Harmonic Perculator

[vanessa]

Two versions? Could it be that they switched over to a two NPN (si) design later to rid the effect of issues with thermal runaway (ge transistor)?

[johngreene]

Two versions? Could it be that they switched over to a two NPN (si) design later to rid the effect of issues with thermal runaway (ge transistor)?

If I had to guess I'd say the two NPN version came first because the board is laid out for it. The board with the Ge PNP has traces for the collector feedback resistor but isn't drilled for it. It also has the pads for the input diode and isn't drilled for it. The 220K resistor uses one of the pads that was for the 1K + 2.2uF. The other pads for these parts are there but not drilled.

--john

[George Giblet]

> I have a feeling that it varied quite a bit over the life of the pedal.

Yes, it definitely looks that way.

> I didn't see 'notanotherschem.gif' anywhere. So I can't comment on it.

It's basically this one.  I've marked up all the corrections.  It is now complete and should at least work.

http://www.geocities.com/george_giblet/effects/percolator_later.png

It does look like this version has patched components onto an old PCB so I agree with you that this is probably a later incarnation.

I think the barge version is correct for that (earlier) version.

http://www.bargeconcepts.com/bp/largeschem.bmp

[johngreene]

http://www.geocities.com/george_giblet/effects/percolator_later.png

In my opinion, your schematic is -the- schematic for what everyone considers the harmonic percolator. You should put your name and a copyright notice on it.

--john

[George Giblet]

> You should put your name and a copyright notice on it.

I purposely left it off!  I can't take credit for it too many people have contribute in some way shape or form - including yourself (thanks for the comments by the way)!

[Dan N]

O.K., I pulled and switched parts to match the later schematic. I left the 20K and used film for the 0.1's. To my ears, it kills the earlier version. Sleeping people keep me from further testing.

Way to go, guys! I believe we got us a Percolator!

edit- Again...

[johngreene]

> You should put your name and a copyright notice on it.

I purposely left it off!  I can't take credit for it too many people have contribute in some way shape or form - including yourself (thanks for the comments by the way)!

Seriously, I think you should. I don't have any problem with it. You made the effort to commit it to schematic. I was too lazy to do it. So you should claim credit for what you publish. It is that simple. If I actually -wanted- any credit, I would have published a schematic myself.

However, if you would think it necessary to give me credit in your 'copyrighted' schematic, I would be honored.

I think this is the true 'goal' behind DIY. It is just to receive credit where credit is due. Nothing more.

My opinion.

--john

[mac]

I wonder what would happen if the input were inverted, incoming signal to q2 base and from its collector the 100nf to q1 base.

mac

[johngreene]

I wonder what would happen if the input were inverted, incoming signal to q2 base and from its collector the 100nf to q1 base.

mac

uh, wut?

[George Giblet]

OK, John you have twisted my arm.

Here's an update with credits.  I've also cleaned a few things up and put the control names on the schematic.

http://www.geocities.com/george_giblet/effects/percolator_later_rev1_1.png

(I will pull the uncredited schematic).

[George Giblet]

> O.K., I pulled and switched parts to match the later schematic.

Cool Dan, thanks for testing it out.



One idea which kind of detracts from getting the schematic right:   I suspect the unit will sound OK with a silicon PNP, it should work although tweaking the 220k on Q1 might improve the sound.

[George Giblet]

I did some simulations to check the bias points, gains and general behaviour:

- The circuit behaviour is fairly independent of the 220k resistor at Q1 C-B.
- 10k to 20k on Q1's Collector seems to maximize the gain.  So the 20k looks pretty good from that respect.
   This resistor does change the swing to some extent so it will change the clipping characteristic and sound.
- The cap between Q1-C and Q2-B tunes the low frequency response.  LF Cutoff is currently around 120Hz.
- The circuit seems to hold-up as it stands with both silicon and germanium transistor in the Q1 position.
- Biasing looks quite stable.  I wouldn't be too worried about changing the 750k to 680k or 820k.

Haven't done any side by side comparisons of Ge vs Si in the Q1 position yet.

[markusw]

Here's an update with credits.  I've also cleaned a few things up and put the control names on the schematic.

Thanks a lot!! 

I compared both versions on breadboard. The 2x NPN with two 2N2222's and the PNP/NPN with different trannies for Q1 and again a 2N2222 for Q2.
Again a bit trial and error.....
A AC128 (hfe 70) or AC188 (hfe 150) for Q1 defintely sounds better for me than a 2N2904 (hfe 75).
Interestingly, with both the AC128 and the AC188 for me it sounds better if the trannie is used the other way round (i.e. emitter to 20k, collector to emitter of Q2). Don't know if it makes any sense to use it the other way round 
Changing the 20k to higher or lower values didn't really change the sound.

Comparing the two versions it's hard to say which one sounds better.....

Regards,

Markus

[RLBJR65]

Thanks Guys!

