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

[John Lyons]

Jordan Bosstone
Ok, So this one is a little out of the ordinary.
It uses NPN and PNP transistors.



R1 is a voltage divider which turns down the signal into the circuit.
C1 forms a high pass filter with R5
R3 and R4 form a feedback resistance and filter along with C2
which also sets the bias for the base of Q1 withe the help
of R5 (R3+R4 and R5, voltage divider).
C3 filters out some high end and possible oscillation.
Q1 (NPN) is feed directly into Q2 with no coupling cap. ("direct coupled"?
Q2 is PNP is set up with the emmiter to 9v through R2.
So does this make Q2 a buffer of sorts? This is where I bog down...
Anyone care to tackle this?

Also. When the Attack pot is lowered oscillation is somewhat common
in the builds I've done. Why is this?

John

[BRingoC]

Not much to add, but my bosstone has oscillation when the gain is over 90% up, there isn't much more gain in the last 10% so I am not missing much.  I had to add a 10uf cap from positive to ground in order to cut down on oscillation and to stop the most annoying thing, it always picked up radio stations, even when boxed up.  This probably wasn't what you were looking for, but there it is.

[aron]

Yes, I remember thinking why did they use a PNP on the end? It's an emitter follower, so an NPN could have been used.

[John Lyons]

I think there is more going on in here.
Even without the diodes this thing is pretty fuzzy.
More than just one low hfe transistor and an an emitter follower
can do. Hmmm....

[doitle]

Just for my own clarification and not really adding something insightful or important here but, with those diodes on the end the output is limited to + - Diode drop right? So no matter how gigantic a signal should be coming out it will be between like .7 and -.7 or .3(.24) if they are germanium?

[brett]

Hi

It's an emitter follower, so an NPN could have been used.

The emitter of Q2 takes 50% (18k/36k) of the output voltage of Q1's collector.  That makes Q2 an active load on Q1 because the laws of physics say that the base of a PNP transistor is about 0.65 V lower than its collector.  This has some really cool effects in this circuit.  The gain is high and for small signals (0.02 V p-p), the slew rate at the output of Q2 is very high.  The results would sound very harsh except for 2 things.  Firstly, Q1 is *barely* conducting, and the base voltage is only 0.6V (or slightly less - 0.58 V ?).  This means that "upward" swings (voltage increases) on the base make Q1 conduct better.  Downward swing causes cut off.  This gives the output (pre clipping diodes) an asymetric appearance (like a FuzzFace).  A sine wave results in a mark-space ratio of about 2:1 for small signals.  A bit like a tube-amp. I was surprised that the clipping diodes do not simply "cut" the signal at + and - 0.65 V.  They leave some overshoot, especially on the low side (maybe because the output impedance is different on the + and - sides?).

The circuit seems quite insensitive to the values of many components.  22k replaces 18k, 1 Meg and a 22uF bypass cap can replace the old-fashioned 2x560k split with a small cap.

The results are quite sensitive to the amplitude of the signal.  All in all, an interesting and attack-responsive circuit.

cheers

PS Here's a LTspice approximation of the Bosstone:
R1 N002 N006 1Meg
R2 N006 0 220k
R3 N002 N004 22k
R4 N003 0 100k
Q1 N004 N006 0 0 2N2222
C1 N006 N005 0.022µF
C2 N004 N002 47pF
C3 N002 N003 0.022µF
V1 N005 0 SINE(0 0.1 880) AC 0.1
V2 N001 0 9 Rser=10
D1 N003 0 1N4148
D2 0 N003 1N4148
R6 N002 N001 22k
Q3 0 N004 N002 0 2N3906
C4 N001 0 22µ V=20 Irms=0 Rser=0.225 Lser=0 mfg="KEMET" pn="T495D226M020AS" type="Tantalum"
.model D D
.lib C:\PROGRA~1\LTspice\lib\cmp\standard.dio
.model NPN NPN
.model PNP PNP
.lib C:\PROGRA~1\LTspice\lib\cmp\standard.bjt
* Jordan Bosstone
;op
.tran 0 10ms 0
.backanno
.end

[fuzzo]

Mine hasn't oscillation but with the pot at max, the sound become really odd . I must keep the pot at 80 per cent , no more, to hear the good fuzz sound.

[jessetrbo]

Brett

"The circuit seems quite insensitive to the values of many components.  22k replaces 18k..." 

I'm assuming that you're talking about R2 from the schematic John posted...  I did find that lowering slightly (16k or 15K)  helped tame some of the oscillation problems people report (while not drastically altering the tone of the pedal).

Jesse

[John Lyons]

Thanks for the insight brett.

I haven't had any oscillation with the attack full up but it's a somewhat common problem with
oscillation when the attack is turned all the way down. Voodoo labs fixed this with a 1k resistor between
the bottom of the attack pot and ground, setting  a minimum atttack amount before osc.
Why does this happen? Something to do with shunting R5 to ground affecting bias at the base?
Or maybe something to do with the R3/R4/C2 filter?

[zombiwoof]

I have an original one and it doesn't have that problem, maybe there is something wrong in the schematics going around.

Al

[brett]

Hi

I'm assuming that you're talking about R2 from the schematic John posted...  I did find that lowering slightly (16k or 15K)  helped tame some of the oscillation problems people report (while not drastically altering the tone of the pedal).

