Petter help

[BruceM]

Thanks for the great info BobG.

I learned a time consuming lesson about bad PF loads and my simple AVR.  My original AC battery bank charger was a bad PF 1500 watt PF load, and on my homebrew AVR the voltage regulation dropped way LOW, low enough that 240V relays were chattering.

My AVR allows manual switchover to harmonic  operation, and I found  that the harmonic setup worked OK on the bad PF load.  So I think you are right on the mark about the current transformer regultion scheme working well for bad PF loads.

Rather than add a microprocessor to my AVR design (PITA and too much software development time) to do real time RMS voltage calculation, I then modified my prototype to add an (expensive)  analog RMS voltage measurement chip.  It worked great but the response time was too slow to do Listerflicker correction and I couldn't stand my shop heat lamps without that.  So I finally abandoned the analog RMS chip, went back to my simple original voltage averaging circuit and then did a major redesign of my 120VDC battery charger to use a motor driven variable transformer instead of a dynamic capacitor switching scheme for current regulation. A lot of work but now my shop heat lamps/lights don't flicker and there is decent voltage regulation.


I don't know if any of the cheap AVRs are true RMS voltage, but it would be something to look for if you were planning on using something with a big switching supply like a battery charger and were shopping for an AVR.  For a production AVR of modern design, I would expect that adding a microcontroller to compute RMS voltage every half cycle and then adjust the excitation timing accordingly would be an obvious solution and fairly cost effective, given the current cost of the microcontroller chip at under $3. But for the bargain AVR's I just don't know how they are commonly designed.