Thanks, I have a stash of VU29, which is a bit like an early 872 in size and have no shield around the cathode. I'll probably build something with them one day, but the costs of filament transformers and electric bills is a little off putting. Looks like the spares will last me a life time. So far I've only tried them in a half wave test circuit using a variac to bring a 6.3v tranformer down to 4v that they needed.
Why go to all the trouble you may wonder - well I've seen a few different people on this and other forums sing their praises in audio circuits so thats temptation enough.
Why go to all the trouble you may wonder - well I've seen a few different people on this and other forums sing their praises in audio circuits so thats temptation enough.
Oliver Heaviside solved the speed limit problem by shewing the cable had to be terminated in its characteristic impedance.
Ideally the termination is a simple resistance. 👍
Hi all,
I built a SE-Amp with the big KT150/170 Tetrodes two years ago using two mercury rectifiers in the power supply. The power tubes work with 450V plate, 300V screen and about 130-150 mA plate current. Together with the pre-stages, total HV current consumption adds to nearly 400 mA. That's the reason why I used gaseous rectifiers (Xenon types would work as well). Obviously I had to take stronger tubes than the 83, but the operational basics are the same, that's why I'm writing here.
Power supply is a choke input layout (mandatory for gas rectifiers) delivering about 460V - the 300V screen supply stabilized down via glow reference (the two 0D3/VR150 tubes on the left) and Mosfet regulator.
I used Edison screw base rectifier tubes, since I have quite a lot of them at hand. These are all more or less 866A-equivalents, one photo shows two VH-550 tubes (with the nice blue glow and the Tung-Sol KT150), the other shows two rectifiers from "La verrerie scientifique" ("the scientific glassworks"), type "7500-0.6", using deep cup-shaped graphite anodes hiding (sadly....😉) the nice blue glow almost completely. Also testing here the new Psvane KT150-J - I like it a lot! Both rectifier types are French military grade production.
Almost all gas-filled rectifiers (no matter if Hg, Xe or Ar), especially the high-power types, need sufficient preheating of the cathode to avoid heavy stripping of the oxide coating during warmup. I do this with a small time relay using the time constant of a R-C-array which switches a relay via a small thyristor after reaching a certain level of charge - see the small sketch. Usually I feed this time relay by the heater or grid bias voltages, adapting the part values accordingly. This adds to system safety since HV can be switched on only if heater or bias voltages are ok. Delay time usually is 3-5 minutes.
Generally, with these comparably low voltages (< 1000V), there are no problems with backfiring or RF hash generation - at least I never encountered them (and this is not my first amp using gas rectifiers).... Also, with decreasing voltages, the current capabilities of these rectifier tubes can safely be increased (for the 866A, usually 250 mA are allowed as max. continuous current at voltages > 5kV). But look at the old battery charger rectifiers, with the same cathode heating power, they deliver several Amperes continuously at low voltages (100-200V). So the 400 mA in my amp are no overload at all. In the meantime, the amp has run for about 2000 hours, the rectifier tubes are still in perfect shape (contrary to other tubes.....).
As I said, choke input is mandatory, the tubes "love" to get ignited smoothly with current slowly rising after they have become conductive. Cap input filters with their high surge currents would lead to a more or less quick erosion of the oxide coating, reducing tube lifetime considerably.
Longtime unused mercury tubes need a special treatment - a "bakeout" prior to first use. Mercury alloys with almost all metals, that's why it sometimes seems to have "disappeared" in longtime unused tubes. Extended heating sets the mercury free again, especially the plate needs to be free from any droplets or condensations prior to use, since these would instantly evaporate at switch-on and heavily risk backfiring. On my amp, there is a small switch in front of the glowtubes on the left, this is the "bakeout" switch which simply switches off the time relay to allow for extended heating (0.5 - 1 hour should usually be enough...).