Yes, R2 and R6 can be changed to 15k or 22k if you don't have 18k resistors.  (and who has 50pF caps?  47pF is the obvious replacement, but it might be worth experimenting with higher values, too).

I'm no expert, but I wonder if the oscillation is because Q1 is barely conducting and because new transistor have high hFE than older ones (making the Beta of Q1 more like 300 or 400 rather than a vintage vaue of 150 or 200).  My thinking is this:  Even without an emitter resistor, the effect of being barely turned on creates an input resistance that is extremely high (Rin = Beta*0.025/Ie, where I is the emitter current).  If beta is 400 and Ie=1uA, then Rin = 10 megohms (!).  That is perfect for picking up some of the output signal.  Therefore, if I had oscillation issues, I'd try lower hFE transistors (TIPs or BD139s) and be careful to separate the input and output traces on the layout (and perferably not use vero).
cheers

[Mark Hammer]

My Bosstone clone performs nicely but starts to produce weird oscillations as the battery wanes.  Not sure why, but it does so reliably.

[jessetrbo]

Brett,

Don't sell yourself short!  You have made many posts concerning the Bosstone.  Not to hijack the thread and make it about oscillation, but... my current breadboarded version uses low hfe transistors and still is prone to oscillation when the attack pot is turned to a particular spot and I hit a particular note. It's pretty specific.   

Mark's post will make me run and measure the battery's voltage-- I have a suspicion that it may be low, and that by lowering the resistance of R2, it helped remedy the situation just enough... 

There's a lot of resonance in the circuit... almost has a twin-t wah filter sound.  There seem to be similarities in the Q1 stages of both circuits (I'm still learning, though).  My original build when going through lower settings of the attack pot sounded like a %^&*ed wah a bit before becoming a fuzz monster.  In that build I used an NTE123A in Q1, and a 2N4125 in Q2.  All of the resistor values were identical to the schematic, as were the cap values (except for a  47pF).  Sounds awesome-- I regret getting rid of it, but will hopefully have another one up and running soon.

Jesse

[Mark Hammer]

Great little fuzz.  I have a 3-position switch in mine that selects between an increase in input cap value (for that quasi-octave-down sound), stock, and a cap in parallel with the clipping diodes for a "rounder" sound.

[edvard]

I built one of these for a friend a few years ago and he never had a problem with it.
Well, except for minor 'tweaks' he was constantly pestering me for... 

There is another version posted at the free information society:
http://www.freeinfosociety.com/electronics/schemview.php?id=933

From ToneFrenzy's site (http://www.tonefrenzy.com/effects/jordan_boss_tone.html) I gather there were two versions of this pedal.
One would plug directly into your guitar like one of Dan Armstrong's 'jack warts' and a pedal version a few years later.

I wonder which schematic is which? 

[brett]

Hi
is my memory playing up, or do I recall the two circuits being called the "Nashville" and the "Dallas" versions (or some similar geographic references?).

I haven't had a problem with oscillations, but I made 2 chages.  Firstly, I used a 1 M resistor to bias Q1.  This would turn is on a bit more and make it a bit less buzzy.  This would reduce the input resistance which might be the source of the feedback (see my post above).  I also put a 10 ohm resistor on the emitter of Q1.  This may have knocked the gain down a bit (and lifted the base voltage just a tiny bit further).

My guess is that the 2 x 560k resistors could be replaced with 2 x 470k without losing much (any?) Bosstone magic, while improving oscillation issues and probably making the circuit less sensitive to the choice of BJT for Q1.

cheers

[John Lyons]

There are at least two versions.
Nashville, CA and one other than I can't think of now...



Here's a "factory" schematic

[earthtonesaudio]

Thanks for the insight brett.

I haven't had any oscillation with the attack full up but it's a somewhat common problem with
oscillation when the attack is turned all the way down. Voodoo labs fixed this with a 1k resistor between
the bottom of the attack pot and ground, setting  a minimum atttack amount before osc.
Why does this happen? Something to do with shunting R5 to ground affecting bias at the base?
Or maybe something to do with the R3/R4/C2 filter?

I think it might have something to do with phase shift.

[John Lyons]

I think it might have something to do with phase shift.
Any further theory on this?

[earthtonesaudio]

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!

Now add back in the other parts.  The PNP follower causes the 18k resistor to have the same signal appearing at both ends, thus less current can flow through it and it "appears" larger than it really is.  This would be good for making an oscillator, as it would make this resistor closer in value to the two 560k resistors that make up the other parts.  However, there's a wrinkle here because the PNP emitter is also where the first cap was located, and on the other end of that cap are a couple diodes and a volume pot.  For small signals coming through the cap, ground is 100k away, not exactly a short circuit.  But for larger signals, the diodes conduct and there ya go.  Then there's still that pesky PNP emitter.  To a small signal it looks like a low impedance input which would tend to stop any signal from proceeding any farther.  This is where my "theory" runs out of steam.  To make a "good" oscillator out of this circuit, I'd probably want to put some series resistance between the emitter of the PNP and the .022 output cap.

This may account for why some people have oscillations and others do not.  Different layouts and different part tolerances might make up the difference.