Page 2 of 14

#16 Re: GM70 PP

Posted: Sun Jul 14, 2019 11:12 am
by RhythMick
What value of virtual cathodes would you use for a GM70 ? I'm not sure it's even practical, based on the tube datasheet I just looked at (20V, 3A heaters, plate current minimum 800mA).

With a pair of 20R resistors, both see 500mA through them but one has the plate current in the same direction and the other has it in opposition. So one resistor sees 910mA the other -100mA. 16W dissipation, so you'd need at least 30W (I normally multiply by 4, so >64W). Plus of course you're making the DC board provide the additional current through the virtual cathodes, so total 4A instead of 3A.

If you pick, say, 100R the current drops to 500mA but the dissipation goes up to 25W.
Screenshot_2019-07-14-11-09-46-1.png
Unless of course I have something wrong, not for the first time.

#17 Re: GM70 PP

Posted: Sun Jul 14, 2019 1:02 pm
by simon
Nuts! You'd convinced me with the space charge scenario! :)

I can't think why a different part of the DC supply might sound different though?

#18 Re: GM70 PP

Posted: Sun Jul 14, 2019 1:05 pm
by simon
800mA through a GM70? Sounds like a factor of 10 too high? Or am I missing something with this topology?

#19 Re: GM70 PP

Posted: Sun Jul 14, 2019 4:00 pm
by RhythMick
Is this the wrong datasheet? I've never used a GM70.
Screenshot_20190714_151710.jpg

#20 Re: GM70 PP

Posted: Sun Jul 14, 2019 4:15 pm
by Paul Barker
Not relevant for our use of the valve. Look at the characteristic curves.

#21 Re: GM70 PP

Posted: Sun Jul 14, 2019 4:22 pm
by RhythMick
Paul Barker wrote: Sun Jul 14, 2019 4:15 pm Not relevant for our use of the valve. Look at the characteristic curves.
Ok fair enough. What mA would you expect through the cathode?

#22 Re: GM70 PP

Posted: Sun Jul 14, 2019 4:33 pm
by Paul Barker
I’m not up to date on gm70 sweet operating position. Depends on so much.

There must exist examples of pp gm70 circuits.

Otherwise you’ll have to create push pull compound characteristic curves put in you’re loafline which links the two and choose you’re operating Point and limits. Should all be in RDH but should be findable on the interweb.

#23 Re: GM70 PP

Posted: Sun Jul 14, 2019 4:40 pm
by Paul Barker
Might find examples on Russian enthusiast forums, using a translator to help you find the talk. Or maybe a search under images will pull up a circuit.

#24 Re: GM70 PP

Posted: Sun Jul 14, 2019 5:04 pm
by RhythMick
Once Steve tells us what Ia he's aiming for I can recalculate, but let's say for the sake of argument it's 80 (based on nothing more than Steve's "factor of 10" comment).

You'd still need to choose relatively high virtual cathodes to make it practical. I multiply the expected dissipation by 4 and regards that as my minimum, which means at least 8W resistors. Personally I'd strongly recommend not using virtual cathodes.

Screenshot_20190714_165846.jpg

#25 Re: GM70 PP

Posted: Sun Jul 14, 2019 6:40 pm
by ed
I was looking at 7k transformer when I was messing with this, but that was se..I'm not sure we're not talking pp here but this is what I had:
GM70.pdf
(155.44 KiB) Downloaded 232 times

#26 Re: GM70 PP

Posted: Sun Jul 14, 2019 7:43 pm
by Nick
1000v, 80ma, -100v on the grid is not a bad operating point.

#27 Re: GM70 PP

Posted: Sun Jul 14, 2019 8:28 pm
by pre65
Re -100v on the grid.

Does that become -80 (or -120) if we take the filament voltage into consideration ?

Cathode resistor taken from + filament connection and NO balancing resistors.

Would filament bias be a STUPID suggestion ? :oops:

#28 Re: GM70 PP

Posted: Sun Jul 14, 2019 8:38 pm
by RhythMick
pre65 wrote: Sun Jul 14, 2019 8:28 pm Re -100v on the grid.

Does that become -80 (or -120) if we take the filament voltage into consideration ?

Would filament bias be a STUPID suggestion ? :oops:
Filament bias would be one hell of a heater lol

I've seen datasheets refer to setting the bias relative to the midway point. I attached the cathode to the +ve heater and adjusted bias by half the heater voltage.

#29 Re: GM70 PP

Posted: Sun Jul 14, 2019 10:11 pm
by izzy wizzy
The amp will be fixed bias like the 813. I'm not into load lines etc as I already have a transformer with configurable secondary. So a bit of adjustment with the fixed bias and wiggle the secondaries round and see what sounds best. My starting point would have been what Nick suggested, 1000V 80mA is about the same as the 813s.

After all the discussion, I think I'll start without the virtual cathode resistors. I'd figured on 22R 10W welwyn w23 for that but maybe won't bother now. Didn't bother with the 4P1L either. I like the simplicity and lack of resistors.

Cheers,
Stephen

#30 Re: GM70 PP

Posted: Sun Jul 14, 2019 10:23 pm
by RhythMick
Surely you're going to try both and tell us which sounds best...?

:o :) :lol: