Simple supertriode stage. Cathode AC voltage.

[Elerion]

Hello!

I'm playing with the idea of doing some supertriode experiments. After reading the couple of good documents from the Japanese sources, and those of Bronski, I decided to go with the CCS approach.

The simulation works. Using a single 12AU7 (just 3mA Iq) and a pentode (very relaxed EL34 in this example), the stage can deliver clean 2 W into 2 kohm load at 250 V, with an amplification factor close to 15, which is on the order to be expected for a 12AU7.

What I don't know, is how to make the triode cathode follow the grid more closely. Attached, the input is 10V p-p, the cathode signal is half that value. The signal gets amplified 15 times, from this halved cathode voltage, obviously, which is what the EL34 grid sees.

Does anyone know how could I improve this?

 

[DF96]

To get the cathode to follow the grid more closely you need to swap the 12AU7 for a valve with higher mu. Then you get less feedback, which seems to me to rather defeat the point of the circuit.

 

[Elerion]

Yes. 12AX7 gets a higher AC cathode voltage, so do 12AT7 and 6DJ8. But I get little or more voltage gain from higher mu triodes, I think, due to the pentode, which has limited gain (switching to an EL84 improves a little bit, but I still get a max gain for the pentode stage itself of around 18).

If we use a nmos instead, we can get a gain close to 30 for a 6DJ8 and 70-80 for a 12AX7. Much closer to mu.

Attached, a 6DJ8 comparison.
- Using an EL34, I get Av=9.57, and a voltage gain Vo/Vk-triode (ac) = 14.8
- Using some "random" nmos (the model used is not suitable for HV), I get Av=29.2 (close to triode's mu) and Vo/Vk-triode (ac) = +1300, due, I think, to the much greater transconductance.