Harmonic Percolator Question

[Paul Marossy]

I did a little experimenting last night.

On the emitters, I had two 22uF caps in parallel instead of the single 47uF cap. I put one of them on a switch. It sounds more "open" to my ears with the 47uF cap.

I also put the output diodes on a switch and used a 6.8K resistor in series with the one diode that had no resistor. I was still getting an octave up sound without the diodes even being in the circuit, but it was accentuated more with the clipping diodes in the circuit.

So I have come to the conclusion that it's because of the Q1 & Q2 arrangement. Looking at stuff on the scope yielded nothing definitive. Someday I need to get a more modern scope.

When I have the harmonics knob at max and I hit the strings hard, it sounds like it's extremely distorted, and I think there is some crossover distortion going on as well. When I play softer, it sounds quasi-tubeish. If I play just a little harder, that 2nd harmonic octave up thing starts to come out.

What I want to do is find a way to dial in whether or not you get any 2nd order harmonics. I don't know if it's possible with the circuit as designed.

[midwayfair]

I think it is two cascaded stages.

I asked on Madbean couple weeks ago and a people suggested that it's a Colpitts VCO. The Colpitts is a nearly identical transistor arrangement, enough to make me think that it's at least where the designers started.

Paul: I get almost no octave effect in mine in the setting with symmetric diodes Fv ~6.5 (OA126). The standard germanium setting (1N695) with the series resistor gives me some octave, but only with the harmonics cranked (mine's pretty low gain overall, so this sounds close to the threshold for the octave you've described). I don't think there's a single part responsible -- it's likely to be some combination of the total amount of gain and the total clipping as well.

Maybe try using a 10K pot in series with one of the diodes instead of a fixed resistor and seeing if it changes the likelihood of an octave happening. OR just a pot in series with ground after both of the diodes.

[Paul Marossy]

I get the octave up without even having the diodes in the circuit.

EDIT: I found some Colpitts VCO schematics. They don't look the same to me...

[tca]

I get the octave up without even having the diodes in the circuit.

That is what the simulation shows, although subtle .

[Paul Marossy]

I get the octave up without even having the diodes in the circuit.

That is what the simulation shows, although subtle .

Mine is not very subtle, at all. But my circuit has been modified quite a lot in an attempt to get rid of the low frequency "waterfall" noise.

At this point for Q1 I have a 2N1110 at about 85 Hfe with 510K bootstrap resistor and for Q2 I have a 2N5089 with a 22K bootstrap resistor. On Q2, the larger that bootstrap resistor is the more noisey the circuit gets. I've tried all kinds of resistor/transistor combinations for Q2 with pretty much the same end result as far as the sound goes, I just get more or less noise. So I opted for less noise. It messes with the voltages at Q1 too, but the sound doesn't really change much at all after I dial in the 100K trimpot on the emitter of Q1 to "the sweet spot".

[Ronan]

Here's a minor variation of this circuit, seperating it so you can see the 2 cascaded stages. I haven't breadboarded it, but I believe it should sound the same, and would give the ability to adjust the DC operating conditions for each stage seperately from each other.

I could be completely wrong too...

The idea is to measure the dc voltage at the 47uF electro on the standard Harm Perc circuit. Then on the new circuit, select Ra and Rb to give you that same dc voltage, and see if it sounds the same as the original. If it does, Ra and Rb could then be tweaked to taste.

[mac]

What I have learned is that there's something with the commons that is doing some sort of cancellation or something. This is why replacing that 47uF cap with different values messes with it, as digit (I'm new, I think his username is something along those lines) experimented with.

A quick way to understand the HP is to notice that the pnp sends a copy of the input signal to the emiter. BUt the npn sends an amplied and inverted signal to the same emiter, producing some kind of cancellation as stated in the original advertisement.

Also, I've read that potentiometers between the emitters makes for some really ugly and gnarly octaves, but I haven't tested this personally.

I suggested this pot when experimenting with an all germ HP, but works with a std one.
You can also make a nice booster by putting a 47k without the 47u and without the diodes.
IIRC there is a post with some maths I did a couple of years ago or so.

mac

[tca]

What I want to do is find a way to dial in whether or not you get any 2nd order harmonics. I don't know if it's possible with the circuit as designed.

I've been playing a litle more with the simulations. You can built a symmetrical variation that would work, and by controlling the amplitude of the input signal you can have a overdrive type distortion to a typical HP (HJ) sound!



You can also reverse the roles of the PNP and NPN transistors, with the proper changes, and take the output from the PNP device and get the same results; the circuit is invariant to this symmetry. The way the original circuit is just permits the large amplitude behavior to occur for small values of the input signal.

