Schaded F6 P-N cascode with very low THD

As a fun exercise to train my newfound LT-Spice skills I set about to create a very low intrinsic distortion Schade feedback F6 output stage. There is no transformer or global negative feedback trough the transformer to help with the distortion in this sim, as in a real F6.
I just wanted to test a simple P and N compounded device Schade feedback F6 outputstage.

R13 and R14 should be 1 kohm multi-turn pots to adjust bias and DC offset.

I have no idea if there is any merit to this overly complex variation of the F6.
I will probably never build it, even though I am tempted. I have some Cree SiC fets lying around..




0.004243% THD at 3,31 volts p-p. No global feedback. A nice balance between the second and third harmonic. I purposefully warped the balance slightly to increase the second harmonic slightly.

V3 and V4 is the secondaries from the Jensen transformer. There is no need for the standard F6 bias circuit with led diodes or zener diodes and some trim pots. The input is DC coupled to the transformer secondaries.

I might just need to order some Jensen transformers...... :)

Last edited:


At 33 volts output over an 8 ohm load and 22 + 22 volts supply, there is a distinct knee in the distortion as the amp starts to clip.
Still only 0,265% THD, even though there is a lot of higher order harmonics in the distortion spectrum. Not bad from such a simple amp with only local Schade feedback around each compounded gain device.



I could not restrain myself.
Even this idea got a dose of IXYS depletion mode fets, greatly simplifying bias, DC offset etc. Very nice stable DC-offset and quiescent current
The bias and DC offset is adjusted with the value of R1 and R2 and/or R7 and R8.
R7 and R8 sets the amount of Schade feedback, and if this amp is to be tuned for slightly higher second harmonic this is where I would adjust it (tilting the balance of the amp by a small asymmetrical Schade feedback).

V5 and V6 is the secondaries of the transformer and I would use the same negative feedback through the transformer as the normal F6.

I don't have any depletion mode fets or Jensen transformers at the moment. I have just bought a handful of IXTK 100N25P (similar to the beefy transistors Nelson Pass uses in the BAF2015 amp but in a TO-264 package) and Cree SiC fets so I will build other amps first.

I am tempted though. I like the results in LTSpice. I know a simulation does not say much of the soundquality, but I know from previous experience that the P+N source to source cascode has the potential to sound very very nice. It is fast and intrinsically linear.

I just post these ideas to share my ideas and stimulate experiments and discussions.