openinverter forum

So I found my CAN dongle also featured a 120 Ohm termination resistor. With that switched off and slight rerouting and shielding of the CAN cables everything is back to normal.
I've closed off the exhaust tunnel and battery pack last Saturday, look really slick Pictures later.
Now I started working on the BMS again. I want to find out whether it balances. The Delta between lowest and highest cell is like 90mV. So now I connected permanent 12V to Pin 1 and bridged Pin 5 and 7 via Nissans safety disconnect gadget. Supposedly that should be enough to start balancing, right?
The heater relay is not connected, could that cause issues?

Some good progress today.
I permanently fitted the Meanwell into the original battery box together with 4 40Ah LFP cells and a chunky 200A 12V relay. The latter is needed to interrupt the connection when the car is off, as the Meanwells output will sink 200mA from the battery. Since the relay sinks 400mA I used a P-MOSFET whose gate is pulled low (i.e. on) by the DC switch OD signal.
Also boxed the DC connection to the Nissan inverter. The original DC cable is held onto the inverter with a bracket that also grounds the shield.

And finally here are some pictures from the workshop when we installed the underfloor cover.

Magically the Nissan BMS now no longer displays the interlock error. The interlock bit is still low on the CAN bus but DTC is gone. Hmm. Not sure how to figure out whether it balances. People on the web seem a bit unsure. For the Leaf BMS fanboys: I'm reading group 6 but it seems to return all zeros.

Well it's not all magic, I actually traced the interlock signal. It is a 3.3333333Hz 50/50 PWM on the source. It is then routed via the various interlock switches back to the LBC, through a transistor to the MCU. Behind the transistor the signal was swinging between 4.5 and 5V which didn't seem right. So I beefed up the base voltage divider. Now a nice 0-5V swing is generated. Not sure what the deal is.

Anyway I'm getting a meaningful charge and discharge limits now. It's about 1°C outside and when I started I had a 3kW charge limit and a 75kW discharge limit. Just by waiting a bit it climbed to 6kW/90kW.

Did some semi-fast charging today It's the remainders of the EMW charger but sporting an inverter mainboard for control. It has a nice 3-phase rectifier and the inductor is good enough for 30A. So about 11kW at 380V.
As there is no charge control I have to put the car in drive mode. Of course one time I forgot an fried the precharge resistor
But then it worked. Basically it logs into the car wifi and reads the Nissan BMS charge current limit from there. Since it's quite cold the limit was 9kW and it dropped to 3kW as the battery filled up.

johu wrote: ↑Thu Dec 19, 2019 6:56 pm Did some semi-fast charging today It's the remainders of the EMW charger but sporting an inverter mainboard for control. It has a nice 3-phase rectifier and the inductor is good enough for 30A. So about 11kW at 380V.
As there is no charge control I have to put the car in drive mode. Of course one time I forgot an fried the precharge resistor
But then it worked. Basically it logs into the car wifi and reads the Nissan BMS charge current limit from there. Since it's quite cold the limit was 9kW and it dropped to 3kW as the battery filled up.

Very good!
I am doiung this every day now, except for those really cold days. I dont have benefit of Nissan BMS and have to gauge temp and decide how to charge. Also with minus °C i have to disable regen also as the battery Ri increases in cold and the whole pack voltage rises with regen and causes inverter to trip overvoltage limit.

Sunny day today, so I cleaned up the wiring a bit. I reused the 10-pole connector for the former exhaust sensor (whatever it sensed) that has 9 poles, 6 of which end up in the ECU, one being 5V hard wired, one GND. The other side conveniently ends up under the front of the car.
So I put CAN, Fuel tank simulator (now runs on ECU) and DC switch signal on there. The latter forks off to the DC switch directly, the remaining go to the fuel tank service hole under the rear seats. Back there CAN connects to the Nissan BMS, and the fuel tank simulation to the fuel tank controller.

End result: no more contention of the CAN bus when spinning the motor, one piece of electronics made redundant (the former fuel tank simulator) and two makeshift cables made redundant (former DC switch and CAN bus wires). Happy with that

Of course all of the fiddling has made one of the battery sense wires come loose, will fix that tomorrow.

Yay, video

Today I tested my heating concept, i.e. using motor and inverter to generate heat.
So I used the newly created manual mode that allows you to manually specify id and iq current. I cranked up id positive, i.e. created a magnetic field that is perfectly aligned to the rotor magnetic field. Field strengthening so to speak. I thought permanently countering the magnetic field with negative id might damage the magnets? I don't know.

