Hi,
I'm currently working on a PP 807/6AU6 pentode amp with mosfet regulated screen grid supplies. I have some 0D3 VR tubes on the shelves, two in series should provide the desired 150 and 300V reference voltages for the mosfets. I know VR tubes turns into oscillators if they are shunted by caps larger than 0,1uF or so, but would it be correct to assume that this behaviour could be stopped by a big resistor between the tube and the cap as in my LazyCad drawing below? (Please ignore the lack of gate stoppers, protection diodes etc):
I'm currently working on a PP 807/6AU6 pentode amp with mosfet regulated screen grid supplies. I have some 0D3 VR tubes on the shelves, two in series should provide the desired 150 and 300V reference voltages for the mosfets. I know VR tubes turns into oscillators if they are shunted by caps larger than 0,1uF or so, but would it be correct to assume that this behaviour could be stopped by a big resistor between the tube and the cap as in my LazyCad drawing below? (Please ignore the lack of gate stoppers, protection diodes etc):
Thanks.
I've searched both here and on Google and there's plenty of information about what happens when VR tubes are shunted by too large caps but nothing about circuits like this. It would be nice to use caps large enough to allow a slow ramp-up.
I've searched both here and on Google and there's plenty of information about what happens when VR tubes are shunted by too large caps but nothing about circuits like this. It would be nice to use caps large enough to allow a slow ramp-up.
I used that kind of circuit in a prototype of my valve DAC and it indeed worked as it should.
Glow discharge tubes have an ignition delay that varies from ignition to ignition and that can vary a lot between different tubes from different manufacturers. It's usually in the dozens of microseconds to dozens of milliseconds range, but I have one 85A2 that requires seconds. It may be a good idea to test the ignition delay of your tubes in darkness, just to check it isn't excessively long. Too long ignition delays can cause nasty voltage overshoots.
Glow discharge tubes have an ignition delay that varies from ignition to ignition and that can vary a lot between different tubes from different manufacturers. It's usually in the dozens of microseconds to dozens of milliseconds range, but I have one 85A2 that requires seconds. It may be a good idea to test the ignition delay of your tubes in darkness, just to check it isn't excessively long. Too long ignition delays can cause nasty voltage overshoots.
Last edited:
Thanks!
100k was an arbitrary value, I guess I can go much higher if needed without affecting the Mosfets bias.
I've read about the ignition delay, practical tests will reveal if it is something that will cause any grief. I have 5-6 0D2s to try, I used one in a 6AV6GA PP prototype some years ago without any notable delay.
I also read that some VR tubes contain radioactive elements to ensure instant ignition in darkness...😱
100k was an arbitrary value, I guess I can go much higher if needed without affecting the Mosfets bias.
I've read about the ignition delay, practical tests will reveal if it is something that will cause any grief. I have 5-6 0D2s to try, I used one in a 6AV6GA PP prototype some years ago without any notable delay.
I also read that some VR tubes contain radioactive elements to ensure instant ignition in darkness...😱
I used an 85A2, a 47 kohm, 1/2 W resistor and a 10 uF aluminium electrolytic capacitor, but there indeed should not be any reason why higher resistance values can't work (especially with a capacitor with low leakage, such as a film capacitor).
Radioactive primers such as tritium or paint containing uranium dioxide are/were indeed used to speed up ignition. There is still a delay, but far less than without primer. I suspect that my 85A2 that has seconds of delay is the only one that is not radioactive. The radiation from tritium can't pass through the glass of the tube, but tritium has a half-life of 12.32 years, so the average ignition delay can almost double every 12.32 years.
Radioactive primers such as tritium or paint containing uranium dioxide are/were indeed used to speed up ignition. There is still a delay, but far less than without primer. I suspect that my 85A2 that has seconds of delay is the only one that is not radioactive. The radiation from tritium can't pass through the glass of the tube, but tritium has a half-life of 12.32 years, so the average ignition delay can almost double every 12.32 years.
True, leakage might become an issue if I go for too high resistor values. I have a decent variety of different caps and resistors so I'll probably find something that works.
Good to hear that I won't get a "dose" just because I think VR tubes look better than zener diodes.
Nelson Pass once defined the perfect audiophile component as "heavy, expensive, obsolete and toxic" (or something like that), adding "radioactive" to the list would be a bit of an overkill 😀
Good to hear that I won't get a "dose" just because I think VR tubes look better than zener diodes.
Nelson Pass once defined the perfect audiophile component as "heavy, expensive, obsolete and toxic" (or something like that), adding "radioactive" to the list would be a bit of an overkill 😀
Probably.
I could always up the ante by adding mercury vapor rectifiers and perhaps a selenium diode in the bias supply, along with the 400 volts on the plate caps it would make an respectable piece of gear🙂
I could always up the ante by adding mercury vapor rectifiers and perhaps a selenium diode in the bias supply, along with the 400 volts on the plate caps it would make an respectable piece of gear🙂
In relation to his HV Tweeter that produced toxic ozone gas which unfortunately was enough to cost him a few days of hospitalization.
Just be very sure that the bottom OD3 ignites before the top OD3 ignites.
1 Meg Ohm might not be low enough resistance to ensure that.
Otherwise, expect the bottom MOSFET to say "Ouch".
1 Meg Ohm might not be low enough resistance to ensure that.
Otherwise, expect the bottom MOSFET to say "Ouch".
"In relation to his HV Tweeter that produced toxic ozone gas which unfortunately was enough to cost him a few days of hospitalization."
Sounds familiar, yes.
"Just be very sure that the bottom OD3 ignites before the top OD3 ignites.
1 Meg Ohm might not be low enough resistance to ensure that.
Otherwise, expect the bottom MOSFET to say "Ouch"."
Ouchies are to be avoided is possible. Would 680k be enough, or should I go even lower? I read something about this the other day but I don´t remember the details.
Sounds familiar, yes.
"Just be very sure that the bottom OD3 ignites before the top OD3 ignites.
1 Meg Ohm might not be low enough resistance to ensure that.
Otherwise, expect the bottom MOSFET to say "Ouch"."
Ouchies are to be avoided is possible. Would 680k be enough, or should I go even lower? I read something about this the other day but I don´t remember the details.
Suppose the tubes have enough radioactive primer to have an ignition delay well below the RC time constant, and that the resistor in parallel with the upper tube either has a resistance much greater than the filter resistance or is not there at all.
When you then start with empty capacitors and quickly apply a 400 V step, the upper tube is bound to ignite first because of the effect of the lower RC filter; the upper tube's cathode is connected to an empty capacitor via a resistor.
This may cause a voltage peak at the input of the lower RC filter, but as long as the filter time constants are much greater than the ignition delays, that peak won't reach the gate of the lower MOSFET.
I therefore don't see the problem, provided the ignition delays are much smaller than the RC time constants - and if they are not, you are in trouble anyway.
When you then start with empty capacitors and quickly apply a 400 V step, the upper tube is bound to ignite first because of the effect of the lower RC filter; the upper tube's cathode is connected to an empty capacitor via a resistor.
This may cause a voltage peak at the input of the lower RC filter, but as long as the filter time constants are much greater than the ignition delays, that peak won't reach the gate of the lower MOSFET.
I therefore don't see the problem, provided the ignition delays are much smaller than the RC time constants - and if they are not, you are in trouble anyway.