Switchable diodes in tubescreamer clone

[MikeH]

I'm working on a Tube Screamer style layout (actually it's a ROG Tube Reamer), and I've been trying out different clipping diodes while I've got it on the breadboard.  I couldn't decide between the smooth, semi-fuzzy clipping from the 1N914s, or the crunchy, mid-boosted clipping I got from a pair of LEDs.  Finally I realized, I've got some 3pdt toggle switches, lets just do both.  But now I have a couple problems that I don't know how to solve:

1.  There is a HUGE volume difference between the 2 clipping styles.  The 1N914s seem to clip a lot more of the signal, and therefore aren't nearly as loud as the LED's.

2.  There is also a loud audible pop when toggling between the 2 clipping styles.  Which I assume is comming from the split second when the switch is not connected to either set of diodes, so the signal is not being clipped, and hence the volume is much louder.

Does anyone know any methods to fix these problems?  The volume difference is my main concern; I would really be a pain to have to fiddle with the volume every time I changed the clipping style.

[MikeH]

The only strategy I can think of to fix it, (and I'm not crazy about this) would be:  since I'm using a 3pdt toggle I could run the out put into the pole I'm not using and have an out put volume for each clipping style.  Almost like having 2 channels.  Which kind of makes me want to put a stompswitch in there, but then I still have the popping problem to deal with.

[bancika]

is clipping style really something you want to switch all time? people usually try few combinations and settle with one permanently.
as for volume difference, if you have spare pole you could use it to add resistor in series with volume pot. Difference in levels is constant so I don't think you need two pots. However, use pot to determine resistor value so you have equal levels on both "channels", measure resistance and replace it with closest value resistor
cheers

[MikeH]

Yeah, normally clipping style isn't the kind of thing I'd want to switch back and forth from (especially on the fly), but I liked both styles and they're different enough, so I figured I might as well build both in.

[Mark Hammer]

Given something I posted before lunch over in the Lounge/OT area, this posting makes me smile as much as it makes me sigh, but here goes...

Just about any diode change that you have heard about likely produces a change in sound more because of what it does to the likelihood of clipping occurring than due to the manner in which it clips when it does clip.  Yes, I'm sure that a wide array of diode types can be shown to produce different conduction time response curves to steady state signals of varying types, but do not interpret that as an indication of what is taking place in your average fuzzbox.  Nope, in the typical distortion unit, what gets changed when you switch between the number or type of diode is essentially what percentage of the time, under normal playing conditions and given the amount of gain in the circuit, your input signal will be above the clipping threshold of those diodes.  Raising the clipping threshold by changing diode type or increasing the number of diodes will always have the effect  of permitting more of the gain that has been applied to the input signal to be heard without any ceiling placed on it.

The challenge this poses to the user/builder is whetehr they want that volume level shift or not, and what to do about it if they don't want it.  Using more diodes or diodes with higher forward voltages will always result in a significant upward level change,  If you want to reduce the level shift then I suggest the following.  Instead of merely switching the diodes from a pair of LEDs to a pair of silicon, switch the diodes and level simultaneously.  The simplest and easiest way to do this is to insert a fixed resistor in series with the 100k volume pot such that it attenuates the output for LED use, and unattenuates the output for Si use. A DPDT toggle can be used to accomplish both diode and level switching at the same time.  For starters, try a 33k resistor on the input of the 100k output level pot.  When half the toggle is set to Si diodes, you arrange for the other set of contacts in that position to shunt/bypass the 33k so that it is simply a straight wire for all intents and purposes.  When you switch to LEDs, you now un-shunt the added resistor.  This will make the volume pot behave as if it were a 133k pot that can never be turned up more than 3/4 of the way.  My guess is that using a 33k resistor will not likely produce equivalent output volumes for the two diode settings, but it will bring them much closer together and reduce the contrast.  My guess is that you probably want to preserve the extra push the LED setting provides the amp anyways, so yuo don't want to take the volume shift out completely.

[MikeH]

Wow.  Thanks for info!  You not only taught me a good lesson on the functionality of diodes, but you also taught me that I knew absolutely nothing about the way diodes work.     So, my rather elementary understanding of this is telling me, that what changes here, is not the way it clips the signal, but rather how easily the signal clips (how likely it will be to clip)?  Either way, the resistor in series should be a easy way to fix the problem.

Does anyone know how to solve the popping problem when switching between diodes?  I wouldn't know where to start.

[Mark Hammer]

It would depend on how you are switching them.

Take a look at the terrific (and clearly written) Warp and Tone clipping articles that jack Orman has on hisLab Notes page at AMZ.  You may get some better ideas about what you can do to change between the two diode types.

http://www.muzique.com/lab/tclip.htm
http://www.muzique.com/lab/warp.htm

[MikeH]

Whoa.  Some cool diode tricks.  I'll have to try some of those.

Since I just have it on my breadboard right now (which is on the small side), I have everything stretched pretty far apart and jumpered together.  But currently it is wired as such:  (please excuse the crude explanation)

Wire from one trace where diodes would normally be plugged in to the center of 3PDT.  Wires from each throw of that pole connect to traces with diodes.  Each set of diodes jumped to other traces (obviously) and those traces each connect to the outer trows of a different pole, of which the center goes back into the circuit where the other side of the diodes would have been.  (Oh god, I don't know if that makes any sense.  I don't even know now).

I realize the 3pdt switch is overkill, and really I could do this with an SPDT switch, but I wired it this way so that things were spread out and a little easier to work with for the time being (It's a pain to be swapping out diodes when they're all jammed up on top of one another).  Would it fix the popping if it were wired with an SPDT, with all for diodes plugged into the center of the switch, and each pair connected to each throw?

Of course I'd used a DPDT switch with the resistor to fix the volume change problem.

[PaulC]

Sometimes the change in output level you get from diode switches can mess up other parts of the circuit. The higher signal levels can overdrive things like the active tone control of a screamer for instance, or other pedals that have a gain/boosting stage following the distortion stages.  Sometimes when when people say the high end got harsh when they switched to LED's is because the higher signal level is now clipping something that wasn't clipped before.

One way to fix this is to have voltage dividers switch in after the different diodes trying to keep the signal levels equal, or at least putting a limit on the signal to keep from overdriving something that shouldn't be. 

Later, PaulC