Screen current, straighten me out

[Randall]

In revisiting screen current relative to plate current and biasing, I did some reading and measuring. I had gotten into the habit with fixed bias of only measuring cathode current, and thinking plate current was a little less due to screen current, and not worrying too much about it. Today I wanted to play that out and get some numbers.

On a Bassman combo I built that has 5881s and a pentode triode switch, I got the following measurements:

Across 1 ohm cathode resistor Vk = 33.5mV (33.5mA cathode current)
Across 466 ohm screen resistor = 0.53v

Which leads me to:
0.53v/466 ohms = 1.1mA screen current
33.5mA - 1.1mA = 32.4 mA plate current

So, this doesn't seem like enough of a difference to worry about measuring each time, but Is it correct? Is 1.1mA screen current normal? I thought it was more like 3 -5mA from my reading? I seen it mentioned it is somewhere around 5 - 10% of cathode current, and Aiken mentions around 5 mA.

Also, this was in pentode mode with the screens connected to the screen supply. I don't understand why the screen resistor voltage drop is only 4 or 5 mV less in triode mode, and the cathode current only drops 2-3 mA in triode mode. I thought it would have been more. The volume does decrease.

[Enzo]

Cathode resistor?

[Randall]

Sorry, 470 ohm screen resistor that measures 466 ohms. Fixed it.

[nosaj]

Sorry, 470 ohm screen resistor that measures 466 ohms. Fixed it.

That's still an acceptable 10% margin established in most tube amp designs, no?

nosaj

[eschertron]

What signal did you use when measuring screen current? 1ma or thereabout does not seem unreasonable at idle.

[Randall]

I measured with no signal, as I thought was proper.

[Randall]

That's still an acceptable 10% margin established in most tube amp designs, no?

nosaj

This is not the question. At first I quoted the Cathode resistor, Enzo called me on it. It should have been screen resistor, an error I corrected.

[tubeswell]

What was the plate voltage and the screen voltage?

[Randall]

Plate voltage is 464.5, screen voltage is 464.3v.

[alexradium]

In revisiting screen current relative to plate current and biasing, I did some reading and measuring. I had gotten into the habit with fixed bias of only measuring cathode current, and thinking plate current was a little less due to screen current, and not worrying too much about it. Today I wanted to play that out and get some numbers.

On a Bassman combo I built that has 5881s and a pentode triode switch, I got the following measurements:

Across 1 ohm cathode resistor Vk = 33.5mV (33.5mA cathode current)
Across 466 ohm screen resistor = 0.53v

Which leads me to:
0.53v/466 ohms = 1.1mA screen current
33.5mA - 1.1mA = 32.4 mA plate current

So, this doesn't seem like enough of a difference to worry about measuring each time, but Is it correct? Is 1.1mA screen current normal? I thought it was more like 3 -5mA from my reading? I seen it mentioned it is somewhere around 5 - 10% of cathode current, and Aiken mentions around 5 mA.

Also, this was in pentode mode with the screens connected to the screen supply. I don't understand why the screen resistor voltage drop is only 4 or 5 mV less in triode mode, and the cathode current only drops 2-3 mA in triode mode. I thought it would have been more. The volume does decrease.

1mA is normal in 6l6 type tubes,in el34s is more like 5,the screens start to draw more when overdriving the power amp,in triode mode its different,they tend to be stressed more,especially with high voltage and low impedance.

[Helmholtz]

The GE 6L6-GC datasheet specifies Ip=58mA, Is=2.8mA (@ Vp=450V/Vs=400V) corresponding to an Is/Ip ratio of 4.8%.
Assuming constant ratio, Ip=32mA would result in Is=1.5mA for a typical GE 6L6-GC.

From the Mullard EL34 datasheet: Ip=30mA, Is=4mA (@ Vp=500V/Vs=400V) resulting in Is/Ip= 13%.

Real pentodes (EL34) have higher screen currents than beam tetrodes (6L6).

[pdf64]

...Is 1.1mA screen current normal? I thought it was more like 3 -5mA from my reading? I seen it mentioned it is somewhere around 5 - 10% of cathode current, and Aiken mentions around 5 mA...

When reading Aiken, it can be helpful to bear in mind that he's a Marshall guy, and when he talks about power tubes / amps, he's usually coming at it from an EL34 or 84 perspective.

Something to bear in mind is that although current tube manufacturers don't specify what rating system they're using, from the late 50s onwards USA manufacturers generally switched from design centre to design max, whereas UK / European manufacturers seem to have stayed with design centre.
I wonder if that may be worth bearing in mind with the 70% limit thing? ie as it was thought up for design centre tubes, should a lower limit be applied to design max tubes, eg 55%

[Helmholtz]

I wonder if that may be worth bearing in mind with the 70% limit thing? ie as it was thought up for design centre tubes, should a lower limit be applied to design max tubes, eg 55%

As we won't really know, it might be a good idea to use the manufacturers' design examples as guidelines (especially regarding dissipation).

[catalin gramada]

I.m not sure if it worth to bother with screen current by itself. I always take it in consideration in respect with voltage for max operation condition. I found by my limited experience screen grid as be more a power concern of wattage supported by the screens. And start to believe voltage too. Keeping screens iddling at 500v around on my latest projects with different new el34 brands seems to not be a problem. As time can be limited by limiter resistors to not overdisipate and pretty slide under operation. I wonder if someone did a test to see what is the dielectric screen strength, at least from a arcing perspective in respect with internal electrodes related to the ground or whatever the biggest potential differences and risk of arcing is.

[mikepukmel]

As we won't really know, it might be a good idea to use the manufacturers' design examples as guidelines (especially regarding dissipation).

Very difficult for someone like me. The old data sheets, from 50's, 60's, have tons of data. The new data sheets, e.g. from JJ, are 1/2 sheet.