Paul and Sabrina's Cheap 3 Phase Inverter (AC Controller) with Field Oriented Control

[cts_casemod]

My IGBT's can handle 60KHz. Model FF600R12ME4, but that would require some beefy gate drivers pumping out 20-30 Amps peak current.

ZVS would avoid the need for the SiC rectifier altogether and would allow a small (inductance) inductor to be used. After all we are looking for discontinuous.

I have some literature for fly-back ZVS and thyristor forced commutation that I may look at. These would be a good starting point. One just needs to calculate the LC resonant frequency.

[P-hack]

It is definately interesting, been looking at the UC3855BDW datasheet and example circuit, though I don't know how they found their way to a APT5020BN or mtp8n50e mosfet, makes me thing mouser/digikey might be inadequate

Thing is, with that spreadsheet and cores I had to tone down the frequency from 100khz to 50khz to keep the temperature rise within limits (AC losses went up unless you lowered the inductance value from the already low 17uf, it appeared to stay in continuous mode at that power level though, ~%30 ripple). Is discontinuous mode a requirement for ZVS (would kinda make sense)? edit, or something like an old-school scr shutoff?

Also, as you noted, that example is PFC, so a relatively constant supply *might* not need be as complicated a circuit.


@paul, the ch-200 appears to stay continuous down to about 2khz, but major current ripple. 10k is about the min.

[P-hack]

I can *almost* follow this paper talking about passive snubbers: http://www.ee.bgu.ac.il/~pel/pdf-files/smnr13.pdf plus some ltspice nomenclature included.

edit: this does not lend itself to a half bridge igbt boost unfortunately (assuming using top body diode). Not sure what bi-directional boost/buck would look like either, it might snap together though back into a half bridge with additional diodes.

[cts_casemod]

One could use a buck converter rather than a boost if it turns out to be more convenient.

[P-hack]

Originally Posted by cts_casemod

One could use a buck converter rather than a boost if it turns out to be more convenient.

Hmm, thought we were trying to overclock some 400+rms motors from a 144v battery, not sure I follow your thinking here. Though it is worth noting that you only "need" this on half your switches if you fulltime svpwm. Could be top or bottom, whatever makes more sense.

[P-hack]

hmm, maybe, 2 stage, 2x the voltage and 1/2 the current each stage. 144*2*2 is almost enough for a svpwm switched 480v motor even (468rms). Though each stage has to handle full power, smaller voltage steps seem to be less stressful.

[cts_casemod]

Originally Posted by P-hack

hmm, maybe, 2 stage, 2x the voltage and 1/2 the current each stage. 144*2*2 is just about right for a svpwm switched 460v motor even. Though each stage has to handle full power, smaller voltage steps seem to be less stressful.

Just to double. Otherwise the boost would be more advantageous.

[P-hack]

the inverting buck/boost can't provide direct power either, everything would have to run through it. Though I wonder if 2 stages could "fix" the polarity

gonna want more than 2x for serious overclocking of a 200ish motor though, otherwise you are back to giant motors.

[cts_casemod]

My suggestion is using just a single buck or a single boost, with the obvious limitations that the buck can only achieve 2*Vin.

For 144V, where one would need voltages higher than 300V boost would be a better topology, but still a viable choice for 300V inputs. Again, the inductor only handles half of the output (the negative mirrored portion).

Here is a paper describing such method
http://www.ti.com/lit/an/slyt286/slyt286.pdf

[P-hack]

I like the idea of overrevving 215/240v motors by at least 2x since a 25hp weighs a lot less and costs less than a rated 50 or a 100hp job. Though rewinding/reconfiguring is probably an option not to be overlooked if you want to keep 144v.