bias: cathode current vs plate current

[johnnyreece]

I have been using precision 1R/1W resistors on the plates of my last couple builds. I like it and understand the voltage danger... and I build them for myself. I might even put some on the screens and the plates of my next build(?).

I hope this isn't hijacking too much, but here's a thread where I mentioned a safe plate measurement from test points. I personally hate taking out a chassis if I don't have to for simple things like tube changes. I also hate the thought of slipping with my meter probe and doing serious damage to myself or the amp. These test points, along with bias test points, tell me pretty much all I need to know for simple measurements from outside the chassis. And, the resistors really are pretty cheap. I'd venture to guess much cheaper than a 1% 1R/1W would be. And, a 450V plate voltage would be reduced to .45V.

https://ampgarage.com/forum/viewtopic.php?f=1&t=29592

[martin manning]

I would be mindful of the voltage rating of those resistors. They will see about 2x the B+. Perhaps the best solution would be two 500k in series instead of the 1M.

[pdf64]

I would be mindful of the voltage rating of those resistors. They will see about 2x the B+. Perhaps the best solution would be two 500k in series instead of the 1M.

Probably even more that that when overdriven into an inductive load

Pete, how do you set output tube bias? What are your criteria?

Haha, good question, I need to find a bias guru I can believe in
Till then, it's probably best not to agonise over it too much; if there's technical info regarding the amp's intended idle conditions, then it makes sense to follow that.
Modern Fender info, including RIs, tends to indicate ~30-35mA idle cathode current for 6L6 in fixed bias, eg https://el34world.com/charts/Schematics ... ematic.pdf https://el34world.com/charts/Schematics ... in_amp.pdf

I find that works out pretty well for most big octals in the fixed bias AB1 conditions common in regular guitar type amps.

The 70% guideline is helpful as an upper limit if the HT is high and 23 or 25W plate tubes are being used.
But, to expand on my thought process, it seems perverse to set it as a target, eg given 2 EL34 Marshalls with the regular 784-103 OTs, one with a 450V HT and the other a 400V HT, why idle those in the 400V amp at a higher plate/cathode current than those in the 450V amp; I just can't see the rationale behind that
Note that the Mullard suggested fixed bias operating condition of 3k5 p-p at 400V HT (I suspect Marshall used that as starting point for their EL34 switch) p2 of http://www.mif.pg.gda.pl/homepages/fran ... e/EL34.pdf idles at ~35mA with a THD of 1.6% @ 54W!

Of course hotter idle means higher stage gain, and sure, I agree that the slight gain boost will most always sounds better as the bias is tweaked to increase plate/cathode idle current. But that doesn't mean that when normalised, the actual tone/response is better. Rather the likelihood is that trying to run a fixed / mixed bias amp at idle plate/cathode currents much higher than the above will compromise the amp's dynamic response and tube life, and at the extremes (eg 6550A idling at 70% in a regular 50W Marshall, 6L6GC idling at 70% in a 5E6), will pull down the HT nodes and reduce the max continuous power output as well.

Anyway, after setting idle plate / cathode current to ~30-35mA, disable the NFB loop if fitted and run at about quarter to half power into a suitable resistive load and monitor the output waveform; with reasonably matched tubes, 0 degree crossover distortion shouldn't be apparent, though the 'gm doubling' kink at the AB transition may well be evident if the signal level is too high (if so, turn the level down).
The amp should be unconditionally stable over a range of test levels and frequencies, and the waveform at the onset of clipping should look reasonable, though the 'gm doubling kink and perhaps a bit of bias shift may cause perhaps as much as 20% THD (compared to a sine wave at the power amp input).
If applicable, re-enable the NFB and recheck for stability; the high power waveform should look a lot cleaner, though still maybe 10% THD.
Recheck stability into an inductive load and if ok, sound test; my FluxTone speaker is a massive benefit with this

For fixed bias EL84 / 6V6 in fixed bias AB1, 15-20mA idle cathode current seems to work well. With higher HT's and lower impedance loads (eg 6k6 or even 4k) often used, higher idle currents can be beneficial in keeping the 'gm doubling' kink from thinning out the high power waveform excessively, so the 70% idle guideline (which I think is best used as an upper limit) comes in useful to ensure that plate dissipation isn't going to get pushed too hard.

Hope that makes sense
BTW, I credit / blame Andy Le Blanc for setting me on the path to the above, as he questioned the validity of using a 70% target a few years ago, putting forward a suggested guideline of 'idle (fixed bias AB1) such that HT winding current is at 50% of its rated max'. That seemed a good idea to me, much more sensible to view the amp holistically, as a system in which power supply loading and the power amp / power supply interaction is significant, but with the problem of chasing down the (possibly non-existant) PT spec.