Ebike issues - strange and unseen
- Thread starter Navarro
- Start date
larsb
1 MW
I did hall status readout from your video. It does not seem to be correct, i can readout 001,011,110,001,011 etc which does not follow commutation table.
Try this yourself.. stop wheel rotation in each position and go 12 states and note the combinations, compare to the table.
This is probably why controller shows hall error. One hall could be offset in position so that the hall is on or off for too long which gives an incorrect readout. :wink:
Try this yourself.. stop wheel rotation in each position and go 12 states and note the combinations, compare to the table.
This is probably why controller shows hall error. One hall could be offset in position so that the hall is on or off for too long which gives an incorrect readout. :wink:
Hello again...
Ok,will do that. If so, what should I do? Should I inverse the hall sensor wires? I don't understand why this hapened as I replaced all 3 hall sensors with corresponding parts. And what is the link with the busted controller?
These kits and their electronics are mental!
Ok,will do that. If so, what should I do? Should I inverse the hall sensor wires? I don't understand why this hapened as I replaced all 3 hall sensors with corresponding parts. And what is the link with the busted controller?
These kits and their electronics are mental!
larsb
1 MW
I also wonder about this connection to your first error. Let's say your first issues were related to a hall error, same as now then it does not explain your measurement with the controller.
Were the halls in the slots or on a separate hall board?
Did you get exactly the same position for the halls as before?
Did you add any part to the circuit like capacitor when you exchanged the halls?
A short circuit somehere would in one of the hall wires could also throw the signals into error.
It's hard to guess when you don't know exactly what everything looks like..
Were the halls in the slots or on a separate hall board?
Did you get exactly the same position for the halls as before?
Did you add any part to the circuit like capacitor when you exchanged the halls?
A short circuit somehere would in one of the hall wires could also throw the signals into error.
It's hard to guess when you don't know exactly what everything looks like..
Yeah... Sadly you are right.
I did not add anything and soldered the new hall sensors in the exact position with the exact pin numbering/positioning.
I will try to remove the wires and add new ones. Basically remove the wires from the hall sensors to the plug - if all, this might be the initial issue. Maybe the original controller got fried because of bad readings? I dunno, I am starting to feel like a techno-Sherlock.
Great, more mind grinding work.
On the other side, apart from losing the KT display, maybe a halless controller is a good idea.
I did not add anything and soldered the new hall sensors in the exact position with the exact pin numbering/positioning.
I will try to remove the wires and add new ones. Basically remove the wires from the hall sensors to the plug - if all, this might be the initial issue. Maybe the original controller got fried because of bad readings? I dunno, I am starting to feel like a techno-Sherlock.
Great, more mind grinding work.
On the other side, apart from losing the KT display, maybe a halless controller is a good idea.
larsb
1 MW
Were the halls on a board (possibly misalignment already before replacement?) or in the slots? (Unlikely that they got misaligned)
pictures would've been worth a thousand words
pictures would've been worth a thousand words
They were in their slots and I extracted them quite hard (some had to be scrubbed out).
I have the exact set up as below, will open my motor again in a couple of days. I see that you suspect a bad logic in my soldering rather than a short in the wires? I'm asking this because rewiring the motor sounds like a horrible thing to do.
I have the exact set up as below, will open my motor again in a couple of days. I see that you suspect a bad logic in my soldering rather than a short in the wires? I'm asking this because rewiring the motor sounds like a horrible thing to do.
captain387
100 W
Navarro said:
The old controller may be fine, it functioned for 800 km with the hall sensor wiring order in the connector. The new controller may have different wiring colours, and it seems to be throwing the same code.
Get yourself a stand to lift the back wheel off the ground.
Check the battery voltage at the battery connector to see if it drops when applying throttle. If the battery is not an issue then....
Sounds like a bad connection or short in the hall wiring.
Or phase wiring order, the new controller may have different colours for different phases. (My phase colours never line up)
Look inside the phase wire connector make sure all pins are ok and do not push out when connecting the two ends.
For the hall sensor short:
A simple test would be to unplug the hall sensor connector, take a volt meter and set to ohms and attach one of the volt meter leads to the motor shaft, and test each individual hall sensor at the connector.
As you are testing move the hall sensor wires around and see if there are any shorts.
Check ohms between all phase wires too.
