Feedback question, Philco 6017 PA amp

[Diverted]

Hi all,

I got this cool little Philco PA amp, recapped and it plays very well. Clean, clean sound even at volume. Since I have no use for the phono input, I removed the phono circuit and am using the phono level pot for an across the line master volume control, to help get some decent breakup at lower volumes.

I'd also like to experiment with the negative feedback loop in this set. Apart from a switch to disconnect the negative feedback which I will try this afternoon, anyone have any suggestions on how to alter this loop to provide a little more versatility? Thanks!

[PRR]

Context. Whole schematic, please.

[PRR]

As is, the amp has very little NFB, and probably is not designed for a lot of NFB. And it is taken from a tap you will never use. Replace the 0.12Meg NFB with a fixed and a pot, and go to a speaker tap.

[DummyLoad]

interesting implementation of a paraphase, have never seen that done with a pentode/triode mix... bet it's got some quirky overtones.


--pete

[trobbins]

As I recall, Philco were in to self-split varieties, and I reckon this is another.  The 'bottom' output half is driven from an amplified tap off of the drive signal to the top half.  So one setup activity should be to tweak the gain of the bottom half (perhaps via the 12k tap resistance) to get symmetrical drive for the class A operation.


And you've got grid biasing of the mic input stage, and of the self-split amp stage (which may be interesting as it will have an attack-decay character to one half of the output signal).


You may have to tweak the 3.3M screen drop resistor of the 6SJ7 to get equivalent operating values to the schematic.


If you've removed the phono circuitry, then you can reduce the 330k mixer resistance from the mic channel vol pot path to just a typical stopper value.


If you remove the feedback circuit you could change that feedback stage from grid-leak bias to more normal cathode bias with bypass cap, which can then get rid of the extra 20nF coupling cap (used for the grid leak bias).

[trobbins]

Ok its not self-split, but rather just the early form of simple paraphase, and it doesn't use the more commonly seen form of self-balancing common-cathode pair of triodes to provide some negative feedback.

[shooter]

bet it's got

that is an elegantly bizarre circuit!
I have seen "driver" tubes post PI, but there's 2!
a way to sell tubes?
Designed to purposely distort a mic?

[PRR]

The only strange bit is that stage 3 is in the same bottle as stage 1.

Triode for low-hiss input gain. Pentode for main gain, and to drive one power bottle. A triode lamed (0.56Meg to 0.012Meg, 56K:12K) to unity gain inverting to drive the other power bottle. This was all common stuff late-1930s and all through the 40s and into the 50s.

The only oddness is that you do not usually bring-back your high level power tube drive to the same area as your sensitive input. It begs to squeal. However with the topology they chose, and the tubes they had, it musta been worth a try. Aside from supersonics sneaking back inside V1, there is no reason it shouldn't work. And apparently it worked.

[PRR]

The grid-leak bias is because the 6SC7 (poor 12AT7 ancestor) has only one cathode for the two sections. Used in sequential stages, any off-ground cathode bias is very likely to feed-back through the shared cathode. While coming from the next-past stage puts the feedback negative, it won't be benign, and that pentode between is a LOT of gain. I'm not sure even an LED would be low enough shared impedance to avoid trouble.

The grid-leak is quite playable even with hot pickups if you can just turn your guitar down to the edge of grid overload. That means not using 9,999-turn pups and not leaving the knob at 10.

[DummyLoad]

The only oddness is that you do not usually bring-back your high level power tube drive to the same area as your sensitive input.

and that is what i was eluding to. especially one that shares cathode with an independant gain stage. i understand why the grid leak bias was used, but why the 6SC7? - 6SC7 was introduced in 1938 - or maybe there weren't other duo triodes with individual cathodes in that era that would have worked.

meh! old shit. bet it's quiky on that merit alone, but then again, the pentode biased on the bleeding edge of gain sure helps.

--pete

[shooter]

the pentode biased on the bleeding edge of gain

I'm mentally probing my imaginary sine wave, at the grids to the PA tubes, am I "seeing" ~ 10X difference between + and - 1/2cycles?

[PRR]

> ~ 10X difference between + and - 1/2cycles?

You mean between push and pull sides?

No, it is exactly balanced (for practical purpose). The 0.56Meg and 0.012Meg divide by about 47.5. The 6SC7 datasheet says a similar condition has gain of 46. 46/47.5 is real darn close to 1. Given resistor and tube tolerances, the math is spot-on, reality won't be too far off.

(Yes, I shudda shot the right end of the rule; slip a stick often enough, you can read either end.)

[shooter]

Ok, I think a lightbulb just came on, but I'm not sure were

So the signal at "A" gets feed back through "B"'s line, amplified, and comes back to "C" ~= to point  A?

Thanks for staying late for us problem kids

[sluckey]

So the signal at "A" gets feed back through "B"'s line, amplified, and comes back to "C" ~= to point  A?

That's it. Classic paraphase. Also classic cheap Philco consumer electronics. 

[shooter]

Again, thanks for staying up late
I like my PI trannies better, 5 wires, done!