soft-clipping diodes: more distortion for bass strings vs. higher strings

[boogieWoogie]

Hi,
I'm playing around w/ Jack Orman's Rex's Fuzz.  It works probably as expected. I'm using it for bass.

The problem is that I'm getting more distortion for low string notes, and less distortion for higher string notes.  I know this has as much to do w/ the pickup output across frequency as it does w/ the frequency response of the op-amp circuits, but I'm wondering if there's a general solution that folks use for this problem in soft-clipping op-amp circuits?  I don't want to roll off the bass (w/ a high-pass) b/c I play bass ;-).  It's more like I want to boost higher frequencies so they hit the soft clipper diodes with a larger signal.

Thanks,
Mike

[Ben N]

If you want more distortion happening on the treble end of the spectrum relative to the bass end, then I think you have to feed less bass into the circuit. There are AFAIK two ways to do this and still keep the bottom: one is to boost the bass after the fuzz to compensate for the bass cut before, and the other is to provide a path for your clean signal (or a clean low pass filtered signal) in parallel with the fuzz, taking care to keep the fuzz signal and the parallel signal in phase with one another, a la the old Voodoo Labs Sparkle Drive. This means more complexity, as the post-boost means adding at a minimum a filter stage after the distortion, and the parallel clean approach requires at a minimum a buffer/splitter in front and a mixer at the end, but there are plenty of schematics around for circuits that do this, either as a separate unit or built in. Look up Buff N Blend, or Splitter-Blend.

[Gus]

I posted this in the past
A simple way to try first
Decrease the input cap value more than you might think you need with a bass
try .0047uF and go up or down for what you like

[Mark Hammer]

One trick is to do what many chorus and flanger pedals do, which is to produce pre-emphasis via dual input paths.
Look at that first op-amp stage after Q1.  You can see two parallel input paths.  The C3/R5 path provides more gain for content above roughly 2.3khz.  If you were to put a 6800pf cap and 47k resistor in series, and put them in parallel with that 100k input resistor, you would have roughly twice the gain for content roughly 500hz and above, or rather half the gain for content below 500hz.

[Fancy Lime]

Hi Mike,

welcome to the forum!

I don't want to be presumptuous but considering that this is your first post and your building a very simple circuit, I think you may be looking for a fairly simple solution. Gus's point is definitely valid. If you decrease the input cap (the one all the way to the left) value from 0.1μF to maybe 10nF, that will reduce the bass before the clipping section and therefore reduce clipping of the low frequencies. But because the diode clipping essentially cuts everything to the same level again, you don't actually loose bass frequencies at the output (unless you choose your cap way too small). This is a good first thing to experiment with.

There is another thing you can try. This is my favorite thing to do with diode clipping for bass guitar because it provides a (relative) bass boost. I am amazed that this is not done more often in commercial pedals except the various Big Muff clones and derivatives, especially because it is such an essential part of what makes the Russian Big Muff sound so popular with bass players. What you need to do is put a cap in series with the diodes. 22nF is a good point to start experimenting, in my experience. Go smaller for more bass boost, larger for more distortion of the bass frequencies and less boost. You can also add a resistor in parallel with this cap if the bass boost is too strong. Start with a 33kΩ, go larger for more bass boost, smaller for less. But you can leave out the resistor completely if you like the sound of that. Here is a quick schematic of just the clipping section:



I really like this mod on bass and would strongly encourage you to give it a try, especially if you have the circuit on a breadboard.

Hope that helps,
Andy

EDIT: P.S. The diode arrangement in the Rex, namely in the negative feedback loop of an inverting opamp, is usually referred to as "hard clipping", although the whole "hard/soft" clipping nomenclature is a bit of a mess anyway.

[anotherjim]

An advantage of Marc's suggestion is that if you've soldered it all up, the extra R and C series circuit can be added across the input resistor (yours is 100k?) which should be fairly easy to achieve.

[boogieWoogie]

Thanks all for the ideas.

Ben N: I was already wanting a clean / wet mixer, so if some of these leads work out, I'll invest in the mixer too (and thanks for the leads on the splitter / mixer circuits).

