What wattage should screen resistors be?

[frank57]

Originally on my Korean Hiwatt the screens were .5 watt. I bumped them up to 1 watt.
Also increased r44 to 1k.
The voltage there is about 374 volts.What wattage is best for screen resistors?
Usually I see 5 watts or 2 watts.

[tubeswell]

Originally on my Korean Hiwatt the screens were .5 watt. I bumped them up to 1 watt.
Also increased r44 to 1k.
The voltage there is about 374 volts.What wattage is best for screen resistors?
Usually I see 5 watts or 2 watts.

Depends on the tube in question.

6V6s can typically get away with 3W, 6L6s with 5W, but EL34s should maybe be chunky 10W. Ideally you should measure the voltage drop across the resistor(s) in question to calculate the current flow, and from there, work out the power dissipated. Allow for a reasonable power rating buffer remembering that under signal conditions, as the plate voltage increases, the screen current will also increase somewhat.

(I'm over-generalising on the safe side a bit)

[jjasilli]

6L6GC Class Ab1 450VDC Plate voltage.  2 tubes draw 22mA screen current @ full signal, or 11mA per tube.  Typical screen resistor = 470R; Ohms law:  voltage drop = 470 X . 011 = 5.17.  Watts = volts X current = 5.17 X .011 = .06 watts.  http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6L6GC.pdf

However, guitar amps may slam the power tubes with double the max signal. So lets say we cause the screen to draw double the current stated by the tube specs.  Then, 470R X .022A = 10.34V; 10.34V X .022A = .23W.  Double that to be safe and .5W at first seems marginally adequate.  But the screen resistor is often mounted on the tube socket, a high heat location.  So bumping up to at least a 1W rating seems appropriate.

If you use higher value screen resistors, different tubes & current draws, etc., then the same calculations will show the wattage needed.

[DummyLoad]

fender used a 2W carbon comp - likely not just for the wattage rating but the dielectric specification as well.

most builders these days use the ceramic 5W bricks. they work.
 
if i have them i like to use the ohmite series 10 3W - they have a 1000V rating, they are smaller than the 5W bricks, and they strap in easily across (pin 6 to pin 4) an octal socket.

http://www.ohmite.com/cgi-bin/showpage.cgi?product=10_series

[tubeswell]

if i have them i like to use the ohmite series 10 3W - they have a 1000V rating, they are smaller than the 5W bricks, and they strap in easily across (pin 6 to pin 4) an octal socket.

Me too. Neat and tidy

[frank57]

I wonder how well the 1 watt in this amp is taking the punishment.
The amp runs about 388, I think the plate voltage was.(pretty high but that's how they designed it)
Biased at about 60% for el84 tubes.
I could check that again. I do have 5 watt wirewound ones  I can stick there although it is pcb so tricky to fit in.
How high off the pcb would I put them?
I'm a little confused by the calculations, but I would think that at least 2 watt should be used?

[tubeswell]

I'm a little confused by the calculations, but I would think that at least 2 watt should be used?

Measure the DC voltage drop (E) across the resistor (R) under idle conditions. The power (W) being dissipated (in the resistor) will be:

W = E x E/R

The resistor rating should be at least double the dissipation.

[jjasilli]

I think the wattage of the screen resistor should be 2X the diss under full signal condition, not idle.

[frank57]

The plate voltage on the el84 today is about 388-389.
Bias is -16.97 to -17.(fixed bias amp but I have an adjustable pot)
(12ax7)Pin 4 to 9 I get 6.79. Pins 4 to 5: 3.84.

On the EL84 screens I get 384. Not much of a drop most of the time.
Occasionally there is a small voltage drop
383.8 to 384. Or the odd 1 voltage drop:383 to 384 .
This is with nothing plugged in to the input.