Tesla Large Drive Unit Support Thread

[mohalkh]

hello everybody

my question is can I replace the brain of the Tesla large drive unit (LDU) that is used on Mercedes b250e with the Tesla™ LDU drop-in logic board that found on the openinverter.org/shop, AND get the same function on the car?

my board brain looks like the on in the attached image

[johu]

You'd have to replicate the CAN communication between the drive unit and the rest of the car. So not plug&play and certainly not a beginners project.

[johu]

At this point I want stress again that contacting the developers privately without prior support time payment is one of the big DONT'S of this forum and hurts your reputation

[mohalkh]

You'd have to replicate the CAN communication between the drive unit and the rest of the car. So not plug&play and certainly not a beginners project.

Will paid support help me to replicate the CAN communication between the drive unit and the rest of the car?

[johu]

Will paid support help me to replicate the CAN communication between the drive unit and the rest of the car?

It is likely quite a rabbit hole with encryption going on and I have no special knowledge around B class. All I know is how to record CAN messages, how to make sense of them and how to configure the inverter to sort of replicate them. Everything else you'd have to find out and one possible outcome is that it just turns out unsolvable - as it is with experiments

[mohalkh]

Will paid support help me to replicate the CAN communication between the drive unit and the rest of the car?

It is likely quite a rabbit hole with encryption going on and I have no special knowledge around B class. All I know is how to record CAN messages, how to make sense of them and how to configure the inverter to sort of replicate them. Everything else you'd have to find out and one possible outcome is that it just turns out unsolvable - as it is with experiments

thanks for replay. I guess I'm in trouble. i feel I'm losing my car

[johu]

If a broken logic board is your issue maybe someone can help you out. Just post in the classifieds forum.

[leman2112]

I have an unbroken logic board for the mercedes b class. I am not using it.

[leman2112]

I doubt his is broken though. probably just wants to use the opensource logic board like me. oh, he wants to fix his mercedes b class and get the whole car running again? I misunderstood the question at first I guess. But yeah I got the mercedes b class board if ya want it.

[leman2112]

anyone know where I can find the pinout for the jst connector that comes off of my opensource board.

JST 24CPT-B-2A

trying to start soldering from the jst connector and harness to my 23 pin ampseal connector but not sure how the pinout is ordered. I do have this information.

[fredd90]

anyone know where I can find the pinout for the jst connector that comes off of my opensource board.

JST 24CPT-B-2A

trying to start soldering from the jst connector and harness to my 23 pin ampseal connector but not sure how the pinout is ordered. I do have this information.

Did you not save the original cable between the old "Logicboard" and ampseal connector? i just replaced the board and all worked fine, do you want to make a new harness?

[jon volk]

anyone know where I can find the pinout for the jst connector that comes off of my opensource board.

JST 24CPT-B-2A

trying to start soldering from the jst connector and harness to my 23 pin ampseal connector but not sure how the pinout is ordered. I do have this information.

Did you not save the original cable between the old "Logicboard" and ampseal connector? i just replaced the board and all worked fine, do you want to make a new harness?

He's messing with a motor from a Mercedes installation.

[fredd90]

anyone know where I can find the pinout for the jst connector that comes off of my opensource board.

JST 24CPT-B-2A

trying to start soldering from the jst connector and harness to my 23 pin ampseal connector but not sure how the pinout is ordered. I do have this information.

Did you not save the original cable between the old "Logicboard" and ampseal connector? i just replaced the board and all worked fine, do you want to make a new harness?

He's messing with a motor from a Mercedes installation.

