I'm getting a bit tied up understanding maximum screen ratings, especially when the max screen rating is a lot different to the maximum plate rating.
For example, the 807 has a plate limit of 600Vdc, and a screen limit of 300Vdc.
As far as I can understand it, it is not just a question of the voltage, but also the implied dissipation on the screen, since screen current is generally a proportion of the plate dissipation.
So that would imply that if voltage and current are known, then maybe screen voltages can be exceeded in some cases.
However, plate behaviour is also dictated by screen behaviour, so if a screen voltage is too high, then plate current can increase and exceed the plate dissipation limits, causing red-plating.
I have read this can be linked to the way the grid is wound, and in some cases it is hidden behind other grid windings, and in other cases it is not, and hence dictates the behaviour of the tube a bit more. I think an example of this is an EL86, which is sensitive to the screen voltage being exceeded.
All of this is static behaviour - the DC operating point. There are also the dynamic conditions, where a screen with a fixed supply can become positive to the plate, so could have large current changes.
Then in some cases people are using tubes with lower screen ratings in UL with, say, a zener diode string to clamp the maximum voltage. Isn't that going to need to be bypassed so that the dynamic behaviour is correct (if feedback via UL is the goal)?
On the 807 PPP UL thread the idea of a tertiary screen winding was bandied about, that is going to be prohibitively expensive for most people.
What I am trying to do is find a home for some 807 look-a-likes I purchased. They sit very nicely in a Quad II, with a B+ of 330V, and with a 390ohm cathode bias resistor, they have a quiescent current of 58ma. The tubes are operated as pentodes, and the screen supply is regulated from a choke, and is at 325V. The cathode is at 22.5V, so the screen is slightly exceeding its limit of 300V.
What do I need to do to be confident I am not abusing my output tubes? I'm thinking a screen resistor might be prudent to measure current? Or can I measure the resistance of the choke, and infer the current from the voltage drop?
As an aside, has anyone experimented with 5B/255M here?
For example, the 807 has a plate limit of 600Vdc, and a screen limit of 300Vdc.
As far as I can understand it, it is not just a question of the voltage, but also the implied dissipation on the screen, since screen current is generally a proportion of the plate dissipation.
So that would imply that if voltage and current are known, then maybe screen voltages can be exceeded in some cases.
However, plate behaviour is also dictated by screen behaviour, so if a screen voltage is too high, then plate current can increase and exceed the plate dissipation limits, causing red-plating.
I have read this can be linked to the way the grid is wound, and in some cases it is hidden behind other grid windings, and in other cases it is not, and hence dictates the behaviour of the tube a bit more. I think an example of this is an EL86, which is sensitive to the screen voltage being exceeded.
All of this is static behaviour - the DC operating point. There are also the dynamic conditions, where a screen with a fixed supply can become positive to the plate, so could have large current changes.
Then in some cases people are using tubes with lower screen ratings in UL with, say, a zener diode string to clamp the maximum voltage. Isn't that going to need to be bypassed so that the dynamic behaviour is correct (if feedback via UL is the goal)?
On the 807 PPP UL thread the idea of a tertiary screen winding was bandied about, that is going to be prohibitively expensive for most people.
What I am trying to do is find a home for some 807 look-a-likes I purchased. They sit very nicely in a Quad II, with a B+ of 330V, and with a 390ohm cathode bias resistor, they have a quiescent current of 58ma. The tubes are operated as pentodes, and the screen supply is regulated from a choke, and is at 325V. The cathode is at 22.5V, so the screen is slightly exceeding its limit of 300V.
What do I need to do to be confident I am not abusing my output tubes? I'm thinking a screen resistor might be prudent to measure current? Or can I measure the resistance of the choke, and infer the current from the voltage drop?
As an aside, has anyone experimented with 5B/255M here?
It is partially that (especially for the smaller tubes) and partially the pinout (potential arc-over problem).
807 (600V plate/300V screen) is quite closely based on 6L6 (400V plate/300V screen), and those extra 200V are there solely because the plate tab is on the other side of the tube.
Zeners are used here mostly to shift the tube to the optimal OP and to lower the THD. And zeners are quite stable and linear whithin their working current range (that's their job after all), so bypassing them with nonlinear stuff like electrolytic capacitors may induce even more problems (I'm talking here of just the zener in series with the grid, not the zener-based parametric regulator).
