GC with LM3875 as a voltage regulator also

my bet would be the protection circuity in the lm3875 would prevent it from being an effective regulator.

without the current limit resistor i think it still will limit current, just at a different value. pretty sure it lists curves or values for such in the datasheet.

why don't you try the lm338 regulator - just like pedja - you might like it......or the lt1083.

GregGC said:

Hey Pedja, maybe you can shed some light here.

Click to expand...

Hello Greg,

Actually I don’t feel so, but as I am named… I have some thoughts.

If you settle on the feedbacked regulator it is easier and safer to use the LM338 (or more expensive 1083/1084) which is the device designed for such purpose. Anyway, the usage of the LM3875 and similar not unity gain stable chips is not good idea since the voltage reference (and its noise) must be amplified (10 times minimum). The OPA541 (or something similar) looks more appropriate but probably can’t drive capacitive load directly. Some series resistive decoupling at the output still might be a way to go (BB offers spice models and some short sims I ran earlier for the OPA549 seemed promising). Using unity gain chip and low noise voltage reference you can probably achieve better noise figure than with the integrated regulators but neither I am sure that will bring much of anything nor I could see some other advantage.

Pedja