Transistor Pwr Amp Pot

[CJ]

what is the action of this 5 K pot, it increases idle current thru the pwr transistors, but how?

Q710 seems to be saturated with 689 mv V-eb and does not change as the pot is turned so we were wondering what in the world is going on, Thanks! 

 

[Winston OBoogie]

How much does it alter the bias?  Doesn't look like it would by much. 
Obviously, we have the three diode package, and the one Vbe drop of Q710 as the bias spreader, which overcomes the 4 drops of the OP stage. 
The Vbe on V710 won't change of course, but current through the series pot and resistor will.  There must be an incremental voltage drop change across that 3 diode package dependent on current which puts the whole bias spreader slightly over the 4 drops of the OP stage.

 

[Bo Deadly]

With the numbers shown, the pot is set at 5.03K. So the bias is maximum hot. Err, what I meant to say was the bias is maxium cold. So yeah, that is strange since the diode drop is 0.69V and Q710 has slightly less than 7mA through it making it's resistance roughly 300R? That seems high. So maybe Q710 is bad and it's on resistance has taken a dramatic turn upward.

 

[Winston OBoogie]

... the action of this 5 K pot, it increases idle current thru the pwr transistors,

 

[Winston OBoogie]

...maybe Q710 is bad and it's on resistance has taken a dramatic turn upward.

It's not clear to me that the amplifier isn't working.  Just that C.J. wants to know how the bias spreader  "semi-rubber diode" (I'm claiming dibs on that last term) works. 
What amlifier is this C.J. ?  Looks similar to an old Sony HiFi type.  And importantly, is it working, or not?

 

[JohnRoberts]

At first glance the pot looks like a typical Vbe multiplier, BUT it is lacking a resistance between base-collector for that, so instead the output stage bias looks like just the 4 diode drops in series. I ASSume the triple diode inside the dotted line is thermally connected to the heat sink? 

The trimpot likely varies the nominal current in the triple diode which will vary the forward voltage drop and output stage class A bias. Unclear what benefit if any that trimpot configuration provides.

I have never seen anything quite like this, don't necessarily desire to again...  ;D

JR

 

[Winston OBoogie]

I thought that's what I said, but then, I suspect I'm on your ignore list so you may not have read it.  Although when this has happened before, it only makes the discussion messy.
Anyway, with a resistor between base and collector, these are, as your post which I took the trouble to  read said, indeed Vbe multipliers.  Sometimes they're called rubber diodes, but since this is not really that, I called it a "semi-rubber".
Obviously I'm a fucking idiot who knows nowt though.

 

[Winston OBoogie]

Too much one upmanship here sometimes.  Ego or bias getting in the way of sharing knowledge. 

 

[JohnRoberts]

Too much one upmanship here sometimes.  Ego or bias getting in the way of sharing knowledge.

excuuuuuse me....  :

JR

 

[Winston OBoogie]

excuuuuuse me....  :

JR

Excusing would take having an open mind on both our parts so, not really up to me alone. 

Edit: I'm just guessing, but if this were PRR, he would show C.J. why this is not a particularly good bias spreader.  I sometimes really don't get why you want to moderate here John, you seem jaded about anything that isn't built with single digit ppm integrated circuits or doesn't fit your own idea of what audio should be.

On that note. I'll find a video or article or something for C.J. on Vbe rubber ducky dinkies...

 

[Winston OBoogie]

Simple explanation of the basic deal:
https://youtu.be/Obh_PIC2qqo

Searching for improved Vbe multiplier:   
https://www.diyaudio.com/forums/solid-state/216385-optimizing-vbe-multiplier.html

There's lots more over there

 

[Bo Deadly]

Too much one upmanship here sometimes.  Ego or bias getting in the way of sharing knowledge.

Yeah, I'm not seeing that. Of course nerds are going to talk over one another a little. There's hardly anything controversial about that. Personally I would want to see multiple independent descriptions of how the circuit works. Especially if they coincide.

 

[JohnRoberts]

Excusing would take having an open mind on both our parts so, not really up to me alone. 

Edit: I'm just guessing, but if this were PRR, he would show C.J. why this is not a particularly good bias spreader.  I sometimes really don't get why you want to moderate here John, you seem jaded about anything that isn't built with single digit ppm integrated circuits or doesn't fit your own idea of what audio should be.

On that note. I'll find a video or article or something for C.J. on Vbe rubber ducky dinkies...

Sorry that I bother you so much, that you have to speculate about what I think and might say.

