openinverter forum

First road trip with the new pack today. I drove about 250 km in one piece with 150 Wh/km. Hypermiling with 100 km/h. Started with 95% and arrived at the charger with 27%.
Unfortunately had a VCU reboot after 100 km. Finally caught the flag: IWDG - watchdog reset. Hmm

Charging was pretty smooth and started at 80 kW. I think I can optimize power some more, it throttled too early. But then it was me who throttled. As I never trust my own wiring-foo I stuck my hand into the HV loom and noticed the B- cable got uncomfortably warm up to the point where it started emitting smells. So dialed back charge power to 15 kW. Cable quickly cooled down and I went back up to 130A/50 kW. That went well enough to finish the session. The inconvenient consequence: I have to drop the lower 4 battery modules and redo the bolt of the upper 4 modules.

Apart from that the modules gained 10°C of temperature during the 250 km drive (8 to 18°C). They heated up pretty equally to 35°C during charging and then cooled back down to 33°C over the remaining 100 km. Outside temp was 10-14°C. So I think for moderate trips this should work well.

Back from the trip and have some data to share.
I drove the car lightly to the first charge stop and also topped up 9 kWh while visiting Udo (that 3.3 kW OBC is really inappropriate now). I started in the morning with 13°C in the battery and arrived at the one and only quick charge stop with 22°C after 284 km. Again the battery reached 36°C from charging 15 kWh with about 60 kW average charge rate. I throttled it straight from the beginning as I had to leave the car alone to take care of my own business.

The temperature then dropped to 34°C while b-road driving. Then on A49 I drove a bit more spirited with 140 kph cruise control and ended up back home with 40°C in the battery. So 110 kph it is for long range driving until I improve battery cooling.

Temperature is spread evenly across the modules, the warmest one was 41°C and the coldest one 38°C.

I'm no longer 100% sure what it is that starts to smell during a quick charge session. It could also be my 50cm extension lead from the MG charge port. Will check that at some point for signs of over temperature. So maybe I don't need to drop the battery for just that. It seems impractical to charge at 100 kW anyway with a lack of battery cooling.

Next longer trip is to Augsburg via Frankfurt, 550 km. That probably needs 2 charge stops and will supply us with more data.

johu wrote: ↑Fri Mar 15, 2024 7:01 pm I drove a bit more spirited with 140 kph cruise control and ended up back home with 40°C in the battery.

Maybe you can install couple of the ATV fans to blow air across the assembly during charging AND driving? I noticed a lot of difference if i removed the battery box lid while driving. temperature actually wend down 6deg!
The reason to use those is they are waterproofed and resistant to outside use.
This is the fan i am thinking of using in my boxes. It draws about 0.6A at 14V.
https://www.aliexpress.com/item/3301573 ... 1802tPuJlT

This blowers would also be worth trying especially as you can more efficiently direct the airstream over modules...
https://www.aliexpress.com/item/3302106 ... ry_from%3A


I also have this one, but even though it blows quite good it draws 2.5A!!!
https://www.aliexpress.com/item/1005005 ... ry_from%3A

I have such a fan installed inside the battery case.
https://aliexpress.ru/item/32797575257. ... 3931858497

Maybe getting some stick on temperature labels will be good for your testing It "logs" highest temps.

https://uk.rs-online.com/web/p/temperat ... ls/0285914

midway wrote: ↑Sat Mar 16, 2024 7:12 am I have such a fan installed inside the battery case.
https://aliexpress.ru/item/32797575257. ... 3931858497

Nice find! I have like 3 separate module locations this would really make it easier to mount... thanks.

tom91 wrote: ↑Sat Mar 16, 2024 9:52 am Maybe getting some stick on temperature labels will be good for your testing

Good suggestion! Was thinking about some 1-wire sensors but this is even easier.

midway wrote: ↑Sat Mar 16, 2024 7:12 am I have such a fan installed inside the battery case.

Nice unit.
Can it blow air out of the case? How do you make sure no water gets in?

johu wrote: ↑Sat Mar 16, 2024 12:41 pm Good suggestion! Was thinking about some 1-wire sensors but this is even easier.


Nice unit.
Can it blow air out of the case? How do you make sure no water gets in?

I plan to route the air ducts under the hood or into the trunk

Hi juhu,

maybe I missed the information in this really huge thread. How did you deal with the two signals of the generator DFM and L? I know that L is used to activate high load devices like seat heating or the rear window heating.

bewo wrote: ↑Fri Mar 22, 2024 9:00 am How did you deal with the two signals of the generator DFM and L

I wasn't even aware these existed. But rear window heating works and also power steering. Power steering just checks for motor rpm, that's why I simulate idle speed of 700 rpm. Maybe heated window does the same?

Despite the replaced wifi module and despite giving the scheduler top priority the VCU restart happened again today. I almost have a hunch it happens after a fixed time now but that is very vague.

The reset cause is IWDG, so watchdog reset. There'd have to be some exception or endless loop to cause this...
If anyone wants to follow along: https://github.com/jsphuebner/stm32-car/tree/touran-meb

Reboot problem solved: viewtopic.php?t=4925

With the above established I want to return to the original questions and the (partly premature) conclusions.

