MXR Distortion+ mods question about Diode placement

[half_smith]

Hello, I made this MXR Distortion+ (GGG kit) and the PCB includes a bunch of "bonus" holes and traces...

The Schematic only mentions D1, D2 (in line with the LEVEL pot and final output) but the PCB calls those D3, D4 (with options for doubling all of them) and the PCB also differs from the Schem with D1, D2 slots earlier in the circuit (around C5, see pic).

Besides the documentation being a little off, I just can't wrap my head around which placements are supposed to do what?

I've tried so many combinations of them that I've lost count, and tbh the pedal sounds kinda best with only 1 diode and nothing in any of the other holes... am I just liking the IC clipping?
I'd like to be able to swap in different diodes and get different effects but I notice a big drop in output with certain diodes (mostly due to different vf?), but I'm pretty clueless as to why I don't get substantially more clipping, fuzz and/or OD with 6 diodes vs 1 or what? 

Some mods ala Antonis:
I have replaced R4 (resistor leading into the Gain pot wrongly labelled R6 on the diagram) with 47k, and I changed R2, R3 to 470K instead of 1M.

Advice is Much appreciated!

here's the schematic and layout I'm working with:

[antonis]

First of all you have to locate items and verify their correct PCB placing..

Try to make a working circuit with the least part count..

Proceed to one at a time distinct circuit mode and reinstae circuit before proceeding to the next mode..

edit:

Some mods ala Antonis:
I have replaced R4 (resistor leading into the Gain pot wrongly labelled R6 on the diagram) with 47k, and I changed R2, R3 to 470K instead of 1M.

Never suggested the first one and 100k change was the second one..

[Mark Hammer]

The board looks like it was laid out in anticipation of using different kinds or numbers of diodes, like a silicon and one Schottky or germanium on one "side" and only one diode on the other.  Or two germaniums in series on each side, or whatnot.  Raising the summed forward, by means of different types or numbers of diodes will raise the clipping threshold, but will also raise the maximum output level.

And yes, I think part of the key to the Dist+'s sound IS the combination of op-amp clipping and diode clipping.  It's not a sin to like what the op-amp does.

The extent to which different numbers of diodes will result in different audible intensities of clipping is a function of the op-amp output itself.  If a person used 3+3 silicon diodes, the likelihood of any audible clipping is slim.  The chip can only swing 3.5V in either direction of Vref.  Three silicon diodes would need a half-wave exceeding roughly 1.8V and that's only likely to occur at the pick attack, and not much afterwards, since there is always a sharp drop in amplitude.  And since you increased the minimum ground-leg resistance from 4K7 to 47K, your highest amplification factor will be a gain of around 11x, which is generally not enough to push diodes into conduction with a typical guitar signal.  If you want to start hearing diode differences, you have to feed them enough.  And you can begin doing so by reducing that 47K resistor to a much lower value.

[antonis]

And since you increased the minimum ground-leg resistance from 4K7 to 47K, your highest amplification factor will be a gain of around 11x,

More closely to 22X, Mark..
(11X should stand for 100k R6 resistor..)

[Mark Hammer]

I confused which 1M resistor was being replaced by 470k.  The math was correct, but the component wasn't.    22x is not gonna get you much clipping either.

[ElectricDruid]

(mostly due to different vf?)

Yes, very likely. If you use diodes with lower Vf, you get more clipping, but a lower level. If you use a higher Vf (LEDs of various types, say, or series pairs of silicon diodes) you won't clip so much waveform (=less distortion) but you get more output level. In this circuit, you can't have both. Some people (Hi, Mark!) have designed other circuits that compensate the different volume levels with different diode selections, but for this one, you'd simply adjust the final level control to suit the diodes you're using.

Your mods have drastically reduced the maximum gain available, which I don't really see the benefit of. You could have got those sounds with the stock values by having the gain knob turned down further. All the changes have done is cut off some of the range of adjustment. Perhaps that's worthwhile if after some time with the pedal, you find you never used the top 50% of the travel, but I don't think that's the case here.

HTH

[half_smith]

I changed R6 from 4k7 to 47K because the GGG build instructions suggest doing that. (I am trying to figure out why)

Antonis suggested I change R2, R3 to 100k (from 1m) but that was accompanied by a big drop in output, I backed that off to 470k and got better output so left it. I guess I will try the 100ks again now that other problems have came and went.

How do increase the output and drive the diodes?

