Hub motor repair: hall sensor edition

[boudin]

Momentarily ignoring the wisdom of buying a cheap bike from the lowest bidder, we took such a vehicle slightly outside of its operational envelope this weekend (a 20" fat bike on +1 "blue" single-track) and, according to the error the controller is throwing, killed one or more Hall sensors. While it's theoretically still under warranty, the vendor has not inspired confidence by deferring my post-sale support to another company.

If I end up on my own, what am I getting into here? If, indeed, it's a Hall sensor failure, what's typically involved in sourcing and replacing those - are they usually a fairly common off-the-shelf component in this sphere? I'm reasonably mechanically inclined - I do all my own mechanic work, have hand-soldered basic SMD components before, and have more than passing familiarity with places like Mouser and Digikey, but haven't ever disassembled a hub motor.

I see mentions on these forums of using a multimeter to test them for failure which I'm certainly equipped to do, but any other pointers? Do we tend to see scorching or other similar signs one does on certain failed components?

 

[E-HP]

If, indeed, it's a Hall sensor failure, what's typically involved in sourcing and replacing those - are they usually a fairly common off-the-shelf component in this sphere? I'm reasonably mechanically inclined - I do all my own mechanic work, have hand-soldered basic SMD components before, and have more than passing familiarity with places like Mouser and Digikey, but haven't ever disassembled a hub motor.

I just bought them off of ebay:
https://www.ebay.com/itm/114505778894

Testing c/o Tommycat:
https://electricbike.com/forum/forum/kits/golden-motor-magic-pie/110168-testing-bldc-motor-s-phase-wiring-hall-sensors-and-wiring

 

[Chalo]

(a 20" fat bike on +1 "blue" single-track) and, according to the error the controller is throwing, killed one or more Hall sensors.

Yeah, those deformed little bikes are at war with themselves. They have tires that make no sense at all on pavement, with drive systems that are incapable of doing anything besides pavement. They don't fit anyone correctly, you can't pedal them worth a damn... they're basically useless and inferior to all but the silliest real bicycles.

You can replace the Hall sensors and get the bike running again, but if you keep using it like it's competent for anything, you'll cook the Halls again, or maybe the phase windings this time. Beware the sunk cost fallacy. My vote is get it running and pass it to another sucker, or else commit to putting a motor/controller/battery on it that can deliver on whatever deranged promises the bike is making.

Yeah... find another sucker.

 

[amberwolf]

If I end up on my own, what am I getting into here? If, indeed, it's a Hall sensor failure, what's typically involved in sourcing and replacing those - are they usually a fairly common off-the-shelf component in this sphere? I'm reasonably mechanically inclined - I do all my own mechanic work, have hand-soldered basic SMD components before, and have more than passing familiarity with places like Mouser and Digikey, but haven't ever disassembled a hub motor.

I see mentions on these forums of using a multimeter to test them for failure which I'm certainly equipped to do, but any other pointers? Do we tend to see scorching or other similar signs one does on certain failed components?

With most problems that indicate a hall signal issue, it's not usually the sensors that are bad, it's usually the wiring or connectors between motor and controller.

Sensors can be killed by severely overheating the motor or by damaging the motor cable in a way that shorts phase wires to the hall signals or power/ground. The latter usually also damages the controller hall signal inputs (so replacing the halls doesn't fix this, controller then must be replaced). The former, if temperatures went high enough and stayed there long enough, also tends to damage the motor windings, and if it is a geared hubmotor it may also damage the plastic gears.


What are your specific symptoms?


Grin Tech http://ebikes.ca has testing procedures in their Learn - Troubleshooting section, and I and others have posted procedures as well, over the years (I dont' have a direct link, but it should show up in a search if you have the time).

If it is actually failed hall sensors, SS41 or SS411, latching bipolar types, are the usual ones used in ebike hubmotors. If you go with other part numbers, make sure they are latching bipolar and that they can run on as low as 4.5v Vcc or lower, and that they are open-collector outputs tolerant of at least 5v pullups, but as high as you like.


If it's a geared hub it may be more complicated to disassemble to get to the sensors than a DD hub, but the actual sensor replacement procedure is about the same. Somewhere in my posts about the MXUS and Fusin hubmotors over the years, there are pics and descriptions of me replacing halls in them, if it's helpful.

 

[boudin]

Yeah, those deformed little bikes are at war with themselves

Yup, they're attractive to a different market than the average Spherester. If it's repairable within reason we'll do so and keep it in town for its original purpose: year-round grocery-getting in a city that doesn't plow surface streets.

In the greater scheme, this just bumps the budget priority of an actual trail-capable [re]build for my wife. The trail was just a lot less casual than what we'd initially believed.

I will admit that it was amusing to see quite a few riders watch the deformed basket-bike pass and reconsider both their >$5k bikes and how difficult they thought the trail was.

With most problems that indicate a hall signal issue, it's not usually the sensors that are bad, it's usually the wiring or connectors between motor and controller.

Sensors can be killed by severely overheating the motor or by damaging the motor cable in a way that shorts phase wires to the hall signals or power/ground. The latter usually also damages the controller hall signal inputs (so replacing the halls doesn't fix this, controller then must be replaced). The former, if temperatures went high enough and stayed there long enough, also tends to damage the motor windings, and if it is a geared hubmotor it may also damage the plastic gears.


What are your specific symptoms?

I'll be sure to triple check the cabling to make sure we have appropriate insulation. The symptoms are that the system starts and appears normal, but once we request power (throttle, walk-assist, or pedal assist), the motor energizes, ticks forward a mite, then the controller appears to shut things down and display an "Info 03" error. The manual indicates (page 13) that this is a hall sensor error, which is all I have to go on at the moment. Haven't even unloaded the bikes from the truck yet, much less begun disassembly.

 

[amberwolf]

I'll be sure to triple check the cabling to make sure we have appropriate insulation. The symptoms are that the system starts and appears normal, but once we request power (throttle, walk-assist, or pedal assist), the motor energizes, ticks forward a mite, then the controller appears to shut things down and display an "Info 03" error. The manual indicates (page 13) that this is a hall sensor error, which is all I have to go on at the moment. Haven't even unloaded the bikes from the truck yet, much less begun disassembly.

The symptoms could be that of a hall signal issue, where the motor attempts to spin but the wrong signals (or no signals) are received by the controller vs what it expects based on what it tried to do.

What happened just prior to the problem starting? (different possible causes may lead to different troubleshooting paths)

 

[boudin]

What happened just prior to the problem starting? (different possible causes may lead to different troubleshooting paths)

In generic terms, we'd just exited a fairly rough descent of 1.25mi, no assist so the motor should have been relatively cool. Began climbing an approximately 10-15% grade on PAS 1 or 2, occasional throttle squirts to help boost past roots or rocks that were curb size or smaller (not braced against them, just adding thrust before contact to ensure continued momentum). It was at one of these bursts that the motor showed the hall sensor error and ceased to function. Speeds were slow, 5-6mph

In very specific terms, we were climbing the hill between mile 7.23 and 8.05 on this trail having completed the loop as-described on the map that far (actually 7.x miles in, having climbed from Mayhem Gulch).

 

[amberwolf]

With only occasional throttle bursts, the motor should not have experienced overheating, so hall sensors shoudln't have been damaged by that.

It's more likely that somethign was jarred loose in the wiring or connections from bumps.

It is also possible though unlikely taht a hall sensor, if not secured correctly, could slip out of it's slot in the stator laminations enough to not read the passing rotor magnets correctly (or at all).

It's probably not applicable, but: there is a phenomenon that has happened to some people with a geared hubmotor (not a DD) when a geared motor wheel is in the air (bump, curb jump, etc) and under power and then the tire grabs the surface, where a shock of torque can hit the clutch or gears with enough force to damage them. If this results in jamming damage to the gears (teeth stuck in other teeth preventing rotation) or clutch (fractured housing preventing engagement) then the motor might either not be able to spin the wheel, locked in position once it hits the jam point, or it might prevent the motor from driving the wheel and just spin inside it. In either case, the controller doesn't get the results it expects and if it's "smart" enough it may error out, though it shouldn't be a hall error, it should be a current or RPM error. (too much or too little). A dumb controller would just keep trying to spin the motor anyway, unless the gearing is locked such that it can't spin and the current limit is reached long enough to cause shutdown...if it's really dumb it wouldn't shutdown and just keep feeding current to the locked motor and overheat it. ).

If there is a gearing or clutch problem, then flpping the bike upside down and manually turning the motor wheel backwards will try to (but be unable to) spin the motor inside the casing. A clutch failure would feel about the same forwards as backwards, while a gear jam would feel very high reistance to turning in the reverse direction.

In generic terms, we'd just exited a fairly rough descent of 1.25mi, no assist so the motor should have been relatively cool. Began climbing an approximately 10-15% grade on PAS 1 or 2, occasional throttle squirts to help boost past roots or rocks that were curb size or smaller (not braced against them, just adding thrust before contact to ensure continued momentum). It was at one of these bursts that the motor showed the hall sensor error and ceased to function. Speeds were slow, 5-6mph

In very specific terms, we were climbing the hill between mile 7.35 and 8.05 on this trail having completed the loop as-described on the map that far (actually 7.x miles in, having climbed from Mayhem Gulch).

 

[boudin]

Although it's a geared motor (CZJB-104C2), it rotates fine in reverse. All connectors to the controller are snug and have no visible insulation failures or scorch marks, so no immediately apparent shorts at this stage.

