[hadron]

I stumbled across Mr. Gilmore's sketch of a dual tracking version of the Diden-Jung regulator recently [http://gilmore.chem.northwestern.edu/kgdidenjung.gif].

Adapting this circuit to supply the Dynahi amp seems doable. Biggest obstacle is the high voltage opamp, needs to handle at least +/-30 V supply, most resonable I've found so far is the LM13CL (LM143 is out of production, SG143 and HA-2640 are both pretty much OEM only). This is opamp is overkill, being rated for 80W, but the specs are good and given what's available rather cheap (comparable military/aerospace type amps are $100-300 each).

Anyway, a simple SPICE analysis seems to indicate that this circuit should work for a +/-30 V supply at 1 A. Though I did use ideal components.

D1-3 & D4-6 are workarounds for a bug in the B2 Spice diode model, combined they give a vj of ~2.1V.
A1&2 are 6.9V zeners.
A3 is a 15V zener, this is just is to keep the analysis simple, actual value will prolly differ, R7&8 will need to be adjusted when changed
X1&2 will most likely be LM13CL.

B2 SPICE circuit

So are there are major factors that I'm neglecting?

 

[Glassman]

nice, but I don't think it's really dual tracking, the opamp in the positive rail knows nothing about what's happening with negative rail.. and I think it's a serious overkill, we're talking about powering Class A amp, there are little to no load variations, so why not make it easier with highly optimised and proven standard Jung reg for positive and negative rails? you can select some nice opamps with ease, no need to hunt for some boutique amps with inferior performance..

just my $0.02, don't take me wrong, I'm glad someone finaly took the plunge

 

[hegestratos]

The layout you're suggesting copies errors and noise on the positive side directly into the negative side. On the other hand, for well-matched rails you only need to find two well-matched resistors R12 and R13. Very similar to the "official" psu for the dynalo.

I second Glassman's opinion. The choice of opamp is not trivial since it affects the stability of the psu. The AD825 seems to be what people are settling on.

If you want well-matched rails (I did on my dynalo) you can add a trimmer pot with the wiper tied to ground between R8 and it's counterpart in the negative half. That way the total voltage difference between the positive and negative side is fixed, but you can shift the ground up and down.

 

[kevin gilmore]

opa452 should work as the opamp. +/-40 volt rails.

opa541's work as the whole thing. 2 of them plus a 5 volt reference.
just like original power supply.

way busy, hope to get to this by the end of 2004.