Strange wiring in Gibson GA-20

[Pedro Vecino]

Thanks! The problem is not in the gain of the circuit around 6SJ7, I think. I also tried connecting in triode mode, with G2, G3 and the plate attached. Basically the volume/distortion obtained with volume at 4 was obtained at 8 approximately. A much better centering in relation to what the amplifier is but nothing else was obtained, so I returned to the original pentode circuit.
I would have liked to try a 6SN7 as a second tube, but I did not have any.
The amp, which brought a collection of different noises is now surprisingly silent. The clean sound threshold is also slightly raised by recovering the original voltages, a better bias setting and other little details.
For the next one I will keep in mind all this.

[Dave H]

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.

Right, so it doesn't have a loud enough clean sound?
Have you measured the clean output power?
A 290V supply is a lot lower than a 5E3.

I ran a similar output stage through LTSpice with a stiffish 290V supply and a 13k screen dropper resistor (as marked on the schematic). The clean output was only 4.5W! Changing the screen dropper to 1k increased the clean output to 9.5W.

[Pedro Vecino]

Right, so it doesn't have a loud enough clean sound?
Have you measured the clean output power?
A 290V supply is a lot lower than a 5E3.

Yes, that is exactly.
I have not measured the power, but by ear is It's similar or slightly smaller than a single ended with EL84 and 260V. Itīs consistent with those 4.5 watts.

I ran a similar output stage through LTSpice with a stiffish 290V supply and a 13k screen dropper resistor (as marked on the schematic). The clean output was only 4.5W! Changing the screen dropper to 1k increased the clean output to 9.5W.

Tomorrow I will do that test.
Thank you, Dave!

[Helmholtz]

Lowering the screen dropper to 1k will increase PS ripple and might even support motorboating. You may need to use higher capacitance filter caps for the screen and preamp nodes.

Surprized that you didn't like the 5V4 rectifier which increases HT by 10% and clean output by at least 20%.

[Pedro Vecino]

Lowering the screen dropper to 1k will increase PS ripple and might even support motorboating. You may need to use higher capacitance filter caps for the screen and preamp nodes.

Surprized that you didn't like the 5V4 rectifier which increases HT by 10% and clean output by at least 20%.

Yes, I thought first that I would need to accommodate the cathode resistor value in the power tubes when making that change, but I'm sure more things will be necessary. Dave talked about "stiffish supply" and if I understand the concept well, this is not the same case.

I associated the effect of 5V4 with the voltage rise in the 6V6 plates. I considered it more convenient not to change it because once the type of compression becomes familiar (itīs good), altering it supposes a high price in this type of amplifier. It would be great that elevating the other voltages could have an intermediate effect.

Thanks again!

[Pedro Vecino]

Iīve already done the tests. I started by lowering the resistor from 10K to 3K2 to test. The hum was raised appreciably. Then I tried with 5K and to eliminate the hum I added 10uF in the second node (screen grids and PI). Silent again.
Voltages have been raised 25V on the screens and 40V on the preamp in relation to the initial voltages with the deviations in the values of the resistors. It does not suppose an incredible change but it allows better control in the pulsation (a wider field) and better presented.
Thank you very much to all.