RH84 SE, pentode driver... advice please

[e_fortier]

Thank you both for answering, I must admit that this is way over my head...I have modified the schematic and came up with this for the driver;

Tube audio is new to me and I'm hooked big time

B+ 300V
Rp = 18K
Ia = 7mA
Va = 170V
Rg2 = 150V
Ig2 = 2mA
Rk = 250 ohm
Vk = 2.25V
Ik = 9mA (Ia + Ig2)

I have lower Rp to 18K and therefore left my Rfb similar to before.

Thanks a lot for your help
Eric

[e_fortier]

Hi,

Well I did some test tonight and at least I have some sound with the 6AU6 installed.

My left channel is with the 6AU6 and the other one is with the 12AT7, I can easily compare ;-)

As is like on the last schematics, the gain is much lower with the 6AU6. I then bypassed Rk with a 220uF and now the gain seems way too high, when I turn the volume it does not take much to start distorting. My Rfb is 150K and Rk is 300ohm.

Is there a way to approximate the resistors (Rp, Rk) values to have a similar gain as the original RH807 ?

Thanks for your help.
Eric

[Michael Koster]

Originally Posted by e_fortier

Hi,

Well I did some test tonight and at least I have some sound with the 6AU6 installed.

My left channel is with the 6AU6 and the other one is with the 12AT7, I can easily compare ;-)

As is like on the last schematics, the gain is much lower with the 6AU6. I then bypassed Rk with a 220uF and now the gain seems way too high, when I turn the volume it does not take much to start distorting. My Rfb is 150K and Rk is 300ohm.

Is there a way to approximate the resistors (Rp, Rk) values to have a similar gain as the original RH807 ?

Thanks for your help.
Eric

Here's an example based on your circuit values.

First look at the 807 loadline to get an estimate of the in-circuit gain. Here we have about 4mA/V into 5K load for a voltage gain of approximately 20. Assuming 150K Rfb, the effective load due to Rfb will be 150K/(20+1) about 7143 ohms.

This is in parallel with the 18K plate resistor for an effective plate load of about 5114 ohms. Plotting this slope through your op point of 170V/7mA gives the 6AU6 loadline. I've highlighted the full signal plate swing in green.

The feedback ratio here is the ratio of 6AU6 signal current through Rfb to the total signal current, or about (1/7143)/(1/5114) which works out to .72 or 72% local feedback.

The sensitivity will be about 2.75V P-P plus 7mA * 300 ohms = 2.1V or 4.85V P-P input for full output. This is a little less than 2V RMS if I did everything right Increasing Rfb will increase the sensitivity correspondingly.

[e_fortier]

Hi Michael,

I'm not sure I understand about the feedback, are my chosen value good or is the feedback ratio at 72% too high ?

I will try tonight the following 2 versions and keep you guys posted, so far I prefer the sound of the original RH807 with the 12AT7 but this could be simply because I did not properly set-up the 6AU6, any help is welcome.

I will change Rk for 180 ohm (Vk 1.7V) and change Rfb for 270Kohm

For fun, I have attached a picture of my RH807.

TIA,
Eric

[Michael Koster]

Originally Posted by e_fortier

Hi Michael,

I'm not sure I understand about the feedback, are my chosen value good or is the feedback ratio at 72% too high ?

I will try tonight the following 2 versions and keep you guys posted, so far I prefer the sound of the original RH807 with the 12AT7 but this could be simply because I did not properly set-up the 6AU6, any help is welcome.

I will change Rk for 180 ohm (Vk 1.7V) and change Rfb for 270Kohm

For fun, I have attached a picture of my RH807.

TIA,
Eric

It just means that 72% of the plate signal gets back to the grid as feedback. It puts the effective plate resistance at about 350 ohms. You should have good speaker damping. Changing Rfb to 270K changes the feedback to 58%, resulting in about 430 ohms effective plate resistance. The load line of the driver goes to about 7500 ohms.

Changing Rk and Rfb should increase the sensitivity quite a bit. The Rk change should get you a little more plate current on the 6AU6, maybe getting into a more linear regime. The input sensitivity should be about 1V P-P grid swing plus about (5.5mA * 180) 1V degeneration for 2V total P-P input sensitivity at full output.

Maybe measure Vk, Vg2, Va on both tubes, just to verify all your op points.

[e_fortier]

Hi,

Here is my schematic (exactly as built...for now...) with some measured voltage, hope this helps to figure out why as soon as I turn up the volume the amp start to distort quite rapidly.

Please help

I have a list of planned action but if possible I'd like tube wizard's opinion first.

Regards,
Eric

[Svein_B]

There is nothing terribly wrong with the voltages.

I do however get somewhat different values with LtSpice simulation. That could be due to the tube simulation model, but it could also be some wiring error or other difference between schematic and the real thing.

Could you please double check that the Rfb is connected as shown, and not to the other side of coupling cap.



Edit: Your voltages and currents do seem to add up. The difference in readings must therefore be my Spice model, or your tube being a bit outside the typical rating.


SB

[e_fortier]

Hi Svein,

For simulation purpuse I have RCA 6AU6WA and the 807 is also RCA.

The only difference I could think is the 166V on 6AU6 Vg2 comes from the 3rd zener of a 5 zener string to get to 275V on the 807 screen, I taped myslf at the third one which gives me 166Vdc. The rest of the schematic is identical.

Thanks all for your help, this is much appreciated.

Regards,
Eric

[Svein_B]

With the 6au6 screen supplied from a 166V source, the other voltages match close to my simulation.

The simulation gives about 9W power before the distortion approaches 5%.


Performance will be slightly better with Rk=180 and B+ about 300V, but this is a minor tweak, once you are satisfied with the basic design.
The bypassed Rk gives a bit more distortion from the driver stage than unbypassed, but lower overall distortion due to some distortion cancellation.

SB

[Michael Koster]

Originally Posted by e_fortier

The only difference I could think is the 166V on 6AU6 Vg2 comes from the 3rd zener of a 5 zener string to get to 275V on the 807 screen, I taped myslf at the third one which gives me 166Vdc.

I wonder if the total current on the zener string is more than ~8 mA on peaks, and the g2 voltages are dropping out. Do you have bypass caps on the g2 connections?