Okay that explains the question

[rjmcdermott81]

Looking to see if anyone has ideas to what is going on. Here is a rundown of where I'm at:

What is working:

• Board is coming online
• Negative contactor throws on ignition
• Precharge throws right after
• Main positive contactor closes on start signal
• Gear selection works
• Throttle gets a response

Issues:

• Motor runs but clunks. Sounds like what people described was happening when the A/B signals were swapped
• I found that two of my encoder plug wires were damaged during my first runs and some of the insulation may have come into contact with the shielding (this is because of several painful hours of tooling with that ampseal connector)
• MAIN ISSUE - every time I run it, it overheats whatever wiring I have connect to the high-voltage side. This last run I had thick wiring so I'm worried there is a bigger problem

Here is a summary of our test runs:
1. 80v lithium ion battery - ran for a few seconds and then blew the internal fuse
2. 12v (thought backing it down would be safer) - melted 20 awg wire from negative
3. 12v - melted 16 awg wire on positive side. In this run I recorded a [1]: WARN - PWMSTUCK (parameter file attached)
4. 38v - unplugged encoder, ran in manual mode. Clunking during turning and smoke coming from jumper cables (this is the video below). No error message.

In the video below I walk through all my wiring at the end. I've quadruple checked all my wiring but I'm a novice at this so if there is something wrong please point it out. Excuse the messy wiring, I made the harnesses to length in the car and they are a mess with this test setup.

I took pictures of all the IGBTs when doing the board installation and I didn't see any discrepancies.

Video of the last run along


Picture of damaged encoder cable

[rjmcdermott81]

Sigh...well I dropped the drive unit out so that I could take the cover off and inspect it internally. I see small bubbles in the gel in the IGBTs, very small. No discoloring but I can see them. It's impossible to get a good picture of them because of glare. I don't think they were there before but maybe now that I'm looking closer I see them, who knows.

I'm looking for any suggestions on my next course of action. I assumed I kept melting wires because I wasn't accounting for the current that the tesla motor would pull but maybe there was another issue brewing. Is it possible that my encoder signal was all sorts of wrong so it kept telling the system to pour more current in to overcome the issue and then this lead to thermally stressing the IGBTs?

As of last test the motor was turning in manual mode, albeit clunkily.

Impossible to see in this photo but there are tiny tiny bubbles in the gel:

Bratitude mentioned I should inspect the HV connections. Cursory look I don't see any obvious issues. I assume if there was a short here it would be present the second I applied HV and not be related to happening only when I press the throttle.

[Boxster EV]

I just posted in your project thread before noticing this.

Is the encoder definitely wired correctly?

Does it still spin okay in manual mode?

[jon volk]

Really needs a solid 40 and 100v dc power to turn over, closer to 100 is probably better. The motor will metallic clunking noise from the gear train if there's no load on it it. Even though you're just testing, put some 1/0 or 2/0 cable on it. Last, set UDCNOM to 0 or close to what your actual voltage is for testing. A large deviation from that is going to be manipulating values in the background.

Small bubbles in the good shouldnt really be cause for alarm. A large quantity of bubbles or ones that look like it captured a pressure wave from an exploding IGBT are reason to be concerned. If one was blown, you'd be seeing a DIN overcurrent fault.

[rjmcdermott81]

Boxster EV and jon volk - thanks for the input. I'll upgrade the wiring and get some more powerful batteries lined up. I've been a little spooked about adding more power but I can see how it is necessary.

I am going out of town for the week which will give me some time off from the project and I'll order up the wiring. Does anybody have ideas about low cost battery solutions? I don't want to spend a bunch before I get my batteries from ElectricGT. I suppose I could buy a single volt battery of eBay.

[collin80]

Boxster EV and jon volk - thanks for the input. I'll upgrade the wiring and get some more powerful batteries lined up. I've been a little spooked about adding more power but I can see how it is necessary.

I am going out of town for the week which will give me some time off from the project and I'll order up the wiring. Does anybody have ideas about low cost battery solutions? I don't want to spend a bunch before I get my batteries from ElectricGT. I suppose I could buy a single volt battery of eBay.

What I do for inverter testing is use a bridge rectifier. You can buy them in a pack that looks a bit like an IGBT module. Here's an example of what I'm talking about: https://www.ebay.com/itm/YueYang-MDS150 ... Sw1XNerxpn

That's even three phase so you could conceivably give it 240V 3-phase and really get some nice power out of it. With single phase 120V AC you get an output of about 167V DC. This is sufficient voltage to spin an unloaded motor and best of all it's fairly fail safe as drawing too much power basically just pops your breaker. I haven't done this with a Tesla drive train but I have with other inverters. For bench testing a rectifier tends to be OK. For any sort of load you can't use one. You do need a reasonable voltage to get the magnetic field going so others are right, 40V is probably even too low, 100 is better, 160V+ with limited current isn't so bad either. The real danger in using a battery for testing is that you probably won't fuse it and then the drive train could draw huge current and blow things up. You just don't have such a big problem with that when you use a rectifier. Worst case, you ruin your $14 rectifier. Oh well.

