Dual rail and voltage divider on screens - MATH....

[EFK]

I could use a little help being as I am somewhat math-challenged. I'm experimenting with a prototype amp involving 6973 tubes and running them off a dual rail. So the plates are diode rectified @ 420V, but I need to keep the screens considerably lower. They are being fed from a 5Y3, and since it is very lightly loaded (feeding only the screens and preamp) I needed to institute a voltage divider to feed the screens since I didn't want to stick huge series resistor inline between the first two filter caps feeding that portion of the amp. So, for the moment and using what I have on hand, the dc off the 5Y3 hits a cap, then a 2.2K resistor, then the screen cap with a 25K bleeder off that screen cap. I like to calculate bias at idle and then inject a signal and check to see what it does under load, and I typically calculate the current being used by the screens and preamp so I can get a very accurate bias measurement for the plates (cathode bias). However, in this case, I think I'm doing something wrong and getting funky results.

My question is, having never done anything with voltage dividers on the screens, how to I calculate the current that the screens are actually consuming? In other words, how do I know what is being dropped across the 2.2K resistor vs. how much is being bled to ground across the 25K resistor? My calculations using just the total drop in conjunction with the 2.2K value alone indicate much more screen current than I believe they actually are consuming, so something must be off here. However, I'm not sure how to interject the presence of the 25K bleeder into the calculations. I'm ignoring the preamp current for the moment for the sake of simplicity.

Anyone? Thanks!

[jazbo8]

First of all, how are you measuring the current? If it is just from subtracting the voltages across the 2.2K, then, that is the total current which includes the screen current plus the bleed current from 25k to ground, so you need to subtract the screen voltage/25k to get the screen current.

[EFK]

Yes, this is my issue - I'm not certain how I am to separate the screen current from the bleeder current. I measure 385 prior to the 2.2K, 335 at the screens after the 2.2K, but I'm not sure how I determine how much of that 50V drop is attributable to the 2.2K vs. how much is attributable to the 25K bleeder. Is there a formula for this? Thanks for responding.

edit: is it as simple as cross multiplying, i.e. 2.2K multiplied by total voltage drop, divided by total resistance (25K + 2.2K)? This may seem overly simplistic but using that method, I get a screen current at idle which is exactly what I would expect......

[jazbo8]

The current of the bleeder resistor is 335/25 = 13.4mA, the current through the 2.2k screen resistor = 50/2200 = 22.73mA, so Ig2 = 22.73-13.4 = 9.33mA. Which tube are you using? it seems a bit high for idle screen current.

[EFK]

Oooooooohhhhhhhhhhhh, I see. Thank you for that! See when I see it spelled out for me I realize how you are applying the math and it makes sense. And the 2.2K is shared for both screens, so that's 9.33/2 for around 4.6 per screen which is right at what I would expect compared to the usual version of the amp with lower plate voltage off the 5Y3. These are 6973 or 6CZ5 tubes, they idle in cathode bias @ 11-12W (@ 90 to 100%) and climb to around 16W dissipation under constant load at 1 kHz. It's beyond spec but they can take it in stride as long as you don't run the screens much beyond the 330V max spec. They're a lot like EL84 in that regard, they sound best run fairly hot and they seem to be able to take quite a bit more than the datasheet suggests.

[J M Fahey]

Remember screen current varies a lot and goes much higher than idle when overdriving the amp, so voltage dividers in practice drop quite a bit more than expected or measured at idle.

[EFK]

Thanks for mentioning that JMF - one of the things I'm trying to do now is see how far I can push the screens at idle vs. how much they drop when hit with my test signal and the amp is completely 'cranked,' and I'm comparing that drop to the screen drop in a standard version of the amp running everything off the 5Y3 (there's some noticeable sag there also). I'm not pushing things too far at the moment because I'm waiting on higher-rated dropping resistors. All I had on hand for the 25K was a 5W, which is pushing it pretty good at idle (its getting hot but so far so good) but I know I really should have a 10W to 12W in there to handle the amp being driven.