VR Tubes

[Carlp]

I'm working with some VR tubes (OD3 and OA3 mostly) and have a few questions I can't seem to find the answers to.

1. Specs say the tube (e.g. OD3) needs 160V min. to start the tube and 185 min. supply voltage. In this case, since the min 185 is above the 160 starter voltage, can't I just apply 185 to the anode and expect it to work properly (if current is controlled)?

2. And...I'm interested in possibly using two in series (to get the proper voltage). If I connect two in series to increase the voltage that's regulated, how do I figure the starter voltage and min voltage? If a single OD3 requires 160 starter and 185 min to operate at it's 150V regulation point, do I need 310V starter voltage and 335 min voltage to operate (add the same extra voltage as for one tube) in order to get the 300 regulated volts out, or something else?

Thanks a ton,
Carl

[Miles Prower]

Originally posted by Carlp
I'm working with some VR tubes (OD3 and OA3 mostly) and have a few questions I can't seem to find the answers to.

1. Specs say the tube (e.g. OD3) needs 160V min. to start the tube and 185 min. supply voltage. In this case, since the min 185 is above the 160 starter voltage, can't I just apply 185 to the anode and expect it to work properly (if current is controlled)?

That will poof it. The whole idea is that the VR tube represents only a small amount of capacitance until it actually starts. Since there is very little current, there isn't much voltage drop across the ballast resistor. Once the voltage is high enough to start it, the current limits the voltage to whatever the reg voltage is. The 185 min. voltage suggestion takes into account, not only starting voltage, but ballast resistor voltage drop.

2. And...I'm interested in possibly using two in series (to get the proper voltage). If I connect two in series to increase the voltage that's regulated, how do I figure the starter voltage and min voltage? If a single OD3 requires 160 starter and 185 min to operate at it's 150V regulation point, do I need 310V starter voltage and 335 min voltage to operate (add the same extra voltage as for one tube) in order to get the 300 regulated volts out, or something else?

That's pretty much it. Two will require twice the starter voltage plus a margin for ballast resistors. To aid starting when doing series strings, you can connect a 1.0M resistor between the plate(s) of the tube(s) below the uppermost VR tube and the DC supply. This will supply the necessary starting voltage so that they all start up, but it pulls very little current when they're going so that nothing poofs and degrades the voltage regulation.

Just remember that VR tubes, since they operate by means of a glow discharge, have a negative resistance characteristic so you can't parallel them with a big capacitor (like you would do with a Zener) without making an oscillator.

[Carlp]

The 185 min. voltage suggestion takes into account, not only starting voltage, but ballast resistor voltage drop.

Ahhhh. Thanks, Miles. That supply voltage spec was confusing me, but now I understand.

To aid starting when doing series strings, you can connect a 1.0M resistor between the plate(s) of the tube(s) below the uppermost VR tube and the DC supply. This will supply the necessary starting voltage so that they all start up, but it pulls very little current when they're going so that nothing poofs and degrades the voltage regulation.

Interesting and very helpful. You're saying the resistor goes between the cathode of the upper and the anode of the lower? Thanks again.

Also, re: bypassing with a large cap, I've read a little about doing this (intentionally, though not in an amp), but I thought I also read somewhere that it's a good idea to bypass with a LOW value cap (e.g. .1uF or lower) for some reason I can't recall right now. And one design I've looked at uses a 100uF + cap BEFORE the ballast resistor, tied to ground (bypassing). Does the resistor somehow prevent oscillation?

Thanks,
Carl

[ray_moth]

Does the resistor somehow prevent oscillation?

Yes but its purpose is to act in concert with the VR tube to provide a regulated voltage at the junction of the two. The load is in parallel with the VR, so if the load starts to take more current than usual, causing a bigger voltage drop across the ballast resistor, the VR takes less current to compensate (and vice-versa).

[Carlp]

Thanks, Ray. I understand the role of the resistor in general. What I was wondering is how does it stop oscillation (e.g. when there's an upstream cap bypass, but presumably also when there isn't), and I think you've explained it if I understand it fully. It's sort of self biasing. Any idea what changes when the value of the bypass cap (bypassing just the VR tubes, not the tube and resistor) is (relatively) small?

Carl

[Tweeker]

As mentioned, cap bypass has to be really small, <100nF, or you get a relaxation oscillator. The change is slightly lower impedance at high frequency, slightly less noise. Gas tubes are quieter than zeners anyhow.

Also, re: bypassing with a large cap, I've read a little about doing this (intentionally, though not in an amp), but I thought I also read somewhere that it's a good idea to bypass with a LOW value cap (e.g. .1uF or lower) for some reason I can't recall right now. And one design I've looked at uses a 100uF + cap BEFORE the ballast resistor, tied to ground (bypassing). Does the resistor somehow prevent oscillation?

The 100uF cap is just a PSU filter element if its located before the VR tubes ballast resistor.

Shunt regulator, in our case we have a VR tube instead of a transistor.
The 100uF cap would be the source on the left. R is the ballast resistor.

[Carlp]

The 100uF cap is just a PSU filter element if its located before the VR tubes ballast resistor.

Ah, shoulda guessed that.

To aid starting when doing series strings, you can connect a 1.0M resistor between the plate(s) of the tube(s) below the uppermost VR tube and the DC supply.

Hmmm. My mistake earlier. Connect the 1M resistor to the DC supply BEFORE the VR ballast resistor or after it?

Thanks,
Carl

[Tweeker]

Doesnt really matter, does it? Its a meg.