To say it again - Mercury is toxic, these tubes are no toys, know what you're doing, have a mercury spill kit ready, keep them out of your children's reach, dispose of them properly! So you can enjoy the blue glow safely!
all the best
Uli
@tubeampuli
I built a SE-Amp with the big KT150/170 Tetrodes two years ago using two mercury rectifiers in the power supply. The power tubes work with 450V plate, 300V screen and about 130-150 mA plate current. Together with the pre-stages, total HV current consumption adds to nearly 400 mA. That's the reason why I used gaseous rectifiers (Xenon types would work as well). Obviously I had to take stronger tubes than the 83, but the operational basics are the same, that's why I'm writing here.
Power supply is a choke input layout (mandatory for gas rectifiers) delivering about 460V - the 300V screen supply stabilized down via glow reference (the two 0D3/VR150 tubes on the left) and Mosfet regulator.
I used Edison screw base rectifier tubes, since I have quite a lot of them at hand. These are all more or less 866A-equivalents, one photo shows two VH-550 tubes (with the nice blue glow and the Tung-Sol KT150), the other shows two rectifiers from "La verrerie scientifique" ("the scientific glassworks"), type "7500-0.6", using deep cup-shaped graphite anodes hiding (sadly....😉) the nice blue glow almost completely. Also testing here the new Psvane KT150-J - I like it a lot! Both rectifier types are French military grade production.
Almost all gas-filled rectifiers (no matter if Hg, Xe or Ar), especially the high-power types, need sufficient preheating of the cathode to avoid heavy stripping of the oxide coating during warmup. I do this with a small time relay using the time constant of a R-C-array which switches a relay via a small thyristor after reaching a certain level of charge - see the small sketch. Usually I feed this time relay by the heater or grid bias voltages, adapting the part values accordingly. This adds to system safety since HV can be switched on only if heater or bias voltages are ok. Delay time usually is 3-5 minutes.
Generally, with these comparably low voltages (< 1000V), there are no problems with backfiring or RF hash generation - at least I never encountered them (and this is not my first amp using gas rectifiers).... Also, with decreasing voltages, the current capabilities of these rectifier tubes can safely be increased (for the 866A, usually 250 mA are allowed as max. continuous current at voltages > 5kV). But look at the old battery charger rectifiers, with the same cathode heating power, they deliver several Amperes continuously at low voltages (100-200V). So the 400 mA in my amp are no overload at all. In the meantime, the amp has run for about 2000 hours, the rectifier tubes are still in perfect shape (contrary to other tubes.....).
As I said, choke input is mandatory, the tubes "love" to get ignited smoothly with current slowly rising after they have become conductive. Cap input filters with their high surge currents would lead to a more or less quick erosion of the oxide coating, reducing tube lifetime considerably.
Longtime unused mercury tubes need a special treatment - a "bakeout" prior to first use. Mercury alloys with almost all metals, that's why it sometimes seems to have "disappeared" in longtime unused tubes. Extended heating sets the mercury free again, especially the plate needs to be free from any droplets or condensations prior to use, since these would instantly evaporate at switch-on and heavily risk backfiring. On my amp, there is a small switch in front of the glowtubes on the left, this is the "bakeout" switch which simply switches off the time relay to allow for extended heating (0.5 - 1 hour should usually be enough...).
To say it again - Mercury is toxic, these tubes are no toys, know what you're doing, have a mercury spill kit ready, keep them out of your children's reach, dispose of them properly! So you can enjoy the blue glow safely!
all the best
Uli
@tubeampuli
A beautiful construction. Yours doesn't seem as if it would stress type 866 rectifiers too much, but for thoriated tungsten output valves, stuff in the "50 Watter" class or bigger "Eimac" class, maybe a kiloVolt B+ and several hundred milliAmps, we'd want to also consider charging current of all the power supply capacitors at turn-on. At the very low frequency of the charging rate the inductors' impedances are too small to effect capacitor charging current - they're pretty much just resistors. Peak rectifier current limits can be a significant issue; series'd "damper" diodes can work well.
Another design issue for any sequenced power supply is hot-switching, and it's extra tricky for gas discharge rectifiers because everything needs to restart from time zero. We built one recently with no semi-con "cheating" and it required two two-pole relays and the Amphenol bi-metal delay switch - can be done, drama free, but a moderately PITA.
