better if i biased Q1, but have fixed resistors here.
2SA101 damn good- favorite up until a few days ago has 34hfe, 55uA.
2SB171 also excellent.  tossing up whether to keep one in the HP (42hfe, 60uA).  slightly darker/more mids than 2N404??
MP21E measured between 30-50hfe and leakages 50-90uA were also bloody good.  I haven't had much luck previously with Russians in this cct, but impressed with these right now.  Best one so far measures 50hfe, 60uA- killer!
**EDIT: just tried 2 more MP21E.  Both 46hfe, one 55uA, the other 144uA.  the lower leakage was again killer, but the higher leakage lost fuzz/gain/octave down.

It's not scientific by any means, but I've noticed that generally when you're 60-90hfe (and low leakage) in my circuit the gain drops and tames the circuit.  Lower hfe and low leakage better.  Higher hfe/higher leakage also better.

[rankot]

I made one HP with Albini values, sounds good, but I didn't take much time for transistor swapping. Since I have another PCB for this, I might take another hit. This discussion is really great!

[iainpunk]

Greetings from Australia.

As it happens, I'm in the middle of another round of HP Q1 transistor swapping/tuning.  I don't think the 2N404 are an absolute must have as I've had great tones from a variety of types, hfe and leakage.  I've tuned my circuit to give the octave down

octave DOWN?!?!?!?!?!?!?!?!
i have build a few Harmonic Percolator bastardizations with octave up sounds, but octave down is a new one!
how do you achieve that?

also, what are your emitter voltages like? have you tried changing the emitter voltages through collector resistor of Q1?
i found that having the same collector voltage with different trannies puts you in the same ballpark, sound wise, might have a bit more gain and grit, but the overall profile seems quite consistent with the same voltage

cheers, Iain

[Rob Strand]

This 1978 fuzz article by D.McCabe makes it clear how biasing can affect the part of the transfer curve you are on.    It seems more applicable to Bazz-Fuzz than the Percolator.    The Percolator biases the transistors with low Vce, a few other old pedals do the same trick.

https://worldradiohistory.com/hd2/IDX-UK/Technology/Technology-Modern/Archive-Practical-Electronics-IDX/IDX/70s/Practical-Electronics-1978-10-OCR-Page-0052.pdf
https://worldradiohistory.com/hd2/IDX-UK/Technology/Technology-Modern/Archive-Practical-Electronics-IDX/IDX/70s/Practical-Electronics-1978-10-OCR-Page-0053.pdf

[rankot]

Just finished another one for my late friend's little son - those are the values. It seems to be singing even nicer than the first one, which was closer to Albini specs. This one is almost classic HP. I used two pots to find sweet spots for collector resistors:
Q1 MP20, hfe 52, Vc=4.67V
Q2 2N6517, hfe 140, Vc=5.84V

Resistor values, following schematic above:
R1 750k
R2 220k
R3 39k
R4 56k
R5 2k2
D1, C5 and C7 not installed, C3 is 1u.

The same comment is here, in similar thread: https://www.diystompboxes.com/smfforum/index.php?topic=99596.msg1208314#msg1208314

[iainpunk]

This 1978 fuzz article by D.McCabe makes it clear how biasing can affect the part of the transfer curve you are on.    It seems more applicable to Bazz-Fuzz than the Percolator.    The Percolator biases the transistors with low Vce, a few other old pedals do the same trick.

https://worldradiohistory.com/hd2/IDX-UK/Technology/Technology-Modern/Archive-Practical-Electronics-IDX/IDX/70s/Practical-Electronics-1978-10-OCR-Page-0052.pdf
https://worldradiohistory.com/hd2/IDX-UK/Technology/Technology-Modern/Archive-Practical-Electronics-IDX/IDX/70s/Practical-Electronics-1978-10-OCR-Page-0053.pdf

wow, thats a really cool article! i might just build it since i have all the components in there, the acompanying article is really nice as well, t really makes me have a different view on the Bazz Fuss, maybe ill add some resistors to the first gain stage of my Harmonic Instant Coffee to have some more sustain, and less of a gating effect when the 2nd transistor is set really nasal and harsh.

cheers, Iain