Let's talk about the Jordan Bosstone. An odd one...

[John Lyons]

I'd like to order the part(s) needed to rid of any oscillations, should they arise (as they seem to be common).

Use low gain transistors. 200ish hfe or lower

[kinski]

Right, 2n2222 and 2n3906 are under 200 hfe, yeah?

Also, will either of those layouts work? They seem to have differences. Like the 2.7k resistor, etc.

Thanks!

[brett]

Hi
modern 2N3906s are way over 200.  Even 2N2222s are too high for older circuits.
Try BD139s or other medium power (2 to 10W) devices instead of old small-signal transistors.  hFE around 150.  TIPs, MJEs, etc are also good.
cheers

[mac]

Just speculating on my part, but for fun follow this path and ignore the other components:

NPN collector, up the 18k resistor, stop.  Over to one side is a .022 cap to ground.  Turn left and head down the 560k resistor, stop.  Hey, another .02 cap to ground!  Continue heading toward the NPN's base along the next 560k resistor, then stop at the base.  With the input control grounded, why look!  There's another .02 cap to ground!  We got ourselves a phase-shift oscillator!

IIRC gain has to be above 29 for stable oscillations. Can you estimate maximun hfe to be below that limit?

mac

[Dingus]

Nashville version:

My apologies for the thread necromancy, but is that a 2n3565?  I hadn't seen a version that used that, maybe I just need to read up on this circuit some more.

[skrunk]

most likely.
from my notes the Nashville used a 2N3565 and a M924/95232, though there may have been other combos used(?).
the gains of 2N3565s can vary quite a bit though, mine go from hfe 150 to 600, so I'd make sure to use a low gain one if you use one.

[Fancy Lime]

It's alive! It's aliiiive!

So Ive been trying to test the versatility of this old favorite of mine by playing around with different kinds of gain and tone control. One thing that may be interesting to some people: I had oscillations in certain combinations of gain, tone and starve control (like in the Trombetta/Rockett WTF, highly recommended mod for this one, sounds fantastic and makes it much more versatile, going from a farty mess to brutal imploding duck-and-swell fuzz to rather nice distortion). Placing a 3n9 cap across the clipping diodes killed that dead without cutting an awful lot of highs.

But there is something that I just now realized I don't understand: What the actual funk does C3 do? If Q2 were an NPN and upside down, then C3 would be between Base and Collector and make perfect sense as a treble cut on Q2 (as seems to be claimed or insinuated in all threads on the topic I found here and elsewhere). But it's not. As it is C3 is between Base and Emitter of a common collector stage, and since V_E = V_B there is no voltage across C3 since both its ends move in unison. So all it does is bypass R6 from the perspective of Q1. This does indeed cause a bit of a high cut above 176kHz. Have I gone completely nuts now, or is this the most woefully ineffective way of filtering out radio frequencies? Wouldn't placing this cap from Base or Q1 to ground or a smaller cap between Emitter and Base of Q1 make a lot more sense? There also seems to be some confusion about the value of C3 in the tracing threads:
http://www.diystompboxes.com/smfforum/index.php?topic=48469.0
http://www.diystompboxes.com/smfforum/index.php?topic=57936.0
Is it at all possible that this is supposed to be a 500pF (470pF) instead of 50pF (47pF) cap? Would make at least a bit more sense as a radio frequency filter in this strange position.

For reference: I am still referring to this schematic:



Thanks and Cheers,
Andy

P.s.: Sorry to those who oppose resurrecting old threads but in this case I think it's useful to have all the info in one place.

[PRR]

Have you figured out the Q1 Q2 connection?

Taxonomy aside, this is a VERY high-gain stage with two similar HF poles. It loves to oscillate MHz.

50p seems small to me, but if it works don't mess.

> a common collector stage

No.

[Rob Strand]

But there is something that I just now realized I don't understand: What the actual funk does C3 do? If Q2 were an NPN and upside down, then C3 would be between Base and Collector and make perfect sense as a treble cut on Q2 (as seems to be claimed or insinuated in all threads on the topic I found here and elsewhere).