Phase wires:
Try all variation
There is a chart around which has all combinations for hall and phase, which you can check off one by one, quite a few combinations.
Because your halls are mounted on a PCB forcing the leads to all go the same way, this is probably not the case, but:
Sometimes the center hall sensor has to be "upside down" compared to the other two. Meaning, with it's writing facing the magnets if the others are facing away from them, or vice-versa.
That can give the wrong combination of hall readings as the magnets pass, if it's supposed to be that way but isn't.
If it is the case for yours, then you'd have to have the three hall wires insulated so they don't touch when they cross (because you'd have to twist the hall wires over each other to get the sensor itself in the upside down position, while still having the leads go into the right holes in the PCB).
Sometimes the center hall sensor has to be "upside down" compared to the other two. Meaning, with it's writing facing the magnets if the others are facing away from them, or vice-versa.
That can give the wrong combination of hall readings as the magnets pass, if it's supposed to be that way but isn't.
If it is the case for yours, then you'd have to have the three hall wires insulated so they don't touch when they cross (because you'd have to twist the hall wires over each other to get the sensor itself in the upside down position, while still having the leads go into the right holes in the PCB).
larsb
1 MW
What slot&pole count do you have? That and a picture of the hall positions will show if you need a flipped sensor. But as long as the commutation signals of the halls don't match the table then i don't think so.
I don't suspect bad soldering or any other cause yet, i am just guessing..
I don't suspect bad soldering or any other cause yet, i am just guessing..
Ok guys, back from work and on to dissasembly.
Checked again the hall sensors:
https://youtu.be/NtURuHe1jJ0
This is exactly how my hall sensors look at the moment:
I did my best to isulate them with motor silicone and taped them to the stator using isolating tape. You can see that they have only one position, not possible for them to be flipped - their hole has a one position only option. I checked them before opening with multimeter using sound option for shortcircuit and no sound. All the wires are ok, no short between hall - frame. No short between hall-hall. Short only between phases (normal, I guess).
For the moment I have the motor opened and laying around, tell me what else needs checked/measured before assembling it again (such a horrible work without extraction tool, I cut my skin again in 2 places).
Checked again the hall sensors:
https://youtu.be/NtURuHe1jJ0
This is exactly how my hall sensors look at the moment:
I did my best to isulate them with motor silicone and taped them to the stator using isolating tape. You can see that they have only one position, not possible for them to be flipped - their hole has a one position only option. I checked them before opening with multimeter using sound option for shortcircuit and no sound. All the wires are ok, no short between hall - frame. No short between hall-hall. Short only between phases (normal, I guess).
For the moment I have the motor opened and laying around, tell me what else needs checked/measured before assembling it again (such a horrible work without extraction tool, I cut my skin again in 2 places).
larsb
1 MW
That picture is not good enough, could you take one closeup for each sensor, in good light and focus?
How many slots?
How many magnets?
the hall legs should be covered better: "liquid tape" would be good (if you don't want to resolder and use shrink tube). Also they are quite bent, i'd worry about cracked wire entry into the hall body and unnecessary induced currents when they are separated such long distance.
Did you count the signals/are they correct?
Are you sure the hall legs cannot touch anything like side covers?
Can you see any touch/scrape marks on the hall legs?
These two hall legs look really close, could they short circuit during rotation/vibrations:
How many slots?
How many magnets?
the hall legs should be covered better: "liquid tape" would be good (if you don't want to resolder and use shrink tube). Also they are quite bent, i'd worry about cracked wire entry into the hall body and unnecessary induced currents when they are separated such long distance.
Did you count the signals/are they correct?
Are you sure the hall legs cannot touch anything like side covers?
Can you see any touch/scrape marks on the hall legs?
These two hall legs look really close, could they short circuit during rotation/vibrations:
I have 51 magnets and 52 slots. For the Hall sensors only 3 slots with dedicated shape (no way to stick them elsewere). No scratches on the hall sensors/legs - only the leftmost one has scratches because I had to force it a bit inside (not so much becase I feared braking the legs). The legs are not staying open to contact like that, another piece of tape glues on the back one (like the 2 glue sides with the legs between them). I'm quite confident that my isolation is good (basically everything was red/black when I opened the motor). Also, the legs have solid connection with the Hall body.
Below pic with the 3 sensors - they are very sturdy, so not on the edge of breaking/separating.
