Watt Sucking Fireball Series

[Zen Mod]

Originally Posted by Roadbagger

Wadaya think, are we the only people in here?

nope ;

I'm having (all the time ) real fun reading you wakoos ....... but , unfortunately , I'm too occupied with other things , so I'm not able to participate with my peanut brain

[eyoung]

Ditto that...

However I think I have a processing unit more like a BB(.177in). I am very interested in the line of conversation so I will just continue to lurk.

Keep up the good work, Elwood

[kenpeter]

I've never seen or drawn nothing like this before. Dunno how I come to it?
I think the tin foil helmet is becoming less and less effective over time...
Anyways, what exactly is this strange topology? And why does it seem
to be working? What more would it take to get this thing fireball worthy?

Needs a JFET up front, or something to provide the initial base current.
Its another unity gain wonder. Wonder how to get some voltage gain
out of it???

[kenpeter]

Q1 never seems to shut completely off, but 100ohms across the
emitter of Q3 might draw a few extra milliamps just to be sure...
Its not in the drawing above, but probably should add it.

The sum of voltage across R1+R2, or maybe its the average...
Is held sorta constant by the series regulated thingie goin on
at the emitter of Q4 (does that makes it an SRPP?)

But then you got Q1 emitter down, Q4 emitter up, so sorta
halfways a diamond buffer too. Or one of them error correcting
something or others. I don't really know if thats how it works?

If the base currents required to drive Q3 and Q4 were equal
and opposite, the current in Q4 would be constant. But both
bases require a nonlinear swing of current, so Q4 is always
adjusting the sum of drive currents to keep the outputs in
proper balance.

[kenpeter]

Agh... Turns out its just another White Cathode Follower variant!
Only thing worse than a circuit that don't work, is one that DOES
but you can't explain why... Makin me even more nutz than usual.
Hopefully I got it figured now, we'll see what RBagger has to say?

Imagine the emitter of Q1 tied directly to the load, So's you can
see its actually behaving as a Sziklai Pair in conjuntion with Q3...

All of this sits atop an identical Sziklai Pair, Q2 Q4. Driven in
paraphase by the summation of errors detected.

For those reading, but not familiar: See Broskie's TubeCAD for
tech and background history on "White Cathode Follower".

The way I got it miswired, half Diamond Buffer style: Luckily
doesn't seem to ruin the WCF function. Q1 - Q4 does cancel
out an emitter voltage offset! I think the diamond buffer is
synergistic, but largely swamped out by the WCF effect.

Whatever... You got a new way to abuse four transistors.

[john curl]

Kenpeter, it seems that Kirkwood is off with his family for an extended holiday. They come in from other countries. I hope he will be back soon.

[Roadbagger]

Back soon.......................

[scott wurcer]

Originally Posted by Roadbagger

Back soon.......................

Kirkwood your email bounces.

[scott wurcer]

Bump, Kirkwood where are you?

[sumaudioguy]

I cannot believe just discovering the writings of RoadBagger- wow. Someone else that actually knows something and even better can say it! It is really good to read someone else writing things I have said for years. Will be watching for more, YEA!

The part about load beta reflection of follower BJT is so true especially at RF designs and so many just do not understand. The reflection from a bad load is sent back into the amplifier to the last gain stage where phase margin collapses to minimum value and the amplifier breaks into oscillation. Most amps I have ever seen oscillate pretty easy because of this effect which results in a blown amp or at the least horrible sound. If phase margin can be maintained then different things happen. Load control can be one solution or amps with very low impedance final gain stages. The amps here when shorted simply blow the line fuse and no SOA protection is required. These amps will not oscillate under any condition that does not blow the line fuse immediately from severe overload condition as an example. When these other amps break into oscillation the bias and feedback both usually quit working and it is a very short path to smoke. Many amps tested here have operational points that are not to good and go into and out of variable gain or bias conditions (also mentioned by RoadBagger) and go into and out of oscillation during every cycle or more likely at a fast large transient of the signal under loudspeaker loads. Naim amp all do this as example and tick away in oscillation which become physiologically addictive to the point where only a Naim amp will do to Naim aficionados .

As RoadBagger indicates the probe at every location of the circuit should be the linear signal and never anything but. Unfortunately "but" is what I have found in almost every solid state amp ever tested. I used to run a repair shop so have seen a lot of commercial power amps and poked at them on the bench even more. If every point is not the linear signal, stop and fix it before moving on! Most will not do this. An amplifier is best thought of as "a filter with gain" and should behave as exactly that in every sense of the definition. Most amps do not.

Would really like to have talked long to the Dr. Slottow but living out here in the middle of dumbville (we do not even have evolution here) just does not lend itself to such activities and my asocial problems pushes me deep into the wilderness. All this fine tuning and great perspective from RoadBagger is wonderful- and what the internet if for. It is delightful to see such clear and concise writing and thank you for that. I hope to learn more and already have seen some very novel solutions to lowered distortion and faster circuits, hurray! With two different amps on the drawing board I am certainly not above learning and using great ideas unless someone says the idea is proprietary. Then of course not.

RoadBagger gets my "Most Fun" award! If more is better then most is best!