Troubleshooting a DBA Fuzz War

[Skajaquada]

Hey guys, I built Johnk's 7 transistor 2 knob Fuzz war, and I'm having problems with it.

Before I start, I'll say that I changed the 20n cap to a 22n because I don't have any 20s.  Also, the 680p cap is a 560p in mine, I think.  Also, I changed the 2n5089s to 2n3904's due to the explanation in the thread that the 2n5089s are problematic.

Basically, here's the short version of what happened so far.

I built the board, wired it up, and when I tested it, the LED came on, but I barely got any sound out.  Seems like there were spikes.  I put it on my test box to eliminate the offboard wiring, and it worked perfectly.  Then I came to find out that the test box was only supplying about 1.3 volts.  I corrected that to about 9.5 volts, and the pedal stopped working again.

I took some transistor measurements, and here's what I found:

Input voltage 1.6V:
Q1 - e - 1.507
b - 0.550
c - 0.087

Q2 - e - 0.019
b - 0.550
c - 0.005

Q3 - e - 0.612
b - 0.533
c - 0.005

Q4 - e - 0.063
b - 0.532
c - 0.004

Q5 - e - 0.064
b - 0.532
c - 0.004

Q6 - e - 0.594
b - 0.511
c - 0.003

Q7 - e - 0.742
b - 0.509
c - 0.002

Input voltage 9.26V:

Q1 - e - 8.89
b - 0.642
c - 0.173

Q2 - e - 0.043
b - 0.642
c - 0.036

Q3 - e - 0.391
b - 0.627
c - 0.043

Q4 - e - 0.050
b - 0.611
c - 0.028

Q5 - e - 0.050
b - 0.611
c - 0.028

Q6 - e - 0.664
b - 0.584
c - 0.022

Q7 - e - 1.509
b - 0.580
c - 0.021

Does any of this information offer you guys any insight as to what my problem may be?  I noticed on Johnk's modded 7 transistor version, the resistors on E and B of Q1 are basically swapped.  Could that be the problem here?

[Skajaquada]

On the other hand, given that the circuit works perfectly at 1.6V, would I be able to solder in a resistor between the 9V and the board to drop it down to approximately that much?

[Muadzin]

From what I can gather the 7 transistor version is notoriously difficult to get it to work.

[tjdracz]

Don't think anyone had much luck with that. Weirdly designed circuit which depends to much on certain HFE/transistors.
Probably you would need someone to measure the original voltages to try to get close without massive changes.
Or just stack two or three Muffs together. Similar enough methinks

[tjdracz]

Also, come to think of it, try to replace them 2N5089 with MPSA18. Seen at least two people mentioning they managed to get it sound right with those.
Actually, I remember some other DBA pedal which used 2N5089 that nobody seemed to get right until they swapped it for MPSA14 or something. Leads me to believe that Ackerman and co. have a crapload of really weird out of spec 2N5089 that make it work alright.

[rocket88]

DBA is an interesting group. Most of their designs when you look at the schematic they shouldn't work, but they do. I have to agree with tj, that they must select specific transistors for their effects, and this may be one reason they went from the 7 transistor to 4 transistor design.

[tjdracz]

DBA circuits make little sense from the good design perspective (and have you seen soldering on those? I mean mine's dodgy as hell but I swear it's better!)
I might try to do the fuzz war on the weekend just for laughs but the way I'd do it is to stick to schematic values, socket all the transistors and mix and match them til it sounds good. Preferably also isolate t the stages to make sure one works before moving to the next.
But again, I think people had good results with MPSA series and 2N3904's so I'd start there.
Would be useful if someone put the voltages of the traced units but seems like those are non-existent.
God, it seems like it's more of a pain than Tonebender!

