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

[MPaulHolmes]

I've almost got a dyno set up. Very soon there will be lots of testing that can be done. Here's a picture from earlier this week.

[MPaulHolmes]

Here's a quick demonstration of the dyno with one motor running in torque mode, and the other running in speed mode. I'm getting some really great data, too. The big question I had was about finding the motor's saliency, and how to automate that later. The motor's "current" can be represented as:

<Id,Iq>, where the 2 currents are at 90 electrical degrees to each other, and the current command is

currentCommand = sqrt(Id*Id + Iq*Iq).

The trick is to figure out how the current command should be divvy'd up between Id and Iq. Think of it as a right triangle. The hypotenuse is the current command, and Id & Iq are the 2 other sides of the triangle. CurrentCommand can be mostly Id, or mostly Iq, or somewhere in between. I have been varying Id from 0 up to currentCommand, and looking at how big currentCommand is in order to maintain the constant speed. The <Id,Iq> combo you want is the one that results in the smallest possible currentCommand, which means you are getting the biggest torque per amp of motor current. The ratio of Id to the currentCommand changes as the current command increases.

https://www.youtube.com/watch?v=1ecuIS3jPQk

[MPaulHolmes]

Here's a first test with a TO-247 mosfet controller.

https://www.youtube.com/watch?v=PU-x...oms2pbk0h00410

[thingstodo]

Originally Posted by MPaulHolmes

Here's a first test with a TO-247 mosfet controller.

Good progress on your proto-type!

I'll echo the questions I saw posted on the youtube video - Are you thinking of a kit, a or bare board with a BOM to solder onto? If so, I'll take 3

Why? I have 4 SEVCON mosfet-based AC controllers that are in need of repair. ... and the cases are too small for the larger AC control board to fix. But they need brain transplants if I can drive the MOSFETs.

3 of them have kinda strange fault codes - motor temp too high, low bus voltage, encoder faults, etc. The brain transplant should avoid the cold solders, corrosion, and various other possible problems.

The last one is complaining about a power stage problem so it may have actual problems that I can't fix with a brain transplant.

It would be easiest (for me) to have a new driver board and a new control board. I might be able to re-use the driver stages ... but that could be where I am having the issues

I was going to use your larger AC controller. And I blew one set of IGBT's by reversing + and - on the battery back ... newbie mistake! ... and I have not gotten back to the project since.

But ... a controller board that will fit inside the case (or looks like it might) would help a bunch. And driving the existing MOSFETs would also help since they appear to be in good shape, and it makes my mechanical how-do-I-make-it-fit challenges smaller

[thingstodo]

BTW Paul, there is a member looking for the gerber files for you ReVolt (that was the cougar, right?). It looks like I don't have those

https://ecomodder.com/forum/showthre...tml#post605654

[Jonisingt35r]

Hello
Im looking to build an ev, and i am looming to build an ooen source controller for an ev
Car.
Just wondering is this one is realiable and has good documintaion ect
I would like to go ac route, rather than dc really as i have lots of experience with inverters
Tbh both routes have their advantages dc, low cost, just after a realiable controller i
soppose.

Would really like to buils an ac inverter as im also interested in solar as well

I see this has been going now since 2008 hope its still running, or maby if its not is there another controller out there open source?

Thank you

[MPaulHolmes]

Hi Jonisingt35r! I can't keep all the names straight of stuff either. Don't even worry about it. I did a controller for an asynchronous and synchronous motor, but I've only done quite a bit of testing on the synchronous motor under load, not the asynchronous motor. I've only ever tested asynchronous motors NOT under load. Well, a friend tested his tesla motor while hooked up to a gear box, and it seemed to work fine, but I don't know about driving around with it.

[vivek]

I hope all are doing great. I am planing to covert my honda city to electric. Previously I have done dc conversion using open revolt motor controller(thanks paul) now I am looking for ac conversion. can someone please send me gerber file and other details regarding Paul 3 phase ac controller. That will be a great help

email - vivek.naw@ live.com

[sparkey]

Hi all,

I'd like to share my progress on my inverter build. I bought a bare board about 18months ago (yeah, i know), and following the instructables, paul's documents and this thread, I finally have what seems to be a functioning inverter. I've followed Paul's instructions almost to the letter, including which current sensors, capacitors etc to buy, even down to the design of the housing and copper plates. The inverter is tested using every test I've read, and it appears to function correctly!

I've attached some photos.

I'm using a lexus GS450H motor as a test motor, and I'm having trouble getting it to spin using the throttle. This is a permanent magnet motor, with an rx-8 hall-effect throttle, and a 48v (nominal) pack

The main values i'm changing to tune are KP/KI, the throttle values (min, max throttle & regen) and angle-offset. Are there any more that I'm not aware of?


So what am i seeing:

1. The throttle seems to have no effect, but I can see the raw values change as expected when streaming (values change from ~315 -> ~800). I've tried every combination of throttle settings I can think of, but with no luck.
2. Running the run-pi-test2 produces inconsistent results: the motor turns briefly (maybe 45 or 90 degrees) when the test is run, and sometimes converges to near 0, but it doesn't converge every time i run it for any settings i use. So when i *think* I've found good values but then it doesn't converge a minute later.
3. The run-angle-offset-test causes the motor to run smoothly (between angles 140->240), so the inverter is capable of running the motor smoothly. I just cannot make this happen using the throttle.

So my questions:

1. For a given KI and KP, should I see a response from the throttle, even if it's a small movement of the shaft?
2. Are KI, KP and angle-offset related to such an extent that they would cause no response from the throttle if it's set up wrong? In other words, which numbers are most important to mess with to verify that I can get ANY response from the throttle.
3. Is there a test I can run which proves / disproves that throttle response is / is not working?

Sorry for the long post, but this thread has been amazing help, and I hope to contribute in future.

[freebeard]

I'm no help, but I hope you get your answer. My best guess would be it needs a DC-DC up-converter to 120+ volts.

Here's a question I posed in another thread*:

Do you know if the GS450h/LS600h uses the Interchange Part Number: 300-65002?

I've got one of these sitting on a bench.


* ecomodder.com/forum: Hot-rodding the Toyota MGR