P.S.(edit)

BTW, I don't think it is based on a Colpitts oscillator. I think is a clever war of putting two class-a amplifier arrangements, using a PNP and a NPN, without using a phase splitter, similar to the JLH class-a amplifier with out the first transistor.

[tca]

What I want to do is find a way to dial in whether or not you get any 2nd order harmonics. I don't know if it's possible with the circuit as designed.

By varying the upper resistor R2 one can control the amount of the 2nd harmonics:



Cheers.

[Paul Marossy]

A quick way to understand the HP is to notice that the pnp sends a copy of the input signal to the emiter. BUt the npn sends an amplied and inverted signal to the same emiter, producing some kind of cancellation as stated in the original advertisement.

Yeah, that makes it a simpler concept to understand. So then as I was thinking, if I were to remove the common connection between the emitters and make them instead each go through their own capacitor to ground, this cancellation would not be happening anymore and it probably would become more like some kind of overdrive/distortion circuit. This probably would also explain this crossover distortion sound I hear when I really slam the strings.

By varying the upper resistor R2 one can control the amount of the 2nd harmonics:

You mean the collector resistor on what you are calling "T2"? I have a 100K trimpot in that location, and tweaking it does have some amount of effect on how much 2nd order harmonics I can hear. And also on how nasty the distortion gets when I slam the strings.

[digi2t]

What I want to do is find a way to dial in whether or not you get any 2nd order harmonics. I don't know if it's possible with the circuit as designed.

By varying the upper resistor R2 one can control the amount of the 2nd harmonics:



Cheers.

Unwittingly, that's what I've done. Just goofing on the breadboard, I placed a 25K trimmer in place of R2. In my set up I find that dialing the trimmer, I can hear when I hit the "sweet spot", which tends to be around 1.0v on Q1 collector.

[tca]

I meant R1 not R2, sorry! The collector NPN resistor. Let me run another sym.

P.S. (edit)

Varying R2 does have the effect of stretching the negative part of the signal down and that is also a signal for extra 2nd harmonics, isn't it?.

[tca]

This probably would also explain this crossover distortion sound I hear when I really slam the strings.

Why do you say that is a crossover distortion, or is "sounds like crossover distortion". With 1V (R2=4.7k) at the PNP collector the sym shows that VBE=-.6V for the PNP and VBE=0.6V for the NPN. They are almost always conducting. If it were crossover distortion you should ear it for small amplitude signals.

[Gus]

The HP is two common emitter stage in series with a few twists

Note that the same current is in both gain stage sections of the HP

The biasing and cap coupling make sure that in each stage the collector and base current sums to the emitter current.

The stages being powered in series then have to have the same current.

Yes there might be some leakage current in the caps but it should be small and not matter

I ran some sims and two caps make an interesting difference in tuning this circuit.
The one from the emitters to ground the larger the value the less interaction between stage.  You get a low frequency bump and more interaction when the cap is reduced in size
For people running sims try values from 1uf to 220uf

The other cap is the one from the PNP colllector to NPN base
large values make it flatter in the bass end, smaller values high pass with a low bump, even smaller values more of high pass without a bump depending on the size of the cap at the emitters
for people running sims try values like 10uf, 1uf, .1uf, .047uf, .001uf

Myself and others have posted about the HP in other threads at this site in the past

[pinkjimiphoton]

i'm gonna try messing with this when i get a chance in a few days....would love to have a little more harmonic feedback..what i ended up with sounds great, but it's a little tame compared to dino's, which really seems to nail that albini thing.

what this does to feedback is a thing of beauty, for sure.

[Paul Marossy]

This probably would also explain this crossover distortion sound I hear when I really slam the strings.

Why do you say that is a crossover distortion, or is "sounds like crossover distortion". With 1V (R2=4.7k) at the PNP collector the sym shows that VBE=-.6V for the PNP and VBE=0.6V for the NPN. They are almost always conducting. If it were crossover distortion you should ear it for small amplitude signals.

Well it gets really ugly sounding, maybe it's severely running out of headroom, I dunno. My circuit isn't really the original circuit anymore either. I like it for certain things. It will probably end up on my pedal board.

[Paul Marossy]

I just realized what it is this morning when I had it plugged into a higher gain pedal I designed - it's intermodulation distortion. It doesn't really come out until you hit the strings hard with the "harmonics" knob from the range between about 3:00 o'clock to all the way up. Sounds like when you play a chord with an octave pedal, that's intermodulation distortion. Except that it's an octave up version of what I've heard on my Boss OC-2.

[tca]

I just realized what it is this morning when I had it plugged into a higher gain pedal I designed - it's intermodulation distortion.

Brilliant, thanks for the insight.

Cheers.

[Paul Marossy]

Yep, no problem. I'm happy that I can at least satisfy my own self on this one.