Intermittent no power problem

rick_p

100 W
Joined
Feb 16, 2021
Messages
260
Location
Los Angeles
Hello,

My ebike, which is powered by a BAFANG RMG06 48 volt 750 watt rear hub motor, sometimes works normally and sometimes it doesn't. When it works normally, it's fast, powerful, and quiet. Here are the symptoms when it acts up. With a fully charged battery I turn the bike on and everything lights up like normal, except there is no power to the wheel via throttle or pedal assist. Here is what is strange, if I roll the bike backwards a foot or so and then hit the throttle, it tries to go, you can feel a little pulse from the motor but then it cuts off immediately. If I repeat this process, but instead of using the throttle I just start peddling, it will try to kick in but it makes a growling noise, at which point I pull the brake lever to kill the power to the motor because that is, it can't be good for the motor. The bike doesn't have a throttle only mode by the way, only pedal assist and throttle mode together. If I give up on trying to get the motor working and just simply peddle the bike, it will sometimes kick in at some random amount of time, and from that point run normally. My first thought was it has a bad controller, but when I discovered the rolling backward thing I started questioning that thought and starting thinking it's a motor problem. Has anyone experienced any of these symptoms? All thoughts and suggestions welcome of course.

Thanks, Rick
 
rick_p said:
If I give up on trying to get the motor working and just simply peddle the bike, it will sometimes kick in at some random amount of time, and from that point run normally.

Do you mean that it keeps working from then on? Does the issue only happen when you fully charge the battery?
 
Sounds like a Hall sensor problem to me. Sitting stationary the controller sounds like it's not getting startup feedback, but when you roll it, they give enough info about the magnet position that it half ass tries to run. Then whatever is shorting goes back to working normal. The fact that it goes back to normal running suggests it's not the core power function of the controller, but just a sensor glitch disrupting the normal running.
It could be glitch in the 5v output from the controller to the sensors also. Can you check that while it's in non running mode?
I've also had times there was a weak connection causing a circuit break, but then after heating up a little started working again from thermal expansion until shut off, when it would cool and disconnect again. That's frustrating to track down🙄

Just plain unplugging and replugging the motor and Hall wire connectors can sometimes fix things too, if you haven't tried that yet.
 
E-HP said:
Do you mean that it keeps working from then on? Does the issue only happen when you fully charge the battery?

Yes, after it kicks in it's like the problem was never there and it runs great. The state of charge on the battery isn't a factor, I only mentioned that to rule out a dead battery issue.
 
Voltron said:
Sounds like a Hall sensor problem to me. Sitting stationary the controller sounds like it's not getting startup feedback, but when you roll it, they give enough info about the magnet position that it half ass tries to run. Then whatever is shorting goes back to working normal. The fact that it goes back to normal running suggests it's not the core power function of the controller, but just a sensor glitch disrupting the normal running.
It could be glitch in the 5v output from the controller to the sensors also. Can you check that while it's in non running mode?
I've also had times there was a weak connection causing a circuit break, but then after heating up a little started working again from thermal expansion until shut off, when it would cool and disconnect again. That's frustrating to track down🙄

Just plain unplugging and replugging the motor and Hall wire connectors can sometimes fix things too, if you haven't tried that yet.

I definitely tried the unplugging and re-plugging trick, every connection on the controller and to the wheel. I couldn't find documentation on the controller, so tracking down the 5v wire for the controller could be tricky, but replacing the controller is an option for troubleshooting. The Hall sensor sounds the most logical to me, I don't know much about them but from what I read, what you describe sounds about right, that it's not able to detect the current state (location) so it doesn't start. How hard is it to replace the Hall sensor in one of these?
 
docw009 said:
Is your display capable of showing error codes, and are there any?

No, the display is the biggest disappointment on this bike, it's not even LED, it's just status lights, on/off, battery charge level, and which PAS mode you're in, that's it.
 
If you have this style Hall plug, it's easier to probe with a multimeter while it's hooked up and running. The skinny black and the red carry the 5v to the Halls.
fd88dc68f1f2edb577fc3a896a285184.jpg

This type plug, where the phase and Halls are built together is a lot harder to probe.
s-l300.jpg

They sell tester boxes like this also pretty cheap, but you usually still have to make an adaptor if you have the newer waterproof style plug.
E-bike-tester.jpg
 
@Volton, thanks for the images and instructions!!

Voltron said:
If you have this style Hall plug, it's easier to probe with a multimeter while it's hooked up and running. The skinny black and the red carry the 5v to the Halls.

