Considering 15A*5= 75A, i don’t see why you should see anything above 1-2V sag at 50A
Sudden sag can cause “weirdness”
Any chance you have another battery to test with?
2 Likes
I do not.
Can you test your battery voltages, atleast to see if it’s balanced
It’s also possible a connector/wire joint is failing at the higher current and causing V sag
Yes. have to open it up measure on the balance plug at the BMS. which I’ve done before. it’s always been balanced to 2 decimal places. checked 3 different times. 1 of which was mid voltage the others were charged.
it’s always had this level of voltage drop.
Then I would really check all the power leads, redo them all if not sure, something is causing the V sag
Is the battery voltage reported the same on both vesc?
metr doesn’t show both voltages in realtime or records. but I have hm-10 on the other one and xmatic.
I don’t think i’ve watched close enough to see if it sags relatively the same. let me test run it and see. but I think they both drop equally.
I’m not the builder of this board. I assume I’d be looking for things that look like poor connections bad solder points that might increase resistance int the pack wiring? all the series connections are bundled up nicely so might be hard to inspect without really getting intrusive.
Assuming the series connection are good… yeah check all solder joints between the battery and vescs, also check inside connectors for any singe/burn marks, as well look if they are legit connector vs copy, at 50A I would expect to see stuff like this with copy connectors
1 Like
wasn’t at full charge but used xmatic -> hm-10 -> second vesc,
there’s a 4-5v drop.
1 Like
Ok, my TODO/Maybe on this problem:
-
drastically raise the motor temp cutoff start and end to avoid problem in the short term.done motor temp cutoff end can be set to max of 120c. did this for the short term. - change motor in order to change temp sensor to see if any other temp sensor does the same noise in metr measurements…
-
get metr to record temps from the other motor to see if has the same noise problems on the other side / different motordone thanks to metr updates.
both motors show the same kinda noise. - try the twisting of phase wires to see if this reduces noise in temp sensors as measured by metr.
- investigate some option for shielding the sensor wires from the phase wires.
- investigate some option for hw low pass filter on the sensor wire.
- maybe try running on ackmaniac to see if situation is handled any different.
-
file bug on bldc about the fault message reporting a temperature that couldn’t trigger the faultdone here
and regarding the voltage drop and possible connection to sudden voltage drop things get weird.
-
verify same voltage drop happening on the other sides vesc.done it does. -
dig through metr records and closely. look for correlation between voltage drop and motor temp fluctuations.done pointless. see below. - do experiment bypassing the antispark switch to see if voltage drop situation is improved.
- double check cell balance at full charge
and low charge.( low charge done, seems balanced 2 of 10 pgroups at 3.18 and the rest at 3.19 )
2 Likes
I think it was @mishrasubhransu that said a while back hi did something to his wires to shield it from noise, not sure
If it’s the sensor, It’s really easy to add another one, I don this to all my motors now, the ones that como with them are on the sensor PCB and are not good at getting the windings temperature, they take a while to heat up
4 Likes
I just kept it physically away from the phase wires. The correct way would be to use copper braided wire around the sensor wire and hook connect it to ground.
3 Likes
Can you describe “physically away from the phase wires” a bit more? or any pictures?
Temperature is not something which can rise by crazy amount instantaneously, so a low pass filter would be awesome! @Trampa
4 Likes
That’s a hack. It just worked for my setup. I repositioned the sensor wires such the temp reading didn’t get affected much when I pulled the trigger.
I would suggest to use a coax cable for the temperature wire with the sleeve connected to the ground. That will take care of any induced voltage. And proximity to the phase wires won’t cause any problems.
1 Like
oh right. I can add that to the list. though I’m doubtful it’ll change temperature readings, and metr can catch that they spike wildly.
That said, there is a lowpass filter in the code. so code is involved in stabalizing the reading.
1 Like
Well, didn’t take long to realize that’s pointless. there’s strong correlation between voltage drop and motor temperature sensor noise in the data. but that’ll be because of the strong correlation each one has to motor current.
1 Like
I thought i was gonna have to move the metr module to the other vesc. but thanks to the metr update it now separately records both temperature values. and according to that, there’s a very similar noise in both motor temperatures.
which probably means I’ll try possibly the shielded cable route next.
here’s this mornings commute.
4 Likes
Hmm, I ordered some 6 conductor shielded cable, to make shielded sensor wires. … for some reason I got 18awg pretty sure I need 24 awg to crimp the jst-ph connectors on the end. and it was too stiff. so little more delayed on that experiment.
2 Likes
if youre bored, could take the shielding off an Ethernet cable and run the sensor wires through it
2 Likes