Hi all,
I've got a dumb question on ratings for power tubes.. sorry if it's been asked / answered a bunch of times already.
For background, most of the power parts I've worked with professionally are solid state (IGBTs and FETs). Their data sheets will have a "safe operating area" plot of VDS vs Id and lines corresponding to time limits - ie, stay under this line for DC, or that one for up to 10us, etc., and your circuit will hold in its smoke. These are "hard" ratings - even real thin pulses poking above them mean that your circuit is probably about to pop.
With tubes, I'll see a maximum plate voltage.. but then the examples in the data sheet run the DC supply right up to that. In a push pull design in a large signal condition, when a tube is headed toward cutoff, its voltage would be headed toward double the "maximum".
So, my question is, are these more like average "DC" ratings, where the plate is actually allowed to approach double that - ie, a 6L6GC would be a 1000V part if it were rated like a transistor?
I see on some data sheets a second plate rating that seems way high - maybe this is the instantaneous one?
Semi related, what do you guys think would be a "safe" B+ to run on modern KT88s and 6L6GCs? I've been reading that the current production KT88s are kinda flimsy compared to the originals, which has got me worried about coming up with a design that's reliable..
I've got a dumb question on ratings for power tubes.. sorry if it's been asked / answered a bunch of times already.
For background, most of the power parts I've worked with professionally are solid state (IGBTs and FETs). Their data sheets will have a "safe operating area" plot of VDS vs Id and lines corresponding to time limits - ie, stay under this line for DC, or that one for up to 10us, etc., and your circuit will hold in its smoke. These are "hard" ratings - even real thin pulses poking above them mean that your circuit is probably about to pop.
With tubes, I'll see a maximum plate voltage.. but then the examples in the data sheet run the DC supply right up to that. In a push pull design in a large signal condition, when a tube is headed toward cutoff, its voltage would be headed toward double the "maximum".
So, my question is, are these more like average "DC" ratings, where the plate is actually allowed to approach double that - ie, a 6L6GC would be a 1000V part if it were rated like a transistor?
I see on some data sheets a second plate rating that seems way high - maybe this is the instantaneous one?
Semi related, what do you guys think would be a "safe" B+ to run on modern KT88s and 6L6GCs? I've been reading that the current production KT88s are kinda flimsy compared to the originals, which has got me worried about coming up with a design that's reliable..
I think of it like a MOSFET SOA DC rating / 2 - and you don't need LN2 to achieve the stated Pd either.
If the 6L6GC was a transistor it would be a 500W Pd part (but only if you ran it on the planet Neptune for cooling reasons).
Also, max voltage ratings can be exceeded by a LOT if no current is flowing.
You'll see ratings like Max plate voltage = 300V, (550V for a cold tube).
I have run "modern" Chinese KT88 and Tung-Sol KT120 at 560V without issue. I probably wouldn't go higher than 600V.
I have run Soviet 6P3S (6L6 copy) at 400V with no issues.
Tubes are robust compares to SS parts. They don't explode when you look at them wrong. 😀
If the 6L6GC was a transistor it would be a 500W Pd part (but only if you ran it on the planet Neptune for cooling reasons).
Also, max voltage ratings can be exceeded by a LOT if no current is flowing.
You'll see ratings like Max plate voltage = 300V, (550V for a cold tube).
I have run "modern" Chinese KT88 and Tung-Sol KT120 at 560V without issue. I probably wouldn't go higher than 600V.
I have run Soviet 6P3S (6L6 copy) at 400V with no issues.
Tubes are robust compares to SS parts. They don't explode when you look at them wrong. 😀
When you say you run the tubes at 400 or 560V, is that the B+ voltage (where the instantaneous Va-k is heading up toward 1kv on big peaks)?
I think that's what you mean, but just wanted to double check ��
I think that's what you mean, but just wanted to double check ��
Yes.. The B+ voltage.
If I remember correctly, that was B+ to ground, not Va-k. The amp was biased with resistors so about 370V and 500V respectively Va-k.
If I remember correctly, that was B+ to ground, not Va-k. The amp was biased with resistors so about 370V and 500V respectively Va-k.
Examples:
1. 6L6GC Quiescent (no signal conditions), DC ratings:
Maximum Plate Voltage: 500V
Maximum Plate Dissipation: 30W (and 30W with signal)
30W at 500V is 60mA
Suppose the quiescent plate voltage is 450V; 30W at 450V is 66.7 mA.
And,
Maximum Screen Voltage: 450V
Maximum Screen Dissipation: 5 W (and 5W with signal)
5W at 450V is 11.1 mA
60mA + 11.1 mA = 71.1mA cathode current; 66.7mA + 11.1 mA = 77.8mA cathode current.
I could not find a maximum cathode current rating. I know I would not run a 6L6GC at more than 70mA DC cathode current.
