Overcoming Miller

[dsavitsk]

What is the formula for figuring the 3db point due to Miller capacitance and the preceding stage's output impedance? And, what is a reasonable HF roll off point? Morgan Jones (pg 84) seems to jump over these.

[SY]

Total input capacitance is (assuming no cathode degeneration) Ctot = Cgk + ACgp, where A is the stage gain. The f3 would then be 1/2piZCtot, where Z is the source impedance of the driving stage.

If there's cathode degeneration, the Cgk term will be reduced, but for most triodes, the ACgp term dominates.

[dsavitsk]

Oh, so low pass formula is the same as high pass?

And, what value am I shooting for? Jones seems alarmed that his 12ax7 -> 12ax7 example rolls off at 29kHz.

[SY]

I think it depends on the application. Inside of feedback loops, rolloffs like that can be real trouble. And many people believe that bandwidth ought to be extended to some multiple of the highest perceivable frequency.