The simplistic Salas low voltage shunt regulator

[Salas]

Originally Posted by klewis

If I reduce the unloaded voltage by .01v, it halves the result.

324mOhm with DVM readings. Tiny things show it enough times up or down easily. Would it drop double 0.02V if it is a 150R for instance? Was the Vref well steady? Just keep it held well. Was there remote sensing to the 330R by the way?

[klewis]

Originally Posted by Salas

324mOhm with DVM readings. Tiny things show it enough times up or down easily. Would it drop double 0.02V if it is a 150R for instance? Was the Vref well steady? Just keep it held well. Was there remote sensing to the 330R by the way?

Salas,

I will try with 200r and 150r, I will also try with double voltage. The Vref was steady, used aligator clips to hold the test leads. Remote sensing was used.

[Salas]

Use less than 5k6 5k6 in the cascode when you need 20V out by the way. That 5k6 value comes from the 46V example. Use 2k2, 2k2.

[klewis]

Salas,

For my application, I would like to get about 60v out of the regulator, maybe I'm stretching it a bit too much, but, thought I would give it a try. What's the worst that can happen, smoke?

I tried jumping the voltage, but, it's a bit high going in at approx 70v. I looked at the simulation and saw that Q3 would be over it's ratings for Vcbo and Vceo, so I put a 2sc3503E in it's place. This is for the positive regulator. The regulator fires up and runs, though the voltage creeps up 10mv every two or three seconds. I've set it at 60v out and put 1k load resistor on it. The load resistor is a 5 watt resistor and gets pretty hot. As I write this, I'm starting to wonder if it's oscillating...

Do I need to put 10k ish resistors in at R14 and R4? Other tweaks or thoughts?

[Salas]

You run a Q3 with different Ft and Cob now. Compensation could be struggling. R4, R14 6K8. Check with scope. If it still gives a DC line on AC mode it maybe ain't acting up. Could be the new BJT's Vbe, Ib, and new point for the Vref CCS jumping up as a system.

[klewis]

Originally Posted by Salas

You run a Q3 with different Ft and Cob now. Compensation could be struggling. R4, R14 6K8. Check with scope. If it still gives a DC line on AC mode it maybe ain't acting up. Could be the new BJT's Vbe, Ib, and new point for the Vref CCS jumping up as a system.

Salas,
I'm actually using the Fairchild version of the 2sc3503 (KSC3503) the Ft are the same - 150Mhz, but the ft for the 3503 is measured at 30v vs. 5v for the 547 the Cob are very close 1pf for the 547 and 2pf for the 3503. Thanks for the values for R4&R14.

[Salas]

There will be ways to tune the step network or the termination if you will have problems with those voltages and substitutions I believe.

[pidesd]

for salas

one of my finest v1.2 build

[Salas]

True hard rock DIY.

[Salas]

I evaluated some easy changes on V1.2 and I think they come across subjectively in a positive manner. At least on my 61dB Low MC phono. Pretty stable on the scope also. So I am calling the usual suspects with builds in hand to evaluate and confirm or not on their client applications. Lets call this the V1.2R(revisited) beta test. The changes are two. First the 33p-1k step circuit is deleted and the BJT error amp's base resistor is made to 1k. Second, the output network across the force lines becomes a combination of one electrolytic, a resistor and a film capacitor. Its not crucial to change the value of the Vref film capacitor or the value of the output film capacitor in existing builds for testing. Only to add the 47uF and change the Zobel's resistor to 0.22R. I attach an example. If you fancy beta testing, let me know your feedback.