Gas regulator tubes

[rkoonce]

I'm looking for some insight into the failure mode(s) of gas rectifier tubes. Here's my symptoms:

Raw B+ is around 295V. 600 ohm R to series 0C3s. Regulated B+ taken at the 600 ohm R. Amp played normally about 15 minutes and I saw a flash in the 82 rectifier and the 250mA B+ fuse blew. I replaced the fuse and the amp played, but the 0C3s didn't strike. Got the amp on the bench and found "regulated" B+ about 40V too high, shooting for 210V. I tested both 0C3s on my tube tester which as no settings for that tube, so I used 0A2, assuming similar pinouts. One tube tested in the "replace" area on the meter and the other tested in the "good" area. I replaced the tube that tested bad and got no strike. I fiddled with the other pair of 0C3s I have and was finally able to get a pair to strike, one of which was the original "replace" tube. The regulated B+ is still about 240V. Should I expect 210-214?

Amp has had fairly heavy use, about 1.5 years at 8-10 hours a day. This is the first failure.

TIA from *HOT* East Texas. 100° again!

Raymond

[Miles Prower]

600R seems awfully small for that ballast resistor. These regulator tubes aren't rated for all that much current, this being 40mA for the bigger octal types, and 30mA for the seven pin mini types.

There is also a potential problem with series operation: the bottom VT might not get enough voltage to strike, especially if the supply voltage is marginal (it is here) and the VTs have had some use, especially if rough, which seems likely given the Vpp and the size of the ballast resistor.

One thing you can do is connect the plate of the bottom reg VTs to the main supply through a 1.0MEG resistor. This will ensure enough voltage to strike without compromising voltage regulation or pulling too much current.

I've always used them as voltage references for active regulators and so run them at ~7.0mA -- ~10mA. Haven't seen any failures yet since these aren't being pushed very hard.

[sandyK]

Why not replace them with a few series connected 5W Zener diodes ?
e.g. 1N5369B 1N5369B 51V 5W Zener Diodes .
Futerlec has them for about 80C each, as well as other voltages.
Earlier LME Crossbar Telephone Exchanges used them for regulation of high voltage test circuits derived from transistor inverters that stepped up the voltage from 52V to around 200V. They were fitted to replace similar to what you are using to greatly increase reliability. I would have replaced 100s of those failing VR tubes for that reason.They were a P.I.T.A.

SandyK

[Arnulf]

Originally Posted by Miles Prower

One thing you can do is connect the plate of the bottom reg VTs to the main supply through a 1.0MEG resistor. This will ensure enough voltage to strike without compromising voltage regulation or pulling too much current.

Alternative which doesn't waste any current would be a capacitor in place of resistor in parallel with one regulator tube. This will allow one tube to strike first with full B+ avaliable, and then the other with B- minus regulation voltage (which is lower than striking voltage). Capacitors are more expensive than resistors but if power consumption was an issue ...

[SY]

I'd be careful about a parallel cap- that can form an unintended relaxation oscillator. Miles's idea about a separate strike resistor for the bottom tube would be a safer bet.

That said, a couple of Zeners will outperform the gas reg tubes, but not look nearly as cool.

104 here yesterday, so I have zero sympathy.

[Globulator]

Originally Posted by SY

I'd be careful about a parallel cap- that can form an unintended relaxation oscillator. Miles's idea about a separate strike resistor for the bottom tube would be a safer bet.

I could get mine (OA2 and OB2 in series) oscillating with any load, resistive or capacitive. In fact the only time they didn't oscillate was completely unloaded, so I borrowed a tip from Wavebourne and used an IRF730 source follower with a zener protecting the gate - the only load that it could take without oscillating was a mosfet gate

rkoonce: sounds like you need a new 82 regulator

[rkoonce]

Originally Posted by Miles Prower

600R seems awfully small for that ballast resistor. These regulator tubes aren't rated for all that much current, this being 40mA for the bigger octal types, and 30mA for the seven pin mini types.

There is also a potential problem with series operation: the bottom VT might not get enough voltage to strike, especially if the supply voltage is marginal (it is here) and the VTs have had some use, especially if rough, which seems likely given the Vpp and the size of the ballast resistor.

One thing you can do is connect the plate of the bottom reg VTs to the main supply through a 1.0MEG resistor. This will ensure enough voltage to strike without compromising voltage regulation or pulling too much current.

I've always used them as voltage references for active regulators and so run them at ~7.0mA -- ~10mA. Haven't seen any failures yet since these aren't being pushed very hard.

600R was the largest value which would allow the tubes to strike. I'll try the idea of the 1Mohm R between the bottom 0C3 anode and the B+ to see if that will allow a rejected tube to strike.

[rkoonce]

Originally Posted by sandyK

Why not replace them with a few series connected 5W Zener diodes ?
e.g. 1N5369B 1N5369B 51V 5W Zener Diodes .
Futerlec has them for about 80C each, as well as other voltages.
Earlier LME Crossbar Telephone Exchanges used them for regulation of high voltage test circuits derived from transistor inverters that stepped up the voltage from 52V to around 200V. They were fitted to replace similar to what you are using to greatly increase reliability. I would have replaced 100s of those failing VR tubes for that reason.They were a P.I.T.A.

SandyK

Zeners don't look nearly as nice.

[rkoonce]

I installed the 1M resistor from B+ to the anode of the bottom VR tube and was able to get a more consistent strike. Thanks for that tip. I now have three working VR tubes, so I suppose the other one died in a flash. Just for grins, I put a 10 ohm R in series with the two VR tubes so I could measure voltage drop and I'm running 15mA through them. That sounds about right to me.

What dies in a VR tube? The thing still glows in my tube tester. Aren't they just a specially designed neon/argon tube? I've got 60 year old neon tubes that still glow brightly. In the tester as I have it set now, a good 0C3 tests in the "replace" range while the bad one almost pegs the meter in the "good" range. Somebody explain the science to me.

Best regards,
Raymond