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

[P-hack]

Originally Posted by MPaulHolmes

All the AC (ok not all!) current is between the capacitor and the IGBTs. Doesn't that mean the skin effect isn't very applicable?

40% current ripple in an inductor isn't unusual, so I would rule of thumb the skin effect is %40 as much of a factor (I don't really know, just guessing). But twisting smaller magnet wire into larger bundles isn't a deal breaker (though how it effects distributed capacitance I haven't figured out yet, it is a function of voltage differential though).

This shows the skin depth vs diameter
American Wire Gauge table and AWG Electrical Current Load Limits with skin depth frequencies and wire breaking strength

so you could theoretically twist together as many 26 gauge strands as you needed at 100khz, or as many 16 gauge wires as you needed at 10khz.

[cts_casemod]

Originally Posted by MPaulHolmes

I had written to XFMRS, and they came back with a design suggestion, along with a .xmls file with all their design steps. I asked about the skin effect when I saw how large of diameter cable it was that they were suggesting, and they said the stranding "helps to mitigate" the effect. Wait! But there isn't (much) AC current in the strands! All the AC (ok not all!) current is between the capacitor and the IGBTs. Doesn't that mean the skin effect isn't very applicable?

Are we talking about the inductor core cable or something else?

All the current on an inductor is AC, for discontinuous boost converters.
For continuous conduction only the ripple portion is AC

[MPaulHolmes]

They came back with a design for an inductor (core, number of windings, type of winding) design based on 10KHz, 600amp continuous. They also ended up with 30% current ripple rate.

[cts_casemod]

Originally Posted by MPaulHolmes

They came back with a design for an inductor (core, number of windings, type of winding) design based on 10KHz, 600amp continuous. They also ended up with 30% current ripple rate.

Thats 120Amps ripple.

Two questions come to mind:

Do we really need 600Amps? What is the input and output current and duty cycles?

Could this be reduced by increasing the switching frequency to 15-20KHz? (will their core support?) This is desirable to reduce the output ripple and most IGBT's are fine with it. It also means the inductor ripple goes way down.

[P-hack]

did you specify input and output voltage? 144 to 340ish? You can probably get away with 290v if you have svpwm I think, unless you want to extend the nominal speed.

edit, plugged in 144 and 290v here @ 600A 30% ripple, 5% voltage ripple:
Switching Converter Power Supplies

Item	Value	Units
Volts In	144	V
Volts Out	290	V
Load Current	600	A
Freq.	10	KHz
Vripple	14.5	V
Duty Cycle	50.638591646531	%
Ipp Inductor	180	A
Ipk Inductor	690	A
Irms	512.64022471905	A
L	40.229547808077	uH
C	2095.3899991668	uF

changed to 20khz:

Item	Value	Units
Volts In	144	V
Volts Out	290	V
Load Current	600	A
Freq.	20	KHz
Vripple	14.5	V
Duty Cycle	50.638591646531	%
Ipp Inductor	180	A
Ipk Inductor	690	A
Irms	512.64022471905	A
L	20.114773904039	uH
C	1047.6949995834	uF

[MPaulHolmes]

They had just gotten a little overzealous. I said maybe 400amp intermittent, with 200amp continuous. Then they asked "what is the saturation current", to which I replied that the peak current would never ever go over 600amp, and would trip a hardware overcurrent circuit if it attempted to do so. So, they made it for 600amp continuous. haha. I just mentioned that I wanted to double the voltage. Specifically, I listed 144 to 288 and 300 to 600.

[cts_casemod]

Ideally if you plan to use an inductor, keep the output regulated.

A 144V pack can sag to very near 100V on a cold day, so keeping the output constant will help with the control loop. Or even better. Keep it off until the V/Hz ratio is reached at 25Hz or similar and then start boosting accordingly.

P-hack, what happens if you keep the same inductance value for both cases?

[P-hack]

Originally Posted by MPaulHolmes

and 300 to 600.

Interesting, @ 300-600v @600A %30 ripple, 10khz I get 83uh (which is also given by the mhw NI curves and #turns, only at ~152A. 600 amps at 25 turns is well off the charts (and inductance goes towards zero). Do you have a datasheet on that core?

[P-hack]

Originally Posted by cts_casemod

P-hack, what happens if you keep the same inductance value for both cases?

re: is 600 amps needed, yes, because it behaves much like a transformer, to get 288vdc @ 300 amps out, you need ~144vdc @ 600 amps in, but I think we are talking more like three phase 208/230v motors, are there lots of those with 100hp capability waiting to be used that don't weigh 2000 lbs? You'd have to overboost a smaller motor probably, and that means even more volts (or a rewind). If you want to run it at dataplate though, it should need a lot less power, but throwing more current at an induction motor doesn't help AFAIK, it needs more volts, so how much current is needed is a good question, but it is basically a function of battery voltage and motor power.

I goofed though, and used 60 amps in that example, fixed:/

Anyway, to keep the inductor the same (40uh) at 20khz, you can reduce the ripple by 1/2 to %15, or you can reduce the current by 1/2 (300A).

[MPaulHolmes]

Here's their design tool. It was using 26u permeability material.