openinverter forum

So I got my hands on a cute motor from a kia e-niro.
Its for a full electric car I think since the motor is so powerfull.
The ide i to use it with my prius gen 2 inverter, but I have to read up on how the resolver sync works in the openinverter software.

Here's some images on it, teh resolver and some data from the kia mechanic I bougth it from.

Resolver setup should be easy with this diagram. You might want to parallel MG1 and MG2 outputs to get most power out of it.
11000 rpm is challenging with open inverter as field weakening stops working when coming OFF the throttle. But up to like 6000 or 7000 rpm it should work smoothly. More investigation on this topic is coming up

Seems like a lot of these motor develop bearing failure sounds, so kia just changes the whole motor to a new and scraps the old one.

johu wrote: ↑Sat Jul 30, 2022 5:31 pm Resolver setup should be easy with this diagram. You might want to parallel MG1 and MG2 outputs to get most power out of it.
11000 rpm is challenging with open inverter as field weakening stops working when coming OFF the throttle. But up to like 6000 or 7000 rpm it should work smoothly. More investigation on this topic is coming up

Ah, yes, the electric connections shouldn't be a problem, but more how openinverter sw handle the resolver syncing, is that automatic, or do I have to feed some offset to the software? (sorry for not having read up on this yet..)

and yes, we will not use the full RPM of the motor in this installation, more like half rpm if even that.

(its maybe going into and old landrover directly on the driveshaft to the differential

Here's how: https://openinverter.org/wiki/Using_FOC_Software

that sounds like low rev application indeed

Ok,so started to prod the kia motor this evening, the resolver resistance is quite different from the prius resolver though.
the kia resolver has as follow:
-/+ 26,3 ohm
S1/S3 84,3
S2/S4 75,6 ohm

Compared to the prius p112 that has 8, 14 and 14 ohms on the winings

I guess that the -/+ is the primary winding, and the other two are cos/sin windings, it's strange that the two later differ so much, while the prius cos/sin-windings are almost spot on 14 ohm both...

The resolver is a semiteck korea OE36562700
Their homepage doesnt really disclose anything usefull.
http://eng.semiteckorea.com/products06-04/

Tomorrow Im going to feed sinewave with a function generator into the resolver and see what comes out on the cos/sin outputs on the scope...

Linda out

ok, been fiddling with this a few hours today, and now I get an angle from the resolver, but it seems to turn over from 365 to 0 several times in one axle revolution, and that doesnt seem right?
I tried incresaing the "respolepairs" to 4 and then it goes down to 365 degrees per 90 degree turning on the axle.
If i try to set "respolepairs" to anything higher than 4 i dont sem to get an angel at all strangley, 8 or 16 on should give 365 degrees on one axle revolution.
or is there any other scaling parameter for the resolver?

/Linda

linda.ljungdahl wrote: ↑Sat Aug 13, 2022 8:16 pm ok, been fiddling with this a few hours today, and now I get an angle from the resolver, but it seems to turn over from 365 to 0 several times in one axle revolution, and that doesnt seem right?

The angle refers to the electrical angle, not the mechanical angle of the axle. Thus, it is possible and maybe also expected that there are multiple electrical rotations for each mechanical rotation. Based on the image of the resolver, respolepairs=2 should be correct.

LeonB wrote: ↑Sun Aug 14, 2022 1:30 pm The angle refers to the electrical angle, not the mechanical angle of the axle. Thus, it is possible and maybe also expected that there are multiple electrical rotations for each mechanical rotation. Based on the image of the resolver, respolepairs=2 should be correct.

Are you sure? The Leaf motor uses a four lobe resolver and usually respolepairs=4 is used?
https://openinverter.org/parameters/view.html?id=10

bexander wrote: ↑Sun Aug 14, 2022 3:10 pm Are you sure?

I’m sorry, I got that mixed up with the 2-pole pair prius resolver – yes, the Kia motor uses a 4 pole pair resolver.

LeonB wrote: ↑Sun Aug 14, 2022 1:30 pm The angle refers to the electrical angle, not the mechanical angle of the axle. Thus, it is possible and maybe also expected that there are multiple electrical rotations for each mechanical rotation. Based on the image of the resolver, respolepairs=2 should be correct.

Ah, my bad then, thanks, then maybe everything is good and I can start trying to do a FOC on it

johu wrote: ↑Sat Jul 30, 2022 5:31 pm Resolver setup should be easy with this diagram. You might want to parallel MG1 and MG2 outputs to get most power out of it.
11000 rpm is challenging with open inverter as field weakening stops working when coming OFF the throttle. But up to like 6000 or 7000 rpm it should work smoothly. More investigation on this topic is coming up

is there any description/guide of how to parallell mg1 and mg2 on the inverter side?
Just connect all i/o in paralell?
how about the current signals from the inverter?

Also, are there any simple way of telling which phase on the motor is which according to the resolver, or is it just trial and error?

/Linda

kia was kind enough to actually press the phase letters into the plastics of the connection block,but here are som images:

linda.ljungdahl wrote: ↑Mon Aug 15, 2022 12:09 pm is there any description/guide of how to parallell mg1 and mg2 on the inverter side?
Just connect all i/o in paralell?
how about the current signals from the inverter?