Regarding switching impulse noise, I'd suggest that the huge ignition noise of gas discharge rectifiers swamps any little wiggles near zero crossing caused by junction effects in semi-con rectifiers. Both should of course be damped by snubbers across the transformer's secondary. Without talking about actual numbers, there's not much to discuss. This in response to quoted text in post 55 above.
MV rectifiers have two additional issues that everyone must be aware of, UV radiation (don't stare at them! when operating) and EMI (electronic interference). Even if the EMI isn't illegal in your location, it can effect others. Maybe even your own (!) hifi sound.
All good fortune,
Chris
Another design issue for any sequenced power supply is hot-switching, and it's extra tricky for gas discharge rectifiers because everything needs to restart from time zero. We built one recently with no semi-con "cheating" and it required two two-pole relays and the Amphenol bi-metal delay switch - can be done, drama free, but a moderately PITA.
Regarding switching impulse noise, I'd suggest that the huge ignition noise of gas discharge rectifiers swamps any little wiggles near zero crossing caused by junction effects in semi-con rectifiers. Both should of course be damped by snubbers across the transformer's secondary. Without talking about actual numbers, there's not much to discuss. This in response to quoted text in post 55 above.
MV rectifiers have two additional issues that everyone must be aware of, UV radiation (don't stare at them! when operating) and EMI (electronic interference). Even if the EMI isn't illegal in your location, it can effect others. Maybe even your own (!) hifi sound.
All good fortune,
Chris
Perhaps a worthy application for Quasimodo checking of appropriate CRC snubber on each power transformer winding section, to dampen the winding leakage inductance that would be interacting with any flicker/fluctuations of the diode current. In addition to some small value C before the choke to soak up and bypass any higher dI/dt, and keep it on the transformer side of the choke.
Also WRT switching impulse noise, any vacuum diode rectifiers will dominate rectifier recovery time and its wiggles, so folks should remember the Tim Robbins improvements/enhancements to the "yellow sheet mods". If not already familiar, this is a series combination of (originally semi-con, but applicable generally) rectifiers and vacuum rectifiers (original version, intended to protect crap modern rectifiers), with the added benefit of reduced peak loads to the vacuum rectifiers. And it only costs a piece of wire. Great stuff and overlooked.
I still have questions about the Mark Johnson criteria for damping. Probably missing something, but I just don't see the value of critical damping of the inductive ringing. More damping is more better? seems to me. Happy to be corrected of course.
All good fortune,
Chris
I still have questions about the Mark Johnson criteria for damping. Probably missing something, but I just don't see the value of critical damping of the inductive ringing. More damping is more better? seems to me. Happy to be corrected of course.
All good fortune,
Chris
I see the damping as a practical compromise, given the physical size and cost of the snubber parts, which are preferably located as close as practical to the winding. Imho, Quasimodo is a convenient test tool for those who can't make their own, and for those who have queries about how to interpret waveforms, as it defines a set outcome to aim for.
For valve amp type power transformers, the winding shunt capacitance can be substantial, to a level where just an RC snubber may be a reasonable outcome, and practical given the cap size and voltage rating and X factor.
Quasimodo test tool looks at the response of the transformer winding to a current step type disturbance, which appears to be the disturbance effect generated by the mercury vapour diode and it's internal arc machinations.
For valve amp type power transformers, the winding shunt capacitance can be substantial, to a level where just an RC snubber may be a reasonable outcome, and practical given the cap size and voltage rating and X factor.
Quasimodo test tool looks at the response of the transformer winding to a current step type disturbance, which appears to be the disturbance effect generated by the mercury vapour diode and it's internal arc machinations.
Slight detour; is the Quasimodo test tool available to purchase? I've done a search without any sales information coming up.
Perhaps best to use the Quasimodo thread - it's long but it answers your query. https://www.diyaudio.com/community/...rmer-snubber-using-quasimodo-test-jig.243100/