Gain of output valves

[martin manning]

When scoping plates, make sure to use a100x probe with a voltage limit >500V peak.

Plates will swing to something like 1.8x B+, so to be safe, 1kV peak is better.

Mark, your numbers seem to add up, i.e. with plate voltage at ~400 just prior to clipping (power tube grids driven to 0V), 66W sounds about right.

[Helmholtz]

Plate will swing to something like 1.8x B+..

You're right.
Originally I had written 1kV minimum, then edited to >500Vp because plate swing is peak-to-peak assuming AC coupling.
But forgot that probe limit typically is DC + AC peak.
Many 100x probes actually are rated for 1.5kVpeak, though.
I try to avoid probing power tube plates anyway, as in rare cases it may trigger oscillation resulting even higher peak voltage.

[Helmholtz]

I plugged the numbers into a loadline calculator (https://www.vtadiy.com/loadline-calcula ... alculator/).
This gives me 48,5W per 6L6 pair, so 97W in total with a B+ of 400V.
(The OT might steal something like 5W).
This result is in line with the general power formula Pout = 2* (Vb - Vsat)²/Raa, where Vb = B+ and Vsat is the plate "saturation" voltage, where the loadline meets the Vgk = 0 curve.
I used Raa = 4.4k for the pair and Vsat = 72V (from the calculator).

[Helmholtz]

The SF Twin is putting out 55vpp on both sides of the P.I. stage. I will check what is on the plates of the output valves.
There is 46vpp on the secondary of the output transformer. I’m running the amp into a 4 ohm load.

Power tube grid signal Vpp needs to be at least 2 x grid bias for full power, so clean 80pp to 100Vpp.
It is normal that output voltage is less than grid signal.
Power tube voltage gain should be around 6.5.
Don't forget that the OT steps down voltage by roughly 12.
When scoping plates, make sure to use a100x probe with a voltage limit >500V peak.

Edit: This is a double post. Couldn't delete it.

[Helmholtz]

I should mention that a Vsat around 70V only holds for good quality/new 6L6GCs.
Also low screen voltage will increase Vsat, resulting in lowered output.

[martin manning]

I plugged the numbers into a loadline calculator (https://www.vtadiy.com/loadline-calcula ... alculator/).
This gives me 48,5W per 6L6 pair, so 97W in total with a B+ of 400V.
(The OT might steal something like 5W).
This result is in line with the general power formula Pout = 2* (Vb - Vsat)²/Raa, where Vb = B+ and Vsat is the plate "saturation" voltage, where the loadline meets the Vgk = 0 curve.
I used Raa = 4.4k for the pair and Vsat = 72V (from the calculator).

My own load line plotter also shows 100W from 4x 6L6 at 400V Va and Vg2. However, a simulation of a complete 4x 6L6 power amp at 400V Va with a typical CLC power supply produces only 68W clean output power into a 4 ohm resistive load. The min Va-k is 115V at Vg1 = 0, likely due to Vg2 sag. Screen resistors are 1k. Lowering them to 470 I can get up to 74W, with min Va-k at 108.

Mark, are your output tubes in good shape?

[Mark]

The SF Twin is putting out 55vpp on both sides of the P.I. stage. I will check what is on the plates of the output valves.
There is 46vpp on the secondary of the output transformer. I’m running the amp into a 4 ohm load.

Power tube grid signal Vpp needs to be at least 2 x grid bias for full power, so roughly clean 100Vpp minimum.
It is normal that output voltage is less than grid signal.
Power tube voltage gain should be around 6.5.
Don't forget that the OT steps down voltage by roughly 12.
When scoping plates, make sure to use a100x probe with a voltage limit >500V peak.

Thanks for your reply, I appreciate you taking time to assist me. Thanks for the tip about the 100:1 CRO lead, I only have the conventional 10:1 switchable lead.

I had a look at the output of the P.I. stage on the 220K grid leak/bias resistors. I’m getting 80Vpp before I see the signal distorting (getting nibs at the top of the waveform and the shape of the waveform distorting, not clipping due to forward conducting.)

The bias voltage is -46VDC and -49VDC (a balance pot is in the amp.)

[Mark]

I plugged the numbers into a loadline calculator (https://www.vtadiy.com/loadline-calcula ... alculator/).
This gives me 48,5W per 6L6 pair, so 97W in total with a B+ of 400V.
(The OT might steal something like 5W).
This result is in line with the general power formula Pout = 2* (Vb - Vsat)²/Raa, where Vb = B+ and Vsat is the plate "saturation" voltage, where the loadline meets the Vgk = 0 curve.
I used Raa = 4.4k for the pair and Vsat = 72V (from the calculator).

I wish I had a better understanding of this. Are there any tutorials that you may get recommend?

[Mark]

This is what I’m seeing on the CRO.

They are labeled. If there is any other waveforms or voltages you wish to see, please let me know.

I don’t have a 100:1 CRO probe yet to show the plates of the output valves.

IMG_0416.jpeg

IMG_0417.jpeg

IMG_0419.jpeg

IMG_0420.jpeg

IMG_0421.jpeg

[martin manning]

Do you know the condition of your output tubes? Are they new, and do they meet specification?
Bottom line, I think you will need 450V B+ to get 100W of clean output power.

[martin manning]

I plugged the numbers into a loadline calculator (https://www.vtadiy.com/loadline-calcula ... alculator/).
This gives me 48,5W per 6L6 pair, so 97W in total with a B+ of 400V.
(The OT might steal something like 5W).
This result is in line with the general power formula Pout = 2* (Vb - Vsat)²/Raa, where Vb = B+ and Vsat is the plate "saturation" voltage, where the loadline meets the Vgk = 0 curve.
I used Raa = 4.4k for the pair and Vsat = 72V (from the calculator).

I wish I had a better understanding of this. Are there any tutorials that you may get recommend?

I can recommend these two Valve Wizard articles:
http://www.valvewizard.co.uk/se.html
http://www.valvewizard.co.uk/pp.html

[Helmholtz]

I wish I had a better understanding of this. Are there any tutorials that you may get recommend?

I learned a lot from books like the RDH4 or Terman's book.

[Helmholtz]

This is what I’m seeing on the CRO.

Nothing really unusual in your pics.
From the display it seems you're using a 50Hz signal? Would also explain the slanted tops.
If so, available output will be lower than at some mid frequency (e.g 400Hz) because of increased magnetizing current and OT losses.

Output voltage should be measured using a good RMS voltmeter connected directly to the OT secondary wires (not across the load resistor).
This avoids possible errors due to contact resistance.

[martin manning]

From the display it seems you're using a 50Hz signal? Would also explain the slanted tops.
If so, available output will be lower than at some mid frequency (e.g 400Hz) because of increased magnetizing current and OT losses.

Good point. Going from 400 Hz to 50 Hz took my sim from 100W down to 75W at 450V B+.

[Helmholtz]

Going from 400 Hz to 50 Hz took my sim from 100W down to 75W at 450V B+.

What's your OT model?
Does it take into account non-linear effects like current-dependent inductance and core losses?
Is your B+ of 450V at idle or at full output?