Strange wiring in Gibson GA-20

[Pedro Vecino]

Hi all. I have a Gibson GA20 that was previously treated (change of coupling and filtering capacitors) in which I have seen something strange: the feed line to the PI is taken directly from the screen grids. Does not match the schematic.
Wiring distribution on the main strip seems to be original but I have doubts because the new electrolytic pair and the own 47K resistor are inserted in the housing of the original electrolytic capacitor.
What catches my attention is that from pin 4 of the 6V6s (screen grids) there is a (old) wire connected to the common point of the PI 220K plate resistors, and at that same point is wired the electrolytic (the corresponding to the screen grids).

The change is exposed in this schematic:



Can it correspond to the original assembly or does it make no sense? If it were, in what way would it affect the behavior?
Thanks!

[Helmholtz]

Can it correspond to the original assembly or does it make no sense?

It could very well be the original assembly. Connecting the PI supply as shown in the schematic will reduce PI and preamp supply voltages by maybe 50V.

[Pedro Vecino]

I will keep it that way.

Danke, Helmholtz.

[Pedro Vecino]

Well, I have it already working. It has dimension and a nice compression but its clean sound threshold is very small.
I have put the voltages I get.
It is equipped with a 5Y3 RCA and I get 290V. With a 5V4, around 320V, but loses compression and sounds more abrupt.
Would there be any possibility of raising the threshold adjusting the voltages/gains in the preamp, perhaps?

[Helmholtz]

I will keep it that way.

Danke, Helmholtz.

No hay de que'.

If you have preamp distortion (low vol. setting) you may try to reduce the plate resistor values by paralleling e.g. 470k resistors to the existing 220k. This should improve the clean headroom of the preamp and will reduce gain.

Also bringing back the power supply filter/dropper resistors to their original values should help a little.

[Pedro Vecino]

A moment ago, before reading your comment I replaced the 220K plate resistor by 47K to see what happens. I have done it on one of the 6SJ7 inputs that already has a cathode bias circuit and that previously sounded practically identical to the original grid leak. I notice a more gradual transition to the overdrive, leaving more margin in the pulsation.
I will recover the values of the resistors in the feed line. I know that this is a very long way to get small bread crumbs.
Thanks again!

[Helmholtz]

I replaced the 220K plate resistor by 47K

THat would be a big step and won't allow for much positive plate voltage swing. I wouldn't go below 100k.

[Pedro Vecino]

I'll keep it in mind. Thanks!

[Dave H]

...its clean sound threshold is very small.
Would there be any possibility of raising the threshold adjusting the voltages/gains in the preamp, perhaps?

If it's clean at low vol pot settings (no first stage distortion) but distorts as soon as it's turned much above '1' then it could be the back to front quick start 5E3 style vol pot wiring. The vol pots could be rewired in the normal way with mixing resistors added.

[Pedro Vecino]

Volume transition between 0 and 3 is fine. It's really good on the whole scale. The problem is that above 3 the distortion appears prematurely in relation to the volume.
A good 5E3 or 5C3 has a much higher clean sound threshold and its overdrive is not so hasty.
Good for slide, some say...

[pdf64]

Try making the paraphase shared cathode fully bypassed to see if that helps. Beyond the linear range, a positive local feedback loop somehow can get created there, causing horrid parasitic distortion. I think the worst of it gets masked by power tube grid clipping.

[Pedro Vecino]

Thank you. Do you mean install an electrolytic capacitor in parallel with the cathode resistor or to use individual resistors on each cathode?

[pdf64]

Yes, fully bypass the shared cathode resistor, it should be a quick easy comparison.
It shouldn’t alter the gain and upset the balance because in normal use, within the linear range, the equal but opposite signals there cancel each other out, so it effectively bypasses itself.

[Pedro Vecino]

I did it through an aerial miniswitch but without results. It sounds exactly the same. I also tried to raise the value of the cathode resistor and effectively lowered its gain but the overdrive ratio once compensated in volume remains the same altering its character to worse.
We are not talking about an ugly overdrive. It's musical. It just appears hastily.
It has always been difficult for me to understand inefficient designs. I see two 6V6 and imagine a Deluxe Reverb

[Tony Bones]

With that 6SJ7 this amp has a monstrous amount of gain. Honestly, I don't know what they were thinking.

A tweed Champ 5E1 has two 12AX7's. The first delivers a gain of about 50. The second, with unbypassed cathode resistor AND added NFB delivers maybe 25? That's just a guess, but we're talking a ballpark maximum gain of 1250 from input to the grid of the 6V6.

In the GA20, the 6SJ7 likely has a maximum gain of over 100 (with the volume control maxed.) The 1/2 6SL7 is probably good for 40. That's a max gain of 4000! More than 3 times as much as a Champ.

I realize the GA50 is PP and the Champ is SE, but with the paraphase PI it's reasonable to compare to SE. (Ignore the bottom 6SL7 and associated 6V6.)

Not sure what to say. There are lots of ways to reduce gain. Perhaps wire the volume controls right side up and add some series resistance before them.