2.5 extra volts are not a problem. Even 30 extra volts (10% of the rated value) are not a problem. Just make sure you don't exceed the rated screen dissipation.
1-10 Ohm current measurement resistor in series with the screen grid isn't a bad idea. Current through the choke will include the last capacitor charging surges - that's not the precise way to measure the screen current.
807 is a Beam Power Tube, or as some say, Beam Power Tetrode with Beam formers.
1. 807 in Beam Power mode, Screen 300V Max.
When the plate swings down to 50V, the screen current is excessive, and it is still at 300V.
Now, try the same plate swing to 50V . . .
but connect the screen to 400V. What do you get: Heat. Melting. Shorting. Dead Screens.
2. 807 In Triode Wired Mode, Screen 400V Max.
Why . . . well, when the plate swings down to 50V, guess what . . . the screens are also at 50V.
And, when the plate swings to 800V, the screen is at 800V, but the control grid is so highly negative that there there is almost no screen current, and very little plate current.
Cool, man!
3. 807 in Ultra Linear Mode . . . I estimate that the maximum screen voltage is 350V.
But that depends on what the UL tap percent is.
A. Example, lets use 350V on the plate and the screen, and a 40% tap.
When the plate swings to 50V, it has moved down by 300V.
But the Screen moves down by 40% of 300V, so it moves down by 120V.
350V - 120V = 230V. Well, that is safe, because the screen was rated at 300V.
B. Now, lets use 350V on the plate and the screen, and a 20% tap.
When the plate swings to 50V, it has moved down by 300V.
But the Screen moves down by 20% of 300V, so it moves down by 60V.
350V - 60V = 290V. Well, that is still safe, because the screen was rated at 300V.
C. Now, lets use 350V on the plate and the screen, and a 10% tap.
When the plate swings to 50V, it has moved down by 300V.
But the Screen moves down by 10% of 300V, so it moves down by 30V.
350V - 30V = 320V. We are exceeding the maximum specified screen voltage of 300V.
D. Use UL and 375V on the plate and the screen. I will let you calculate whether that is safe for UL taps of 50%, and for UL taps of 75%.
E. Know what percentage UL tap you are going to use.
Know the B+ you are going to use.
Know that the screen voltage is higher than the plate voltage, because the UL tap DCR has less drop, than the Plate tap has voltage drop (UL tap drop, plus the rest of the DCR to the plate).
Get the idea?
Now you know how to find out if it is within the maximum specs or not.
All maximum specs can be exceeded.
Some 807s are more rugged than other 807s.
But remember . . . Your Mileage May Vary.
1. 807 in Beam Power mode, Screen 300V Max.
When the plate swings down to 50V, the screen current is excessive, and it is still at 300V.
Now, try the same plate swing to 50V . . .
but connect the screen to 400V. What do you get: Heat. Melting. Shorting. Dead Screens.
2. 807 In Triode Wired Mode, Screen 400V Max.
Why . . . well, when the plate swings down to 50V, guess what . . . the screens are also at 50V.
And, when the plate swings to 800V, the screen is at 800V, but the control grid is so highly negative that there there is almost no screen current, and very little plate current.
Cool, man!
3. 807 in Ultra Linear Mode . . . I estimate that the maximum screen voltage is 350V.
But that depends on what the UL tap percent is.
A. Example, lets use 350V on the plate and the screen, and a 40% tap.
When the plate swings to 50V, it has moved down by 300V.
But the Screen moves down by 40% of 300V, so it moves down by 120V.
350V - 120V = 230V. Well, that is safe, because the screen was rated at 300V.
B. Now, lets use 350V on the plate and the screen, and a 20% tap.
When the plate swings to 50V, it has moved down by 300V.
But the Screen moves down by 20% of 300V, so it moves down by 60V.
350V - 60V = 290V. Well, that is still safe, because the screen was rated at 300V.
C. Now, lets use 350V on the plate and the screen, and a 10% tap.
When the plate swings to 50V, it has moved down by 300V.
But the Screen moves down by 10% of 300V, so it moves down by 30V.
350V - 30V = 320V. We are exceeding the maximum specified screen voltage of 300V.
D. Use UL and 375V on the plate and the screen. I will let you calculate whether that is safe for UL taps of 50%, and for UL taps of 75%.
E. Know what percentage UL tap you are going to use.
Know the B+ you are going to use.