The right one
Middle one
Left one
Middle and left one (to confirm same serial):
How do I count the signals?
Below pic with the 3 sensors - they are very sturdy, so not on the edge of breaking/separating.
The right one
Middle one
Left one
Middle and left one (to confirm same serial):
How do I count the signals?
larsb
1 MW
Light on hall tester =1
Light off=0
For each sensor write down each state combination.
Compare 001..011.. to commutation table. If it's not the same then hall signals are not correct.
You didn't have to extract the sensors, i wanted to see their location. They need to be aligned with the stator surface and glued in position.
Light off=0
For each sensor write down each state combination.
Compare 001..011.. to commutation table. If it's not the same then hall signals are not correct.
You didn't have to extract the sensors, i wanted to see their location. They need to be aligned with the stator surface and glued in position.
Roger. But in this case I believe it's easier to just try other combination with the wires? I am not that good at being careful at 52slots/51 magnets. 
Also, the pic with them in position. The leftmost one sticks a bit outside but once fixed with the isulator tape stays completly inside (how I found it).
Also, the pic with them in position. The leftmost one sticks a bit outside but once fixed with the isulator tape stays completly inside (how I found it).
larsb
1 MW
52poles, 26 e-revolutions per mech revolution
51slots: mech angle between slots 7.1 deg
Halls to be placed at 120 degrees electrical
120/26=4.62 degrees spacing minimum between sensors, little more than a half slot spacing (could be a multiple of this also).
sensor distance seem to be at 1.5 slots, 10.6 mech deg spacing which is not really correct as it's not a multiple of 4.62, it's actually 2.3x that.
If halls are at 60 degree electrical then it's supposed to be 60/26=2.31 deg between
10.6/2.31=4.6 which is not a multiple either..
Can you confirm if the halls have a 1.5 slot spacings between them?
That's 0.31 deg off which is 0.31*26=8 electric degrees too long spacing.
At least you don't need a flipped sensor
51slots: mech angle between slots 7.1 deg
Halls to be placed at 120 degrees electrical
120/26=4.62 degrees spacing minimum between sensors, little more than a half slot spacing (could be a multiple of this also).
sensor distance seem to be at 1.5 slots, 10.6 mech deg spacing which is not really correct as it's not a multiple of 4.62, it's actually 2.3x that.
If halls are at 60 degree electrical then it's supposed to be 60/26=2.31 deg between
10.6/2.31=4.6 which is not a multiple either..
Can you confirm if the halls have a 1.5 slot spacings between them?
That's 0.31 deg off which is 0.31*26=8 electric degrees too long spacing.
At least you don't need a flipped sensor
larsb
1 MW
Why would anyone mill seats for sensors in the wrong spacing.. Idiots
I think it's not enough to throw the signals off, root cause to issues are something else,
i'd now guess that it's hall isolation issues at vibration or movement. If you take care to fixate them and isolate all exposed metal that might solve the issue, otherwise you could try to place new sensors at 120 deg mechanical spacing, 17 slots apart. they'd be in a perfect placement
I think it's not enough to throw the signals off, root cause to issues are something else,
i'd now guess that it's hall isolation issues at vibration or movement. If you take care to fixate them and isolate all exposed metal that might solve the issue, otherwise you could try to place new sensors at 120 deg mechanical spacing, 17 slots apart. they'd be in a perfect placement
larsb
1 MW
If not then don't know what to check. Hall wires or hall connector..?
Did you ever have any twisting/movement on the torque arms or axle?
Did you ever have any twisting/movement on the torque arms or axle?
amberwolf said:
From the board traces, you can see the signal lines are on the left on all three sensors, so they should all be facing the same direction.
If there is any spot in the rotation where you get 000 or 111, this will trigger the hall error.
larsb
1 MW
That was smart Fechter
Didn't think of the trace directions, i'll remember it for another time
Didn't think of the trace directions, i'll remember it for another time
larsb
1 MW
Navarro said:Roger. But in this case I believe it's easier to just try other combination with the wires? I am not that good at being careful at 52slots/51 magnets.
You only need to check 6 positions but do 12 so you get them repeated once. If it doesn't follow table it's NOK
Which is why I said "Because your halls are mounted on a PCB forcing the leads to all go the same way, this is probably not the case"fechter said:
but wanted to point out the potential issue as it has been the case for others.
