i3 controller hacking from diyelectriccar.com

[MrX]

Simple issue after all! Looks like my SCK signal was being voltage divided between RN1 and RN2 on the Mainboard, meaning that my "Low" level was 1.7V (3.36V / 2 on my board). Therefore I wasn't getting a SCK down to the i3 gatedriver as it was constant-high signal after going through the line driver of the adaptor.

I just bypassed resistor number 3 of RN1 (pin 3 to pin 6) on the mainboard to drive the signal rail to rail.

Well, we have DC bus voltage and phase temps now. Next onto some tuning!

Resistor 3 bypass RN1

voltage divided SCK before bypass

MOSI and *CS signals

[johu]

Oh... but do you have the board with 3 line drivers (5 pin chip next to header)?

[MrX]

Oh... but do you have the board with 3 line drivers (5 pin chip next to header)?

Yes, this is the release adaptor PCBA, it arrived a week ago, after a few months getting to the other side of the world. *edit* I migrated the missing parts from old board to the new on Tuesday and have been tinkering since. However, I suspect this is the issue I was having previously with the prototype version adaptor PCBA as well.

I will say that I measured SCK before and after the line driver with the scope. The “low” level was too high and causing the line driver to be in a constant “high” state without the bypass wire. I’m planning on putting a 100 ohm resistor in parallel with resistor 3 of RN1 to provide some impedance protection to the output of the STM32, but still low enough so that the voltage waveform is okay for the line driver “low” input level.

[MrX]

It seems like the gate driver power consumption from the HV side keeps collapsing whenever I try to switch the IGBTs. Johannes, just in case you had something specific in your setup, was there a minimum voltage you had on the DC bus to make the inverter work? Also, do you have any information on the approximate power consumption on the 12V system when switching? My standby current seems very low, less than 0.5A on the 12V circuit.

I have a benchtop supply at 65V, 3A max for supplying the DC bus for initial testing, but I'm wondering if I need to go up in DC bus voltage to get things working.

[johu]

So it's not the 32V supply that collapses but DC bus voltage? Are all IGBTs good? deadtime set to 63? All gate drive signals sane when they arrive on the IGBT board?

[Scott166]

Hi folks
I'm working in making the inverter package fit in the Mini, I'll have some progress up soon
A quick question with this HV filter assembly, do I only need to keep the two circled connectors?
Can I get rid of the rest?
Tia

[MrX]

Hi folks
I'm working in making the inverter package fit in the Mini, I'll have some progress up soon
A quick question with this HV filter assembly, do I only need to keep the two circled connectors?
Can I get rid of the rest?
Tia

That's right Scott, although you should try to keep the filter capacitors as well if possible (as shown in this edit).

So it's not the 32V supply that collapses but DC bus voltage? Are all IGBTs good? deadtime set to 63? All gate drive signals sane when they arrive on the IGBT board?

Sorry for the delay Johannes: The 32V rail is pretty solid, but with some ripple on the DCDC. I didn't get to measure gate drive signals yet, but I'm going to mod my mainboard for the SCK line tomorrow (parallel 100R on R1) and measure them while I have it on the bench.

[Scott166]

Thanks Xavier

[Scott166]

Are those filter capacitors absolutely necessary?
Or is there a smaller alternative I could use as I can't fit it in and it no longer needs to connect to the DC side of things

[EV_Builder]

What do you mean with not to the DC side of things? The capacitors are there for the DC bus...
The idea is that when the current flows and suddenly needs to stop it wants to go somewhere...by having those capacitors it will absorb that energy...

[Scott166]

If anyone's interested there is a new vid on my channel of my attempts to shorten the inverter housing.
Still some bits to work out, how to get the coolant to flow correctly, how to prevent coolant from going everywhere.
I tried the aluminium brazing sticks but the heat dissipated too quickly to melt them on the inverter base and starting to melt the top before the sticks melted!
I also cut the top too short one side but it's not critical.

