I meant only the VCC..obviously UN is not shorted to the ground per your results.

Salut Lionel,

   Sounds promising, that would be a lucky case indeed if a single optocoupler change would fix the issue. Meanwhile it may be also worth checking the UN port on the power module to make sure it is not shorted to VCC (VN1) or ground (VNC). I think the fault detection of the module only monitors the output side.

/ Azzuro

Salut,

   Luckily there is only one version. Some of the later controllers have an additional small module that is aiming to completely eliminate the few mA idle current from the battery but this is not really relevant in my view - particularly in case of Li-ion batteries.

/ Azzuro

Salut Lionel,

   Sounds like a plan, I think you are on the right path. These motorcontrollers are fairly reliable compared to their complexity. Indeed the rush current caused by instant connection of the battery can cause issues with the supercapacitors and their fuses however the IGBTs should be less sensitive to this. Let me know what you've found based on your measurements :) 

As plan B you may also consider to look around for an MC on Leboncoin, just in case. In an optimal scenario if you could fix the current one you'd still have a hot spare.

/ Azzuro  

Salut Lionel,

   No worries at all, in fact I just finished the captures and shared the resulting two videos with you (apologies for the low resolution, just noticed it). In the first video I have captured the native 15 kHz PWM signal for the power lines A (red) and B (yellow). I only have a dual channel oscilloscope but line C is similar anyway. Notably the signals are 50-50% PWM having the amplitude of the battery voltage as we could expect. The revving was zero to 40km/h equivalent rpm then zero again. 

The second video shows the signals sampled behind a pair of makeshift low pass filters (R=470k / C=100p) - using 5 and 10 ms resolutions having two 0-40-0 km/h runs again. These videos shows the actual sinusoid PWM based drive signals, even the 120 degree difference between the two waves could be recognized.

As you could see my results are significantly different from your yours which I could not fully explain - I see you have used software based filter, that could be a contributing factor for example. 

Worth noting in my case regardless the rpm the A and B (and also C) signals are basically identical as the 3 phrase motor driving logic dictates. Accordingly you may consider to focus on this aspect namely the assymetry between the 3 lines in your measurements. In my view a possible cause is a partially defective IGBT module. Curious of your opinion.

/ Azzuro   
    

Salut,

   Luckily I am an electrical engineer therefore a low pass filter should not pose a problem. The few days delay is caused by the fact I actually need to use my primary scooter that is in daily use to check the power lines - since I found my spare scooter having old NiMH batteries is HS. Back to you soon.

/ Azzuro

Salut, pas de soucis - meanwhile I have checked the details in question; 

Regarding the calibration I've initiated the command in VectrixDiag and indeed there is no status message at the end of process. The orignal value on my scooter has been 65 (by reading manually with the application). During the calibration process which took only a few seconds the counter started from 1 and quickly reached the value of 62 while the wrench light on the bike has been also illumunated then went dark after a few seconds once the process has finished. I've repeated the calibration 2-3 more times and the results have been identical. If yours was similar (apart from the value) then this aspect should be fine I'd say.

Unfortunately concerning the power signals I could not retrieve consistent results yet since I only see a 15 kHz high voltage signal which obviously slightly changes with RPM as expected. This also explains why you've decided to use a low pass filter (which I don't have one at hand). Accordingly this part is still "work in progress".

/ Azzuro

Hi Eric,

   Nissan batteries are a bit old and not the biggest capacity but technically they are fine. For more capacity one of the best solutions today are the ID.3 cells but they come with a price. 

There is no firmware for the ESD charger that supports Li-ion as far as I know, only the EVPS Runke V2 (gold) and TC chargers are compatible. 

The BMS question is tricky since - apart from the standard battery management tasks - it must talk to the other units on the CAN bus as well therefore only a very few custom solutions exist. To my knowledge there are 3 teams in Europe who are capable to address this, located in Spain, Italy and Hungary. 

/ Azzuro 

Cool, so the question of firmware is off the list. I've looked once more at your video showing the power lines driving the motor and started to wonder why the shape of the A,B and C signals became quite different while the motor was revving up. Honestly I would expect them to be pretty much identical regarless of the rpm or load. Let me check these on my spare scooter and come back to you with the findings.

Similar remark for the calibration process, I've done the last one long time ago but I can't recall it just quitting during the process. I'd rather assume there was a message confirming the success at the end of the process. Let me check this as well (btw I haven't seen a link to your video showing this one). 

Please allow me a few days to collect the forementioned results.

/ Azzuro  

Salut,

   Thank you, I've looked at them at first glance they look good (earlier I've assumed the Hall signals are not in use, but wasn't completely sure). 

In fact in the background of the video I hear the wheel spinning, so what is the problem exactly ? Is it the power not being sufficient to drive the scooter on the road under real life conditions ?

Has been the firmware issue addressed yet ?

/ Azzuro