HAFLER OUTPUT STAGE

[Stee]

this is good for hybrid application

[djk]

Only if there is a feedback loop.

[Stee]

like this?

[djk]

Sort of.

Your method would give a very low input impedance and has room for improvement.

Some thoughts, in no particular order:

You probably don't need the diodes in the bias circuit. especially with lateral FETs and high bias.

You don't need a high voltage front-end, full output will only require about ±5V drive (the grounded output stage has voltage gain).

You probably only need a one-stage voltage gain from the tube section, the outputs have gain. This would allow a non-inverting feedback scheme to the cathode of the input tube.

You should probably have protection diodes on the gates, a 10V zener and a 1N4148.


You should probably have protection diodes on the gates, a 10V zener and a 1N4148 for each output.

You might consider a single tube section with an inter-stage transformer, or a single tube section with a cathode follower and a current source.

You only need about 8mA per output pair for good square-wave performance, less if you just need good sine-wave performance to 80Khz or so.

[djk]

This forum won't save my edits.

You should probably have protection diode on the gates, a 10V zener and a 1N4148.

Don't forget the gate resistor, probably 470R.

One of the current source resistors for the gates needs to be adjustable to null DC off-set.

The grounded output stage works best with a EI type transformer with side-by-side windings. A toroid should have a screen between the primary and the secondary.

A note on the secondaries. If you don't have a screen between the windings, expect a little crosstalk, or go dual mono supplies. A common supply for the front-end is fine.

If you choose to use a non-center tapped secondary, use a voltage divider to establish a DC ground point (see a QSC schematic).

[juma]

Originally Posted by djk

...especially with lateral FETs and high bias...

2sk1530/2sj201 are not laterals. They have positive tempco and some kind of temperature compensation is needed.

[Stee]

the question is
it's possible have negative feedback from bias resistor?
the partition from 17 V power supply is auround 1/10

[Stee]

like this

[djk]

On an inverting summing node the impedance is effectively 0 ohms, so you have an input impedance of 360ohms.

Can your tube stage drive 360 ohms?

(there is another issue with your last schematic but it takes too long to explain and since it won't work right anyway I am not going to discuss this other issue)

juma is correct, if they are not true lateral FETs you will need bias compensation. Although if you bias into class A this may not be an issue (see various N. Pass designs).

[Stee]

important is ratio
probably need 33K ON 3K3

SAME FOR JFET I THINK
K170 AND J74