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

[JRoque]

Hey SGC, is yours a different design than Paul's or are you contributing to his power section?

You might be aware of this but proceed with caution with protoboarding IGBT/FETs - they tend to be touchy to inductance. Keeping the driver and power bus close by helps quite a bit. There's also current loops to watch out for in your layout - I've been bitten by this before and it can be tricky. Star ground, single returns as much as you can.

Best of lucks and post your schems when you get a free min.

JR

[TheSGC]

Originally Posted by JRoque

Hey SGC, is yours a different design than Paul's or are you contributing to his power section?

You might be aware of this but proceed with caution with protoboarding IGBT/FETs - they tend to be touchy to inductance. Keeping the driver and power bus close by helps quite a bit. There's also current loops to watch out for in your layout - I've been bitten by this before and it can be tricky. Star ground, single returns as much as you can.

Best of lucks and post your schems when you get a free min.

JR

Paul is using a premade 1.5 HP power section, so mine is different. However, I believe I am using the same dsPIC, and I am designing the power section to hook up directly to the PWM output pins of the dsPIC, so what ever logic Paul comes up with should work just fine, but I would need to check my papers. I do have some information on the power section Paul is using, but haven't really read into it yet.

Protoboards are pretty awful, so I usually take the results with a grain of sand. I am just going to start off with one of the phases and make sure I can turn the IGBT on/off when running a small LED and check it with the scope.

Of course, first come replacing the battery pack in my EV and then cleaning up my work bench so I can actually prototype.

[ecorsa]

Hi, New to forum, this is a great project.
I have just read through all the posts and think i have some info to bring to the party to help clarify some of the points raised so far.
1) off the shelf 3 phase induction motors which are avalable new for cheap and 2nd hand for next to nothing are generaly available in standard frame sizes. there is a number in the size which refers to the radius from the centre line of the motor shaft to the mounting line of the feet (if the motor has them). there are 3 standard mountings, face, flange and foot, the latter can be in combination with the former 2. this link to TEC's catalog will help with this. the aluminium ones TEC-A are the most relevant. (add www to the following)
bearingboys.co.uk/uploads/Tec-cataloque.pdf
2) there are lots of benefits to running higher voltages, here is my take on it
a) less I^2xR losses (power loss is proportional to current squared) so high current is bad news all arround. anything which can reduce currents is a good thing
b) Back EMF. As RPM increases so does back EMF, which works against the applied voltage to reduce the eficiency of the motor. the higher the operating voltage, the higher the RPM can go before this becomes a show stopping problem
c) Off the shelf (cheap) motors are set up to run higher voltages as they come, basically for the above reasons.
d) there are advantages to running more batteries of a smaller size i) this makes packaging easyer (can fit in lots of nooks and crannies) ii) if one craps out its cheaper to replace a small one. iii) cheaper hook up due to smaller cables.
Safery is obviously a concern with higher voltages, but you really need to consider anything over about 30V potentially lethal anyway.

[JRoque]

Good list ecorsa!

Drawing from my extensive experience with 3PH motor specs amassed over the last....er... 2 weeks :- ) I can add these to consider when looking for a motor:

1. Light weight, aluminum (aluminium) frame if possible.
2. Rating: "vector" preferred, "inverter" could work. This is mostly related to the high insulation needed to spike the motor into delivery full torque at low RPM.
3. High speed rated bearings. Most industrial motors run at <3600 RPM and their bearings are rated for that speed also. Some can go higher but don't expect continuous >10K RPM
4. TEFC, totally enclosed, fan cooled or better. Yes, many DC motors have exposed commutators but dirt and water play a nice number on those. Drip proof might also work but TE or even wash down is better.
5. Double end shaft. Even if your controller will use flux vector with current sensors, you might need to measure RPM at the motor shaft. Or maybe you decide to mount a pulley there and drive a pump, etc.

Ok I'm out. Anyone else? and then back to Paul so he can tell us how close he is to posting the controller's power block...

JR

[TheSGC]

I have started to clean up my work bench and gather my protoboard parts. I am going to pick up a few NAND gates and other parts tomorrow, then see what I can come up with as a gate driver.

I am going to try to get access to a lab at school tomorrow so I can test out the power section. While I have a scope at home, I currently don't have a power source or a signal generator to create a duty cycle for testing, since I have not worked on the dsPIC section yet. After I am happy with my power section, I will move onto the logic, which will probably take me eternity to figure out the coding.

[JRoque]

After I am happy with my power section, I will move onto the logic

We work the same way on that point.

Are you just throwing things on a protoboard and then come up with a schematic? I couldn't do it that way - I'd forget half the stuff if I tried. Have you given thought to using an off the shelf driver? Those can be expensive sometimes but not when compared to a high power IGBT. Most drivers have things like shoot-thru prevention, boost pumps and inverters which make life easier.

Good luck with your lab work!

JR

[TheSGC]

Originally Posted by JRoque

We work the same way on that point.

Are you just throwing things on a protoboard and then come up with a schematic? I couldn't do it that way - I'd forget half the stuff if I tried. Have you given thought to using an off the shelf driver? Those can be expensive sometimes but not when compared to a high power IGBT. Most drivers have things like shoot-thru prevention, boost pumps and inverters which make life easier.

Good luck with your lab work!

JR

I have been working on a schematic because most prebuilt IGBT drivers do not work with the large IGBT modules. I have actually designed a driver a few years ago that worked with a single large IGBT module, but I have modified the design to function with a dual IGBT module. I can't actually try out the dual setup until I get two isolated DC/DC converters.

[adamj12b]

Originally Posted by TheSGC

I can't actually try out the dual setup until I get two isolated DC/DC converters.

Dave,

What are you using for DC-DC converter??

You should strongly consider the VLA106-24151 by Powerex. Digi-Key - 835-1061-ND (Manufacturer - VLA106-24151)

These are very high grade converters for a decent price.

-Adam

[TheSGC]

Originally Posted by adamj12b

Dave,

What are you using for DC-DC converter??

You should strongly consider the VLA106-24151 by Powerex. Digi-Key - 835-1061-ND (Manufacturer - VLA106-24151)

These are very high grade converters for a decent price.

-Adam

OOhh those do look nice. I started protoboarding the driver this afternoon but I had to change up my design to work with the parts I have. I actually have to make a quick and simple 5v pulse generator and a load to test the IGBT. Unloaded, the driver looked good in the scope but I need to get the setup with a load (12v bulb) and check it again. And using a button connected to 5v supply isn't much of a duty cycle test.

[ecorsa]

I've been looking into using an off the shelf AC drive for a while now using 48 X12v 22AH batteries (570V) and a 480V 3 phase motor. Out of the drives I have investigated the NXI series from Vacon seems to be the closest match to a battery vehicle setup. the NXI range run from a DC bus, intended to maximise efficiency in factories with lots of motors, saving energy by feeding current back to the DC bus when the motor is on overrun. I am not sure how the drives would react in a vehicle situation in sensorless mode, aparently the drive can calculate the rotor speed from its measurements of the outputs to the motor, but i dont really understand how this would work. There is a speed control for the drive which could be connected to a throttle pedal, but in the car situation we are not looking for a set RPM as such, we are looking at varying the torque output of the motor.