[antonis]

How do increase the output and drive the diodes?

Either by using higher supply voltage or using lower forward voltage drop diodes..
(e.g. 12-15V PS or Ge/Schottky diodes..)

[mozz]

I would put all values back to stock, not ggg values. Get the problems worked out first.

[Mark Hammer]

I changed R6 from 4k7 to 47K because the GGG build instructions suggest doing that. (I am trying to figure out why)

Antonis suggested I change R2, R3 to 100k (from 1m) but that was accompanied by a big drop in output, I backed that off to 470k and got better output so left it. I guess I will try the 100ks again now that other problems have came and went.

How do increase the output and drive the diodes?

The 4K7-to-47K change was likely suggested to turn the circuit into what is essentially a mostly-clean-but-a little-dirtyish booster.  When I implement the circuit, I tend to use a 100k gain pot, because  it's not really until you start going above gains of 20x that you even begin to hear much coloration and grit.

Remember that when MXR originally came out with the Distortion+, simple boosters were uncommon.  There was the EHX LPB-1, and that was pretty much it for the brunt of music stores.  So the use of a 500K Gain pot allowed for gain to be dialed way down to produce something like a clean boost (though the pedal NEVER gets pristine clean).  I would think that the "clean boost" function has been pretty much farmed out to other pedals by now, so there's no real point in building it into a Distortion+.

If a person wants BOTH good output levels without having to dime everything, PLUS the ability to get good crunch, then the following are worth observing:

1) As suggested, make the Gain pot 100k, with a C-/reverse-log taper preferred.  Any taper will "work", but C will provide the best dialability.
2) Make the 4K7/.047uf pair in series with the Gain pot 0.1uf-.22uf and 3K3-3K9.  Those changes will yield more gain and more bass at highest gains.
3) If you have and like germanium diodes, then use a 2+2 set, instead of 1+1.  Alternatively, do what DOD did with the 250 and use a pair of silicon diodes instead of germanium.  Use of silicon OR more germanium will raise the clipping threshold and allow for roughly double the maximum output volume.
4) MXR did precious little to constrain unpleasant fizz.  Put a 47-100pf cap in parallel with the 1M feedback resistor, and increase the .001uf cap in parallel with the clipping diodes to something like .0033uf.  That will still give you all the bite you want, without fizz.

Happily, the board accommodates all of these mods, with the exception of the 47-100pf added cap.  But given how small such ceramic-disc caps are, that can simply be tacked on to the pads on the copper side of the board.

[half_smith]

1) As suggested, make the Gain pot 100k, with a C-/reverse-log taper preferred.  Any taper will "work", but C will provide the best dialability.

I will try a 100kC when I get one in.

2) Make the 4K7/.047uf pair in series with the Gain pot 0.1uf-.22uf and 3K3-3K9.  Those changes will yield more gain and more bass at highest gains.

I did this and it is much gainier (and bassier) thanks for the suggestion!

3) If you have and like germanium diodes, then use a 2+2 set, instead of 1+1.  Alternatively, do what DOD did with the 250 and use a pair of silicon diodes instead of germanium.  Use of silicon OR more germanium will raise the clipping threshold and allow for roughly double the maximum output volume.

what do you mean by 2+2 or 1+1? I have found that I don't like the sound the Diodes in D1, D2 (on the PCB not Schematic) sound so I'm leaving them out, but I've had cool results with various combos in the "clipping" area. Right now settled on two N207s and a GAZ17 (total of 3), long sustain and gritty.

4) MXR did precious little to constrain unpleasant fizz.  Put a 47-100pf cap in parallel with the 1M feedback resistor, and increase the .001uf cap in parallel with the clipping diodes to something like .0033uf.  That will still give you all the bite you want, without fizz.
Changed the .001uf (parallel the diodes) to .0022uf, my output is noticeably darker after the changes

I am not sure where to put the 47pf cap
which 1M on the Schematic is the "feedback resistor"?
Can I just sister the 47-100pf cap right onto the same pads as the 1M feedback resistor?


Did I mention I'm a potato?

[antonis]

which 1M on the Schematic is the "feedback resistor"?

R8..(on schematic)

Replace C5 (10pF) with 47 - 100 pF..

[Mark Hammer]

Can I just sister the 47-100pf cap right onto the same pads as the 1M feedback resistor?