The tesla drive train might require so much magnetizing current that it pops the breaker just with that. But, I kind of doubt it. The drive train can use a lot of power but it shouldn't take too many amps to get the field going.

[FJ3422]

Can you measure the resistance between the motor phases, and the HV+ / HV - rails ? To make an accurate comparison between the six banks, it can be useful to temporary loose the motor phases. Large differences could indicate blown igbt’s.

I got my LDU running fine at 36V (24+12V lead-acid), current was around 2-4 Amps with the motor running at around 1000-2000RPM. Unfortunately I can’t add a video here as evidence

*edit*

The noise you hear is the oil splashing around in the gearbox.

[fransoa]

Hi,

I have a problem with an LDU in cold environment. It is about -6°C and the inverter indicates these temperature values (see photo). When the temperature is positive it works much better. I didn't find the software part that was a problem. Have you got an idea ?

[johu]

Indeed, a bug in the temp calc module. The lookup table only reaches down to 0°C for the Tesla heat sink. When lower, it should just display 0. But instead i does funky stuff. Will be fixed in the next release.

[rjmcdermott81]

Issues:

• Motor runs but clunks. Sounds like what people described was happening when the A/B signals were swapped
• I found that two of my encoder plug wires were damaged during my first runs and some of the insulation may have come into contact with the shielding (this is because of several painful hours of tooling with that ampseal connector)
• MAIN ISSUE - every time I run it, it overheats whatever wiring I have connect to the high-voltage side. This last run I had thick wiring so I'm worried there is a bigger problem

A quick follow-up for the community on my issue in case others have the same symptoms. I did a support session with Johannes this morning and it was incredibly helpful. He connected to my board from the other side of the world and we did some quick troubleshooting. (Johannes chime in if I get anything incorrect)
The root cause was that my board wasn't getting any signal from the encoder at all. This was a quick fix because I was suspicious that I didn't have the pins fully engaged. It's possible that because there wasn't a signal coming from the encoder, the motor kept pulling more current to spin which overheated my under-spec'd wires.

We ran the test with 60v (5 x car batteries) and 2/00 wiring on the HV side. Everything is now working great.

In hindsight, if I had to do it all over again I would pay for a few hours with Johannes to be there for the first test runs. Really great resource to have. Thanks Johannes!

[johu]

Thanks for the testimonial Yes that was the root cause. All parameters were correct.
It's really cool to be able to offer world wide support while sitting in my office.

[Boxster EV]

Issues:

• Motor runs but clunks. Sounds like what people described was happening when the A/B signals were swapped
• I found that two of my encoder plug wires were damaged during my first runs and some of the insulation may have come into contact with the shielding (this is because of several painful hours of tooling with that ampseal connector)
• MAIN ISSUE - every time I run it, it overheats whatever wiring I have connect to the high-voltage side. This last run I had thick wiring so I'm worried there is a bigger problem

A quick follow-up for the community on my issue in case others have the same symptoms. I did a support session with Johannes this morning and it was incredibly helpful. He connected to my board from the other side of the world and we did some quick troubleshooting. (Johannes chime in if I get anything incorrect)
The root cause was that my board wasn't getting any signal from the encoder at all. This was a quick fix because I was suspicious that I didn't have the pins fully engaged. It's possible that because there wasn't a signal coming from the encoder, the motor kept pulling more current to spin which overheated my under-spec'd wires.

We ran the test with 60v (5 x car batteries) and 2/00 wiring on the HV side. Everything is now working great.

In hindsight, if I had to do it all over again I would pay for a few hours with Johannes to be there for the first test runs. Really great resource to have. Thanks Johannes!

Is the encoder definitely wired correctly?

So I was right for once.

Glad you got it sorted.