Gus: The 0.1uF cap that's in the design gives a high-pass w/ Fc = 15Hz.  Very low, prob. TOO low.  0.047uF would yield a 212Hz 3db... maybe a bit higher than I want for bass, but I'll experiment w/ values in between.

Mark H: thanks for the dual-input pathing idea.  I'd seen that on many a schematic, but never knew what it was accomplishing.  Makes good sense though.  W/ my 10K input to the 2nd op-amp, it'd be easy to put a cap in series w/ a lesser R so that the high-F gets more gain.

Fancy Lime: I like the idea of a cap in series w/ my clipping diode(s).  I had one in parallel w/ the diodes to reduce high-freq hiss, but it was probably exacerbating my loss of high-string distortion.  Gus too mentioned reducing the input AC-coupling cap.  I think 10nF is too low for my needs: would yield a 1KHz 3db point, which is too high for bass.

anotherjim: love the consideration for doing these mods to a PCB, already soldered.  I'm fortunately on the breadboard now, but that'll be a nice technique to know about for easy on-the-fly mods to existing pedals.

Thanks all!  Will return to the breadboard tomorrow eve.

[Rob Strand]

You should listen to the output of the first opamp and see if the objectionable distortion on the lows is coming from there.   One solution would be to increase input 100k to 220k to 330k (*or* decrease the 1M feedback resistor to 470k down to 330k).  That will decrease the gain and stop the first-stage clipping.   

Beyond that would be the usual thing of reducing the 1uF at the output of the first.

Maybe you should just built a single non-inverting stage like *first stage* of son of screamer and play with C2.
The non-inverting structure has a little blending of its own.  The inverting structure  like the Rex Fuzz doesn't.
http://www.muzique.com/tech/scream.htm

[bushidov]

What you need to do is put a cap in series with the diodes. 22nF is a good point to start experimenting, in my experience. Go smaller for more bass boost, larger for more distortion of the bass frequencies and less boost. You can also add a resistor in parallel with this cap if the bass boost is too strong.

Andy, how would that same idea look for a non-inverting op-amp?

I have a Bass OD that I am looking into modding this way. It currently looks like this:

[Fancy Lime]

What you need to do is put a cap in series with the diodes. 22nF is a good point to start experimenting, in my experience. Go smaller for more bass boost, larger for more distortion of the bass frequencies and less boost. You can also add a resistor in parallel with this cap if the bass boost is too strong.

Andy, how would that same idea look for a non-inverting op-amp?

I have a Bass OD that I am looking into modding this way. It currently looks like this:

Pretty much exactly the same. The technique works in any negative feedback loop and also on diodes to ground. Just disconnect the anode of D1 and the cathode of D2, reconnect both of those to a cap (22n or so) and attach the other end of the cap to the negative input of the opamp. Add a resistor in parallel to that cap if you like. I usually start experimenting without the resistor and then add it once I am satisfied with the cap value. You can use a pot wired as a variable resistor (100k log). This will give you a a variable bass boost control. Not a bad option if you don't mind the extra control.

Instead of a resistor, you can also use another diode. Depending on the diode orientation, this will give you either more or less asymmetry on the bass frequencies compared to the treble. Or use two diodes for a clipping threshold one diode drop higher on bass than on treble. That's what I dd here (D3, D4):
https://www.diystompboxes.com/smfforum/index.php?topic=124063.0

Hope that helps,
Andy

[bushidov]

I figured, but just wanted to be sure. Thanks again for the tip!

[amz-fx]

I'm playing around w/ Jack Orman's Rex's Fuzz.  It works probably as expected. I'm using it for bass.

Try this out:



regards, Jack

Edit: reading back through the messages, I see that Gus and I were thinking along the same track! 

[Mark Hammer]

The caveat with that sort of component change is that the rolloff is continuous from the corner-frequency down to DC.  So, with that first cap = 4n7 the rolloff starts around 338hz, is down by 6db at 169hz, by another 6db at just under 85hz, and so on.  The dual-parallel input I suggested doesn't really have a bass rolloff, just more gain above a certain corner frequency, such that bass is retained, but just not as prominently.

Not intrinsically better or worse.  I suppose it depends on what the user wants.