Sounds like I'll get to break into the housing eventually and start identifying where the sensors even are. Thanks for your help thus far!

 

[amberwolf]

Although it's a geared motor (CZJB-104C2), it rotates fine in reverse.

Do you mean that it rotates under power in reverse, or that you can hand rotate it in reverse with normal resistance? (and hand rotating in the forward direction should have little or no resistance in comparison)

All connectors to the controller are snug and have no visible insulation failures or scorch marks, so no immediately apparent shorts at this stage.

You wouldn't see scorch marks except in some pretty severe kinds of damage and in those cases the controller would probably not be operational at all (not even turn on), based on failures we've seen around here.

Unless there is wire damage from an impact or snagging the motor cable on something, shorts are not likely to be the cause of the problem you see. But open circuits could be, and you can't usually see those; even an apparently snug connection can still be open, or high enough resistance to cause signal problems. If you unplug and visually check each side of the motor-to-controller connector (the only one that should give a motor hall failure error), and there are no visible signs of contact failures (spread out barrels, bent pins, etc), then it would require electrical testing to verify operation or failure. You can't see problems with contact to wire crimps, unfortunately, especially inside the overmolded waterproof connections, so for open-connection suspects, it requires access to both ends of the cable, meaning inside the motor and the controller, to do continuity / resistance testing. Because of the risk of damage when opening things up, it's better to test the hall signals themselves first.

If it's a connection issue at the plug, it's possible simply unplugging and replugging will fix the issue (but if it's a crimp-to-wire or spread-barrel problem it could come back).

Sounds like I'll get to break into the housing eventually and start identifying where the sensors even are. Thanks for your help thus far!

Before opening it up, you should test the hall sensors electrically. If it is a waterproof connector, then you won't be able to test with the controller connected (which is the simplest way, as shown in the Grin Tech documentation), so you will need your own 5v power source to connect between hall 5v and ground, *and* a pullup-resistor (5k-10kohm) to connect between that source and the signal being tested. Since the motor probably doesn't engage when hand spinning in the forward direction (normal for geared hubs) you'll need to very slowly manually spin it in the reverse direction to generate hall signals for this test. (the gearing multiplies the motor rotation rate vs wheel rate so if you don't spin very slowly the reading will change too fast for the meter to keep up).

The sensors are almost certainly in the same place as in other hubmotors--embedded in the stator laminations in grooves cut just for them. Some use PCBs to mount them, and some just run wires to the leads directly. There's pictures of this sort of thing in various hall-sensor repair (and other motor repair and modification) threads, such as these, to give you some guidance:
https://endless-sphere.com/forums/viewtopic.php?f=12&t=114863&p=1713424&hilit=hall+sensor%2A+motor#p1713424
https://endless-sphere.com/forums/viewtopic.php?f=2&t=111478
https://endless-sphere.com/forums/viewtopic.php?f=1&t=108669&p=1591559&hilit=hall+sensor%2A+motor+replac%2A#p1591559
etc.

 

[boudin]

Do you mean that it rotates under power in reverse, or that you can hand rotate it in reverse with normal resistance? (and hand rotating in the forward direction should have little or no resistance in comparison)

That I can hand-rotate it in reverse with nominal resistance. I can't yet check under power.

Before opening it up, you should test the hall sensors electrically. If it is a waterproof connector, then you won't be able to test with the controller connected (which is the simplest way, as shown in the Grin Tech documentation), so you will need your own 5v power source to connect between hall 5v and ground, *and* a pullup-resistor (5k-10kohm) to connect between that source and the signal being tested. Since the motor probably doesn't engage when hand spinning in the forward direction (normal for geared hubs) you'll need to very slowly manually spin it in the reverse direction to generate hall signals for this test. (the gearing multiplies the motor rotation rate vs wheel rate so if you don't spin very slowly the reading will change too fast for the meter to keep up).

I'd certainly prefer to avoid cracking the case. The exterior connector is a waterproof 9-pin barrel strongly similar to this one. Inside the controller enclosure, that waterproof connector terminates on a traditional 6P JST type connector, so I may be able to follow the Grin Tech docs anyway. I'll give them a read.

The whole thing (down to the font used to engrave the motor model info) broadcasts "Bafang RM G060 clone".

 

[boudin]

Minor update: had to reassemble to take photos and video of the behavior for their warranty department. They're shipping me a new motor and I suspect they won't be asking for the old back, so I should have the opportunity to repair that one at my own pace. If the replacement ships from CN though, I may well be repaired before then. Back to the helpfully posted Grin troubleshooting page.

 

[boudin]

Interestingly enough, they shipped an unthreaded motor, then had me remove the stator package from both and swap them. The motor's working again, so now I've got both a working bike and hall sensors to debug.