/diy/ - Do It Yourself

>>1225121
One of the others could also be blown or degraded. Can you check the voltages across the emitter and base on all of them?

>>1225135
Don't apologize, I've been missing the high-brow conversation. What's with all the fish, then?
>There's nothing wrong with this idea and it's probably more efficient anyways, right?
Indeed, you could do just that. The lamp itself might shift color slightly at different I(f) but nothing as bad as an incandescent.
>What's the smallest voltage division that conventional power supplies can output anyways?
1.25V is a fairly common but not universal reference voltage in power supplies and one generally can't go below that reference voltage. Pic related in >>1222896 takes the clever tack of supplying a negative voltage source from a large-ish negative supply using two diode drops to ground as regulators, thus allowing an LM317 a continuous output range that includes zero volts. It looks like you could use the same sort of trickery with the 2596.
You probably could make a high-current power supply using the 2596 as a controller, but it wouldn't be the optimal device. You would still need an external current limiting arrangement similar to the one on that board. dc-dc controllers designed for that purpose include such things as FET drivers with bootstrapping (higher Vgs for quicker rise time and less FET dissipation), current-mode control, soft-start, synchronous conversion (replaces Schottky diode with another, controlled FET for super efficiency), and usually much higher switching frequencies. The TPS40304A, for example, has a Vout(min) of 0.6V, a 25A maximum design output, a 600kHz switching frequency, and synchronous operation, if you can work with a fine-pitch VSON package. The LM27402 has similar specs with a 30A Iout(max) and high-side current sensing, in a TSSOP package which might be easier to reliably reflow.

>>1225137
A buddy of mine used an old waffle iron as a dummy load.