Know that the screen voltage is higher than the plate voltage, because the UL tap DCR has less drop, than the Plate tap has voltage drop (UL tap drop, plus the rest of the DCR to the plate).
Get the idea?
Now you know how to find out if it is within the maximum specs or not.
All maximum specs can be exceeded.
Some 807s are more rugged than other 807s.
But remember . . . Your Mileage May Vary.
Last edited:
I have a PPP fixed screen 6L6 with the plate voltage at 436V while the screen at a fixed 327V (B+ minus OC3 105V).
That thing has been working since 1952 and is a gem.
6A3 is right, with a B+ at 330V and a 50 V down swing with a screen at a fixed 325V, you could reach the point where the screen is at a higher voltage then the plate.
I do not think the screen will be happy.
That thing has been working since 1952 and is a gem.
6A3 is right, with a B+ at 330V and a 50 V down swing with a screen at a fixed 325V, you could reach the point where the screen is at a higher voltage then the plate.
I do not think the screen will be happy.
I can say I ran 1625s (12 volt 807) ultralinear in a car amp for about two years without problem. That was with about 410 volts on plate and screen.
Triode operation would probably handle more.
I bought them for $4.50 a pop on sale around 1991.
Triode operation would probably handle more.
I bought them for $4.50 a pop on sale around 1991.
Make a Hi Fi amp, and a Guitar amp with the same voltages and same load.
Running the Beam Power tubes in a Hi Fi, and not running them into constant clipping is one thing (usually we do not like to listen to constant clipping).
Running the same Beam Power tubes in a Guitar amp that is turned all the way to full clipping all the time is another situation.
All Beam Power tubes are equal.
But some Beam Power tubes are More equal than some other Beam Power tubes.
How you treat your tubes voltages, and how you treat you tubes to completely different amplitude levels may give them different lengths of life.
High heat levels integrated over long periods of time kills electronics.
You can probably rev your car engine up to 7,000 RPM for short bursts.
But most commercial cars are not going to work if you run the engine at 7,000 RPM All Day Long!
Running the Beam Power tubes in a Hi Fi, and not running them into constant clipping is one thing (usually we do not like to listen to constant clipping).
Running the same Beam Power tubes in a Guitar amp that is turned all the way to full clipping all the time is another situation.
All Beam Power tubes are equal.
But some Beam Power tubes are More equal than some other Beam Power tubes.
How you treat your tubes voltages, and how you treat you tubes to completely different amplitude levels may give them different lengths of life.
High heat levels integrated over long periods of time kills electronics.
You can probably rev your car engine up to 7,000 RPM for short bursts.
But most commercial cars are not going to work if you run the engine at 7,000 RPM All Day Long!
Last edited:
Thanks for the replies. I’ll fit the screen resistor so I can keep tabs on the current, then calculate the swing from the plate, and then I have the data to confirm I am within tolerances.
OldHector,
You have to look at the tube screen current curves with a fixed screen voltage (beam power mode), versus the plate voltage.
And you have to either do some estimating, or integral calculations to do.
Ultra Linear mode is harder to estimate or calculate.
Triode mode is just easier to estimate or calculate.
And, you need to know the Actual UL tap percentage (%).
Of course, if the amp never approaches constant clipping, the answer changes.
The easy way to make an amplifier more reliable is to stay away from maximum tube ratings, especially the quiescent operating points. Violate those ratings, and it may not matter whether the amp is clipping or not.
Just pick a more powerful tube, or parallel tubes, if you need more power.
Designing on the edge may cause the amp to fall off the edge.
What are your "807 Look Alike" tubes?
What is the link to their data sheet?
You have to look at the tube screen current curves with a fixed screen voltage (beam power mode), versus the plate voltage.
And you have to either do some estimating, or integral calculations to do.
Ultra Linear mode is harder to estimate or calculate.
Triode mode is just easier to estimate or calculate.
And, you need to know the Actual UL tap percentage (%).
Of course, if the amp never approaches constant clipping, the answer changes.
The easy way to make an amplifier more reliable is to stay away from maximum tube ratings, especially the quiescent operating points. Violate those ratings, and it may not matter whether the amp is clipping or not.
Just pick a more powerful tube, or parallel tubes, if you need more power.
Designing on the edge may cause the amp to fall off the edge.
What are your "807 Look Alike" tubes?
What is the link to their data sheet?
Last edited: