Nissan Leaf Gen3 inverter

[magelan]

Is this inverter also with 800amp igbts?
I am wondering, is it the inverter or ecu that limits the motor to 110/130kw power?

Really looking forward to order a brain board for gen3 inverter as soon as it’s released.

[johu]

Yes, this one: https://rocelec.widen.net/view/pdf/bt5e ... e&u=5oefqw

We will see how they tuned the desat (overload shutdown) as soon as somebody actually puts it in a car with a sufficient battery pack.

[magelan]

Nice

With an adequate cooling for both motor and inverter it should be possible to push this drivetrain to 300kw or more just using a plug in brain board, or 3rd party Vcu if the inverter doesn’t have factory limits, but I doubt that... should be a number of security fail safe option is place, I hope they are not so insane, like those used in Toyota inverters.
My drive train is coming from Japan by truck, another week to wait until I can dismantle it.

[Kelju]

First challenge was to get it running from the board Kelju had sent. I thought the inverter chip was broken because all outputs were 0 when the inputs were also 0. Anyway, I removed it and bridged inputs to outputs, then set control board to active low mode. Which didn't work either. Then I found that the original control board outputs 5V on two pins where there was no 5V on the adapter board. So simply added a wire. Now the drivers came alive and sat at 5V. And 3.8V with the control board connected. I'll never understand the concept of those "5V tolerant" pins.

Sorry for this. I totally forgot to include the fix to get the 5V pullup for the IGBT driver input diode. I have it implemented on the adapter that I am running.

Really looking forward to seeing how running the motor works on your end.

[SciroccoEV]

At 255 deadtime (15us) it still pulls 100mA so I'm thinking that is not shoot through but some other gate drive draw or something. The shortest dead time that still yield 100mA draw is 160 or 2.6 us. The output looks good also. At that sort of slow switching I wonder if the inverter better be run at 4.4 kHz. At 8.8 kHz we are loosing almost 5% amplitude to deadtime.

Is your code running deadtime compensation?

[johu]

Implicitly yes, the PI controllers will just command more dutycycle until they see the desired current. The sine software does not compensate for deadtime but I suppose it's tuned out with boost and fweak

[SciroccoEV]

Implicitly yes, the PI controllers will just command more dutycycle until they see the desired current. The sine software does not compensate for deadtime but I suppose it's tuned out with boost and fweak

I should have said "deadtime distortion correction".

[arber333]

At 255 deadtime (15us) it still pulls 100mA so I'm thinking that is not shoot through but some other gate drive draw or something. The shortest dead time that still yield 100mA draw is 160 or 2.6 us. The output looks good also. At that sort of slow switching I wonder if the inverter better be run at 4.4 kHz. At 8.8 kHz we are loosing almost 5% amplitude to deadtime.

At 4.4 kHz the total draw from 12V is 0.75A, at 8.8 kHz 1 A and at 17.6 kHz 1.8A. Also the driver chips start evaporating their conformal coating, so thats definitely too much
Also 4.4 kHz is much quieter than at 8.8 which is unusual.

A little practical application...
I noticed 1st gen Leaf motor really likes 6kHz a lot. I used 3.1us deadtime with Volt inverter for a year now and yesterday a friend wanted more power from his Leaf setup. We discussed some of the pros and cons of that and we agreed to try at least for testing with wheels up.
I changed deadtime to 2.6us at 6kHz. Response was immediate and very powerful. We decided to go for a drive.
There really was a difference! Car was performing really well. We could burn tires on takeoff and at speed! Mazda RX8 became really lively car.
While using the infineon equasion for deadtime i estimated 2.6us smallest safe margin for reliable operation. Sure you get 2.4us math result, but you need to consider some overhead for real life too.

System is running on 96S 120Ah pack.
Other settings were 600A phase current limit, 400A battery limit, 7000rpm limit and 8K erpm/s limit on acceleration.

[Jack Bauer]

Just got a delivery of a gen 3 inverter. Mine has a kangaroo stamped into it for some reason...

