BMW I3 Project

[johu]

ah that was you You should still try to use the original current sensors, they should be more accurate than the Melexis ones that come with the kit.

[lumax]

ah that was you

That was actually not me... Karl made the order, he is the owner of all these motor and drive, I'm only the guy who will help him build his electrical dream!

I will sure have a look at the original current sensor, Thanks!

[willywonka]

Bonjour lumax

Thanks for your effort! I'd be very interested in the outcome too as I'm about to start a project myself. I could currently get a complete i3 which was totalled in an accident. But with it I would have all the electrical components I would need for the conversion project – except that there's not much info available on how to use these i3 components. Now, while I have enough electrical know-how to poke around and follow instructions and am also quite comfortable writing software, I don't think I would get anywhere trying to figure out that stuff by myself. So I will keep a close eye on your project and will let you know in case I should decide to get that i3. Then I might be able to support you with some simple tasks... And you could even instruct me in French but rest assured: Your English is much better than my French.

Cordialement,
Will

[lumax]

Nice to hear that Will and welcome aboard!

I have mostly dismantled every electrical/electronic box in this car... and it's way to much complicated...
I really like the " Simple thing for simple task" expression... I'm confident that with the Huebner controller the I3 motorization will be unleashed

[Jack Bauer]

Precharge system is in the battery pack. Personally I would want to use the OEM igbt driver board where possible.

[willywonka]

Hi @lumax

Did you already start to disassemble the battery pack? If so could you share some detailed photos + measurements of the modules? I would like to create a CAD model to see how I could fit them in my project – or if I have to go a different route after all. Thanks!

[lumax]

Hi @lumax

Did you already start to disassemble the battery pack? If so could you share some detailed photos + measurements of the modules? I would like to create a CAD model to see how I could fit them in my project – or if I have to go a different route after all. Thanks!

Yes, I will take measurements and pictures and share it

Precharge system is in the battery pack. Personally I would want to use the OEM igbt driver board where possible.

Yes make sense, I have ordered a boost converter to supply the 32V for the gate driver board... thanks to your note it will be easier, at least no reverse engineering is required. I have received the brain , so I will start the project soon

[Scott166]

Hi Lumax
Good luck with your project, I was thinking of using an i3 motor for my conversion as they are reasonably easy to get hold of. Also I own an i3 so would be nice to have an i3 powered project too.
I’ll be following you closely to see how it goes.
How much power are you looking to get out of the motor and battery pack?
Cheers
Scott

[lumax]

How much power are you looking to get out of the motor and battery pack?

Maximum power with an original 33KW BMW battery

Today I have started to connect everything together... And I decided to give a try to the original IGBT driver board... It was a pain in the a** to connect the wiring to the original tinny connector, but I managed to do the connection with 1/16W resistor lead welded to a stranded wire... not perfect but good enough... 24 gauge solid wire was too big to fit the connector

I will keep you in touch !

[oelbrenner]

thinking of using a i3 motor for a future project, any chance you can grab some measurements on the back side of the motor.

mostly looking for these measurements, but if you have more accurate / complete ones that would be great !

[lumax]

thinking of using a i3 motor for a future project, any chance you can grab some measurements on the back side of the motor.

Screen Shot 2020-06-07 at 8.37.14 PM.png

mostly looking for these measurements, but if you have more accurate / complete ones that would be great !

Your long arrow dimention is roughly 13.75'' (overall outside)
Your short arrow dimention is roughly 6'' (c/c)

But the bolt patern is not even... some holes are at 6.75'', 6.25'', 7.25''...
In the final assembly of this project I will post the complete template with all measurent... when i get there

Hi @lumax

Did you already start to disassemble the battery pack? If so could you share some detailed photos + measurements of the modules? I would like to create a CAD model to see how I could fit them in my project – or if I have to go a different route after all. Thanks!

A single pack is 12.25'' wide x 15'' long x 5.5'' tall

[willywonka]

Thanks for the measurements! Although... I thought you're Canadian? What's with the inches?!

[lumax]

Little update...
After almost one year, the project was held to a dead stop, since I got problem with the original brain, (no PWM output) and no more time to play... Johu, kindly send me a replacement board to replace the defective one... After a couple hours of setup and config, I now have PWM signal !!!

