Yep the servo approach allows you to use bigger rg. If a voltage develops across rg then the servo will steer to correct this down to the negative supply. So you are better protected from thermal runaway. 6550A says 47k KT88 says 220k. I would not use 220K with a KT88. I think most damage is exceeding screen grid ratings. The screen melts and bad things then happen.
Not sure about easily. I was driving 2 6550A on each leg so 24k. The servo allows 100k rg = 50k so half the current in the drivers. It meant an e88cc cathode follower was used. With 24k I think an ecc99 is needed. You could even use the same rg as the cathode bias as long as you have enough negative supply. There is no spec for the maximum g1 leakage current with would be more useful.
Last edited:
The led does play a role biasing the bottom transistor as a CCS. As far as seeing what's going on you can take the output of each bias circuit (at the transistor output not after the 100k to the grid) to a rotary switch and to a good old-fashioned voltmeter. If you want to see the operation pop it into a simulator.
Okay, I'll leave this leds then.
Btw, 300uA is too low to light it up. And 1,8V is red/yellow led and blue leds are used in AB-Q.
Btw, 300uA is too low to light it up. And 1,8V is red/yellow led and blue leds are used in AB-Q.