It’s been a while since I posted an update on this project and figured it was time for another one. Not a lot has happened on it over the last year (other priorities), but I did get a prototype of the pre-regulator designed, built, and tested. Overall, it worked well with just one “minor” bug (otherwise known as stupid designer error).
The schematic mostly follows the original design, except for substituting some of the old and obsolete devices for more modern ones. In addition, surface mount components are used extensively (the original design was all through-hole) and the +/-12V auxiliary power supply has been replaced with an integrated RAC04-12DC module from RECOM.
The PCB design:
A prototype was built this last summer and tested. It assembled without incident and hopes were high. <dramatic pause> Until it was turned on for the first time. No, there were no spectacular fireworks or scorched engineers. It just didn’t work.
Turns out triacs are not as bi-directional as I thought. Apparently I didn’t pay close enough attention to which terminal was MT1 and which terminal was MT2. It’s an AC circuit, right? Well, in fact, both were backwards in the design and this matters. Oops.
After a little PCB hackery…
That above is the circuit outputting 31.8V, via a step-down transformer + bridge rectifier + giant cap. The Lambda PSU in the upper right is providing a 5.1V bias to the optocoupler input. It’s not truly in regulation since the loop from the output to the optocoupler is not closed, but the principle has been demonstrated.
Although it works, air-wired line voltage is probably not a smart idea to keep around. My next step will be to fix this little pin mixup and generate a new, clean board. Then the fun begins by closing the loop with an actual regulator.