Circuit analysis of a Boss HM-2 clone... got a couple questions

[bushidov]

Hi All,

I got another classic circuit that I am trying to wrap my head around. Thanks, by the way for the folks who've helped in on the past analysis questions (Boss BD-2, Op Amp Big Muff Pi, etc)

Below is my working schematic for a Boss HM-2


Here's what I believe I already know:
Signal goes into a JFET buffer with an approximate 1M input impedance set by resistor R4.
Then it goes into a distortion section that I have some questions on between Q2 and Q3, which then form the input and part of the feed back to op-amp U1B.
U1B appears to be a non inverting op amp with some clipping and negative feedback RC filter (R18 and C9)
Then some more diode clipping into an op amp buffer U1A
Then going into an op amp gyrator section for tone control, two if which are tied together, which have the bands merged at op amp U3A
After that is a little filtering and volume control, finalized with a recovery transistor stage.

Is this generally correct?

Now, for my specific questions:
1. I usually see this on high gain distortions for clipping, but what kind of clipping is going on with D5 and D6? Is that soft clipping, inline?
2. That hole distortion stage with Q2 and Q3? What is going on there? The Q2 stage looks a lot like a Big Muff Pi gain stage without the clipping diodes and Q3 looks like the same thing, but PNP'ed (if that's a word). Is this correct?

Thanks for all your time.

[Mark Hammer]

1. I usually see this on high gain distortions for clipping, but what kind of clipping is going on with D5 and D6? Is that soft clipping, inline?

No.  In fact it's not even what one could call clipping, since that term applies to the highest and lowest excursions of the signal.  D5/D6 provide what is referred to as "crossover distortion", because it applies only to that region where the signal swings above  or below "zero"; i.e., the point/zone where it "crosses over"

This has two impacts.  The first, and simplest, is a kind of gating effect.  Until the signal exceeds the forward voltage of the diodes, nothing passes.  So if one were to use a diode pair with a 300mv forward voltage, anything below that voltage would go no further.  That can be helpful for blocking noise in pedals that have a tendency to generate hiss.

But bear in mind that guitar signal is not like an oscillator, that maintains an instantaneous and steady level.  The signal has a very brief "rise" when you set the string in motion, and  decays as you let the string die down.  There are also little perturbations while the string is vibrating, and every cycle necessarily means the signal swings above and below the zero point.  What the diodes do is "chop the sides" of the signal.  Not by a whole lot, but enough to have an impact on the harmonic content, as a very distant relative of what is normally called "pulse width" in synths.

I implemented some crossover distortion in a modified clone of the old Gretsch Contrafuzz.  The knob in the upper right hand side introduces progressively greater amounts of crossover distortion in tandem with any "normal" clipping.  Lousy sound but you'll get the idea.  I find that crossover distortion really loves bridge pickups, much moreso than neck or anything else.

[bushidov]

Thanks Mark on that one. I didn't know what it was doing, but usually, when I see diodes paired that way, I assume clipping, which was a bad assumption. The cross over distortion is actually a sound I kind of dig. I'll probably be applying that idea on a couple things.

Any ideas to the Q2 and Q3 sections?

[thetragichero]

q2 and q3 are both common emitter amplifiers (yes, similar to a big muff with their feedback loops). q3 is pnp hence why it is "upside down"
output of q3 goes to the noninverting input of the op amp stage, output of q2 goes to inverting input of op amp stage through distortion pot. third is where my analysis of that section putters out lol

i personally dislike the gating effect of the two crossover diodes, so one of the first mods i do on this type of pedal is jumper them

[Mark Hammer]

Series diodes are neither good nor bad, just different.  Note that they do not "clip", in the sense that we are used to.  The dynamics of the signal are preserved, albeit with reduced amplitude because the diodes remove a bit.  But they ARE a matter of taste, and I can understand why someone might want to forego them.

My own personal bias is that Schottky or germanium diodes may be best for this application, since they remove less of the signal.

I'll note that the gating aspect can come in handy in some instances.  I made the Craig Anderton Ring Modulator circuit from EPFM-2.  A decent ring mod sound, but I found it impossible to get rid of the modulating tone completely with a 10-turn trimpot.  So, I stuck a back-to-back pair of diodes in the ring-mod path to block anything below their forward voltage, and adjusted the gain of the output stage to compensate for the signal loss.  Works like a charm.

[teemuk]

This type of high gain effect will also produce moderately 'square wave-ish' clipping. Due to rapid rise of waveform edges in such signals there will be proportionally very tiny amounts of gating due to crossover  distortion. (Small rise time reduces crossover area).

[PRR]

> The dynamics of the signal are preserved, albeit

Series "crossover distortion" has little effect on loud dynamics but, as you note, can nearly drop-out the low end of the dynamic range.

For LOUD players the only effect noted might be less hiss when you drop your pick (go silent a moment).

It is a major fault in Hi-Fi, masking or garbling the soft passage of the concerto, but that is a Different World.

[Mark Hammer]

Correct.  One always needs to be careful about generalizing from the guitar world to the music reproduction world, and vice versa.  In the hi-fi world, the objective is to reproduce the original source with NO discernible changes.  In the guitar world, the goal is often to make the sound as different from the original as possible.

[bushidov]

Alrighty, I am learning quite a bit here. I am getting the cross-over distortion portion (and actually digging the sound or it). I am still puzzled why the two transistor gain stages Q2 and Q3 go into the input and inverting input of the op amp U1B. I could see one transistor stage going into the next and then going into just the input of U1B, but otherwise, I just don't have enough experience or know-how on this one. Any help?

Thanks,
Erik

[antonis]

>the two transistor gain stages Q2 and Q3 go into the input and inverting input of the op amp U1B.<

Actually, they don't..

Q3 output is in phase with U1B output, which is fed back to Q3 Base..
(you may consider it as paralleled with R17+R11..)

[teemuk]

Note that wiper of the pot is grounded so there is no feedback path through the pot.

[ElectricDruid]

I am still puzzled why the two transistor gain stages Q2 and Q3 go into the input and inverting input of the op amp U1B.

Like Antonis and Teemuk said, they don't really. The key thing is that distortion pot. While it's wired in a slightly unusual fashion, it still provides a variable-resistance connection to ground for the feedback around U1B. That's a standard non-inverting op-amp arrangement, although it's not immediately obvious to see.

The reason it's not immediately obvious is that the *other half* of that distortion pot is also used, providing a variable AC path to ground for the output from Q2. I'm not very familiar with what that would do to that Q2 circuit, so I'll leave that for others to fill in.

Imagine if the schematic were redrawn with the two halves of the Distortion pot done with a dual-gang pot with each individual section taken to ground. Then the two circuits would clearly not be connected. That's actually what we've got, electrically speaking! It's just that it's not straight-forward to see.

[teemuk]

The other half of the pot effectively forms a voltage dividing attenuator for the output signal from Q2 stage. The division is roughly that of source impedance and resistance of potentiometer half and series resistor. 10 uF is too large capacitance to introduce any significant filtering so its purpose is only AC coupling.