Once you go past 150A current starts to show up on the DC side. With 250A I got around 6A DC. As no movement is generated, this should mean 2.2kW of heat. And indeed the motor started to heat up, Supposedly to 60°C (it's -1°C outside). Unfortunately not much warm air came of out the air vents. I mean it is slightly warm but nothing like the direct air heating in Polo which is also 2kW. When I switch on recirculation no air comes out at all, I wonder if something is stuck? There is no manual control but Climatronic, not sure how far it opens the valve on the heat exchanger.

johu wrote: ↑Sat Jan 25, 2020 12:25 pm Today I tested my heating concept, i.e. using motor and inverter to generate heat.
So I used the newly created manual mode that allows you to manually specify id and iq current. I cranked up id positive, i.e. created a magnetic field that is perfectly aligned to the rotor magnetic field. Field strengthening so to speak. I thought permanently countering the magnetic field with negative id might damage the magnets? I don't know.

Once you go past 150A current starts to show up on the DC side. With 250A I got around 6A DC. As no movement is generated, this should mean 2.2kW of heat. And indeed the motor started to heat up, Supposedly to 60°C (it's -1°C outside). Unfortunately not much warm air came of out the air vents. I mean it is slightly warm but nothing like the direct air heating in Polo which is also 2kW. When I switch on recirculation no air comes out at all, I wonder if something is stuck? There is no manual control but Climatronic, not sure how far it opens the valve on the heat exchanger.

Try to see if you have a stuck flap somewhere. Maybe you are using external air so the heater doesnt have enough energa to go past the water Cp.
Other cause could be plugged up heater. Those things have a spiral wire inserted into its tubes. In case you use various brands and mixtures of glycol sometimes pieces of fat and debris get stuck inside the heater and it is not functioning. Do you get warm liquid comming out of the heater tubes on the other side?

When i installed webasto on my Pug i happend to put a right angle tube on the internal heater from the outside and this tube was return flow to the webasto heater. Well i pulled too much tube towards webasto and i pinched it just below the heater output alu tubes. This effectively closed the flow and i was wondering why i dont get any heat inside. Also webasto was shutting down to protect itself. When i found out i refixed the tube and everything was ok.

A

Hi Arber,
it seems that water flows through the heater. I have the de-airing pipe return to the reservoir on the heaters exit. And that happily pumps water into the reservoir.
The flaps make continuous noise though. Maybe something is stuck. Will have to see how to access that stuff.
EDIT: you can actually here the flap noise in the video when I explain the plot.

On other news the BMS is now fixed and I immediately started a charging session:

Maybe it is not the motor that is stuck but there is something behind flap obstructing the movement.
My Pugs flap for knee valve wont go into position fully but only partially. So i get the knee airflow only if i fully close the dash valves. I am inclined to think that car was static for so long that something crawled in and died in there.
It happened with our airplanes parked outside lots of times. One time a pilot took a whole mouse family for a trip and when he landed he saw them running across the runway .

Did a bit of test driving today in 3rd gear. While I was at it I created a power plot as reported by the BMS. Reached the original Leaf power now, running 450A motor current Still improvements needed in the higher rev range, the car becomes sluggish. I could however record the positive influence from the second field-weakening controller. It increases -id if iq becomes too low (unwanted regen). It certainly dampens the effect. Without it I got 100A of unwanted regen above 7000rpm (setpoint 0A). With it set to -70 that is dimished to 20A.

Also fixed cruise control. All throttle action is now governed by the ramp. that way cruise control can't accelerate so hard that the current limit is hit.

Nice! Is 450A what the Leaf pulls stock? I always thought it was 200A due to battery limitations. There's at least a 200A fuse in the disconnect, but I'm not too familiar with the behavior of fuses at this size.

I'm assuming they pull something around this (or more) till 3000 rpm and then start limiting. 450A is AC, only a fraction governed by the sine amplitude shows up on the DC side.

The German (or European) regulations say, if the converted car has ABS/ESP the traction motor must be integrated.
So there are two aspects: do not brake the wheels too strongly with regen and do not generate excess wheel spin.

Since all 4 wheel speeds are on the CAN bus this is not too hard to do. The inverter already has a "throtmax" parameter that lets you limit maximum throttle. I added throtmin that limits regen throttle.

