MK11 Tonebender (fist PNP build !!)

[analog kid]

RG or anyone else... I pose another question to you. Is it possible that there is NOTHING wrong with this circuit IF>>
I have been choosing my transistor gains by using the DMM plus 1Meg resistor method that Steve Daniels posted at S.B., Now I thought I was getting accurate results HOWEVER I just got the little circuit put together from the GEO site (except using two 1.2k for the 2.472value where I only get 2.4 exactly) and am getting WILDLY lower true hfe gains using it!!! So is it possible to have a circuit's biasing to behave this way if the transistor gains of Q1-Q3 are far too low??
I am getting pretty much lower than 60-65 for ALLTHREE positions!!! I realize this would surely cause a problem but would it cause THIS problem?
Only one way to find out , I 'll keep going til I find a suitable set of gains to stick in there and see if the effect behaves any differently.
PS: as far as the GEO method , Am I right to believe that an 'Open Base' (leakage) reading of .741v would mean a leakage of apprx  300uA ? ?
(and  .988v = 400uA etc.....)?  ?

[R.G.]

Sorry - I've been pulled into other things and only now had a chance to actually think about this.

I actually think you may have solved your own problem. The chances are that Q2 is too leaky, which turns Q3 essentially off.

You are correct - if the base of the transistor you're measuring is open and you read 0.741V across the collector resistor, then there is about 308uA of current flowing in it. That means that it would drop 308uA times 100K = 30V across the 100K collector resistor. Obviously, there's not that much voltage available, so it simply goes as close to ground as it can and stays there.

You can test this theory by shunting the 100K collector resistor of Q2 with a 15K resistor and seeing if you then get some signal through. The gain will be way down, but it might work.

Yes - test more transistors. The gain you get isn't all that crucial at this point, but you really have to get leakage under 100uA and preferably under 30 for this circuit to work well.

GOOD THINKING!!!

[analog kid]

Ok , couple of questions.
First off what did you mean by SHUNTING the 100k resistor with a 15k resistor??? I take it you don't simply mean REPLACE ?

Second , I don't have a battery of exactly 9vlt for my tester so I  was using one of 8.20vlt. I then went and bought a new one which of course became the only time I can think of that I had a battery that was TOO GOOD! It test at 9.6volts. So would the 8xx battery or the brand new one be best AND how far off does this throw the readings?? ( as how does the 2.4k resistor value rather than exactly 2.472 affect them??

Lastly, I don't think the Q2 was the tran. I was referring to reading 300uA leakage but the three in there were all between 200 and 300microamps on Steve Daniels test AND your GEO method> (WHICH BTW I GET ALOT LOWER HFE's WITH THAN STEVE'S) The bad news about that is that of ALL the different Ge's I have and have been testing , These are actually some of the lower leakages!!! Out of AC188s , AC123's , AC125's and OC75s

 I thought that 300uA was to be suspicious of but OK to use and that 200uA and under would definitely be OK. (?)

Matter of fact I don't think I 've came across but ONE that came under 100uA leakage which was a Raytheon Tseries which I have but one of and a gain of 135 ( I'll use it in Q3 I guess) BUT I FEEL I'll be hard pressed to find any other from my stock that are below 100microamps leakage and If I find any close to that I'm pretty sure the gains would be quite lower than 60  ... which I've been throwing to the side.
 SO which is truely more important in this, or any other F.F. , Very low leakage OR the 60-130 hfe gain range??

[petemoore]

Around 7 volts things may begin to act funny...misbias or start cutting out. Around 14v I start worrying about overvoltage on the 16v caps I generally use.
 Anything around 9v, just figure 1/2v for biasing a FF...
 1/2 of 9 = 4.5
 1/2 of 9.6 = 4.55 this small a difference in bias [like if you're running a battery and the voltage drops over time] you probably wont be able to notice much if any differnce in function.
 1/2 of 13v = 6.5v...

[petemoore]

Around 7 volts things may begin to act funny...misbias or start cutting out. Around 14v I start worrying about overvoltage on the 16v caps I generally use.
 Anything around 9v, just figure 1/2v for biasing a FF...
 1/2 of 9 = 4.5
 1/2 of 9.6 = 4.55 this small a difference in bias [like if you're running a battery and the voltage drops over time] you probably wont be able to notice much if any differnce in function of a FF.
 1/2 of 13v = 6.5v...

[R.G.]

SO which is truely more important in this, or any other F.F. , Very low leakage OR the 60-130 hfe gain range??

Yes.
.
.
.
.
.
You can't really sacrifice either, really.

The problem with leakage is that it's a DC current that you can't control. That means you have to warp other things in the circuit to work around it.

