Schade84

[revintage]

Instead of foolishly using the triode drivers Ri for feedback(high driver THD), this must be a better way to do it. Used ca 10% local NFB as proposed by Schade. Surely the circuit can be fine-tuned. I get indications of below 1% THD both in the driver and total circuit at 4W/8ohm.

[revintage]

Comparision and two possible modifications to make the first circuit acceptable with respect to THD and clipping.

[revintage]

Had made the last two circuits too complicated. Removed the cap in the feedback.

[jerluwoo]

I bought a Motorola phono console a while back that used this feedback in its EL84 se output stage. It was setup like circuit 3 above. It sounded horrible.

[revintage]

I am impressed Motorola used 6C45 ! So you probably didn´t look close enough!

The interesting thing is that when you go from circuit 1 to 2, distortion is cut almost in half when the local NFB ala Schade is adjusted to give the same total gain as in circuit 1.

Forgot to add the simplified circuit from post #3:

[Bandersnatch]

The change from 1 to 2 is simply adjusting the source impedance of the driver. Swap a pentode in there, or perhaps an ECC81 like the RH84 and the impedance is dominated by the plate impedance of the driver again.
cheers,
Douglas

[revintage]

Hi Douglas,

Mentioned the Ri in post #1.

Maybe you missed the point. Distortion is halved from 1 to 2.

The method used in 1 only works with pentodes.

Look at the point where the feedback resistor returns to the EL84 grid. If we had infinite gain from the pentode this would be a virtual earth point. Due to finite gain we just have a very low impedance. A triode driver will have gross problems.

With a shunt/Schade inverse-feedback as in 2 and 3 we a have a definded input impedance of in, this case, ca 47k. Now we can find a triode driver that works OK!

[Svein_B]

An interesting variation of the shunt feedback Lars.

Inserting a serias resistor before the EL84 will make the load on the driver more "benign". The voltage swing required will however increase by a factor of about 4 due to the voltage divider represented by the 47K and Rfb (Rfb being in effect about 470K/33).

So it seems that whatever way we do it, there is a very delicate balance between feedback ratio, gain, and distortion.
What do you recon is the ovarall gain of the modified topology, and what would be the (simulated) output impedance?


Atached loadlines is my rough estimate of the two variants using 12AT7/ECC81.

Svein

[revintage]

I prefer to call it "Schade" as this variation of shunt feedback around a output tetrode was proposed by him.

I choose the shunt feedback to give exactly the same total gain and output impedance is on par. THD is halfed in the Schade.

The variation recommended by you and others, is IMHO no good enough and with about double the THD compared to the Schade. A triode driver does not work so good under those circumstances.

Started the thread as I couldn´t see the benefit of the earlier variation when using triode drivers.

[revintage]

Normalized the three variations with identical workingpoints and identical gain. Ua of drivers just above 150V on all three.