Simplistic NJFET RIAA

[Salas]

What do you mean 317,337? You use them as a bench supply? 28.1 to 31.2 is not enough for the drop across the leds. If the buffer is working ok, then the Vout adjusts as without it. Else its a sign of oscillation. Presents a comb across bandwidth on FFT. I suggest you feed it from raw rectification and capacitor at 35V. Maybe the prereg is the problem. The bandwidth alters with the buffer. You did not tell us you got more electronics in the loop before.

[disco]

I experienced the same (oscillation problem) with a HV serial reg feeding a HV shunt reg. Would limiting the bandwith of the serial reg cure oscillation of the shunt reg?

[Salas]

Somewhere they turn the phase beyond 180 deg in combination. You got to know specifics.

[RCruz]

I did not have anything but raw 34v in the first experiments.

My last post refers to this unique ocasion where I tried it on the bench... regulated +-15v and it was able to put out 28v from 31.

Unfortunatelly the issue persists.... I just installed it in the riaa and the output droped to 27.5v... the worst thing is that the whystle appeared again.

In the other channel I have a "normal" shunt and the input is 34.2v for 28.1v out.

Do I need more than 34v for this particular vout ?

[Salas]

You need Vout + at least as much as your 3 leds drop.

[ikoflexer]

I assume you didn't try base resistors as I suggested.

Can you please measure the current going through the shunt mosfet? There's a chance too much current sags the input voltage and that can also lead to oscillation. You may want to decrease R1 a little for testing purposes.

Edit: I just noticed your Vin. I think it's too low, you need a better margin of safety there.

[RCruz]

I have 28.1vout + 5.7ledvdrop = 33.8 so 34.1 should do it.... At least the "normal" shunts work fine.

My unregulated bench psu outputs 50v... can I use it to test ?

[Salas]

Take R1 up to 15 Ohm and if your CCS Mosfet sink can take 3.3W do it.

[RCruz]

Originally Posted by ikoflexer

I assume you didn't try base resistors as I suggested.

Can you please measure the current going through the shunt mosfet? There's a chance too much current sags the input voltage and that can also lead to oscillation. You may want to decrease R1 a little for testing purposes.

Edit: I just noticed your Vin. I think it's too low, you need a better margin of safety there.

I did not have the oportunity to try the base resistor change.

The current is 190mA as I read 1.9v on R1 (10r).

If I decrease R1 can it not increase the current ?

[ikoflexer]

Originally Posted by RCruz

I did not have the oportunity to try the base resistor change.

The current is 190mA as I read 1.9v on R1 (10r).

If I decrease R1 can it not increase the current ?

Wait, don't decrease it yet. How much does your load draw? I just want to make sure your shunt mosfet has enough current to work with. You can use 38V from the supply, ONLY if your CCS mosfet has enough sink to dissipate 1A.