In this case, you have a 100K collector resistor, and you want that collector sitting at a couple/few volts. Let's say we want 2-3V at the Q2 collector. That means 6-7V across the collector resistor, and that's a current range of 60uA to 70uA. Right at this point you realize that a transistor with a leakage of more than  70uA will not work in that circuit with that collector resistor.

So you modify the circuit. First impulse is to run the collector resistor down (that is what I proposed, paralleling the 100K collector resistor with 15K). That gets the voltage across the new 15K collector resistor down to 15K * 300uA = 4.5V.  But you have sacrificed all the gain that the 100K would have given you.

Another thing that works sometimes with germanium is to suck base current out of the base. That's commonly known as leakage biasing. It's actually leakage un-biasing. You put a resistor from the base to some lower voltage to suck some of the leaked current out.

In this case, you have a gain of about 65, a leakage current of 300uA. The leakage from the collector that's causing this mess is 300uA /65 = 4.6uA.  If you could ground the base, it would most of that current out the base. But you can't really do that and also expect gain because it also sucks the signal to ground. If you use a resistor sufficient to suck some of this current out the base, you can get a lower effective collector leakage, but again sacrifice gain because you load the signal down at the base. In this case, you need to get something like 3uA out of the base at the normal base voltage of maybe 50mv, so that would be something like 50mv/3uA = 166 ohms.

One thing that experienced germanium designers did was to use a voltage more positive (in a positive ground system) than the emitter. With a suitable back bias like this, you can use a much larger base-sucking resistor. That's not too practical here either.

You begin to see the problem. Never mind that leakage doubles ever ten degrees hotter, either.

So the answer is - you select germanium transistors for characteristics you want, and never assume that just because you have a circuit that's worked with other germaniums, it will work with the unselected ones you have.

When we started into this, I think I mentioned that germanium was harder. I had not fully realized just how close the edge was where a leaky device could simply not work. I never worked with germaniums in a production design; I was taught to worst-case design with silicon, and picked up some of the nuances of germanium from the (I thought) gray-haired engineers in the department. Now that I think about it, they were probably only in their 30's and working with germanium made them gray.

[analog kid]

Alright RG, this was a good guess at it (and could still be the problem ) BUT  not much changed even when putting in the lowest Leakage Germ.s that I have As Well as the proper matching gains for the circuit.
I"ll list exactly what I put in there, and what I got . but first...
Don't think I"m dismissing this theory completely , These are just the lowest leakages I can find and I assume that with these in there SOMETHING would improve!! I will still try to do the resistor mod and see what happens BUT I just want to make sure if YOU"RE NOT OUT OF IDEAS for the problem IF this isn't IT???(I'm sure you're NEVER out of them though!)
Here's  what I got and what I used
Q1  65hfe leakage of 120uA  (AC188)
Q2  85hfe leakgae of 180uA   "      "
Q3 130-140hfe leakage of 75uA  (T59247)
Battery voltage 9.44v
Q1 E 8.18v
    B 0.06v
    C 0
Q2 E    0
    B  .050v
    C .083v
Q3 E  .027v
    B .082v
    C 9.0 to 9.38v (This fact that the Q3 Collector trimpot STILL will not bias or hardly work at all is really what makes me wonder if there's still a circuit problem and not to do with the transistors. Although it does work better with a Silicon in Q2/Q3. What do I know!!?)

What would just putting an 8.2k resistor in there on the Q3 Collector instead of the trimpot do

[petemoore]

Just stick in a 'big' resistor, largest value you think you might want, and a 2 connect socket connected to the R, paralleling it.
 say you start with 18k for the hardwired resistor, by paralleling an 18k [socket or otherwise] you now have = 9k resistor
 to lower R value on Q3C, just lower the parallel resistor.
 To find out about what resistances your Q3R 'circuit is providing use the DMM...and another 18k with various smaller than 18k value resistors parallel to it, or pull Q3 and measure across Q3CR.

[petemoore]

Just get a trimpot..or
 1use a regular pot...to set bias,
 2 then remove and measure pot resistance [don't turn the knob during step 2]
 3 replace where the pot was with a resistor about the same value as the measurement in step 2
 You can probably work with a pot between 10k - 100k...using a resistor to taper it...if you have such a pot available post it's value.

[petemoore]

Still I have the hunch there's maybe a wiring error on the board.

