degeneration

[Andy Le Blanc]

I like the tone and playability of simple push pull amps with paraphase
inverters that have no global feedback. Smoothing out the amps response
without a feedback loop isn't a problem if you use the natural degeneration
in the powerside. Combination bias with a unbypassed cathode resistance
works well, or choosing a critical bypass capacitance value with self bias.

I found another approach that pops up repeatedly in old text and in old schematics
for consumer auto and home radios, take a look at "C" in the attached scan

I've been playing with it and a combination bias, around .005
either a shared cathode resistor or separate resistor on each push pull leg.

a 5d5 or similar based powerside with combo bias and the right degeneration
instead of FB is a marvel when pushed over the edge.

[Tone Lover]

Andy this is very interesting I dont think I fully understand it though.
Thanks Bill

[Colossal]

Andy,

Interesting discussion. I'm presuming that in 'C' the bypass cap shown from plate to cathode is a small value film or non-electrolytic tuned to the treble-dump frequency desired?

[Hellhammer]

Yes very interesting! So what ratios of combo bias have you been playing with? I'm guessing the bigger part is fixed bias? Are we talking something like 10% cathode bias? How does it affect output power, Za-a etc?

[Andy Le Blanc]

I've been doing some more digging...

this comes from early radio, in simple regenerative sets you had to dump
the radio energy generated by your own receiver so that it wouldn't reduce
the audio output to the headphones or your speaker. A standard design
feature in nearly home receivers was a cap to ground from the plate of the
power tube, it was also a common reason for repair, the old paper cap shorts
and the radio has no output.

the cap value ranges from .002 to .01

late 30's early 40's you start seeing larger values and a rheostat being
exploited for tone control in the power side.

time goes by and the control moves to the "pre-amp".

the design application gets dropped for audio, hi-fi wants to extend the
usable upper and lower range.

The best guitar amps to my ear always seem to be LO FI. a lot of even
order harmonic distortion, limited and unstable frequency response

musical tone color

with no global feedback, it's becomes a way to tone shape
the only other feedback is degeneration through the cathode resistor.

I remember my grand folks saying they always wondered why radio
never sounded as good as it did in the early days, I'm pretty sure that
this would very easily limit the top end of the output, but that's not
necessarily a bad thing, much rather things be warm and sweet
than hard and brite

I've been fiddling with pretty small cathode resistance and cap values
just enough to hear a difference, and I havent noticed a reduction
in apparent power yet, I'm going to increase the cathode bias around
a .005 cap, and see what happens. its nice the have a fixed bias for control.
it sets the current in the PT at a consistent place, despite tube variations.

[andrew]

Hi Andy, I would like to experiment with this idea on my PP cathode biased 4 EL84 amp. The cathodes all share a 160ohm to ground. Would I place a cap between the plate and cathode of each tube and have them all share the 160ohm resistor? Thanks for any help.

[Andy Le Blanc]

howdy..

4 tubes means there'll be two on each side of the PP circuit sharing a leg
should be able to use one cap on each side between plate and cathode right
on the tube socket, simple mod, make sure the powers supply is discharged

heres a single ended scheme, pretends its PP, you get the idea

I've seen from .002 to .02 as values start maybe .005 or just go for it with what you got

[passfan]

Somewhat along the lines of a conjunctive filter.

http://www.cs.cmu.edu/~ehn/ax84/images/conjFilter.gif

[markr14850]

I'm confused... Given the presence of Ck, how are Cb any different than a shunt to ground?

[Andy Le Blanc]

your not confused.

remove Ck, and you get the most degeneration over Rk

its never as simple as a shunt, you choose the values critically

don't assume that you have to have a bypass cap that completely bypasses

the cathode resistor, thats where it gets fun, its an opportunity for tone shaping

[tubeswell]

That's interesting Andy. I was chatting with Simcha Delft (here in Wellington) the other day whilst demo-ing one of my amps, and Simcha recommended the set-up in 'A' for a similar reason as what you were after. In the end I found another solution that suited me better for that particular amp. Merlin's 1st book has an extensive chapter on dealing with NFB in which he offers a number of solutions for beating HF oscillation.

[Andy Le Blanc]

I found it interesting that theres a development of tone control in the power side
at some point it stops being just to suppress radio frequency oscillation
and begins to tone shape and if you can do it with out relying on a feedback loop
there's a better chance at a warmer more musical response.

[tubeswell]

I found it interesting that theres a development of tone control in the power side
at some point it stops being just to suppress radio frequency oscillation
and begins to tone shape and if you can do it with out relying on a feedback loop
there's a better chance at a warmer more musical response.

That's what Simcha said. I s'pose I better try it again.

[Andy Le Blanc]

heres what I've been noodling with... with inverter, self balancing paraphase.

Its not a clean amp, even before clip, the inverter works, it has a lot of gain
depending on tube type driving it but not a stable bandwidth, its instability
makes it sound like theres several tremolos going at different frequency
bands and at different rates and depths at the same time, crazy tone color
push the power side over the edge and OMG.

I still had the presence issues that one usually addresses with a feedback loop

but I didn't want to "fix" the amp with FB, or effect the response with a vox
brite control...

I'm still fiddling with the cathode resistor and cap value, even beginning at low
values for each there was a difference in the presentation.
still looking to fall into the pocket

Its the opposite approach to a guitar amp, simple, old fashioned
pick apart the assumptions of bias, g2, phase inversion, feed back
look back to see what you can find.

[Andy Le Blanc]

found some bench time and worked the way up to 10r and .02.

the amp got louder before I noticed a reduction in the top end.
Makes me want to rethink HF oscillation, another assumption to get after.

I've see this in a lot of TV schematics and old regenerative radios
If you need to maximize a simple audio output in a consumer product
in the simplest way possible the extra cap is well worth it, cheaper than
bypassing the cathode resistor in a SE amp