So I calculate the average axle speed (front, rear). Since it is a front wheel drive car, if the front wheel turn faster than the rear wheel we have wheel spin, otherwise we have too much braking.

I defined a maximum allowable wheel spin and erm... wheel-unspin and if that is surpassed accleration or regen is limited, respectively.

Tested on the wheelspin part on the driveway, it does the job. Will test on a free road stretch later.

Code on github: https://github.com/jsphuebner/stm32-car

So I decided to put the ChaDeMo port where the fuel filler used to be. Decided against putting it in the front because:

• Wasn't really wheather proof
• Looked strange
• Would have needed to turn the main contactors on for charging

So now it's in the back. I had to make some sacrifices. It's twisted 90° because it wouldn't fit otherwise and the lid doesn't fully open.
Is the twisted socket a problem at the charging stations? And do I need to work on the lid to make it open fully? Would it be a problem if I remove the lid?

Sideways socket, I think no problems. A part from stress to connector.
The cables are usually long enough.
But also quite heavy/stiff.

But the plastic chademo port lid must open near 90degree for the chademo connectors to fit.

Could you fit it in 45° angle?

johu wrote: ↑Sun Feb 09, 2020 12:54 pm Is the twisted socket a problem at the charging stations?

The CHAdeMO cable is very heavy, i'd be surprised if this survives especially is someone knocks it.

johu wrote: ↑Sun Feb 09, 2020 12:54 pm And do I need to work on the lid to make it open fully?

Yes, some of the connectors are huge.

johu wrote: ↑Sun Feb 09, 2020 12:54 pm Would it be a problem if I remove the lid?

You'll need to keep the connector pins dry... the charger isolation test is quite sensitive to moisture.

Thanks for the quick reply

Yes I can imagine that these handles are heavy, just looked it up:



So I guess I'll need to suspend it with with a string... No way to turn that socket the right way.

Concerning the lid I should be able to chop away some plastic

Der Panzer has a Chademo port at a right angle. No problem using it at public charge points. What about a Chademo "exhaust" ? There is usually alot of free space under the rear where the back silencer lives.

I like that "exhaust" idea. If we can provide weather protection it may become standard.

I would rather put a L2 Mennekes socket in place of the fuel lid. Pretty much all fuel filler necks are similar size.
https://leafdriveblog.wordpress.com/201 ... onnection/

johu wrote: ↑Sun Feb 09, 2020 1:26 pm So I guess I'll need to suspend it with with a string... No way to turn that socket the right way.

If you assume someone will walk into the cable and design something to take the strain I guess you'll be ok. Damien has a right angle CHAdeMO connector on Der Panzer but that's very close to the ground so doesn't have the full cable weight trying to yank it out of the car. Of course the trip hazard potential is another issue that the courts will eventually decide

Oh I completely forgot the exhaust charge port
I guess I'll just hang a rubber band or some rock climbing gear from the roof rail should I consider the load to heavy. I'll mostly charge at home where I can make myself a lighter ChaDeMo cable.

Just worked on the battery cover a bit more. That middle rail will be replaced by a smaller one.

Cleared up the last logical details for ChaDeMo and a Guerilla charging port inside the car.
The 5V that power my custom ECU/VCU happen to be on the CAT5 cable that connects to the Nissan BMS. So if I inject 5V there the ECU wakes up and starts communicating with the BMS. Current draw is 500mA.
Also some other stuff seems to wake up that is on the 5V bus so some warning lights light up if the car has just been turned off and the instrument cluster is still powered. Oh well.
So now I can use the ECU to do the CAN part of ChaDeMo or, alternatively, talk to my homebrew charger via wifi.

At that time only the Leaf BMS messages and some custom inverter message should be on the CAN bus as everything else is powered off. So I'll make a simple CAN bridge consisting of two back-to-back CAN transceivers. That way a new physical bus is spawned that is terminated in the ChaDeMo charger and on the bridge. I hope chargers won't mind the extra CAN ids on the bus. Any experience?

I will cut power to the main relays with the proximity signal. That way the car can't drive away while charging. The ChaDeMo port relays will be closed with the charger provided 12V and GND. I don't see any use case where I would want to open those relays by myself.

johu wrote: ↑Wed Feb 12, 2020 4:51 pm
So now I can use the ECU to do the CAN part of ChaDeMo or, alternatively, talk to my homebrew charger via wifi.

Great stuff!
Is this planned to appear in a product of some kind?
Like something from JLCPCB maybe?