looking for 17 watts diss

[scole]

bruce says try el34's at 17 watts dissipation, so i will. its a 5f4 cathode biased. 379v on the plates of groove tube el34's. 28.7v across a 360 ohm cathode resistor. im calculating 13.2 watts dissipation, id like to be at 17 watts. i need some help figuring what value to use for the cathode ressistor.

[MWJB]

Cathode biased? You can go higher than 70%, 70-90% should be fine depending on your taste. Simplify the situation and try 250ohms and 300ohms, ensure max plate current/dissipation isn't exceeded and see which sounds best.

Ideally measure your plate current with some bias probes, I doubt you'll ever get an accurate figure using the method that you describe...not without a lot more math.

[tubeswell]

tube dissipation on cathode biased amps = (plate voltage minus cathode voltage) multiplied by cathode (tube) current i.e.; (374-22) * .044 = 15.5W

Cathode (tube) current = Cathode voltage divide by cathode resistor value (For a pair of tubes in push-pull using a common cathode resistor, you have to divide the resulting tube current measured across the single cathode resistor by half to get the current per tube (i.e. a common 250 Ohm resistor at 22 V yields .088A, so the current per tube would be 44mA)

Since tube current includes both plate current and screen current, you need to deduct a nominal amount for the screen current in order to get a more accurate reading of plate current for biasing purposes. One way of calculating screen current is to measure the voltage drop across each screen grid resistor, then subtract that from the tube current calculated at the cathode.

[MWJB]

But first you need to measure exact values of cathode resistor to make the math work, then you also need to deduct screen current (screen & B+ supplies are typically tied together in a 5F4, so unless Scole has added screen grid resistors at the tube sockets, you are most likely to estimate screen current) ...it's honestly quicker & more accurate to use probes, multiply plate current by plate voltage less cathode voltage, as you say.

250/270ohms are the common values & should work fine at Scole's voltages, 300 may be worth trying? You may end up with 45-55mA? We're biasing the amp primarily to make it sound right, rather than to hit a magic number, so as long as the cathode resistor does that, without exceeding plate dissipation, that's what we want.

[tubeswell]

Right you are MWJB - I forgot to say about the exact measuring of the cathode resistor. For measuring the screen grid current, could he not put in say 1 x 5W 1R screen grid resistor for measurement purposes? (assuming he had measured the exact resistance of the resistor), or a 5W 10R resistor? (or would that have too much effect on the screen voltage?)

[MWJB]

I personally would add 470R 4-5W screen grid resistors, 1K would make the math simpler. Screen voltage could stand to come down some.

Or, the other route would be to allow 1-1.5W from the total dissipation as a "rule of thumb" for 6L6s, if trying to establish a datum from the math.

Cathode bias is relatively forgiving, as long as you're sensible (under 60mA) a few mA here & there isn't going to be make or break, just my philosophy is if you're going to bother to measure something you may as well measure it, rather than take a bunch of other measurements & guesstimate.