I cannot find any information on the response rate of glow tubes.
Do they react to voltage changes quickly enough (at 60 Hz) to replace the 2nd or 3rd cap in a power supply for ripple suppression? (they cost less than a large cap)
Do they react to voltage changes quickly enough (at 60 Hz) to replace the 2nd or 3rd cap in a power supply for ripple suppression? (they cost less than a large cap)
Large caps nowadays are pretty cheap.
However, the Cheap-O-Air commercials scare me when it comes to air travel...
However, the Cheap-O-Air commercials scare me when it comes to air travel...
There are many types of "glow tubes".
However the "Voltage Regulator" types like 0A3 were *made* for DC smoothing.
I did not think they were cheaper than a large cap--- they have not been made in decades, while everybody in Asia makes caps.
The load current can not be large, or very-variable. The open-circuit voltage must be some higher than the regulated voltage.
They will reduce but not eliminate ripple.
They hiss real bad. You *still* need a cap to reduce this. The cap can not be directly across the glow-tube or it is a "Relaxation Oscillator" (see page 4).
They have a pretty glow.
However the "Voltage Regulator" types like 0A3 were *made* for DC smoothing.
I did not think they were cheaper than a large cap--- they have not been made in decades, while everybody in Asia makes caps.
The load current can not be large, or very-variable. The open-circuit voltage must be some higher than the regulated voltage.
They will reduce but not eliminate ripple.
They hiss real bad. You *still* need a cap to reduce this. The cap can not be directly across the glow-tube or it is a "Relaxation Oscillator" (see page 4).
They have a pretty glow.
You have to limit the current through them, so it is not higher than 40 mA, otherwise you'll see a firework instead of voltage stabilization. When it ignites, it's dynamic resistance drops significantly, so if you just use it instead of a cap, it can be damaged. "response rate" at 60 Hz is irrelevant, it works up to much higher frequencies, you should not worry. But if you power from it very sensitive input stages, you must smooth after it using a rc filter, otherwise it's noise may siphon to the signal path.
I just few minutes ago soldered one, with 51K resistor from 450V source. It will give a reference voltage to the source follower that will power screen grids of output stage, so I have the link still open:
https://frank.pocnet.net/sheets/155/v/VR75-30.pdf
A few questions for those of you who have used these:
1. Frank's RCA and GE data sheets prohibit use of the tubes in parallel, but his Sylvania data sheet specifies (on page 2) that they can be used in parallel with a 100 ohm resistor between each. So which is correct?
2. Can they be cascaded in series working down from the OD3 to OC3 to OB3 and OA3 to provide taps for 150V, 105V, 90V, and 75V? They seem to be spaced so that the regulated voltage of each would be higher than the minimum voltage of the next in such a circuit.
3. If so I guess that the circuit could double the tubes in series for voltages of 300 210 180 150 105 90 and 75?
At $1 or $2 each this is a lot less than a multi-tap transformer!
1. Frank's RCA and GE data sheets prohibit use of the tubes in parallel, but his Sylvania data sheet specifies (on page 2) that they can be used in parallel with a 100 ohm resistor between each. So which is correct?
2. Can they be cascaded in series working down from the OD3 to OC3 to OB3 and OA3 to provide taps for 150V, 105V, 90V, and 75V? They seem to be spaced so that the regulated voltage of each would be higher than the minimum voltage of the next in such a circuit.
3. If so I guess that the circuit could double the tubes in series for voltages of 300 210 180 150 105 90 and 75?
At $1 or $2 each this is a lot less than a multi-tap transformer!
Like leds these tubes shouldn't be used in direct parallel. With a load balancing resistor you might get away with running them in parallel but it's possible that one tube will strike before the other.
You should be able to cascade them. Striking voltage of the lower voltage stage will be lower than the one before it so they should all fire.
Series regulator tubes are fine.
You should be able to cascade them. Striking voltage of the lower voltage stage will be lower than the one before it so they should all fire.
Series regulator tubes are fine.
There is an unpredictable delay before they ignite. It can vary from microseconds to seconds depending on the type of tube, whether it is dark or light and whether the manufacturer put a radioactive primer in the tube. I measured average delays from 25.79 µs to 7.675 s on four 85A2 voltage reference tubes from different brands in darkness. An ordinary neon indicator lamp had a delay of about 10 µs with some ambient light, on average 21.6 ms in darkness.
Once they are ignited, they can easily keep up with 50 Hz...120 Hz ripple.
When you connect two in series, shunting one of the two with a 1 Mohm resistor helps to keep the total ignition voltage low.
Once they are ignited, they can easily keep up with 50 Hz...120 Hz ripple.
When you connect two in series, shunting one of the two with a 1 Mohm resistor helps to keep the total ignition voltage low.
Personally I would not use them for removing ripple, but I would use one for removing mains voltage variations from an RIAA preamp.
If you need to connect two in series to get a higher voltage then there are circuits which wire them in parallel for firing and then in series for regulating. One diode and two resistors do the trick.
If you need to connect two in series to get a higher voltage then there are circuits which wire them in parallel for firing and then in series for regulating. One diode and two resistors do the trick.
I like the 1 Mohm shunt for starting the tubes in series - an elegant solution.