So I'm gonna keep following the config and start-up procedure... So far so good!!!

[johu]

Haha, that stuff is often slow cooking
Also look here: viewtopic.php?f=20&t=1296

[lumax]

Thanks for the link!

[lumax]

Already run into a problem

Tested all PWM signal, and adjust the dead time to 1.5us, connected the IGBT (no DC link), power it back on, start inverter and BAM!!! Blow my 1A fuse on the brain... Triple check all my wiring, everything was good.

Unplugged the driver from the IGBT and test all gate one by one... I hade no problem to turn them ON and OFF with my finger resistance...

Put back everything together and test again but this time with only one IGBT connected to the driver (all 3 driver connected to the brain), the current draw from the brain alone and the 3 gate driver is around 260ma... with one IGBT connected and working (PWM ON) the total current draw is around 700ma (450ma for one IGBT), and when 2 IGBT are connected I run around 1.2A and with 3 IGBT connect I got an over current protection error that stop the PWM (current should be around 1.6A)

The gate capacitance look to be to large for the driver and the brain PSU... From the scope I can clearly see rising edge is curved due to gate capacitance, unfortunately I did not take a picture

Is this kind of current draw is normal?
Don't think so...

[johu]

The BMW driver board takes in 11W from the 32V supply voltage when running all 6 IGBTs at 8.8 kHz. This would translate to 1A at 12V. The 5V architecture might be less efficient than 32V. Can you try supplying the gate drivers separately with 5V? They consist of 6x2W DC/DC converters. Assuming 80% efficiency they would already pull 3A from the 5V rail when fully utilized. That would probably overload the 5V regulator.

[lumax]

Yes, I will give it a try and I will be back with the results!

[lumax]

Here are the results

When I got this over current problem on the brain the PWM frequency was set to 17.6Khz...

Gate driver (all 3) idle current @ 5V = 400ma
3.1A @ 5V @ 17.6 Khz (bad wave form)
2.0A 2 5V @ 8.8 Khz (good wave form)
1.2A @ 5V @ 4.4 Khz (excellent wave form)




So I will stick to 8.8 or 4.4 and keep the gate driver power connected to the brain

By the way I got a lots of success yesterday...

Open loop motor rotation wirh SINE DONE!
Closed loop motor rotation with FOC DONE!

Now ready to tune the FOC and put back everything together

[lumax]

The test rig

[lumax]

FOC firmware have been tuned and sincofs is good enough for now , but I think there is a problem in the firmware... I might be wrong to...

My I3 motor is a 12 poles (6 poles pairs) and the resolver is a 6 poles (3 poles pairs)

The motor only turn when I set it to be a 3 poles pairs, it does work perfectly in closed loop (accel-decel-regen-FW-REV) but since the computed poles pairs is not matching the real poles pairs I got a wrong display speed and probably all RPM based calculation must be wrong... I'm still testing at less than 100V @ 50A, the real motor RPM is max out at 1800 (brain display 3600) should be limited to near 11 000 RPM, this must be doing weird things with V/f, advance and field weakening computation

As soon as I change it for 6 poles pairs, motor lockup no matter what I do ???

Is this related to the firmware or I totally miss understood something ???

[johu]

Did you also change respolepairs along with it?

[lumax]

No, there is only 6 lobs on the resolver so it should be a 3 poles pairs... but I did not think about this one... could also be a 12 poles!!!

For now the inverter is disconnected, I'm putting everything back together and do the final assembly and wiring of the inverter/motor, as soon as everything is done, I will test it

Thanks for your reply!

[johu]

What is respolepairs set to right now? If you double polepairs you'll have to double respolepairs as well. The ratio must remain the same.

[lumax]

OK, it make sens... I tough the resolver poles pair was absolutely related to the number of lobs on the resolver rotor... 6 lobs = 3 poles pairs... but in this motor it's a 6 poles pairs on 6 lobs!

Resolver was set to 3PP and motor was set to 3PP (working but haft real speed)
But in real Resolver is a 6PP and motor is a 6PP (to be tested)