This type plug, where the phase and Halls are built together is a lot harder to probe.

I found a great video tutorial on how to test the Hall sensor(s). https://youtu.be/vZUXqfdzQb0
Between this video and your instructions, I should be able to definitively test both the controller and the Hall sensors because even though I have the waterproof connector at the wheel, it has the non-waterproof type connector at the controller:
Hall-sensor-wires.jpg

I should be able to do these tests tonight and report back the results. Obviously I'll have to make sure it's in the not-working state before I test, but it happens often enough that it should be easy enough to make it happen before I start.
 
Test complete, and if I did it right, which I believe I did, one of the Hall sensors is bad. There is 4.7 volts on the hot lead, so that rules out whether or not the Halls are getting power, they are. When I tested each Hall sensor, two of the three responded as expected per the YouTube video I posted a link to, I was getting 3 volts on and off as I moved the wheel, but the last (blue) sensor was dead, I got nothing when I moved the wheel. In the video, he gets closer to 5 volts returned from the working sensors, but it's not the same brand motor, so my guess is 3 volts is correct for this motor. One thing I found unusual though, was none of the sensors return any voltage when I move the wheel in the forward direction, they only return voltage only when rotate the wheel backwards, but that may be normal?

As @Voltron pointed out, with the small wires being inside the small connector, it''s difficult to attempt to hold the probes in the tiny holes and move the wheel at the same time, so I made this simple (temporary) fabrication to my probes so they would slip inside the connector allowing me to let go of them and free my hands.

I have been inside a hub motor before, but not one with planetary gears. Thus far I have not had any luck finding a video that shows how to disassemble the hub to gain access to the circuit board that has the sensors. I'm hoping I can contact Bafang for both a manual and a replacement part. I'd prefer to replace the whole circuit board than attempt soldering on a new sensor.

probes.jpg
 
A quick follow up to my recent post about the test results, I watched the video again and I'm pretty sure he is getting voltage returned when he moves the motor in either direction, which mine did not do. I'm not sure what that means exactly, but may be an indication that there are other issues in addition to the one known bad sensor.
 
With an internally geared hub, there's usually a clutch also, so the motor doesn't spin while the power is off along with the wheel, for more bike like pedaling. That's prob why yours only sends a signal one way. Sadly, sounds like time for a new motor.. It's fixable in there depending on how far you like to dig into things, and maybe it's as simple as a broken wire on the way to the circuit board inside the motor for the Halls. But it might be nice to still be riding on a new motor while you tinker with it, if it isn't immediately fixable.

Kudos btw on combining your own research with backup from here, and making probes etc, and letting people know what it was.
That makes you the kind of person people like to help on here👍🤣
 
Here you go...
Random ones from my broken bin... The more assembled one shows the clutch plate and gears, the other one shows there really isn't a board like you see in a direct drive, but rather the board is the whole circular disc, with the sensors showing around the edge where the three clusters of three each black shrink wrap wires lead. The actual sensors are tiny, like 2 grains of rice size, and embedded in the steel laminations underneath. These particular ones, the cover plate for the outer shell is threaded on the edge, and needed a special tool to engage notches in the plate to unscrew. Some others just bolt together, but some variation of this setup is pretty common.
IMG_20210218_091352210.jpg
IMG_20210218_091344977.jpg

If you open it up this far, continuity check from the plug to the sensor end of the wires is prob next, if the sensor isn't obviously broken out of it's socket or something.
You can see in the second pic, the far left sensor wires got squished and the insulation damaged, not sure that's what killed this one, but that's the sort of obvious stuff to look for.

Edit- these ones, you can't even really replace the board, as the winding are wrapped thru it also.
 
Voltron said:
With an internally geared hub, there's usually a clutch also, so the motor doesn't spin while the power is off along with the wheel, for more bike like pedaling. That's prob why yours only sends a signal one way. Sadly, sounds like time for a new motor.. It's fixable in there depending on how far you like to dig into things, and maybe it's as simple as a broken wire on the way to the circuit board inside the motor for the Halls. But it might be nice to still be riding on a new motor while you tinker with it, if it isn't immediately fixable.

Kudos btw on combining your own research with backup from here, and making probes etc, and letting people know what it was.
That makes you the kind of person people like to help on here👍🤣

Thanks for the insight into what I might be facing, especially the part about the halls not being on a circuit board. Let's hope I don't have to replace the motor though because this one is super low miles, probably only a few hundred miles on it. It's out of warrantee though, I bought the bike used from someone who hardly rode it.

