I was just reading about Class B circuits online and found out that they create crossover distortion, and it got me thinking about how this could be controlled and used as an effect. This is the page:
http://www.allaboutcircuits.com/vol_6/chpt_6/10.html
If you scroll down a bit, there's a schematic of a single opamp driving a pair of transistors with the caption: "If we were to use the resulting output signal to drive the base terminals of the push-pull transistor pair, though, we would experience significant crossover distortion"
...sounds promising! I guess all you'd need then would be a "drive" pot to...change the bias of the transistors to control the width of the crossover point?
I dunno. Anyone got any ideas?
Oh...I see Zvex beat me to it :x
or, as an easy way to do this: a back-to-back diode pair *in series* with the signal. The funny thing is that increasing the signal decreases the distortion.
Sorry if this is a dumb question Mike, but do you mean "back to back" as in the first pic, or the second?
---|<-->|---
|->|-|
---| |---
|-|<-|
How would you go about controling the amount of distortion created?
http://www.geocities.com/tpe123/folkurban/fuzz/snippets.html
Check out TMK*. It has a variable x-over distortion control, amoung other things.
I have found that 2 GE diodes sound the best, with SI's and LED's providing too much. The Boss HM pedal uses this along with the standard diodes in the feedback loop and diodes to ground.
Some folks like it, some folks don't. 8)
I've played with crossover distortion awhile back. Just off the top of my head there are four different ways to create crossover-like distortion.
1. Use a device whos output design is prone to crossover distorion like certain types of opamps.
2. Back to back diodes in series on the output. Not shunted to ground like many distortion designs. Tim's T.M.K. and Square Wave Shaper are examples of this.
3. Create a discreet class A/B output section that is not biased properly. This can be done with a driver section feeding a NPN and PNP complimentry transistor output. Yet again Tim's has variation of this called the Triple Fuzz.
4. Use a deadband generator which in somewhat simple terms is a full wave rectifier in the feedback loop of a device like an opamp with the required bias pull up/downs where needed for signal conduction. As I remember this was a real fun build.
Just some thoughts.
Andrew
Ahhh, I see. :o
So, could it be as simple as using a pair of back to back diodes in series after a gain stage? Kind of like a distortion+ on crack?
Zvex's blurb on the Machine mentions non-selective frequency tripling, and I recall seeing a pair of ge diodes in a gut shot once, so maybe the Machine is like a Triple Fuzz with a pair of diodes in series before the output. (Not that I want to rip him off -I'd prefer the simpler gain stage + series diodes if it would work).
Don't know about the Machine, but you got it with the Distortion +. 8)