I've been working on a full-on adjustable power supply in SPICE and thought I'd share what I have before I start building it.


Notes:
- I put R7 in to limit the inrush, I think with an NTC thermistor and an additional 2.2mF capacitor I can get better results out of the real thing. The transformer I'm using is a 115 to 24V with a listed Imax of 1.5A and 36VA power
- R6 is meant to be a potentiometer
- C5, R3, and R4 provide a delayed startup for the smoothing capacitors to charge up

Shown below is the Vin and R2 voltages, the oscillation is the current through R7. I tried to pick good models from the ltspice library to limit ideal components and out of several permutations this has given the best result!

(I'll share the full simulation when it finishes if anyone is interested)

 

I put R7 in to limit the inrush

R7 limits how much power you can get out. You can start out with 24V on C1, that will help inrush current. "initial condition"

 

R7 limits how much power you can get out. You can start out with 24V on C1, that will help inrush current. "initial condition"

That's why I was thinking of using NTC Thermistors on the real circuit to reduce losses. Is it ok to drop 24 VAC across higher than 2mF capacitors with no current protection?

I'm also worried about the 1.5 amp specification on the transformer I'm using.

 

R7 limits how much power you can get out. You can start out with 24V on C1, that will help inrush current. "initial condition"

I removed the resistor and set an IC on the input of the LM2596 to 33V. The amperage spikes still hit 3A, although I'm sure the rms value is much lower... I'm still unsure if this is a safe way to build my power supply.

 

I'm still unsure if this is a safe way to build my power supply.

Look for schematics of power supplies. Every one has large inrush current. The resistance in the transformer helps limit the current. Look at the data sheet for the diodes. They are good for much current for one cycle.