Yes.  The nice thing about mods like this are that no changes or cuts or new holes are required for the board, and such small caps are able to lie flat enough that they can fit easily.  And if you don't like what the added cap does, or if it doesn't do enough of what you want, you can just unsolder it and use a different value.  Of course you could unsolder C5 and replace it with the new value.  But if you're seasoning to taste, then tacking on additional capacitance to the pads on the copper side is easily.

[half_smith]

thanks Antonis! I modified my last post after making some changes (and starting to understand what Mark was suggesting).

The 100ks in R2, R3 are seeming good now, Lots of power at first, then I loaded up a bunch of diodes and the gain is within usable range.
Is it expected that putting more diodes in line will eat up gain?
Some of the diodes I have are higher VF and they seem to stay super clean no matter what, not what I want here, the 1N270's from Guitar Pedal Parts seem to be my favorite for crunch and gain.

[antonis]

Is it expected that putting more diodes in line will eat up gain?

Let's make some things more clear..

Gain is set by 1 + R8/(R6+R7) and has nothing to do with diodes..
IC1 amplified signal is initially clipped due to high gain and low headroom and afterwards due to diode(s) forward voltage drop ..
(e.g. for +/- 3.5V op-amp output and Green or Blue or White LEDs for D1/D2, there should be NO further clipping..)
The lower the diode(s) forward voltage drop the lower the VOLUME on the output (signal amplitude) but the "harder" the clipping..

[Mark Hammer]

Although increasing gain makes it possible to produce greater distortion, "gain" and "distortion" are not the same thing.  Don't feel bad; the two get confused and treated as equivalent by MANY.

The diodes to ground set the ceiling on the signal level.  If I have a room whose ceiling is only 4 feet high, even schoolchildren will bang their heads into it.  If I raise the ceiling to 6 feet, some people will bang their head, but many won't.  Most basketball players probably will, though.  If I raise the ceiling to 8 feet, very few people in the world will hit their head against it.  If everybody is in bare feet then what I described will be true.  If everybody puts on winter boots, then even more people will be banging their head into the 6ft ceiling though still only very few will hit their heads against 8ft.  If I make everyone wear platform shoes like Gene Simmons, adding 6 inches to their actual height, then most people will bang their heads against the 6ft ceiling and some basketball players will bang against the 8ft ceiling as well (though still not very many).

The forward voltage of the diodes sets the height of the ceiling.***  More diodes in series, or use of diodes with higher forward voltage (like LEDs) raises the ceiling, such that only the odd peak will get clipped.  Increasing the gain is like giving the signal platform boots, bringing it closer to the ceiling set by the diodes.  If a person wants maximum clipping for a long segment of a picked note (remember that the string decays in level over time), you lower the forward voltage by using germanium or schottky diodes, and increase the gain of the op-amp so that the signal stays above the forward voltage most of the time.

Of course, lowering the ceiling via the diodes also means the maximum output level will also be lower.  A person sacrifices output level for distortion intensity.  The ideal is to be able to get decent volume levels without having to dime everything.  The DOD250 used silicon diodes, which gave it a hotter output than the germanium diodes in the Distortion+ provided.  It also meant the maximum distortion was not as intense as the MXR unit.

A compromise is provided by using a little more gain - the basis of my suggestion to reduce the 4K7 resistor to 3K9 and even 3K3 - and 2 germanium diodes or a germanium and schottky in series (in each "direction")  to raise the ceiling.

)***So does the power supply and how the op-amp behaves with this or that supply voltage, but we'll ignore that for now since the diodes are most important here.)

[half_smith]

You guys are seriously really helping me out. It almost makes sense.
I'm going to read and reread this thread over the years until it is like my own personal Prufrock.
Thank you.

One other quick question if I may,
On the PCB there are two areas with diodes, the Schematic doesn't show the other group at all,
What is the purpose of the optional diodes in the other location (in the feedback loop?) ?

areas highlighted:

[Mark Hammer]

The layout allows one to configure the circuit as a "soft" clipper (like a Tube Screamer) OR a "hard clipper" (like a Dist+, DOD250, Rat, etc.).

Realistically, the diodes clip the way, and at the forward voltage, that they clip.  But when placed in the feedback loop, they tend to prevent the op-amp itself from bashing its head against the maximum "voltage swing" it is capable of.  So, for "soft clipping" you get ONLY the effect of the diodes, where in "hard" clipping you get the signal produced by the headroom limits of the chip, when is THEN clipped by the diodes at the output.