Outlander FEMCU (front inverter)

[joe4agze]

Thanks for the information people. This is very interesting.

[jamesn]

Finally got the front drive system working.

I had to give up on the Huebner boards in the end, the first board's 3v3 power regulator went haywire and fried the STM32. The second board I finally got working and managed to tune Syncofs to a pretty good spot with it spinning up fine on manual ID/IQ but when I switched over to using the throttle it went into DESAT and since then any attempt to make it spin even in manual mode resulted in DESAT.
So anyway I had to move the boat and rather than stripping things down, ordering new boards, etc. I just went down the CAN control route using an Arduino on a new inverter:





Cruised for well over a mile with no issues from Meiden inverter or Arduino code. Happy days

[arber333]

Really nice! Congratulations.
Can you share your Arduino code? I would like to begin work on rear inverter and it seems better to begin from a known step.
What limits do you take for torque? Do you stop at 200Nm? This would then mean 0x39 and 0x2A for 205Nm.
For precision we would need to count from B towards A. Like when B reached 255, step counts 1 point into A and backwards...

tnx

[tom91]

Really nice! Congratulations.
Can you share your Arduino code? I would like to begin work on rear inverter and it seems better to begin from a known step.

tnx

I helped James reach this point by sharing a capture from my testing and timing of required messages, I would not advise using his code as a start it contains quite some errors, but will get motors spinning.

Main lacking item is the way the torque command is generated is very crude.

As I have mentioned in other posts, you only command torque from the motors. Thus direction is done via the commanding of positive or negative torque.

"Forward" = Positive torque is motoring, negative torque is regen
"Reverse" = Negative torque is motoring, positive torque is regen

[arber333]

Main lacking item is the way the torque command is generated is very crude.

Yes i see one would have to have really good throttle encoder to run motor with precision .
But maybe 10bit A/D from throttle pedal would be sufficient if i would only use 1/4 torque for testing first...
Regen could be used in larger steps...

[tom91]

Main lacking item is the way the torque command is generated is very crude.

Yes i see one would have to have really good throttle encoder to run motor with precision .
But maybe 10bit A/D from throttle pedal would be sufficient if i would only use 1/4 torque for testing first...
Regen could be used in larger steps...

No, as in the translation from throttle to torque is handled crude NOT the reading of a pedal, thats the easy part.

[jamesn]

Main lacking item is the way the torque command is generated is very crude.

Yes i see one would have to have really good throttle encoder to run motor with precision .
But maybe 10bit A/D from throttle pedal would be sufficient if i would only use 1/4 torque for testing first...
Regen could be used in larger steps...

No, as in the translation from throttle to torque is handled crude NOT the reading of a pedal, thats the easy part.

It would be far from acceptable for a road application. In the boat it works way better than I thought it would for some code I threw together in a couple of hours.
I have started on refining it with independant control but I have some more pressing tasks first including getting control of the e-golf BMS...

[Jack Bauer]

Is the info for getting the front inverter / motor running being shared?

[Jack Bauer]

No? That's ok I just reinvented the wheel :
https://github.com/damienmaguire/Outlander-PHEV

[Jack Bauer]

[chuuux]

Damien, could you be so kind and tell us the pinout for the front inverter's plug (power supply and CAN bus) The colors of the wires will be enough. I still have the 10-pin connector here. I'm almost sure this is Toyota 90980-12380.

[arber333]

There be one Jag battery module in the background there .

Seriously Damien. Can you tell us more about the HV/LV turn on sequence for the inverter?
Like do we need to connect 12V supply before main contactor/precharge is on or not? Or can inverter be plugged in permanently to HV without harmfull effects?
I need to test my Outlander rear motor setup and i just rather be informed .

tnx

Arber

[Jack Bauer]

The inverter is very like the Toyota in that it doesn't care about the power up sequence, Hv, Lv, can etc. You can stop sending can and it sits there until you send can again etc. I'll lookup my notes on the wiring but it just needs +12v, gnd and can.

[arber333]

The inverter is very like the Toyota in that it doesn't care about the power up sequence, Hv, Lv, can etc. You can stop sending can and it sits there until you send can again etc. I'll lookup my notes on the wiring but it just needs +12v, gnd and can.