[analog kid]

Alright RG, this was a good guess at it (and could still be the problem ) BUT  not much changed even when putting in the lowest Leakage Germ.s that I have As Well as the proper matching gains for the circuit.
I"ll list exactly what I put in there, and what I got . but first...
Don't think I"m dismissing this theory completely , These are just the lowest leakages I can find and I assume that with these in there SOMETHING would improve!! I will still try to do the resistor mod and see what happens BUT I just want to make sure if YOU"RE NOT OUT OF IDEAS for the problem IF this isn't IT???(I'm sure you're NEVER out of them though!)
Here's  what I got and what I used
Q1  65hfe leakage of 120uA  (AC188)
Q2  85hfe leakgae of 180uA   "      "
Q3 130-140hfe leakage of 75uA  (T59247)
Battery voltage 9.44v
Q1 E 8.18v
    B 0.06v
    C 0
Q2 E    0
    B  .050v
    C .083v
Q3 E  .027v
    B .082v
    C 9.0 to 9.38v (This fact that the Q3 Collector trimpot STILL will not bias or hardly work at all is really what makes me wonder if there's still a circuit problem and not to do with the transistors. Although it does work better with a Silicon in Q2/Q3. What do I know!!?)

What would just putting an 8.2k resistor in there on the Q3 Collector instead of the trimpot do

[analog kid]

Ok pete I take it you've saw a put in the lower leakage trannies and really didn't get a change in bringing the voltages up and still have the Q3C pot not really working,  so DO I take it correctly that you don't think the Q2or Q3 being too leaky was/is the cause of this??? I can't imagine that it ismyself , That's why I was hoping that RG , and yourself aren't out of idea And it looks like you've given me a couple good ones.
SO..... putting a larger (22k or higher??) resistor in ? Is that IN PLACE of the trimpot?? SO I need to remove it., put this resistor in and then a parallel socketed resistor along side it,  Is that correct?
IF that's what you're saying, first I don't know what this is for to tell me (which I'll leave to you guys until I understand) and second IF the trimpot adjusted .5 to 21k is not allowing a change in bias voltage of the collector THEN WHY will setting up this variable resistor on it give adjustment??
Lastly , I don't understand your post about adjusting bias with a regular pot or trim:  "Just get a trimpot..or
1use a regular pot...to set bias......"
Now I do have trimpots other than the 20k that is in there. I have 10k, 50k , 100k right now.  Is this a separate idea from the parallel socketed resistor idea and either way are you telling me I should put a different value trimpot in the place where the 20k is. ? I'm sorry I just got a bit lost there.
thanks

[analog kid]

Alright!! We must be on the right track. I did what RG said and put two 8.2k resistors in series across the 100k Q2 C/Q3B resistor,  which brings it to about 14 to 15k.  I tested the Q3 collector bias first and BAM!!! IT Now adjust as it should from between .4v to about 8vlts!  Q3 Base is at .812v BUT the Q3 emitter is only as low as .615vlts!!
Q2's Collector is reading OF COURSE .812 as well BUT AGAIN it's BASE is only at 100mv!! I know this base current is normal on Q1's base but I've never saw a reading this low for Q2's Base.
Q1 is still the same 7.52 v for the Collector and 60mV BASE
So since this gave some results do "we" now know what the problem is most likely and what should be done?? If the relationship between Q2 and Q3's  Base /Emitter actually are alright which I'm not sure especially on that small of a voltage differential of Q3 Base/Emitter. I guess the only question on Q2 is if the 100mV at BASE is actually enough above 0v Emitter to let the thing amplify , I guess (?)from what I understand.

Is it ONLY too high of leakage since this worked??!! If so I don't understand how under 200uA on all 3 could cause such a problem. and if so you'd think it'd be documented (by Phillip) and everywhere that the MK11 will NOT tolerate hardly any leakage.!!   But what to do now if so, I don't know where I'll find lower leakage Ge's and I don't imagine it's Practical to keep this resistor at 15k!??
If that's the only option is there an in between that will remedy it, I notice the Marshall Supra Fuzz (Identical besides a couple values) uses a 47k , whichI thought of doing before! WOULD That NOT be enough to get these vltges though??
 I'll let you guys mull this over

[petemoore]

Don't exactly knw, but suspect those transistors are 'workable' to good to great...
 I did up a TB a week or so ago, and it has a Q3C trimpot, the thing pretty much fired right up, and sounds about best.
 These things are close enough to 'right' and then you know it...measuring voltages after the point of great sound is less important IMO moot.
 Sounds like you have enough good transistors to work with, and when a TB acts like it's ready, you'll know it.
 Plan A Trimpot
 Plan B Pot used to Find the right resistance, then replaced with a resistor.
 Plan C 'relatively large' resistor for the application, using same size resistor to halve the resistance in parallel or progressively smaller resistors in parallel to 'adjust' resistance 'there'.
 I prefer plan A by a longshot.
 If you have a 50k pot, stick a 22k resistor across lugs 1 and 3, then use that {= of 20k pot or close enough for trimming] pot to set bias, using lugs: wiper and an outside lug.  Leave it in or replace it's resistance with a fixed resistor[s.