I get up much earlier than anyone else in the house so I whipped up a layout and built one this morning. I used perf but followed this PCB. http://aronnelson.com/gallery/albums/album133/Harmonic_Perculator.gif

I have not had much time to experiment with it and probably won't until next weekend. Sounds good so far though.

Here is what I used.
Q1 - 2N404A Hfe - 76
Q2 - 2N5088 Hfe - 486
D1 & D2 - 1N60
R4 - 22K
Had to make the 91K and 750K came up with 90.8K and 750.1K. Can't get much closer than that
No 50K pots so I used the closest I had an oddball 30K . Shouldn't matter much with this circuit right?

[mac]

Haven't done any side by side comparisons of Ge vs Si in the Q1 position yet.

I've done it, but it seems that no one read my post

mac

[Dan N]

I've done it, but it seems that no one read my post

I think we all just have to try it ourselves. DIY fun and games...

[George Giblet]

>I've done it, but it seems that no one read my post

I read the link to your old post,
http://www.diystompboxes.com/smfforum/index.php?topic=49035.0

but I thought you were changing Q2 with Si and Ge?

There's some interesting results about the leakage.  (Using references from Alphonso's circuit) You mentioned that R5 and R7 only have a subtle effect, they appear to me only to have a subtle effect too but I haven't spent a lot of time with it.  If biasing changes due to leakage were the cause of the sound differences I would have expected the R5/R7 adjustments to compensate to some degree.  However the leakages you quoted are quite large so perhaps the leakage can be an issue.  The base currents in the transistors are going to be in the order of 0.5uA to 1uA.  If you take your 2sa101 the base referred leakage will be about 151uA/42 = 3.6uA!  whereas the 2sa53 would be 0.38uA.  Your 2sd352 NPN will have a base refered leakage of 151/50 = 3uA  which indicates the leakage is relatively high compare to the current through R5.  I suppose in summary since you like the high leakage Q1 configuration better it might be worth lowering the 220k resistor for low leakage transistors (only speculating).

I noticed you played with the diodes too - that's another area to try!

One interesting thing is Alphonso's V1 has a low R1 (on his circuit) and V2 has a high R1.  The new version I put up is in between these.

[mattpocket]

I havent made any circuits that needed biasing before...

What does it mean?

When I have seen it on other circuits, ge fuzz's and the like, it has been done by tweaking a trimpot to change the resistance between two points. What does it normaly involve? Something across the tranny or something? Enlighten me please!

I had a quick look at the schem and I cant see a trim pot is it set with the 220k resistor across the tranny?

Matt

[tcobretti]

I havent made any circuits that needed biasing before...

What does it mean?

When I have seen it on other circuits, ge fuzz's and the like, it has been done by tweaking a trimpot to change the resistance between two points. What does it normaly involve? Something across the tranny or something? Enlighten me please!

I had a quick look at the schem and I cant see a trim pot is it set with the 220k resistor across the tranny?

Matt

Biasing is setting up a network of resistors to make the power run the right way thru a transistor so it will work.  Different types of transistors have different biasing req's.

You should probly read these articles. 

http://www.diystompboxes.com/smfforum/index.php?topic=34611

http://www.diystompboxes.com/smfforum/index.php?topic=48578.msg360118\#msg360118

[George Giblet]

Biasing is simply the setting of the DC conditions of a circuit.   The DC conditions is the DC voltages and DC currents in the circuit.  When people talk about biasing they are usually focussed on one particular aspect.   For example in a tube amp, biasing often refers to the DC current in the output tubes.  However depending on the discussion biasing can refer to the DC voltages on the plates of the preamp tube.  There are of course other DC voltages and currents elsewhere in the circuit, and these are bias currents and voltages, but they aren't  of primary interest.  In effects biasing often means the adjusting something in the circuit to get a particular (desired/wanted) voltage on the collector of a transistor or drain of a JFET.   Sometimes you see people talk about adjusting or changing the biasing of a circuit, all that means is you are changing *something* which will affect the DC conditions in the circuit - eg collector voltage or collector current.

Here's a few examples of transistor circuits to show how playing with the parts can affect the biasing and behaviour of the circuit.

http://www.geocities.com/george_giblet/effects/bias_example.png

In the rightmost circuit the 470k resistor is changed to 220k which causes the transistor collector current to increase and the collector voltage to drop from 4.7V to 3.6V.  The gain of the circuit just happens to stay the same as the leftmost circuit.

In the bottom circuit the 4.7k resistor has been increased to 8.2k.  The collector voltage is about the same as the rightmost circuit.  So the "collector bias voltage" is the same.  However changing the bias in this way also changed the gain of the circuit.

What all this means is:
- Biasing is affected by the various parts of the circuit.
- The behaviour of a circuit depends on all the parts even though a particular aspect of biasing is the same.  In the bottom circuit the transistor collector current (which can be called a bias current) is lower than the rightmost circuit.  The reason the gain changes is *not* because the transistor current is different but the fact that *in that circuit* the gain is determined by the external parts and not the bias conditions.