Thanks for the accolades, I'll continue to do research and will continue to post findings and outcomes. My fingers are crossed for finding a bad wire and not having to dive too deep into the repair. I'm not afraid of it, in fact I enjoy it, but I don't have as much time for it as I wish I did.
 
rick_p said:
When I tested each Hall sensor, two of the three responded as expected per the YouTube video I posted a link to, I was getting 3 volts on and off as I moved the wheel, but the last (blue) sensor was dead, I got nothing when I moved the wheel.
Are you testing with the controller connected, or using a separate power source and pullup on the hall signals?

If you are testing with controller connected, try the test on the controller end with the motor disconnected. If you still get no voltage at the blue sensor wire on the controller, it means the connection from controller to there is broken.

This is actually more likely than a bad sensor (though sensors do fail).

If the wire was broken on the motor side of the connection (if your probe is on the controller side) you would read the full pullup voltage (3v) all the time, with the hall never grounding it.

That's also what you would usually see with a bad hall, if the hall didnt' fail stuck "on". (if it fails "on", it'll read like you saw, with around 0v).

The hall sensors dont' output any voltage on their own, all they do is (when a magnet passes them) ground the "pullup" voltage on their signal lines, provided by the controller typically (and thus only visible when connected to the controller). That's why this:
In the video, he gets closer to 5 volts returned from the working sensors, but it's not the same brand motor, so my guess is 3 volts is correct for this motor.
can happen, because different controllers use different voltages for the pullups. The higher the pullup voltage, the cleaner hte hall signal will be to the controller, so you can see even 12v pullups on some controllers, so teh hall signals toggle from around 0v to around 12v while hall power is around 5v.

3v is pretty low, but not unheard of. It's usually about 5v, and is often a tiny bit higher than the 5v actually supplied to the halls for power. for unknown design reasons. (I have speculations)


One thing I found unusual though, was none of the sensors return any voltage when I move the wheel in the forward direction, they only return voltage only when rotate the wheel backwards, but that may be normal?
As Voltron replied, yes, because the clutch in the motor prevents you from turning the motor forwards, which makes it easier to coast or pedal the bike when the motor isn't being used, than if it were directly connected (like the DD hubmotors do).
 
amberwolf said:
Are you testing with the controller connected, or using a separate power source and pullup on the hall signals?

If you are testing with controller connected, try the test on the controller end with the motor disconnected. If you still get no voltage at the blue sensor wire on the controller, it means the connection from controller to there is broken.

This is actually more likely than a bad sensor (though sensors do fail).

Wow, great tip!! Yes, I was testing with the controller connected and the power on, but I didn't think to check if 3v is being delivered to the Hall sensors. I just ran this test by first checking continuity of the wires in the cable that runs from the connector at the controller to the waterproof connector at the wheel, and that checked out, the wires in the cable and connectors are fine. I left the waterproof connector disconnected, turned the power on, and then checked the voltage being delivered to the waterproof connector. On the first try I got 3.3v on each of the Hall wires, so I connected the plug, turned the throttle, and the motor worked as normal. So, then I powered up and down a few times until I got it to fail with no power from the throttle or PAS, disconnected the waterproof plug, rechecked the power being delivered to the Hall wires at the plug, and low and behold, the blue wire was dead. So, it would appear that I may have a bad controller and not a bad Hall sensor!

To further test the Hall sensors, I did a continuity test with no power on each of the three wires, and got consistent resistance on all three sensors. I'm not sure if that is a definitive way to test Hall sensors, but between the two tests, I think it's worth investing in a new controller at this point, don't you? That question goes out to everyone on the thread.
 
Sounds like it is time.
Maybe look for one that has dual mode sensor/sensor less running too, they're nice for just this reason .
And then open up the bad one later and see if it's as simple as a loose solder joint that needs to be reflowed or something.
 
Each hall signal line has a pull-up resistor inside the controller. Either you have a bad connection on that or simply where the blue wire attaches to the board might be bad.

