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

[Piotrsko]

Your warp sounds exactly like the KOSTOV except for size and operating volts.

The 335@72v is peak power advertising. Motor will do that but for only about 4 -6 minutes. No contradiction there. Misleading? Perhaps.

Kevlar banding wont take much above 250c because the bonding agent melts which would immediately thrash motor. I haven't heard of that ERGO: comm stays below 250c.

Batteries should* do what you want unless they are 3+ years old.

Have fun. I think you want to do this in 12v increments. Runaway dc motors are not exciting for long.

[thingstodo]

Originally Posted by Piotrsko

The 335@72v is peak power advertising. Motor will do that but for only about 4 -6 minutes. No contradiction there. Misleading? Perhaps.

They have stated 400A for 5 minutes, so 335 is maybe 10 or 15 minutes? Not sure, but I'd like to know.

Kevlar banding wont take much above 250c because the bonding agent melts which would immediately thrash motor. I haven't heard of that ERGO: comm stays below 250c.

Excellent information. That banding is on the armature, and holds the rotor bars in, right? Or is that also used on the commutator?

Have fun. I think you want to do this in 12v increments. Runaway dc motors are not exciting for long.

I am going to stick with 24V for now, without a DC controller.

When I can put a DC controller in there, the idea is to go to 125VDC and feed both the DC controller and AC controller from the same battery pack. Then I hope to test at 3000 - 4000 rpm, where the PWM looks more like AC to my metering equipment.

Neither the DC controllers (that are broken) nor the AC controller (that I'm testing) will take 1000+ amps for long, but I'd like to SEE!

[Astro]

Originally Posted by MPaulHolmes

One option is milling channels somehow haha, and then bolting a thin aluminum sheet to the base, somehow sealing the channels so they don't leak, and then somehow attach a hose to the base plate edge.

Hi all, if it is just for a test rig then how about a coil of copper pipe, some thermal grease and some lumber (we call it wood ) to hold the copper pipe in good contact with the controllers base plate. Run some sort of coolant through the copper pipe with a small pump. Could use water, coolant, oil, whatever is easiest.
As long as you don't use too much pressure holding the copper pipe onto the controller plate then after the tests you should be able to recover the copper pipe undamaged and use it again elsewhere. There wouldn't need to be any high pressures so just some garden hose between the ends of the copper pipe and the pump and a big container of coolant should be ok.
For long tests just increase the volume of coolant. If it is cold outside and you have a pool then that would make an ideal reservoir.

[thingstodo]

Originally Posted by Astro

Hi all, if it is just for a test rig then how about a coil of copper pipe, some thermal grease and some lumber (we call it wood ) to hold the copper pipe in good contact with the controllers base plate. Run some sort of coolant through the copper pipe with a small pump. Could use water, coolant, oil, whatever is easiest.

This is a simpler backup plan - I like it.

The primary plan is sitting the controller on a BIG hunk of aluminum - about 100 lbs. It should take a while to heat up ... longer than the base of the controller at least.

[e*clipse]

Here's a quick work-in progress of the round controller:



Obviously it's not complete, but it looks like it will be possible to fit everything within the 265mm 2 layer circuit board restriction. I still have to see whether I can actually do this with only 2 layers, but there is a good possibility.

The four large holes are the output pin/current sensor holes. They also double as the connectors for the high side switches' emmitters and the low side switches' collectors.

On either side of the output holes are separate drivers for each set of three mostfet/IGBT/SiC switches. It's basically the same as Paul's original design with surface mount parts. While it has been VERY TEMPTING to use smaller parts, I've restricted parts to be SOIC and about 1206 minimum. While there are a lot of parts, they will be hand-placeable.

In the center is (I think) barely enough space to fit the control stuff and one big connector. Dang that connector takes a lot of board space...

So far the design includes all the stuff the original design includes, especially all the safety override circuits.
24V power supply input
Relay controls for the precharge relay and main contactor
2 CAN busses and the drivers for them
2 RS232 serial busses - standard and Rasp. Pi.
4 current sensors, 1 for each output and 1 for a boost controller
Quadrature encoder interface OR resolver input + motor temp sensor
Temperature sensor
3 independant throttle inputs - why 3?? OE throttles have 2 inputs, to make sure they are working correctly. Also, if one wants to do some form of throttle override using an external microprocessor or a proportional regen braking setup - use the 3rd throttle.

