previous: lion Wait a minute RG – let me get this ... -- 1/10/2000 9:14 PM

Re: Cathode follower drive to output tubes

1/11/2000 4:04 AM

R.G.

Re: Cathode follower drive to output tubes quote: "Does it mean that I can use an IRF820 – no, two IRF820’s it must be – hooked up with the drain’s to 450V B+ and –50V at the cathode’s"   Yep, kind of. You can think of a high voltage power MOSFET as a pentode that has a 3-10V positive bias on the "grid" and a big "grid"-"cathode" capacitance. In cathode followers, the local feedback cures the problems the big grid-cathode capacitance would otherwise cause you. I've used the 820 in preamp followers several times.     I didn't know your B+ was 450V. The IRF820 is rated 500V, so it gets kind of marginal. Works great in 300-400V preamp circuits. You'd have to either use a dropping resistor/capacitor or get a power MOSFET rated for a higher voltage. They're available up to about 1kV. The IRF820 is probably marginal in this application if you don't do some dropping, just B+ is at its breakdown voltage, but another type would work. Look for maybe 700V and 2A.     Given that you don't exceed the drain-to-source voltage rating and don't overheat them with too much current, yes, you just plug them in. In this application, you would do the cathode resistors either from the resistance you need to bias the output tubes right or if that didn't limit you, you'd size the source resistors to keep the MOSFET dissipation down, with all that B+ above it.     If you sub them for tubes, you may have to take into account the difference between the few volts negative bias a tube needs and the few volts positive that a MOSFET needs. That might be critical if you're direct coupling the sources to the grids of your output tubes and need to get the bias DC voltage right.     In circuits where a cathode follower is direct coupled to a plate, you just pop the MOSFET in and it works. The cathode rides a few volts lower, no big deal in that circuit.

 

Replies:

lion Re: Ah! I wish I wasn’t so ignorant! -- 1/12/2000 11:10 AM