[analog kid]

You're confusing me a LITTLE bit Pete. A trimpot or the other options that you mentioned are a great solution to a problem of biasing and getting these voltages right, BUT the problem is that I obviousy HAVE TO knock the size of the Q2/Q3 Collector-Base 100k resistor all the way down JUST TO GET a trimpot TO WORK !!!!!!!
that's why after discovering this , I asked if maybe I can compromise and just knock that value down around half or so to correct the problem AND to still have decent output volume.
AND there's the question of whether or not those voltage "divides" for Q2 and Q3 Base and Emitter were wide enough?

(also Pete:  You said use a 50k trim and a resistor to get a 20k value, You do realize that a 20k trimpot IS what I've had in there all along don't you?? I wondered if maybe I just didn't follow your meaning?)

[petemoore]

A 20k trimpot on Q3C should do nicely...but...if the transistors isn't biasing with 0k-20k, I suspect the bias problem isn't with the 20k trim...it's probably somewhere else in the circuit.

[analog kid]

Well as the voltages I last posted showed , the adjustment of the Q2 Collector resistor down to 15k Really made a difference!! Matter of fact , this is almost where I want it aside from a few things. One being of course the gain/volume diffence between effect and bypass with that low of a resistor value. I wonder where the cutoff point is to "let this thing work" if 100k is obviously too big ? Since the Marshall version of teh MK11 uses a 47k there I wonder if that would be enough to correct the biasing problem and yet still let enough volume through??
My other problem with it may go along with the volume difference from the low value I don't know, but the fuzz doesn't really clean up until the Guitar Volume is backed WAY off.  And THIS may be inherent to the Tonebender but it doesn't even clean up much when the FUZZ control is backed ALL the way OFF , and at the opposite end when FULL UP the noise sounds like an ocean roaring , cut it back to 9 and it's almost gone.

Oh , last problem:  I noticed my battery started out last night at 9.2v , when I  woke up today and actually plugged in and tried out the pedal the voltage of the battery dropped to 8.7volts after about 15minutes of playing and felt a little warm. !! > Is that a bad sign , I've never had a PosGround effect.   (I've an led with a 2.2k res on it)
Any comments on those voltages? Or what's the best solution as alternative to that 15k Q2 collector resistor??

[petemoore]

Measure the current draw...open the + side of battery clip and battery, insert DMM to ... 'complete' the circuit...see if you can figure out how to read small' current readings.
 Daft Bender...I'm off again, mines' seemingly bias drifting again, I turn the gain knob down, and after a Minute I can turn it up again, but only to a certain level, anything much past that certain [just enough gain fortunatley so far] point, and it'll start 'actin' up' again [osci-boating], Next time I'll try the trimpotss inside [Q's 1 and 3's collectors...
 How does that anti drift diode go in...what type was it...the B/E drift fix diode IIRC...supposed to help stabilize the Ge's.
 Tonebenders and FF's for years I'm STILL figureing it out it seems.
 Does anyone know which of the TB transistors is most prone to drifting [if there is one]? I'd guess: Q3, sees the most current and sets bias for Q2?

[analog kid]

Pete, If I took that reading correctly (red meter lead to circuit ground and black to the battery POS, although that seem backards to me since it's a Pos ground crkt?)  The current draw (in milliamps) that I got was .10mA

Is that supposed to tell me why my battery drained so much??

And I don't understand the other part of your post. . . ?  Is that the experience YOU"RE having with a Tonebender you built??

NOTE::  I just replaced the 100k Q2collector resistor with easy change sockets(since it seems to have control over my Problem Area) and initially I tried the 47 k value that other versions of the MK11 use, this didn't give much responsiveness to the Trimpot, I then went to 33k and got a bias adjustement on Q3Collector trimpot from 4.75 to 9vlts  BUT it would'nt allow the B and C of q3 and C of Q2 to come up past a few hundred mVolts.    
I final drop to a 22k resistor brought me a trimpot adjustment range of 2k to 9k brought the previous transistor legs up to voltages I got and posted most previously (.584v B  .449C on Q3  and 92mV on Q2 Base)
I WAS GOING TO ASK WHY the resistor value on Q2 C would be able to affect HOW Q3's collector will bias... but I think I'm undersanding now that Q3 won't bias properly AT ALL  UNTIL the current to Q2's collector is high enough!!  Is that correct??
   If so what would trying to put a 100k trimpot on Q2's collector accomplish?

[petemoore]

well I've tried 100k's on FF Q2's...I use 33k-47k on Q2 of Ff collector [same as Q3 on TB].
 probly works with a 100k, I think I've seen FF derivitaves use that higha resistor...I'd stick with the [is it 47k on the TB schematic?]
 Tonebenders...they can be tricky...yep I'm talkin' about my latest TB encarnation...seems to drift a bit...as of today, was quiet as a mouse till today...hmm