While I've got a -decent- general understanding of tube theory, I am struggling to wrap my head around dealing with direct heated pentodes.
Cliffs: I've got a couple of amps based around the Murder One which took a bit of work to get going - to cut a long story short, the bias circuit (using a LT1054) didn't work. LT1054 requires a capacitor from the negative output (pin 5) to ground (obviously upside down) to give an output voltage that should be roughly equivalent to supply minus a scooch, and my assumption is the 1044/7660 are the same. According to the original schematic it (apparently) works with no cap, and for some reason happens to inverted half of supply. How this is meant to work I've got no idea (maybe some quirk of current draw + it being floating?) but with a LT1054 it absolutely didn't work.
This modified schematic is what I came up which does work - it has the cap, meaning the negative supply is roughly -11.6v. It uses grid leak to draw current through R11, which drops the bias to roughly -6v quiescient - it works, but I've had the occasional LT1054 that hasn't played along and had heaps of output noise. Changing IC's fixed it. I don't want to be doing this all the time, so I'm looking at more...elegant ways to achieve the same thing. B+ is approx 80-85v. The amp does work and it does sound pretty decent for what it is, but I'm always looking at ways to improve.
I have a few questions, both for terms of knowledge and in terms of this exact situation:
- Does the 1.25v dropped over the heater factor into the bias (ie. if grid is at -6v and pin 3 is at 1.25v referenced to pin 5/ground, do you add these together for the effective bias?
- Instead of relying on grid leak, could I slap a zener or negative linear reg off the inverted output (assuming I set a little bit of parallel resistance to prevent the voltage from floating) and have regulated, fixed bias going to the grid?
- Failing all that, if I split R11 into two smaller resistors - one before the filament and one after, still giving me 1.25v over the filaments and the intended bias voltage over the second, does this work for cathode bias? If so, can I just add in Ck as normal to improve gain/tailor frequency response of the stage?
