VW GTE 3.6Kw charger 5QE 915 682 AF reverse engineer

[Heye]

I can now reliably start a charge after resetting the OBC, it ends after a few seconds though. It seems to make unhappy noises while pumping quite a few amps into a 12V battery, once I get my EV street legal, I will test with its battery. It will take at least another two weeks up to that point. Also, due to life circumstances, I'm currently very limited in the time I can invest into this project.

I probably can't post the code I am using now here due to copyright issues and I don't want to endanger the forum. If you write me a PM, I can give you my E-mail-address.

[Heye]

I hereby declare this charger (682 BG) hacked!

No joke, it worked on the first try, once I had hooked it up to the HV battery. It seems that it does indeed have a minimum voltage, but 375V is higher than that voltage!

I will have to do some cleanup, some documentation and I need to find out which components can be muntzed. It would be great, if the charge door emulation were not actually necessary.

But so far, this seems really, really promising

[Heye]

Alright, here is a short update on the 682 BG with a few findings:

• For now, neither the door lock emulation circuitry nor the CP/PP signaling can be removed
• It's fine to always have 230V on the input, it doesn't need to be switched off
• Only these CAN-IDs are actually required to get the charger going (I didn't look at ODIS though and I don't know the exact coding at the moment): 0x97, 0x187, 0x6a3, 0x124, 0x3af, 0x3a6
• Damiens code works fine for my charger except for one message (ID 0x3af). I opened a Pull Request. It would be great, if folks with other chargers (of the MQB variant, obviously) could test the changes
• I don't know the current coding, but I will fire ODIS up again soon and then look it up
• Same goes for the exact circuitry and code of the door lock emulation. I think timing is important there
• The charger seems relatively easy to control. It doesn't require any ramps or startup procedures. It doesn't seem to keep much state except for when it's being provoked so much it requires a power cycle to resume work
• The voltage sensing on the charger is quite dodgy. Even if the feedback mechanism works, it should really get precisely measured voltages from the BMS.
• Again, at least for the variant I have, there is a minimum voltage and possibly a minimum current? Mine won't work in an open circuit or on a 12V battery, but it works fine on the 104S battery (I haven't tested the upper voltage limits yet)
• For continuous charging, water cooling definitely is a requirement. I Frankensteined a PC AIO water cooler to the OBC and despite the pump being really weak, it would stay around 40°C (as measured by the charger, likely actually higher) while charging at 2800W, compared to >60°C (at which point I stopped testing) without cooling at 1200W

For inspiration, here is the current structure in which I keep the charger. Water is below the electronics (except for the "reservoir"), the small compartment is for control circuitry and the large compartment on top of the OBC is for storing cables.

[RetroZero]

I like your frankenstein inspiration setup! Thanks for picking up & getting this to move along!

[Heye]

Thanks! It's been (and still is) a lot of fun. Especially now that I can observe the effects of my changes directly. Two more notes:

• I was wrong earlier. An 11k resistance between the charge port lock sense wires seems to work fine, the whole emulation circuitry isn't needed. I think it's possible to make the charger error out by requesting the charge port to lock.
• I replaced my EVSE with an Arduino-based PWM generator. Seems to work fine and makes the whole setup a lot cleaner, although I haven't yet found a better way to create -12V than with another 230V->12V adapter.

I will try to find out if switching to CV mode works, hopefully tomorrow

[RetroZero]

Did you try a charge session with a big resistance instead of a battery?
I'll try (over weekend) to look at what happens if charge port lock is requested and report back .
I can only lock the charger port lock. It doesn't return the signal that it is closed, so cannot be opened....i can never get actual value into lock status....will poke around with that if I have time.
Btw, I also get the 3rd LED fault along with charger flap malfunctions, and that is with using a full stock charger port with harness and lock mechanism, so wondering if its not due to other missing messages that it throws these faults.

[Heye]

I don't have a suitable resistor, but I would be quite surprised if it works. In an open circuit, I would have expected to see a least a blip of some voltage on the oscilloscope (or a change in the maximum observed value), but that never happened. This makes me think that this type of charger requires at least some voltage on the output (although, weirdly, it would work for a short time on 12V, so I might be wrong).

Testing on a real battery was the breakthrough for me, so I definitely recommend doing that if it's possible in any way.

For me, the lock is behaving weirdly and not like it should, according to the pinout in the Wiki. Now, that I think of it, maybe I was experiencing the same problem? I remember that I was only ever getting 12V on one of the lock/unlock pins. If there was no mechanical change in the lock (is your harness and charge port from the same vehicle as the OBC?) it's probably likely that some messages are missing/wrong.

[RetroZero]

My lock & harness are from the same make/model vehicle as the OBC according to cross references.

I've got a prius battery that sits at about 170vdc , but hope to do tests with resistance first.