Hi
one problem with my bike is still open which is:
when I do a instant startup (0-100%) the (12kg BLDC) motor starts and the red LED is blinking and the acceleration stops and a fault seems detected with no load attached, so only the 12kg wheel has to be accelerated.
When I look into my PSU I see that shortly (100ms) it consumes a max of 13A and not more (20A is defined as Battery max, while motor max of 50A is left by default, the PSU is unrestricted and provides 120A with the 30V I tested) so overcurrent does not seem the fault at all.
As my VESC does not successfully test the BC=BLDC mode, but only FOC I cannot test the other one to see a difference.
When I do a load test (variable friction test) I can if started slowly (10-15sec) accelerate up to the max of 95% just as defined. The sensored erpm is up to 2500 (which means about 1/3 of max erpm)
Any comments? Are there any settings in FOC to play around?
Cheers
Sam
Is your power supply a switch mode? Sometimes they don't like the load to change too quickly. VESCs like low resistance power supplies with lots of capacitance for the peak current.
Hi TechAUmNu,
the PSU (EA-PSI 9080-120) is of course switched, but especially build to simulate electrical storage and checks its parameter about 15.000 times a second, protocols all data and even compensate the cable losses (feedback). The internal resistance is lower than 3mOhms and the rise time from 5% to 95% load is less than a 150us. It's of course buffered with 4700uF at the plugs. The PSU is surely not the cause and the VESC does not even come near the 120A max. of the PSU, the voltage is stable to 10mV at the plugs at any load.
The error started in the real field with my normal LiIon Battery (30V/15Ah, 35mOhm Ri) and was then further analyzed in my "lab" with better ways than in some garage available. In the "lab" I did see the red LED blink until I return the throttle to 0 input, then I'm ready to start again.
Are the fault codes stored somewhere or some settings to tweak with just to get to the likely cause?
Cheers
Sam
Pedelec usage in combination with a GoldenMotor MagicPie3 BLDC
Can you not just plug in a laptop and see what the error is when it happens in VESC Tool? It will either most likely be an ABS_OVERCURRENT error but it will let you see better what is happening. Then you could post a screenshot of the graphs on here and we can take a look.
Hi TechAUmNu
problem is simple: the VESC is integrated into the HUB of the MP3 so no direct connect or readout. As I need to get further I'll disassemble it today and get the axle mounted on a plate (like the test assembly of the last month) and read out the tool. I'll report back
Cheers
Sam
Pedelec usage in combination with a GoldenMotor MagicPie3 BLDC
You could use one of the Bluetooth modules to connect to it while installed.
No way;it's fully metal encased (Hub) so Mr. Faraday rules here. I'd have to sacrifice the HALL outlet to get any wire through the axle.
I'm already disengaging the Fork (hell of a lot of work as the hub is not screwed but encased).
Pedelec usage in combination with a GoldenMotor MagicPie3 BLDC
Hi TechAUmNu
I got a connection to the VESC. You were right about the overcurrent fault (in accleration and in braking). As I looked into the values (63A max) I was astonished to see such a high value as my PSU did not show anything in that range (13A max). As I had not limits set the 120A could fully be used. So I put in a fuse with 15A rating medium reaction and fault again, no fuse triggered. When I looked into the PSU logbook I see the 13A just vor about 10ms and that's it. I triggered a softfuse with 3us trigger time (the fastest setting in the PSU) and it was not triggered as well. I did fiddle around for about 30 tests and settled then.
Then I found this article:
https://www.vesc-project.com/node/660
and began to dive into the settings of FOC mode. When I figured the Ri and lambda and L is very important for the setting. I did a few tests with the wizard. In 5 tests with different erpm and duty and ampere I received very different values. As I'm using only 4mm^2 from start to end I wondered were this different values came from (wrong detection algo, high tolerance in parts, high resistance in the wiring, ....). I put in the second VESC, importet the same xml files form #1 and: same fault behavior.
My next test was doing the same tests a few times: with 5A, 80erpm, D=0,5: in ten tests I received values from Ri=81mOhm up to 186mOhm which is quite a range which should not happen. I plugged the VESC out, used the PSU to generate an 1,00A current and measured the delta config of BLDCs and received 105mV. The true Ri of one phase should then be about 2/3 of it.
Because I was not sure what the VESC is really measuring, I thought about doing the same tests for 30 times. In this series I middled a Ri of 86mOhm, with mostly in the 100mOhm area. Because I was not sure about the other values I did then repeat the tests until the VESC did measure a 86mOhm Ri and let the rest be on autodetect. Guess what: in 20 tests with this values the acceleration and braking works like a charm. The first value of Ri of my first setting was 176mOhm (I don't know the other values right now) I think the variable Ri detection was the fault. I checked double the cables for high resistance points, but the 1A current test worked as I'd expect it to.
Well the Ri, L and lambda is the key for the FOC mode. If I had more time (I spent about 75 days experimenting with the MP3 + VESC and other controllers, waiting for replies mostly) I'd like to test the BlockCommutation mode (BLDC labeled in the VESC tool) as well manually as the wizard did not work at all.
Thanks so far.
Sam
Pedelec usage in combination with a GoldenMotor MagicPie3 BLDC