The value of C3 came up about 15 years ago on this forum, someone reported it was in fact 47pF (and the other ceramics were 20n).  It does bugger-all to the tone.   

I have a feeling it might have been added to stop the thing oscillating as it helps roll the gain off to a safe level in the same frequency region that stray capacitance would cause oscillation.  Maybe not the best way to do that but it is what it is.  Looks like even 1n would work.

[IIRC, problems with that effect inspired Aron to start this group.]

[Rob Strand]

a common collector stage

No.

It presents a larger collector resistor to Q1, which increases the gain.   You wouldn't be able to get such a high gain without it.  Note the DC biasing is still "good".

[Fancy Lime]

Hi Paul, hi Rob,

> a common collector stage

No.

No? Why not? Since the Collector is common to Q2's input and output, I thought that would make it a Common Collector, regardless of the role the transistor plays otherwise. Well, maybe that's not how the nomenclature works, then.

50p seems small to me, but if it works don't mess.

Is that a dare Challenge accepted...

Anyway, here is how I think the Q1-Q2 arrangement of the Bosstone works, please correct me where I'm wrong:
If we take out Q2 and R2 the whole thing is a completely normal Common Emitter stage with voltage divider bias (with AC cut for higher gain and impedance but that's not important here) and clipping diodes at the end (also not important). Almost like a Linear Power Booster with Electra Distortion clipping. Putting Q2 and R2 back in, we bootstrap this thing by moving the positive supply at the R2-R6 junction up and down in unison with the output of Q1.
When B_Q1 goes low, Q1 closes, C_Q1 goes high. Now Q2 tries to keep its Emitter and therefore the R2-R6 junction at the same voltage as C_Q1 by closing.
When B_Q1 goes high, Q1 opens, C_Q1 goes low. Now Q2 must also open to pull the R2-R6 junction down.
Is that so? Please enlighten me with the correct terminology for this sort of thing.

If my analysis is correct and if C3 is supposed to kill oscillations, wouldn't it make more sense to make C3 go from the R2-R6 junction to ground, or in the negative feedback of Q1, or from the R2-R6 junction to B_Q1, or from B_Q1 to ground? I have had some problems with oscillations on the breadboard despite a 68p mica cap as C3. Sticking in a 3n9 cap parallel with the diodes (which is the same as from the R2-R6 junction to ground for high frequencies) took care of that without noticeably changing the sound. I'm just wondering if the placement of C3 in the original is one of these "it sort of works well enough so lets not think about it anymore" things that could actually be done better. Since many people report oscillation problems with the Bosstone, I would think that this is not as moot a discussion as it might seem at first.

Now that I spell it out like that, this also explains why this circuit reacts so well to replacing R2 with a (100k-250k) pot since R2 controls how effectively Q2 can do it's thang on the up or down swing.

It presents a larger collector resistor to Q1, which increases the gain.   You wouldn't be able to get such a high gain without it.  Note the DC biasing is still "good".

I'm afraid I don't quite follow. Could you elaborate on the "larger Collector resistor" analogy/explanation? It "feels" like this might be a different way to look at what I described above but it seems to me that the "virtual Collector resistor" made up of R2, R6 and Q2 is not merely large but changes value in accordance with the signal. Also: What is "good" DC biasing, in this case and why wouldn't it be good?

I have a feeling it might have been added to stop the thing oscillating as it helps roll the gain off to a safe level in the same frequency region that stray capacitance would cause oscillation.  Maybe not the best way to do that but it is what it is. Looks like even 1n would work.

Is it just me, or does the Bosstone have more post sentences starting "I have a feeling it might..." than most other effects? This really is a quite unique beast and I wonder why. It sounds fantastic and is enormously versatile yet it seems to have very few descendants expanding on this topology. There are clones galore but they mostly just plain copy the old original (often right down to using ceramic caps and carbon resistors) or make modest changes rather then developing upon it as has been done very extensively with Fuzz Faces and Big Muffs. Curious, that, isn't it? Lack of a "Technology of the Jordan Bosstone" article, maybe?

