Tonebender biasing - the bain of my existence!

[johngreene]

Maybe it is about what GFR has written at
http://www.geocities.com/sunsetstrip/studio/2987/britface.html
look at the end

That could very well be.
It's been a while but I seem to remember that the main advantage to collector feedback bias is the stability it gives over temperature and variation in Beta. The smaller the feedback resistor the more stable it is at the expense of lower input impedance and lower gain. So when you try to maximize gain and input impedance, you increase its susceptibility to temperature and Beta variance.

[math]
The equation for collector current (emitter grounded) is Ic = (Vcc - Vbe)/(Rc + Rf/Beta). You can see that if Rf is small and Beta is large, the effects of Beta are dramatically reduced. If Vcc is much large than Vbe, then variations in Vbe over temperature are minimal. However, we have a Rf of 500K +/- and Rc of 28k. If Beta is 100 (for example) then the denomiator becomes 28K + 5K. Or the feedback is responsible for 15% of the bias. A 50% fluctuation in Beta and you have a 7.5% fluctuation in collector current. That diode isn't going to help you here.

In the case of the TB, if I use Vbe of .2V and a Beta of 100 as a generic Germanium transistor specs and a mid-value of 250K for the pot, I get Ic = (9 - .2)/(28K + 260K/100) = 288 uA. That would put the collector voltage at 9 - (28K*288uA) = 1 volts. Not ideal. So lets look at the boundaries.
Pot at minimum:
Ic = (9 - .2)/(28K + 10K/100) = 313 uA
Vc = 9 - (28K * 313 uA) = 231 mV

Pot at maximum:
Ic = (9 - .2)/(28K + 510K/100) = 266 uA
Vc = 9 - (28K * 266 uA) = 1.56 V

Hmmm. Seems that for the nominal Germanium values, you are not going to get much more than 1.5 V on the collector of the output transistor.

The higher Beta is, the lower your collector voltage will be and the less effect the 500K pot will have.

If Beta is at the low end, say 75:
Ic = (9 - .2)/(28K + 510K/75) = 253 uA
Vc = 9 - (28K * 253 uA) = 1.9V

[/math]

I just can't see how this thing is going to bias up very high unless you change the collector resistance. 

--john

[Gus]

I do wonder what a known good TB original transistor operating points are.

[johngreene]

I do wonder what a known good TB original transistor operating points are.

After crunching the numbers, so do I. 

[gmr1]

Me too! I've seen the numbers quoted differently on several sights. I wonder if they'd be the same for this circuit, as it's slightly different than the original.

I've managed to get the Q3 voltage to just about 4v, .5v shy of the magic number. Brought the 18k resistor down to nothing, and went back to the 500k trim pot. I'd love if I didn't have to crank the trim to get the voltage, as it's still farty the more bassy the note played. It's just ok. With a light touch it sounds quite nice, but when you dig in, fart city. Really compressed and choppy. I'm going to build a fuzz face this weekend, and see if I have better luck with that!

Thanks again for the education y'all.

If you ever need an all-grain recipe for a Baltic porter or something along those lines, lemme know... that's more my field of knowledge!

[kismet78]

I'd love if I didn't have to crank the trim to get the voltage, as it's still farty the more bassy the note played. It's just ok. With a light touch it sounds quite nice, but when you dig in, fart city. Really compressed and choppy.

I've never actually played one, but that description sounds like every audio sample I've ever heard. Maybe you just don't like the sound of a Tonebender?

[mac]

I experimented with different transistors and diodes,
http://www.diystompboxes.com/smfforum/index.php?topic=64677.msg512877#msg512877

mac

[johngreene]

If you ever need an all-grain recipe for a Baltic porter or something along those lines, lemme know... that's more my field of knowledge!

I have a great recipe for an Oktoberfest. Even had it published in a brew magazine.

--john

[gmr1]

I experimented with different transistors and diodes,
http://www.diystompboxes.com/smfforum/index.php?topic=64677.msg512877#msg512877

mac

This is a good read. The Q3 transistor has almost no leakage (I have a DCA55 that I checked them on - it registered close to zero). Could that actually be a bad thing in this case? I've read that silicon is a no go, because the circuit requires some leakage, but didn't think that would be an issue with any GE transistor.

I have a friend who I play music with that is a hell of an amp repairman/pedal builder. Stan Day, he builds the Coleman pedals - his "Astro Fuzz" is pretty well know around Boston (and maybe elsewhere), and he said from listening to the pedal, that bias isn't the issue, and he thinks my transistor choices are suspect. He suggested a matched set from Small Bear as a reference point, then I know I have at least 1 set of transistors that are "right". I have tons of GE trans, so I'm sure I have plenty that are good... maybe just not matching them well, but more likely I screwed up the circuit somewhere. He also wasn't keen on the way the 3-know versions controls work, but I think that's just a personal thing.

Turns out it's killer as a bass boost with a jazz bass. Tons of boost, with a touch of compression and only gets fuzzy when you really slam the input.  A happy surprise. I might just start over, and use this one for bass...

John - congrats on getting a recipe published! I've avoided the competition scene so far, but may enter a few this year. Of course everytime I brew a new recipe, people are telling me how it should be in stores - if only it was that easy!

[mac]

Try a (50 - 70; 50ua - 150ua) Ge at Q3.
BTW, I do not like the fuzz control. Also the tone control gets to trebley, I prefer to increase the 2.2nf to 4.7nf or 10nf.

mac

[Gus]

 Found specs for the 1n270 Gold bonded with reverse leakage of 100ua at 50V.  Question does leakage in Ge diodes go up at lower reverse voltages?

Maybe the circuit is meant to bias at lower collector to emitter voltages and trying to get it to 4V or more is changing the original sound to something else.  What did the real one use for Q3?

[mac]

Voltages from an original circuit should be nice.
But we could made some good estimates if we know a little about average gain and leakage of those old Ge, OCxx IIRC.

I was thinking of a biasing trick: replacing the diode with a large pot and vary it until it sounds good, measure the pot value, replace with a thermo resistor (neg or pos?).

mac

[johngreene]

Found specs for the 1n270 Gold bonded with reverse leakage of 100ua at 50V.  Question does leakage in Ge diodes go up at lower reverse voltages?

Maybe the circuit is meant to bias at lower collector to emitter voltages and trying to get it to 4V or more is changing the original sound to something else.  What did the real one use for Q3?

Now that is interesting! It can't be that high at lower voltages otherwise that transistor would be biased full on.....  The base current without the diode is around 3 uA if I figure this right. The diode can't be leaking much more than a microamp or adjusting the 500k wouldn't do anything.