Anyway, I'll be trying to run this via can so (assuming i am successful) will take measurements of the oem igbt drive signals to aid modboard development.

[mdrobnak]

Destined for AU region perhaps?

[johu]

Mine is kangaroo #2
Same stamp on the Gen2

[Kelju]

I'll be trying to run this via can

Sounds good. I have had the same intention after facing serious issues with running the motor with the V3 mainboard, just have not got around to it.
Anyway, here is the pinout as far as my reverse engineering is correct:

[Jack Bauer]

Cool. I grabbed a few bits from nissan tech info last night. Abomination of a website. Had to use windoze and ie11 to even register Anyway I don't want to post the actual doc the forum just in case but here is the pinout :

1-12v battery pos
2-nc
3-resolver s2
4-gnd
5-can h
6-12v supply (ign?)
7-motor temp sense
8-nc
9-resolver s4
10-gnd
11-can l
12-coolant temp sensor
13-nc
14-resolver s3
15-coolant temp sense gnd
16-resolver r2
17-nc
18-power on power supply
19-resolver s1
20-motor temp sense gnd
21-nc
22-resolver s1

Nissan are notorious for mirroring the pins in their diagrams (don't ask how i know!) so I think your diagram is correct but nissan diagram is mirrored.

[Jack Bauer]

Got the gen3 inverter on the bench today and did a can log :
https://github.com/damienmaguire/Nissan ... 3/CAN_Logs

This dbc from Dala does decode some of the info:
https://github.com/dalathegreat/leaf_can_bus_messages

[SciroccoEV]

For those who've cracked their Gen3 inverter open, what's the voltage rating on the main filter cap? Is there anything else obvious that would limit the bus voltage. I've picked up a set of 18 6S1P CALB modules, which would give me 394 Volt nominal. I've got the possibility of obtaining more and pushing the voltage even higher.

[paaa]

There is some more pictures of a damaged inverter and PDM here:
https://www.diyelectriccar.com/threads/ ... 31&slide=5

[celeron55]

Interesting. The PDM looks electrically the same, but they've moved everything around including PCB locations and shapes. Maybe to make it a bit flatter?

[johu]

Some good weather today to I went out to the shed. Powered up the inverter and supplied the DC bus with a humble 40V/5A power supply. Connected an asynchronous motor and spun it up in open loop mode to 20 Hz. That pretty much maxed out the supply but I can say at least that worked.

Next I'd have to find the resolver connectors for the Prius transaxle, use batteries as supply and then run the Prius motors back to back. I should also at least fill the inverter with coolant.

[drprox]

Here are some EV-CAN logs I took from a 2018 Leaf if it helps development:

https://github.com/doctorprox/Leaf_CAN

[celeron55]

Here are some EV-CAN logs I took from a 2018 Leaf if it helps development:

https://github.com/doctorprox/Leaf_CAN

Very nice. Do you have any notes about what you were doing with the car at which points during the logging?

[drprox]

I do for 'evcan3.txt":

Code: Select all

Start Log
Turn on vehicle
Accelerate
Regen
Accelerate
Regen
Turn off vehicle
Stop log

I still have the car so can do some more if there's anything in particular needed.

I'm curious as to whether this one is torque/power limited in the firmware or if its the same inverter and FW in the 'E-Plus'

[celeron55]

I think that should be enough for figuring out the inverter protoocl.

Have you considered logging AC and DC charging? That would allow reuse of the charger components in a conversion.

[drprox]

Got some more logs today of both AC and DC charging.
File size is too big for Github so just need to find an alternate. Any suggestions?

In the meantime, I plotted a nice acceleration to give a bit of an idea of what the torque/power output of this motor/inverter combo is. It seems torque limited up to about 3500rpm. Not sure what that is in vehicle speed yet, probably around 30mph.

Pretty consistent at 110kW output as rated, with approx 120kW draw on the battery.

[celeron55]

The logs might compress better using 7zip using its LZMA format. Worth a shot.

[Jack Bauer]

Sent you a pm for wetransfer. I'll split out the logs and upload to my github.