Cheers and thanks,
Andy

[antonis]

No? Why not? Since the Collector is common to Q2's input and output, I thought that would make it a Common Collector, regardless of the role the transistor plays otherwise. Well, maybe that's not how the nomenclature works, then.

Quite right (and quite wrong)..

If you consider Q2 stage as "stand along" you could name it CC but if you consider it as a "load" to Q1 (roughly h_FE times R2) you can see the role of C3 as Q1 Collector -split resistor- feedback capacitor..

What is "good" DC biasing, in this case and why wouldn't it be good?

Ahaaaaa...!! 
(Any biasing better than another "bad" one should be consider "good"..)

To be more serious, in such a high gain arrangement any Collector current variation (other than nornally expected due to circuit fix set gain) should result in respective Gain non-linearity (distortion) in a non musical way..
(waveform curve shouldn't be a "smooth" line..)

[Fancy Lime]

To be more serious, in such a high gain arrangement any Collector current variation (other than nornally expected due to circuit fix set gain) should result in respective Gain non-linearity (distortion) in a non musical way..
(waveform curve shouldn't be a "smooth" line..)

Well, the Bosstone is not exactly a HiFi phono preamp, is it? And isn't "gain non-linearity" sort of the point of a fuzz?

Back to the C3 mystery:
So, I swapped around some parts and fiddled with the knobs to see if I could get it to oscillate so I could try different countermeasures. Without C3 I quickly managed to get a high-pitched squeal that was tunable with the gain knob and some Russian talk radio. Putting a 68nF mica cap as C3 in it's traditional space made absolutely no difference whatsoever. Neither did replacing it with a 1nF cap. A 1nF cap from B_Q1 to ground killed the radio but not the squeal. The 68nF mica between B_Q1 and C_Q1 brought the silence: no squeal, no radio. I cannot hear any difference in sound other than the lack of noise even in clean settings (lowest gain of the modified gain control and no clipping diodes), which I find a bit surprising, considering the relatively large value and the large 560k+560k high side bias resistors. Guess I found my setting then.

Cheers,
Andy

[antonis]

And isn't "gain non-linearity" sort of the point of a fuzz?

If you presume so, change your query about "good biasing" to a more apropriate like "fuzzy biasing".. 

My bad about previous note: 
C3 actually by-passes part of Q1 Collector resistor, redusing Gain (and oscillation) in high frequencies..
(doesn't act as "classic" feedback cap..)

[Fancy Lime]

Hi,

so here's my current variant:


For the most part a standard Nashville Bosstone plus Mark Hammers Stupidly Wonderful Tone Control. The clipping diodes are switchable because this thing sounds nice, different and interesting with 0, 1 or 2 diodes. 1N4001's seem to sound a bit grittier than 1N4148's, which I attribute to them being a bit slower but that may be imagination. LED's are a bit of a middle ground between the 1N4001's and no diodes. The Gain control is now implemented as a variable Emitter resistor AC bypass for better noise characteristics in combination with the MPSA18, which worked better than I expected. I can no longer make out any self noise of this thing through my already very quiet headphone amp. Nice. The feedback-cut cap (now C3) is switchable because I was missing some better low gain sounds from the original design. It is  connected via Schottky diodes, which I originally added as a noise gate on my last build where I used the traditional pre-gain control and a 2N2222 as Q1. With the new Q1 this is no longer necessary for the self noise but still kills any hum coming from the guitar, which is nice in extreme gain settings, without affecting even the softest picked notes. One could make the Boost switch an on/off/on type and have it both ways, if one wanted but for me there are no downsides to having the diodes although the benefit is not as great anymore as it used to be. That's pretty much it. Oh yeah, proper power supply filtering seemed like a good idea because of the insane amount of gain available. And of course a metal can for Q2 for extra style points

Tone-wise this is by far the most versatile fuzz I know. If I was only allowed one fuzz, this would be it, no discussion. Goes from fat raunchy "transistor overdrive" through the classic 60's tones (think, Acoustic 360 fuzz, Orpheum, Mosrite, Shin-Ei Companion), most of the better Big Muff tones and high gain metal distortion all the way to insanely compressed, implosive duck-and-swell sounds that fold in on them selves before expanding and filling the room. And every combination of those.