I'd take apart the controller and try tracing the blue wire. Check the solder job where it attaches or for any pinch points. If nothing obvious, try tracing the path on the board to the next nearest resistor and check that.
 
rick_p said:
On the first try I got 3.3v on each of the Hall wires, so I connected the plug, turned the throttle, and the motor worked as normal. So, then I powered up and down a few times until I got it to fail with no power from the throttle or PAS, disconnected the waterproof plug, rechecked the power being delivered to the Hall wires at the plug, and low and behold, the blue wire was dead. So, it would appear that I may have a bad controller and not a bad Hall sensor!
Actually, you probably just have a bad cable. Wires and connectors break much much more often than anything else in electrical systems, especially ones on bikes, vehicles, etc., as they get vibrated and jounced and exposed to the environment (in many cases, "waterproof" really means "slightly less absorbent than a sponge". ;) )

Intermittent problems are *almost always* wires / connectors. Almost never anything else. (sometimes broken or cold solder joints or traces on a PCB...but those are also wires/connections! ;) ).

The cable could be bad anywhere along it's length, but the most likely point is where it comes out of the controller or wherever it bends (especially sharply!) in the routing from controller to motor. Or where it goes into the connector housing. It could even be a bad solder joint inside the controller where the blue wire solders to the PCB.


To further test the Hall sensors, I did a continuity test with no power on each of the three wires, and got consistent resistance on all three sensors. I'm not sure if that is a definitive way to test Hall sensors, but between the two tests, I think it's worth investing in a new controller at this point, don't you?
The way to test the sensors without a controller is to power them with something around 5v (3 AA, AAA, C, D batteries in series, an old USB wall charger supply for a phone, etc.), and then to wire up your voltmeter's red test lead with a resistor (anywhere from 1kohm to 10kohm is fine), so the resistor connects one end to the red test lead, and the other end connects to the + of the "5v" powering the sensors. That acts as the pullup from the controller during the signal test.

Then put the black meter lead on the - of the "5v" powering the sensors, and use the red lead on each of the hall signals in turn while slowly hand-turning the wheel (backwards in the case ofa geared hub, either way in a DD hub). As long as the signal toggles from about the same as the 5v supply reads by itself, to around 0v-1v, and all three halls give the same readings, then they're generally fine.


Regarding a new controller...it *would* solve the problem, and also give you teh chance to upgrad eto one that does have a display that shows the thing syou want to see (since it sounds like you wish your bike had that). But unless you just really want to upgrade then it's worth a poke around to see where the bad connection / wire is. :)
 
fechter said:
Each hall signal line has a pull-up resistor inside the controller. Either you have a bad connection on that or simply where the blue wire attaches to the board might be bad.

I'd take apart the controller and try tracing the blue wire. Check the solder job where it attaches or for any pinch points. If nothing obvious, try tracing the path on the board to the next nearest resistor and check that.

I agree, it can’t hurt to triple check the wires, connectors, and even have a look inside the controller before replacing the controller. With that said, my last ebike, which got stolen, also had a controller go bad, and when I opened it up it had an epoxy-like coating over the entire circuit board to prevent vibration damage. I could see where something had fried, but I don’t think a repair was possible on that one. I was able to get a factory replacement from the bike manufacturer for that one, but this manufacturer said they no longer have replacements and that I will need to find one myself. I’m not impressed with that customer service but that is neither here nor there at this point. I’ll report back my findings on this one. Thanks for the tip though, much appreciated.
 
amberwolf said:
Actually, you probably just have a bad cable. Wires and connectors break much much more often than anything else in electrical systems, especially ones on bikes, vehicles, etc., as they get vibrated and jounced and exposed to the environment (in many cases, "waterproof" really means "slightly less absorbent than a sponge". ;) )

Intermittent problems are *almost always* wires / connectors. Almost never anything else. (sometimes broken or cold solder joints or traces on a PCB...but those are also wires/connections! ;) ).

The cable could be bad anywhere along it's length, but the most likely point is where it comes out of the controller or wherever it bends (especially sharply!) in the routing from controller to motor. Or where it goes into the connector housing. It could even be a bad solder joint inside the controller where the blue wire solders to the PCB.

Yeah, it seems that I’m jumping to conclusions here and that I should really be sure it’s not a wiring or connection failure before replacing the controller, so I’ll take that advice before ordering a replacement, especially since I looked into that already and had limited success. There are dozens to choose from, but none have the same connectors as mine, so if I do decide to replace it, I’ll have some work ahead of me to make some of the connections.
 
There won’t be any ability to check any connections inside here, whatever that material is, it was poured in and solid enough that any attempt to remove the circuit board from the housing would certainly destroy it. On to just testing cables and connectors.
5FF392BD-5C3E-4385-8481-BF1C9D074311.jpeg
 
Back
Top