[Skajaquada]

I actually did replace the 5089s with 3904s for that very reason.  It's driving me crazy though.  It doesn't make sense that the circuit would work when I feed it 1.6 volts, but not when I feed it 9...  I'm almost at the point of soldering in a trimmer or something to reduce the voltage going to the circuit.  At least it would work then :P

[rocket88]

Fucking DBA. I had one opened and tj is spot on about their quality. It was one of their >$300 effects and the soldering looked worse then when I first learned to solder. Shit I bet if I taught my gf to solder she'd do better then those guys. It's really disappointing that they get away with charging that much for shit work, and designs that shouldn't work.

Ska, what hfe are the transistors you used? I'm wondering if you try those that are lower then normal or higher then normal of it will effect the pedal. If I'm right, the gain of the transistor is dictated by the supply voltage, so if it works with very low supply voltage, then maybe low gains will work better because they will reach the right gain. But then again, their effects make zero sence to me, since they shouldn't work.

[Skajaquada]

I'm still newb enough to not know what hfe is or how to measure it, so I couldn't tell you exactly lol...

[Travis]

Hfe is basically just a number that quantifies how much "gain" a transistor has

Many digital multimeters can measure the hfe of silicon transistors. I know the super cheap one from Harbor Frieght does. It's an essential tool..

Even better is something like the Peak Atlas DCA55 semiconductor analyzer. This can not only measure hfe of silicon transistors, but measure hfe and leakage of germanium transistors, as well as identify pinouts. It's a very useful tool

[Skajaquada]

My DMM is rather old and cheap, and doesn't have an hfe setting...  Is there a way to measure it without that particular function on a DMM?

[tjdracz]

Skajaquada wrote

My DMM is rather old and cheap, and doesn't have an hfe setting...  Is there a way to measure it without that particular function on a DMM?

There is. Look here:
http://www.eevblog.com/forum/testgear/how-come-transistor-beta-hfe-test-is-only-on-cheap-dmms/msg441097/?PHPSESSID=5a635031a2f28685ae111fbd7c3f1e49#msg441097

Yet, for the hassle, I would just buy cheapest Digital Multimeter. In UK you can get one under £5 and it will be good enough and it will have sockets for HFE measurements

[Skajaquada]

Thanks for that link.  However, I don't quite understand that circuit...  It has two transistor sockets (I'm assuming so that you can test either type)

Would I be correct in assuming that I could leave out, for example, R2, R3 and the NPN socket if I were ONLY testing a PNP?  Or vice versa?

This looks like something I could accomplish on a breadboard, so I don't have to actually solder together the circuit to make it work...  I could even get the 3V right with my test box... hmmmm...

[tjdracz]

Skajaquada wrote

Thanks for that link.  However, I don't quite understand that circuit...  It has two transistor sockets (I'm assuming so that you can test either type)

Would I be correct in assuming that I could leave out, for example, R2, R3 and the NPN socket if I were ONLY testing a PNP?  Or vice versa?

This looks like something I could accomplish on a breadboard, so I don't have to actually solder together the circuit to make it work...  I could even get the 3V right with my test box... hmmmm...

Yes, if you leave those two resistors and a socket, you can still measure PNP.

This is how it's done on DMM usually. The additional E is just there for sake of avoiding bending the legs when testing ECB transistors (many japanese types and so)

[Skajaquada]

Alright, cool.  I'll breadboard that when I get home, and I'll see if I can get some readings to pass on.  Thanks again!

[Skajaquada]

So the hfe seems to be around 180 each.  Does that tell ya anything?

[Travis]

That seems a bit low for 2N5089

Just to clarify, earlier you said you only need to measure PNP transistors, but 2N5089 is NPN. Are you sure you're measuring these properly?

[Skajaquada]

I stated earlier I was using 2N3904s.  I measured them as NPNs in that circuit, and left out the PNP parts.

[tjdracz]

Actually.. maybe try something different. Have you got anything high gain like BC550C or MPSA18? If yes, try that.

DBA's are sure pain in the ass. They use reverse beta property of transistor which makes them super dependent on magic parts. You can have a read here: http://www.muzique.com/lab/reverse.htm if you're interested in the theory behind it