So we start with a perfectly working car and a terminal class that remained largely unchanged for years. After fixing a completely unrelated problem with the battery the VCU starts to do spurious resets with the same software as before.

I then suspected the power supply of the VCU to be the issue as some cables got caught between battery and chassis and were shorted to ground. But never found any further evidence. I then scoped the supply rails of the VCU (3V3, 5V, 12V) only to find nothing spurious at all.

Meanwhile the software got a major overhaul to support the MEB BMS. The problem did not go away.

Then I swapped the ESP module for the new one and seemingly the problem was gone. In reality it only became rarer and I explain this with the poor wifi performance of the new module. It just didn't manage to refresh every 500 ms.

One thing I still don't quite understand is the behaviour of the ESP module. Normally, when the STM32 resets and the ESP does not this results in the ESP expecting 921k baud rate and the STM32 running at 115k. So the dashboard should stop updating in this case. But it keeps updating after the reset. This lead me to believe in the supply rail issue. Namely I thought the ESP got into some state where it briefly shorts out the 3V3 rail (it must dip below 2V for the STM32 to reset). But shorting out an 800mA LDO takes some grunt. So maybe the STM misbehaving led to the ESP resetting as well? Why?

Also the supply rail theory was contradicted by the watchdog flag being presented as a reset cause. I did a bench test were I simulated supply rail dips to find out whether this could falsely be detected a watchdog reset. But couldn't find proof for that.

Well the 500 km drive to Augsburg will show whether the problem has really gone away now. Anyway it was a good chance to harden the terminal.

Yay, road trip reporting time Another 1000 km covered going to Augsburg and back.

Key takeaways

• The VCU restarts no longer occured
• CCS charging was absolutely flawless except one session getting interrupted when someone used the other outlet on the same charger. Could restart no problem
• Battery rapid-gated badly especially on the return trip

Drove the first leg (about 200km) to Frankfurt and while waiting for Janosch recharged the car to 90% (added 33 kWh in 45 mins). Then came Stau time and we were stationary for like 90 minutes total. Surprise: contrary to oily propaganda an EV does not discharge while stationary

Went for another charge stop about 200 km later and added 18 kWh in 20 mins before rapid-gate. Charged another 10 minutes before arrival because I wasn't sure of the accuracy of my SoC calculation and we thought it was a bad idea to test it at night and freezing temps On the return trip I did a nice 28 kWh charge session in 30 minutes heating the battery from 18 to 40°C. After that it hardly cooled down because consumption was rather high. Rain, headwind, 3°C ambient and high-ish speed (120 kph) did not help. So the second charge session heated the battery to 50°C and completely ramped down charge rate to 0 already at 45% while I was eating a burger. So instead of going straight home I had to insert another 5 minute stop once the battery had cooled down a bit.
I have also moved the derating point for the discharge current from 50°C to 52°C so the car is still somewhat drivable after overheating from charging.

More on the event that we intended will be published in the Events section

Did some tests with different wheels:


On other news I've replaced the rear springs and front shocks and springs as well as the crusty hatch

Not sure where this post fits best as it is kind of a road trip, EV improvement and CCS mix.

So, I went to Hamburg to meet Mathieu who has done various conversions like the Mustang and Toyota Land Cruiser. Plan was to get CCS up to speed on the latter and fix a few things on Touran.

I wanted to make it to Hamburg (300 km) on a single charge but 40 km before my destination and with 18% left I chickened out and plugged in for a few minutes. When I arrived the battery had time to rest and let the SoH estimation run: 100.5% That is with 155 Ah configured (158 Ah is spec). So the 18% were accurate and I could have made it on a single charge. Also I didn't start at 100% but 91%.

Anyway, I met Mathieu the same night, we updated firmware on his Foccci and went to Ionity which hadn't worked previously. Now it did: Now he can use up the 1500 kWh that he won on the 2022 e-cannonball

The next day we fitted a pull-up resistor to 5V to the PP line and now Tesla also worked: There was no lack of spectators, Tesla staff from the adjacent Tesla service center.

Then we changed the CAN mapping to use 9 bits for the current request to be able to charge at more then 255A which also worked: We saw 300A briefly but the battery was too full.

On Touran I redid the extension cables from the CCS port, they are 50mm² all the way now. Then I cut open the battery box a bit more to reach the upper battery pole bolt that I suspected wasn't tightened. We made a special T30 tool and I gave it a slight twist. It should be at around 10 Nm now which is spec for an M6 bolt.

Finally I fitted a fan that is supposed to flow air in through the inner two BMS cable holes and exit through the outer ones. I stuck some corrugated tube into the outer ones as an air guide. Indeed there is a small air flow when running the fan. To be tested.

On the way back I drove in heavy rain and encountered 3 accidents. With improved confidence I emptied the battery down to 9% or 3450 mV per cell. Perfect balance down there, just 3 mV delta, likewise at the top. So they are really well matched.