Regarding the throttle, I made the circuit work for 3 wire pots or hall effect sensors that are driven by +5V. I've ditched the 2 wire throttle sensor to keeps things reasonably straight forward.

So far, the I/O listed above has been accounted for in the output connector. There are a few spaces left.

That said, I think using two big connectors to make more pins available will be tough, at best. So, I think it might be smart to consider sharing common pins - Control Ground and +5V would be good candidates since they need to pass very little current. It might be smart to separate the analog input stuff from the digital I/O stuff.

Thoughts? Suggestions? Questions?

- E*clipse

[P-hack]

Originally Posted by e*clipse

Those boost switches would probably need to carry more current than a phase leg.

That is what I was thinking when I saw 4 power sections

the chunky prius booster switch has to carry the conversion current for all three phases, i.e. if the converter output is 150a @600v, and the input is 300a @ 300vish, the inductor sees 300a, the top diode sees 150a semi constant (depending on converter design) and the bottom switch swings/chops between 300ish and zero. Meanwhile the inverter switches only have an average of 50a each to deal with.

[e*clipse]

Originally Posted by P-hack

That is what I was thinking when I saw 4 power sections

the chunky prius booster switch has to carry the conversion current for all three phases, i.e. if the converter output is 150a @600v, and the input is 300a @ 300vish, the inductor sees 300a, the top diode sees 150a semi constant (depending on converter design) and the bottom switch swings/chops between 300a and zero. Meanwhile the inverter switches only have an average of 50a each to deal with.

Yes - that is definitely a serious issue. The boost current is directly proportional to the difference in voltages. Essentially, the power output must be constant. As you state, amount of boost would be limited by the current handling capability of the boost stage. Another issue may be the resulting asymmetrical balance of current in the inverter - resulting in asymetrical capacitor loading and asymetrical thermal problems.

It may seem a bit of a cop-out, but I've convinced myself that it would still be useful for a less dramatic boost than the Prius sees. This would allow someone to use a 650V motor like a Prius or Highlander Hybrid motor with a battery pack with a less stratospheric voltage. This could significantly reduce the cost and weight of the battery pack.

Since the same controller runs all of this, it would be possible to adjust the amount of boost depending on conditions. Up to the point where the BEMF = battery voltage (or something close) the boost could be 1:1, requiring no extra work from the boost stage.

From what I've seen from various IGBT/Mosfet/SiC spec sheets, the TO-247 package can handle about 100A > 120A. ( Even though the manufactures claim over 200A for some models - ) A Correctly designed inverter would balance the design for a PEAK 300A output of the boost stage. For example, a 75kW inverter controlling a 650V motor would be limited to a minimum battery voltage of 250V, assuming constant power, neglecting losses. I plan to use a Nissan Leaf pack, which has been tested to 360V nominal. Thus, the boost stage would be seeing something like 208A, or 70A per switch.

Another thing I'm playing with is a double H-bridge DC controller using this design. I'm building an electric snow-cat that could use this capability.

- E*clipse

[P-hack]

Originally Posted by e*clipse

It may seem a bit of a cop-out

Actually I lied, the output boost diode bounces between roughly 300A and zero, 180 "degrees" out of phase with the bottom switch bouncing between 300A and zero, with a typical 30% inductor ripple anyway (i.e. 166uh here @ 20khz, ~%50 duty cycle).

It isn't a cop-out, it is just puzzling, why the form of a capacitor is dictating and severely limiting the boost circuit design. asthetics?!? do you see divinity in symmetry? Fibonacci's three phase controller? the boost switch/leg needs to be at least 3x larger for "hotrodding" purposes, sorry.

[P-hack]

Originally Posted by e*clipse

Another thing I'm playing with is a double H-bridge DC controller using this design.

or for a two phase motor, for whatever long series of mistakes might lead one to want to build a two phase motor.

[e*clipse]

Originally Posted by P-hack

or for a two phase motor, for whatever long series of mistakes might lead one to want to build a two phase motor.

Just wondering: I mean it seems you make quite a few mistakes yourself when reading my posts.

I said nowhere "two pase motor."

I did say double H-bridge DC motor.

DC motor.

DC motor. Yeah, those old-fashionedee thingees with brushes.

The double H-bridge is a method of providing 4 quadrant control for a DC motor. I was merely playing around with this; there's nothing in the design specifically for this. It just one of those things that work out really well, if you're aware of things.