Hi Damien
I am finally working on motor-inverter-charger-heater combo from Outlander and i have noticed there is none of the usual DC voltage jolt if i connect inverter directly. It remains weirdly indifferent. It would just report its status on CAN.
What do you think? Can i just use main DC contactor to pull in HV from the battery? Since rear inverter, charger/DCDC and heater are all connected together this means i will have to open DC contactor in every case, be it heating, charging or DCDCing .

[OutlandishPanda]

FYI, I bought this female connector kit for the front inverter comms:

https://www.aliexpress.com/item/1005001 ... f19c2GpBFL

Took about 2 weeks and fits perfectly and looks to be reasonable quality. How do they sell these so cheaply!?

[Bigpie]

Anyone used one of the front gearboxes with both motors in a car yet? I've got one in the garage, but just use a single motor on my existing gearbox.

[mjc506]

Not yet, but I plan to. I think using the second motor will require controlling the clutch. (Hydraulically actuated I think, so requires the front wheels to be spinning to build sufficient pressure).

May be possible to replace/modify the clutch so it's locked full-time?

[Bigpie]

I thought about opening it up and looking if it can be welded.

My concern would be over speeding based on the different ratios.

Motor - Differential ratio 9.663
Generator - Engine ratio 2.736
Engine - Differential ratio 3.425

[mjc506]

I think the PHEV runs in parallel hybrid (ie, clutch locked) at motorway speeds, so should be good. Gen-diff ratio is only 9.37, so a little less than the motor-diff ratio.

I guess a welded clutch would give you full torque from standstill. The stock arrangement you'd get full torque from both motors from a few mph as hydraulic pressure builds but you'd probably want to allow some margin to avoid slipping the clutch if you apply power to the 'generator' too early. (Easy to detect slipping - keep an eye on the ratio between motor and generator speeds)

All speculation of course I don't have a unit in front of me yet!

[Bigpie]

Would be interested in how you get on, the single front motor is a little wheezy on my car, once it gets spinning it's a little better. Temptation is to go with a rear motor instead, then I don't have to bother with the oil cooling.

[marucha79]

Hello @jamesn. Could you please share details about connecting Johannes board v3 to FEMCU? I'll try it despite yours experiences. I'm using Bosch SMG 180/120, maybe it will be easier.

[pxxmo]

where did the idea that they sit there a34j? There are drivers in my inverter that I can't identify

[nathaniel]

hi all,
i'm working on a conversion with the front inverter with generator as a motor. i can control the motor via the CAN using the arduino project of damian as a base. now i want some feedback information from the inverter about his status. i have seen in the STM-vcu project and the mini-E-VCU project that the rear inverter is putting out some information, and if i'm running the front inverter it is spitting out some CAN messages. now my question is, has anyone have already decoded these messages? i suspect that they are a bit the same as the messages from the rear inverter but with a different ID.

[marucha79]

I was trying to calibrate the current sensor with sinus load 9A (50Hz) and got such plot. Is it correct? I think that level of jitter (first there was no load) is too high, isn't it?

[Romale]

an interesting feature, today I combined the outputs of two power modules in the front outlander inverter into one, but I left the current sensors for only one module! thus, I deceived the system, the brain of O.I. V3 thinks that, for example, only 100 amperes at phase 1, but in fact exactly the same number of amperes is added from the second power module.
ideally, now this controller consists of a dual power module with a nominal value of 650 volts and 1200 amps, it seems to me that you can get a lot of power.
a little later, I will come up with all the necessary switching more culturally and a new cover for the inverter, install it on the car for testing and create a separate topic on this occasion.

Now I want to draw attention to one oddity - when the motor is connected to one half of the inverter (one power module), the motor has 2300 rpm and the syncofs 47700, if the motor is connected to the second module, it will be the same.
but when I connect two in parallel, the syncofs require 50500 and even the maximum speed is 2550 rpm!!!
in all these cases, the attenuation current is 0

one brain controls two power modules in parallel by taking current information from only one current sensor.

by the way, no spacers are required, if you solder the wires directly into the board, then everything fits perfectly under the native board of the inverter.