If you use the 6L6GC at 30W plate, and 5W screen, it will run very hot (40W).
Plus there is also the filament, 6.3V at 0.9A = 5.67Watts; 45.67W total in the glass envelope.
Run the tube that hot if you want, but it may not last as long as running the same tube a little less hard.
2. 6SN7GT Quiescent (no signal)
Maximum Plate Voltage 300V (used as a Class A amplifier)
With signal, the Plate can rise to 600V at plate current cut off.
The same 6SN7GT, when used as a Vertical Amplifier, The maximum plate voltage is 300V, but with maximum plate voltage during signal, and with the plate current at cut off, the Maximum Pulse Plate Voltage is 1200V.
The 6SN7GT maximum cathode current is 20mA per triode.
And the maximum plate dissipation is 3.5W, but the total of the 2 plate dissipation is only 5W (not 7W).
7W in the 2 plates would run the tube too hot.
The filament is 6.3V at 0.6A (another 3.78W in the same glass envelope; 8.78W total).
1. 6L6GC Quiescent (no signal conditions), DC ratings:
Maximum Plate Voltage: 500V
Maximum Plate Dissipation: 30W (and 30W with signal)
30W at 500V is 60mA
Suppose the quiescent plate voltage is 450V; 30W at 450V is 66.7 mA.
And,
Maximum Screen Voltage: 450V
Maximum Screen Dissipation: 5 W (and 5W with signal)
5W at 450V is 11.1 mA
60mA + 11.1 mA = 71.1mA cathode current; 66.7mA + 11.1 mA = 77.8mA cathode current.
I could not find a maximum cathode current rating. I know I would not run a 6L6GC at more than 70mA DC cathode current.
If you use the 6L6GC at 30W plate, and 5W screen, it will run very hot (40W).
Plus there is also the filament, 6.3V at 0.9A = 5.67Watts; 45.67W total in the glass envelope.
Run the tube that hot if you want, but it may not last as long as running the same tube a little less hard.
2. 6SN7GT Quiescent (no signal)
Maximum Plate Voltage 300V (used as a Class A amplifier)
With signal, the Plate can rise to 600V at plate current cut off.
The same 6SN7GT, when used as a Vertical Amplifier, The maximum plate voltage is 300V, but with maximum plate voltage during signal, and with the plate current at cut off, the Maximum Pulse Plate Voltage is 1200V.
The 6SN7GT maximum cathode current is 20mA per triode.
And the maximum plate dissipation is 3.5W, but the total of the 2 plate dissipation is only 5W (not 7W).
7W in the 2 plates would run the tube too hot.
The filament is 6.3V at 0.6A (another 3.78W in the same glass envelope; 8.78W total).
Last edited:
In the two extremes (1) voltage and (2) current with the max power dissipation of the plate and grid you will see different risks.
With voltage high, you risk voltage stressed breakdown.
With current high, you risk depletion of the electron cloud around the cathode to grid interface, and over stressing the cathode for emission ability. This becomes a wear out effect as the cathode emission degrades with time and is in part dependent on current level.
With voltage high, you risk voltage stressed breakdown.
With current high, you risk depletion of the electron cloud around the cathode to grid interface, and over stressing the cathode for emission ability. This becomes a wear out effect as the cathode emission degrades with time and is in part dependent on current level.
Is that the same effect that (I've heard) causes some kind of damage to the cathode, if Va appears before it's warmed up?
All of these effects fall into the category of "lore", e.i. historical artifacts brought into current wisdom through retelling. In general, published ratings for valves built during an era when they were built in large numbers, by experienced people, with commercial feedback, by large companies with reputations at stake, with specialized and proprietary materials and a market size allowing investment, won't apply to modern valves. But "lore" will allow us to derate our demands as needed.
Separately, cathodes are potentially damaged by "stripping" (material damage at short term high voltages without negative grid voltage) and "poisoning" (long term material damage at too little cathode current, sometimes called "interface"). But electron valves don't have secondary breakdown, and their huge areas and large localized thermal masses relative to semicons made for similar service gives robustness.
A 6L6GC would be a 1000V part without the octal base. Versions with the anode brought out the other end, called type 807, 1625, etc. do that without breathing hard. They also survive reversed voltages without major issues. You do have to watch heater-to-cathode voltages.
All good fortunes,
Chris
Separately, cathodes are potentially damaged by "stripping" (material damage at short term high voltages without negative grid voltage) and "poisoning" (long term material damage at too little cathode current, sometimes called "interface"). But electron valves don't have secondary breakdown, and their huge areas and large localized thermal masses relative to semicons made for similar service gives robustness.
A 6L6GC would be a 1000V part without the octal base. Versions with the anode brought out the other end, called type 807, 1625, etc. do that without breathing hard. They also survive reversed voltages without major issues. You do have to watch heater-to-cathode voltages.
All good fortunes,
Chris
Last edited: