More Bazz Fuss diode experiments...

[LucifersTrip]

continuing the experiments here...
http://home-wrecker.com/bazz.html
===============


My friend asked me to help him build a simple distortion/fuzz as a 1st project. The BF was an obvious choice and a great opportunity to sout through some of the unsorted (mostly vintage) silicon diodes I have laying around. 

Since I don't do much with silicon transistors, I decided to give him a couple germaniums and at the same time give him a short intro to darlingtons.

The BF won't fire up with a single germanium, so I used 2 extremely low leakage ones with hfe ~ 75 for the darlington pair.

We went through a load of diodes, using the stock 914 as the control.

These are the duds. The volume was way too low and the fuzz weak:


These had a very similar fuzz, but a little lower volume...just not up to snuff:


These were very similar to the 914...couldn't tell most of them apart:


The were hotter, a little louder with more sustain and pretty much the same fuzz, but tighter:


...and surprisingly, the winner was....a modern silicon transistor with one leg broken off that I threw into my misc diodes bag.
hotter, sharper, good fuzz, better sustain and richer sound on the higher strings:


A close second was the JCAB above.

I'm not sure what I learned here, since the same forward voltages (all measured with the Peak) produced sometimes good and some bad sound. There's more to a diode than just forward voltage.

Though, the different qualities of sound did have similar groupings of voltages.

Ironically, the sound I liked the best was in the middle of the voltage range that home-wrecker did not report (.57v-.78v).

[mac]

As I wrote in lots of posts about the BF, you can't expect the diode to bias the transistor and clip at the same time.
And if you are using germs, the leakage, even if it is small, will be multiplied by the second transistor. And temperature will be also an issue.
And germanium diodes also leak and depend on temperature.

You can check my gallery for a solution to the ALL GE BF.
Besides, you can see how a big resistor feeding the base can help to fine tune the circuit and get the best tone even if you use GE diodes.

mac

[LucifersTrip]

As I wrote in lots of posts about the BF, you can't expect the diode to bias the transistor and clip at the same time.
And if you are using germs, the leakage, even if it is small, will be multiplied by the second transistor. And temperature will be also an issue.
And germanium diodes also leak and depend on temperature.

You can check my gallery for a solution to the ALL GE BF.
Besides, you can see how a big resistor feeding the base can help to fine tune the circuit and get the best tone even if you use GE diodes.

mac

I had absolutely no problems at all here...just a lot of fun. The sound with my final choice is as good or better than what I got with the stock MPSA13 & 914 combo.
When using ge diodes, I got the best sound with 3 in series.

[Earthscum]

Since you're on it, try out the diode compression trick. Just stick a diode anode to the base, take a 100k to ground, and cathode is now input. I found a ton of new sounds with it, and I was able to use a couple germs in feedback/bias that didn't work before, and schottky's worked great, I remember.

Easy, give it a quick shot. I tried all manner of diodes, and of course the 4148's are the ones to use for the comp itself. I probably was fooling myself, but I kind of remember trying a matched pair of 4148's and it gave a smoother sound, but like I said... I may have been fooling myself with that one.

[allesz]

Hallo I would like to share my humble bazz fuss diode experience:
I experienced an octave up effect using a 5mm led. The circuit was the standard bazz fuss but I was using a standard transistor (bc547 or bc109, can't recall exactly) not a darlington. 

[mac]

Since you're on it, try out the diode compression trick. Just stick a diode anode to the base, take a 100k to ground, and cathode is now input. I found a ton of new sounds with it, and I was able to use a couple germs in feedback/bias that didn't work before, and schottky's worked great, I remember.

So now you have 2 diodes, the compression diode and the clipping diode?
If so, the clipping diode goes from C to B, or from C to "new input"?

mac

[Earthscum]

Basically you build your BF like normal, but between the input cap and the base you put the diode and 100k.

Cap => (100k to Gnd) => (C)Diode(A) => Base/Diode Junction

Also, as it was mentioned earlier, the BF doesn't actually "clip" using the diode, like normal clippers. What the diode does is allows the transistor to bias the collector and base about the same, i.e. E-B is about 1 diode drop, or .6V. The collector to base is being held at the diode drop you selected. The diode will turn off, but only for a brief moment. If you take a square wave into the BF input, the output looks like one sharp spike, then goes straight to ground, and then starts to rise back up to bias until the other half of the wave forces it back to ground, if that gives you a better visual of what's going on. The "spike" is lessened with amplitude. A single Darlington BF responds well to a smaller signal input, while a single med. gain transistor can handle some more boost at the input before completely sputtering and sounding like trash.

If I get a chance, I should fire up one of the scopes this weekend and feed a BF some signals and put a post up on it. The damn circuit is so simple... it's amazing that it wasn't really discovered until, what we can consider, fairly recently. To me, it is right alongside the FF as one of those circuits that are so simple that they have a soul (if that makes any sense). 

[mac]

Basically we both have added a resistor path to feed the base and let the clipping diode CD to do its job better.
Your method connect the catode of the CD to gnd via a resistor and another diode. This let the CD to conduct a little more feeding the base of the transistor with more current than the original design.

My way is to connect a big resistor Rb between Vcc and the base. As Vc equals Vb then ic*Rc = ib*Rb, so you get the bias condition:
Rc = Rb/hfe

Also, as it was mentioned earlier, the BF doesn't actually "clip" using the diode, like normal clippers. What the diode does is allows the transistor to bias the collector and base about the same, i.e. E-B is about 1 diode drop, or .6V. The collector to base is being held at the diode drop you selected. The diode will turn off, but only for a brief moment. If you take a square wave into the BF input, the output looks like one sharp spike, then goes straight to ground, and then starts to rise back up to bias until the other half of the wave forces it back to ground, if that gives you a better visual of what's going on. The "spike" is lessened with amplitude. A single Darlington BF responds well to a smaller signal input, while a single med. gain transistor can handle some more boost at the input before completely sputtering and sounding like trash.

My view is... was that in the original BF under DC conditions the diode let pass just a tiny bit of current to ignite the transistor, but not enough to have 0.6-0.7v across the diode. IIRC there is no DC voltage across the diode. When extra AC current enters into the base the collector can move up/down one diode drop over the base, for portions of the signals with enough power. Thoughts anyone?

mac