One question though: Why does this thing produce an octave down effect in some settings (high gain setting, medium-high impact, both diodes in)? It is actually a really cool synth-organ-like effect and tracks better than some "real" octave down effects. But why? And how? I've read about this phenomenon before (I think Mark wrote something about it) but never heard an explanation.

Cheers,
Andy

[Rob Strand]

I'm afraid I don't quite follow. Could you elaborate on the "larger Collector resistor" analogy/explanation?

There's a few ways to view it, all of which are equivalent.  The easiest and most intuitive way to understand it is to think of the two 18k's and Q2 as a bootstrap connection.  This thinking also gives you an idea of how the designer come-up with the otherwise obscure connection.  If Q2 was an NPN boot-strapped common-collector stage it would have a bootstrapping cap from the output back to the tap point between the two resistors.  Using the PNP lets you remove the cap.  Amplifiers using boot-strapped collector loads to get more gain are quite common, in fact a large percentage of power amplifier use this idea in the main gain stage.

When I said "good" biasing I meant the DC at the collector of  Q1 is near mid-rail.   If you tried to build a CE amp with such a large collector resistor it would operate with a very low collector voltage and have little output swing.

A 1nF cap from BQ1 to ground killed

For oscillation problems it's sometimes a good idea to have a smallish resistor in series with the cap.  The resistor allows the high frequency gain to be reduced but has less HF phase shift, which makes it less prone to oscillation. Tweaking two parts by hand is harder but is still doable.

Guess I found my setting then.

68n is probably audible in that it will roll-off some highs.

With RF and oscillation issues caused by stray capacitance, often a small cap across the input, perhaps after the 100k pot, is best.   For this ckt a small cap across the output might help too.  Ideally one cap should dominate in the sense that it just affects the audio and the other is a "helper" for the HF zone; sometimes using both caps can cause issues.

You want to tame the HF region without affecting the audio region.

[Rob Strand]

Putting a 68nF mica cap as C3 in it's traditional space made absolutely no difference whatsoever.

FYI:
- 1n is the point where you might hear something.
- Any more than about 2.2n appears to achieve nothing as far as suppressing oscillations.
- Above 10nF starts to make a band-pass filter.   68n  produces a band pass with very large narrow peak; quite angry looking.

[rnfr]

keeping an eye on this thread as im currently venturing again into boss tone territory.  its a fun one. will post when I've got something nailed down.

[PRR]

> Collector is common to Q2's input and output

Input appears as a floating voltage across R4, Q2 B-E. There is no NFB as in an emitter follower.

Q2 appears funny because the load is connected at the emitter. But look at *loops*. Q2, R5 etc, and power supply is a loop. It works the same in any order. Q2 is common-emitter.

Two CE stages is VERY high gain. It wants to oscillate. The 50pFd is correctly placed to shift one of the poles and reduce trouble.

[Rob Strand]

Q2 is common-emitter.
Two CE stages is VERY high gain..

I "have to" (cough) disagree.

If that was the case the signal gain from the input upto the collector of Q1 (=base of Q2) would only represent part of the total gain.   In reality the collector of Q1 shows *all* of the gain.  Q2 is in fact a CC amp but it's connected as a bootstrap.

The common form is like this, (I'm sure you have seen it)
http://www.aqpl43.dsl.pipex.com/ampins/discrete/2Q-VEM/2Q-VEM%20pic1.gif

I know the mechanism/trick you are talking about.  That view works when you analyse a one-transistor Colpitts oscillator; You make the emitter the "ground" and take it from there.   I don't think you can do that here because the input and output voltages are referenced to a common ground.