Single-ended mosfet-tube follower

[KlausB]

Hi there,

although I browse a lot through this forum and learned quite a bit reading the posts, I never posted very much myself. However, I would like to ask for your opinion of my suggestion of a single-ended tube-Mosfet follower, before I think about transferring it from SPICE into the real world.

The idea is based on a 2007 publication in the tubecad journal and is based on three components: A triode, a power mosfet and a current source.

The current source is the load for the mosfet and isolates it from the unavoidable ripple of the high current power supply (an idea that Andrea Ciuffoli and Stephen Moore implemented). The voltage for the triode will be smoothed carefully through a C-L-C filter.

The ECC88 and the IRFP240 form a 100% feedback loop. The signal appears at the 1k anode resistor. The ECC88 plate impedance is 2.6k. Hence the impedance to drive the gate plus the Miller capacitance of the IRFP240 is around 720 ohms.

From the drain the signal is fed back into the ECC88 cathode plus it is fed into the output capacitor. The triode is supposed to control the whole circuitry. According to the SPICE simulation the amplifier should be able to deliver 20W into a 4 ohm speaker.


Critical aspects:

The drain follower topology may induce distortion

There is a feedback loop.

Driving the Mosfet with the rather high tube impedance may be critical

There is an electrolyte output capacitor


Any comment is greatly appreciated

Cheers

KlausB

[woody]

I see several problems. First the gate voltage needed for that
bottom mosfet will varry from device to device and also on how
hot it gets so it needs to be done with a pot.

Then the current source may have a problem. That 100 ohm
resistor from the 15v supply will let the transistor and diode
pull ~110 ma and I believe that diode is only rated at 100ma
change it to something like 5k. Or better still look at how
Nelson does these 2 things in his ZEN amplifier!

[Wavebourn]

Good idea, but why don't let the tube to provide some voltage gain instead of requiring higher swing from the preceding device?

[Lumba Ogir]

KlausB,
not bad, coming up with something like that right away.
What is the preceding stage?
I don`t find those critical aspects especially bothersome (OK, some elaboration would be needed here).
Crss=160pF
Ciss=1275pF

I would try to modify it in this manner:

[woody]

Lumba, in the circuite you posted you left out a voltage source
for the irf240 to set it's drain at 15v. Without this the device is
100% off.

But back to the first circuite what advantage does this have over
a standard mosfet power follower other than higher input impedience?

[Lumba Ogir]

woody,
you are entirely right, expectantly, IRF240 is now at least biased correctly.

Due to a tight collaboration, the tube linearizes the FET and puts its characteristic on the output stage.

[Wavebourn]

Originally posted by woody

But back to the first circuite what advantage does this have over
a standard mosfet power follower other than higher input impedience?

Now it is voltage controlled instead of charge controlled. Still, I would add a resistive divider in the feedback loop. Just my humble opinion.

[Lumba Ogir]

Wavebourn,

I would add a resistive divider in the feedback loop.

It can be done, however, in this mode, using 100% feedback, it is a voltage follower, just in case someone is of the opinion that output stages should not have voltage amplification.

[Wavebourn]

Originally posted by Lumba Ogir
Wavebourn,

It can be done, however, in this mode, using 100% feedback, it is a voltage follower, just in case someone is of the opinion that output stages should not have voltage amplification.

What happened Lumba? I thought an output stage should not have a feedback?

[Lumba Ogir]

Wavebourn,
I really like when you ask legitimate questions, I will answer later.