If it makes you feel any better I went back and re-read you original analysis and it appears accurate as far as you went.

I still do not see any obvious benefit from this biasing scheme.  There are probably easier ways to vary the diode string current and I do not see any apparent temperature tracking from tweaking base current, base voltage maybe.

@ CJ I am still a little curious about whether triple diode is heat sink mounted.

Since this does not appear to be a Vbe multiplier I am not sure how that will inform about this particular amplifier.

JR

PS: I am still unfamiliar with "rubber diode" nomenclature, but Vbe multiplier seems adequately descriptive.

 

[Winston OBoogie]

Yeah, I'm not seeing that. Of course nerds are going to talk over one another a little. There's hardly anything controversial about that. Personally I would want to see multiple independent descriptions of how the circuit works. Especially if they coincide.

I was referring to my experiences a whole, not just this thread.  Discussion involves listening to one another and that just doesn't happen sometimes.  If I see a post of yours that I like or agreed with, I acknowledge that.  Likewise if I had a different opinion, I refer to your post and try to offer something else.
But if there's no acknowledgment of what was said before, and I can think of more than a couple of posts where this happened, it's just a bunch of blokes pissing. 
Whatever, I come here to help and learn, I can get my fix of that here no matter what.
Peace.

 

[Winston OBoogie]

PS: I am still unfamiliar with "rubber diode" nomenclature, but Vbe multiplier seems adequately descriptive.

Yep, it's a used nomenclature but maybe it's a limey thing? 
I agree this scheme is sub optimal.  It looks like a budget amplifier though so maybe it was a budget designer.

 

[JohnRoberts]

Yep, it's a used nomenclature but maybe it's a limey thing? 
I agree this scheme is sub optimal.  It looks like a budget amplifier though so maybe it was a budget designer.

I have seen many budget amplifiers without any bias trim at all (I worked at Peavey). That amp's 0.22 Ohm emitter degeneration resistors could be thermally stable without a trim (with adequate heat sink cooling ).

There may be something subtle going on, or not. One common mistake they avoided is that a failed open circuit trim pot defaults to minimum output bias so that suggests either an experienced designer, or good luck.

JR

PS; I've known some Brit boffins, including one digital amp designer but he didn't use class AB in his amp designs. At least I learned a new term today (rubber diode). I think I will keep calling it a Vbe multiplier tho. 

 

[Winston OBoogie]

I try not to use the term rubber diode in public, sounds a bit Sesame Street,  so yes, I don't blame you sticking with Vbe.
I hear you on putting the variable resistor in the base emitter so that a failure falls low, I did it otherwise once, that was enough.

C.j.  the bias you show on that output stage is really low.  If we're trying to make the best compromise possible by keeping small and large signals as much the same as we can, to minimize crossover distortion when it transitions to class B, then a good place to start would be going by what Barney Oliver of Hewlett Packard looked at which was to make the external RE equal to the internal re' of the transistor.    This is satisfied by the 26mV/I that determines re'.  So 26mV across the emitter resistors giving 118mA Iq.
You may find that THD is lower when bias is higher than this, or lower than this but, as a starting point it seems to be a decent spot.

P.S.  I ignored the fraction of an ohm intrinsic re' at these currents in the above simplification.  In reality, Qv will vary slightly but not enough to care.  Also, i found a paper on my hard drive by B. Oliver, see attached:

 

[moamps]

I have never seen anything quite like this, don't necessarily desire to again... 

I agree this scheme is sub optimal.  It looks like a budget amplifier though so maybe it was a budget designer.

Marshall Leach
https://groupdiy.com/index.php?topic=42219.msg525059#msg525059
used this type of VBE multiplier.
https://leachlegacy.ece.gatech.edu/lowtim/2ndstage.html
https://leachlegacy.ece.gatech.edu/papers/lowtim/feb76feb77articles.pdf

 

[Winston OBoogie]

A wee picture showing the gain differences as a function of current into the load of an amplifier OP stage with your .22 ohm emitter resistors.  There's the optimum 118mA bias, and also half and double that.  118mA is, as is obvious, still a compromise.  But better than drastically under or over biasing.

This was the story of the ickle-pickle, wubberly- diode that wasn't.

 

[Winston OBoogie]

Marshall Leach used this type of VBE multiplier.

Thanks Mo . I certainly respect Leach and I'll read the pdf and see what he says.  I see that there is a resistor also in the collector-base diode string path but, hopefully he let's us know what benefits there are to the arrangement.
Appreciate the upload dude