Yes just connect the PWM signals in parallel and make sure to pull enable signal GSDN to 12V. Ignore the second set of current sensors. You could adjust the gain of the first set (MG2) to kind of accomodate for the extra current supplied by MG1. I suggest running a second set of phase leads to the motor.

linda.ljungdahl wrote: ↑Mon Aug 15, 2022 12:09 pm Also, are there any simple way of telling which phase on the motor is which according to the resolver, or is it just trial and error?

That is sorted by setting pinswap=5

Got it running yesterday with just hooking everything up and resolver poles set to 4 !
Short video of running and setup:

Next thing is doing a FOC tuning.

So yesterday I tried to do a bit of FOC-tune according to the FOC-video, but got mostly strange results.
no matter what the sync offset is the motor vill normally rev up to 3000rpm at full throttle at around 200v DC
When I did Id-current test, i got the following:

voltage:193v

syncofs Id-current movment/dir
1000 15 none
2000 15 none
3000 15 none
4000 15 none
5000 15 none
6000 15 none
7000 15 R
8000 2 R
9000 2 R
10000 1 R
11000 1 R
12000 2 R
13000 1 R
14000 1 R
15000 1 R
16000 1 R
17000 1 R

30000 2 R
32000 2 R
33000 5 R
34000 15 none
35000 15 none
36000 15 none
37000 15 none
38000 15 none
39000 10 L
40000 2 L
41000 3 L
42000 1 L

Not shure hats going on here...
I'm wondering if its something strange with the resolver after all, but the resolver plot looks ok I think?
if the sin/cos-inputs where thrown around it would not work at all right? (wrong direction)

Not sure what is the problem... So syncofs is probably around 3000. So now you give it some throttle and it revs up to 3000 rpm, which normal no load behaviour - it is torque controlled, not speed controlled. Does it rev down when you release throttle?

johu wrote: ↑Thu Aug 25, 2022 9:19 pm Not sure what is the problem... So syncofs is probably around 3000. So now you give it some throttle and it revs up to 3000 rpm, which normal no load behaviour - it is torque controlled, not speed controlled. Does it rev down when you release throttle?

ah, but in the FOC-video Damien says that the "trough" in Id you should look for is the one between where the motor changes direction? (
Well, the motor is turning, so i guess we'll put it into the chassi, get a battery and see what happens with some load...
Thanks for the help Johannes.

Here is a cover for the phase connections I made in tinkercad also.

final drive with diff for this motor is part no:4450018EA0
the motor and final drive is the same between kia niro and hyundai kona.

linda.ljungdahl wrote: ↑Thu Nov 03, 2022 5:03 pm final drive with diff for this motor is part no:4450018EA0
the motor and final drive is the same between kia niro and hyundai kona.

and yes, we will not use the full RPM of the motor in this installation, more like half rpm if even that.
(its maybe going into and old landrover directly on the driveshaft to the differential

Interesting project but are you now thinking of using the Kia final drive in the landrover?
As the E Niro final drive ratio is something like 8.206:1 - so if this is connected up to the live axles in a series landrover which are something like 4.7:1 you'll get a total drive train gear reduction of 38:1! which is quite a lot - great for rock crawling!
When they mount Tesla drive units longitudinally in the centre of 4x4's they also change the gear reduction to something like 4:1 from 9:1 to also take into account the reduction in the axles.

Maybe i've missed something - not for the fist time

SuperV8 wrote: ↑Tue Nov 08, 2022 3:12 pm Interesting project but are you now thinking of using the Kia final drive in the landrover?
As the E Niro final drive ratio is something like 8.206:1 - so if this is connected up to the live axles in a series landrover which are something like 4.7:1 you'll get a total drive train gear reduction of 38:1! which is quite a lot - great for rock crawling!
When they mount Tesla drive units longitudinally in the centre of 4x4's they also change the gear reduction to something like 4:1 from 9:1 to also take into account the reduction in the axles.

Maybe i've missed something - not for the fist time

hehe, no, the idea is to connect the motor without its final drive to the land rover diff...

johu wrote: ↑Sat Jul 30, 2022 5:31 pm Resolver setup should be easy with this diagram. You might want to parallel MG1 and MG2 outputs to get most power out of it.
11000 rpm is challenging with open inverter as field weakening stops working when coming OFF the throttle. But up to like 6000 or 7000 rpm it should work smoothly. More investigation on this topic is coming up

Hi J

I can offer to test some further hgh speed OI functionality. That would be a good project during winter if you like.
My hardware:
1.Hyundai HSG motor with resolver setup. Now this is very approprite fo testing. It has a belt drive that i could connect to some load. I can use my Volt inverter and gen2 OI board
2. Prius AC compressor motor. I could take compressor apart and used motor in a belt drive setup for it to resist HSG via a diode bridge. It is a BLDC motor in the end.
3. As for the serial connection i have both wifi ESP module and TTL cable.

It would be a small setup and not much expenses invested for potentially huge gais:).

Ah actually the above quote no longer hold true With firmware 5.24.R FW is fully working.