It seems that the major limitation of these tubes is current capacity. The obvious solution would be to use multiples in parallel but this is forbidden in all references except Sylvania with their 100 Ohm resistor fix. If the solution to enable tubes in parallel is as easy as adding a small resistor how could the other manufacturers missed this? They could have considerably increased the range of applications and sales of their product by promoting this simple fix, communicated with one sentence in their data sheet.
It seems that the major limitation of these tubes is current capacity. The obvious solution would be to use multiples in parallel but this is forbidden in all references except Sylvania with their 100 Ohm resistor fix. If the solution to enable tubes in parallel is as easy as adding a small resistor how could the other manufacturers missed this? They could have considerably increased the range of applications and sales of their product by promoting this simple fix, communicated with one sentence in their data sheet.
They were designed and sold as voltage regulators. Putting several in parallel wastes current. Better to make a proper voltage regulator using a gas tube as the reference if you need more control. Maybe Sylvania found a way of equalising the striking voltage - or ignored the problem.
I see you still don´t get it, sorry.
Please read this:
1) Constant voltage sources (such as glow tubes, Zener diodes, amplifier outputs, etc.), can NOT be put in parallel.
And that´s what all manufacturers say, including Sylvania.
2) If you add series resistors to a constant voltage source, it is not a constant voltage source any more. 😱
3) the better you match the source voltage, the smaller a resistor you need.
But you still need them, they never suggest: "our tubes are so well matched that you just can straight parallel them" by any means.
Last edited:
I don't see how it can work at all unless the resistance is so high that when all current flows through one branch, the maintaining voltage plus the drop across the resistor exceeds the ignition voltage of the other tube.
For example, suppose you try to push 80 mA through the parallel connection of two 0A3 tubes with series resistors. Depending on which tube happens to have free electrons and ionized atoms available first, one of the two tubes will ignite. The voltage across it then drops to 80 V or so, a bit higher than the nominal 75 V because the current is higher than normal. The other tube needs at least 100 V to ignite, so the drop across the series resistor has to be at least 20 V. 20 V at 80 mA means 250 ohm, so the resistors have to be at least 250 ohm. At lower current levels, they would need to be even higher.
When both tubes are ignited, the current per branch will halve. You then still have as much as 10 V of drop across the resistors, much reducing the regulation.
For example, suppose you try to push 80 mA through the parallel connection of two 0A3 tubes with series resistors. Depending on which tube happens to have free electrons and ionized atoms available first, one of the two tubes will ignite. The voltage across it then drops to 80 V or so, a bit higher than the nominal 75 V because the current is higher than normal. The other tube needs at least 100 V to ignite, so the drop across the series resistor has to be at least 20 V. 20 V at 80 mA means 250 ohm, so the resistors have to be at least 250 ohm. At lower current levels, they would need to be even higher.
When both tubes are ignited, the current per branch will halve. You then still have as much as 10 V of drop across the resistors, much reducing the regulation.
That´s exactly how it works 🙂
Sadly Greg sticks to his mis-interpretation of what the Sylvania datasheets says ... which does not contradict what other Manufacturers say, it simply describes a different situation ... which of course degrades voltage regulation by introducing a resistive element.
If you need higher current, best is to use a single glow tube as a reference to drive a pass tube ... or transistor ... or Mosfet, pick one.
Or go the full way and use a Zener 😉
Last edited:
Um… I know it is anathema to the Tube Forum… but if you're in the market to play with pretty glow discharge tubes and you want them to be employed to a really useful purpose in-circuit, why not a chain of them as a quite-stabile voltage reference, which drives the gate of a MOSFET series source-follower regulator? You get as much current as any downwind circuit might possibly deign to desire, and you get better than –40 dB ripple reduction at the same time. Supposing, of course, that the capacitor stack on the “left hand side” of the circuit is also good.
Anyway, as I said, anathema to the Tube forum, I suppose.
Just saying,
GoatGuy
C₁L₂C₃M₄C₅ … my favorite configuration.
C₁ — to be the front line pulse-smoothing from rectifiers
L₂ — to chop out the annoying buzz of rectification and triangle waves
C₃ — to significantly “be the reservoir” in primary ripple reduction
M₄ — MOSFET series regulation. –40 dB ripple reduction. Excellent stability.
C₅ — downwind impedance buffering. Modest size, transient handling.
C₁ — to be the front line pulse-smoothing from rectifiers
L₂ — to chop out the annoying buzz of rectification and triangle waves
C₃ — to significantly “be the reservoir” in primary ripple reduction
M₄ — MOSFET series regulation. –40 dB ripple reduction. Excellent stability.
C₅ — downwind impedance buffering. Modest size, transient handling.
Anyway, as I said, anathema to the Tube forum, I suppose.
Just saying,
GoatGuy
It's not clear to me which of the dozens of Sylvania datasheets Greg referred to in post #7, but if it is a 0A3 datasheet, how can they get away with resistors of only 100 ohm, rather than 250 ohm at maximum current, even more at lower currents?
I had an old Boonton ACVM. The glow tubes could take minutes to an hour to come up even with full bench-light on them.
- Status
- Not open for further replies.