[EV_Builder]

Yes! That's what I meant!
Now for the coolant you could make two plates 12mm thick to cover up the holes and if you need to bridge/ make a passage you add two tube barbs (90 degress knee) and a piece of hose in between.
The welding is best using TIG since it's casting you need a balance of like 60% to 'clean' the weldpool...

[Scott166]

What do you mean with not to the DC side of things? The capacitors are there for the DC bus...
The idea is that when the current flows and suddenly needs to stop it wants to go somewhere...by having those capacitors it will absorb that energy...

Sorry for the confusion, I meant the DC to DC has been removed leaving only the power supply to the inverter

Would the capacitors in the inverter not do the same job? Sorry for my ignorance

[MrX]

If anyone's interested there is a new vid on my channel of my attempts to shorten the inverter housing.

Damn! You made short work of that Scott. I don’t think the filter caps in the HV connector junction will be absolutely critical. I would say leave them out for now. Looks mean I would say that you may need to try welding a blanking plate on, but castings are usually too porous for welding. Brazing will be the best way to go, just try finding a brazing rod with a lower melt temperature.

[Scott166]

Once I'm happy with everything I'll pop some cut lines and dimensions up here so you can all learn from my mistakes

I like the idea of blanking plates and elbows to join the coolant up but might look to glue the blanking plates in place

[SciroccoEV]

I would say that you may need to try welding a blanking plate on, but castings are usually too porous for welding.

I'm far from an expert welder, but have welded a fair few cylinder heads sucessfully.

BMW tend to use high pressure die castings, which should wel well.

[EV_Builder]

It depends on the composition of the casting and the settings of your equipment. @scott any markings on the parts about it's material?

I did the rear Alloy subframe with TIG; Try a high Silicium Rodd and AC Balance of 60% and clean/brush your weld mega clean before you start. If the blankingplate is choosen thick like 6/8mm welding becomes far more easier since you can direct towards the plate when it gets tricky.

[Scott166]

I couldn't see any markings showing info about the material

[EV_Builder]

I couldn't see any markings showing info about the material

Mhhh you sometimes see it near the stamp with the arrow about production.

Best thing is to test/try then. I would start with a high grade% Silicium rod.

[aesthetect]

Is it possible to alter the BMW i3 IPM motor/controller to run it in reverse with full functionality, efficiency, max speed, etc? Would be hoping to use the OEM motor controller + openinverter board and software but would be open to other options if need be. Would it be the same vector control algorithms as forwards?

I have seen that the OEM hardware limits reverse speeds but I am wondering if the unit can be reconfigured as a whole.

(The reason for the question is because I would like to use two of these on two opposing wheels to get a torque-vectoring setup, but the only way I can see getting them to fit is with one rotated 180.)

[johu]

openinverter firmware has no reverse speed limit. Motor and inverter don't care either. Only if you use the stock transmission you'll likely run into issues running it reverse

[johu]

On another note I think there is an error on the mini mainboard resulting in lack of CAN communication. The slew rate pin is floating but needs to be tied to ground with a resistor. I scratched the GND plane a bit and soldered in a 10k resistor. Can be any value from like 0R to 20k whatever is in your parts bin (the value controls the slew rate)

[aesthetect]

openinverter firmware has no reverse speed limit. Motor and inverter don't care either. Only if you use the stock transmission you'll likely run into issues running it reverse

Thank you very much for the quick response and confirmation. Would I be able to simply flip polarity somewhere? Or just within openinverter I would command it backwards?

Indeed the gearbox is the other question. Looking into it, axial loads in the wrong direction from helical gears running backwards would seem to be the biggest issue. Getting some confidence from the fact that its software limited to 50 kph. Would love to hear any other main suggestions to consider, but not to take off topic..

At least having that confirmation makes me a bit more willing to purchase one and start testing, but will be looking into gearbox some more first. Thanks again.

[johu]

Just select reverse direction

[aesthetect]

Thank you again. Back to happily and eagerly lurking on y'all's experiences with these.