[WIP] Volvo V50 upgrading to Brain board

[JanJouke]

I have an broken electric Volvo V50. This car is converted in 2011 and the company that did this does not exist anymore and the provider of the motor and inverter (Magtec) won't provide information. I got myself an Brain board rev 3 and am working on putting it into the car. Below the video I made so you can see what I have been doing. I got the motor spinning and the delta light test working.



My next challenge is to find out what kind of resolver or encoder is used on this Magtec P160/161.
It looks like 10 wires. 4 wires look like they are from 2 temperature sensors. The others I can't find what they do. They don't behave like Damien describes in the openinverter FOC Tuning Tutorial. Does anybody know how to find out what it is? Has anybody converted such an motor already to an Brain board?

It is an fun project, but very challenging for me.

[johu]

So that's what these brainboards are still being used for Very nice.
I guess the low boost was the culprit for now light, very little swing in the PWM.

I used one ULN2003 with pull-ups plus CD4049 6-channel inverter for converting 3.3V to 15V for some older Semikron drivers.

[JanJouke]

I tried with an ULN2003 and an CD4050. I did not got it working. That is why I went to the optocoupler.
By now I am connecting the rest of the wires. Is it normal that the brainboard loses most of the changed settings when it is shut down?
And even with direct power from the 12v battery the web interface keeps locking up and sometimes the wifi stops working and reboots after a while. The leds on the board stay working when this happens. And the brainboard keeps working (I can already spin the motor a little via the throttle pedal, even with an locked up web interface).

What would be a good option to measure the HV voltage? Should I use something like an HCNR201 (for the galvanic isolation) or just an simple voltage divider?

Has anyone an idea how I can figure out how the resolver or encoder works?

[johu]

Wifi lockup is often caused by instable power supply. You can unplug it and bend the pins a bit, that makes a tighter fit.

Damien made a video about resolver. To identify it you measure the 3 pairs of feedback wires. One will be like 10 Ohm, the other two 2x or 3x that. Those are the feedback coils

[JanJouke]

I don't think it is a resolver. Or at least not one like in the video.
Measuring the wires gives me:
Black gives no reading
Purple - Brown 429 K Ohm
Yellow - Brown 1,09 M Ohm
Yellow - Red 426 K Ohm
Orange - Brown 1,09 M Ohm
Orange - Purple 426 K Ohm
Red - Brown 1,09 M Ohm

Does anyone recognize this? The Magtec motor is from around 2011.
I think I have to open the motor. Only it is difficult to acces it in the car. I will have to remove a lot of parts to get to it.

[JanJouke]

I have removed the motor to open it for the encoder/resolver.
Does anybody recognize this? The ring is magnetic.
And the other big question is: Can this work with the Main board rev 3? Or should I try to fit another resolver?

[johu]

Hmm, could be hall sensors. Is it 3 pairs of wires?

If so it is unfortunately not supported

[JanJouke]

There are no obvious pairs of wires. I also found out that an magnetic encoder is not very accurate.

I will look for a resolver. There is space inside the motor where I can fit one. Or is there inside the motor to much interference? What specs do I need to look for to get it compatible with the main board?

[JanJouke]

I have mounted an MLX90830 on the motor and the magnet on the shaft. When dry testing it performed well.
But in the car some unexpected things happen.
All settings are default except the boost. When the boost is up and the motor in in forward or reverse, the motor starts spinning. It wont't react on the throttle.
When I look at the angle (from the resolver) it is stable in neutral and when I put it in forward or reverse, the graph shows that the motor is rotating. Even when it is not rotating (with low boost). It does not matter if I choose single, sin/cos ore something else. When I look at the direct signal from the resolver (with my scope), they are constant.
The only difference with the bench test is that I connected now 150V .
I updated from the 5.20 sin to the 5.32 sin, made no difference.

Does anybody have an idea what I am doing wrong?

[johu]

Are you running an IPM motor with sin firmware?
sin firmware always advances angle to create slip. It's not at all suitable for sync motors.

[JanJouke]

I updated to the FOC firmware, and now it is spinning with manualid. So my mlx resolverchip is working.
Next is tuning the motor as described by Damien.

This is the last video of my progress.

[johu]

Great progress! You can also read here https://openinverter.org/wiki/Using_FOC_Software (I just updated it slightly)

[JanJouke]

I still have issues with the communication. By now I have a separate power supply for the wifi module and the tx and rx are connected with wires. It worked for a while and than it started to give the same loss off connection. My laptop has connection with the wifi module, but the wifi module does not communicate with the brainboard. Got a little scary when I could not return the manual id back to zero with spinning up wheels.

I have for an other project the ESP 32canlite and the ADC BMS. It communicates with the BMS (only for parameters and not other things like the spot values). But when I connect it to the Brainboard, it finds nothing. I looked on various node id's, expecting it on 1. I shorted out resistor R6 (made no difference). Do I have to activate can on the brainboard or something like that?
Does anybody knows a way to get the communication stable? Via can or wifi?

An other strange thing is that when I start the car, the brainboard goes to "run". And when I turn of the ignition it stays on "run". I would expect an status like "wait to start". And when I disconnect the power on the positive side, it stays powered on. Removing ground and it goes off. This only happens when the status is "run".

Thanks

[johu]

Do I have to activate can on the brainboard or something like that?

You need firmware at least 5.27.R and canspeed 500 . All else doesn't matter.
You should also update the ESP32 firmware itself, it has some reliability improvements https://github.com/jsphuebner/esp32-web ... an-backend

An other strange thing is that when I start the car, the brainboard goes to "run"

So you must be pulling the start input high somehow