Sogged motor, possibly dead halls

[Oakwright]

So, my bike malfunctioned. It is an ampedbikes motor. It has this wonderful jerky sensation. If it were a gas motor you would say that one of the cylinders is missing. After a bit of research, I found other descriptions of what it is like when a hall sensor dies. I had hoped it was just a simple busted wire, and a little bit of solder would have me fixed up.

But then I opened up the motor, and water came out.


I let it dry out, and inspected some of the lovely rust.


I then put it back together once it was dry, hoping things would be okay once the water was dry. No luck. If anything, it is worse. So I took it further apart, and found massive amounts of rust.


And some of the rust appears to be attacking the hall sensors.



So... Now I'm trying to figure out "now what?"

Option 1: Clean out the rust, inspect the halls carefully, test the snot out of them. Most likely at least one will be dead. Replace dead hall sensors. (note: I have no idea where to get a hall sensor)
Option 2: Send it back to ampedbikes and demand that they fix it since they claim their motors are water proof and it clearly is not, and hope he doesn't point out that I voided my warranty by popping the cover off. (my experience with that will be that he will say warranty is voided, try to sell me a new motor, and it will take two months to get it)
Option 3: View this as an opportunity to get an eZee geared from ebikes.ca (technically I would be upgrading my wife's bike to an eZee and take the ampedbikes motor off of her bike and put it on mine)
Option 4: Something I haven't thought of yet

I kind of like the idea of option 3, but I also feel it's kind of wasteful to just say "eh, the motor is busted" and toss it. That and getting a new motor is a lot of money...

 

[amberwolf]

First, see this thread about the rust:
http://www.endless-sphere.com/forums/viewtopic.php?f=30&t=34135&hilit=rust

The halls you can test, but since you've already got the motor apart you'll need a separate magnet (like one from a speedo sensor on yoru spokes, or a namebadge) to verify if they toggle or not. There's a few threads around discussing testing halls that a search should find. I think some links may to them may be on the wiki, too:
http://endless-sphere.com/w
though it is still under construction.

It's easily possible to just be a corroded wire or connection on a hall or phase.

It is possible it is a phase wire short, too, either corrosion on windings shorting them together or to laminations, but I doubt that as it probably would've toasted the controller or at least prevented it from turning the wheel for you (while a bad hall might let it turn but not correctly).

Unfortuantely virtually no motor is truly completely waterproof. Most arent' even remotely close to water-resistant, even when "sealed". See this video for why:

[youtube]gwlbAJLzI_w[/youtube]



BTW, if you decide it is too much work and feel like chucking it, then depending on cost (I have very little spare $ right now, but a teensy bit) I'd be willing to buy it and have you ship it here to Phoenix, AZ, as I have never fixed a rusty motor yet, and could definitely use it once it's fixed.

 

[ohzee]

Nice to see you over here Oakwright.

That thread linked above by KingFish is some pretty awesome info.. he did a 2k trip in Cali last year on an
ebike and it's a pretty epic read in of itself.

Good luck getting your motor fixed. Welcome to ES.

 

[ZOMGVTEK]

That rust really isn't a big deal. You can replace the halls fairly easily with Honeywell SS41's, and then conformal coat the stator and magnets to help prevent any additional rust.

It should run for quite a while longer with just a quick repair. That is, assuming a hall is dead. Could be a wire issue.

 

[Oakwright]

Yeah I've seen the video before, great stuff! And good links all around here.

The links on the rust ones are great. Makes my rust problems look tiny!

How certain are you that the Honeywell SS41 will be an equal replacement to the halls on my motor? Any advice on where would be a good quick place to buy them from?

 

[amberwolf]

If you need halls, mouser has them
http://www.mouser.com/Search/ProductDetail.aspx?qs=%2Ffq2y7sSKcK%2F%252bhfJuO3wrA%3D%3D
and either the ss41 or ss411 should work; i'm using 3 of the ss41 linked above on this project:
http://www.endless-sphere.com/forums/viewtopic.php?f=30&t=32838
and they work fine there.

The ones actually in your motor are almost certainly some generic Chinese clone of the SS41 or 411. If you look closely you can see 41F or 411 on them. (your pic is not high enough res for me to tell for sure which one it is).
This si the 41F
http://www.mouser.com/ProductDetail/Honeywell/SS41F/?qs=%2ffq2y7sSKcKYHwRh5bjaxw%3d%3d

 

[neptronix]

This just barely started to happen with my MXUS DD motor ( same as yours, and nearly identical to the 9C ), little tiny pits here and there. After 2 rides in very light rain. Completely unacceptable :|

I wonder why still don't have a solution after a decade of hub motors.. or maybe it is just sloppy engineering..

 

[999zip999]

I going to buy a Lyen tester for hub motors. Hall, throttle, ect. It's in the new for sale section. 29usd I think.

 

[dnmun]

you shoulda tested the halls while the motor was intact. it is hard to do when open.

like AW said, connect up the controller to the 5 pin connector, verify there is 5V between the red and black wires, then measure the voltage on the GYB hall wires as you pass a strong magnet past each hall sensor. the magnet should cause the hall sensor to toggle the voltage from 5V to 0V and then back again to 5V when you pass the magnet again. using the hub magnets to switch the halls is easiest which is why you should test the halls first before opening the motor. measure the voltage by putting the voltmeter probe into the back side of the 5 pin connector so you can contact the pin on the back end where it mates to the wire. you have to do this while the controller is active and supplying the 5V to the halls, and the 5 pin connector must be connected to the hall sensor plug side.

 

[Oakwright]

I tried testing the halls... but my testing setup is a bit shoddy, and I can't find magnets quite good enough (bucky balls don't appear to be strong enough). I think I'll reassemble it just to test the halls.

And yup, the halls say on them "41F 030", so sounds like I need the Honeywell SS41F.

I'm liking the idea of that Lyen tester too... that could make this process a LOT easier. If I can't get any good test results in a day or two I might just get one of those.

 

[dnmun]

you don't need anything special to test them. if you already have it open, then you had to take the plastic connector apart already to get the hub over the end. you can just stick the pins with the wires into the other side without reassembling that side of the plug.

it may be worth the effort to check them where you are, with a strong magnet, so check around for a strong enuff magnet, to use to switch the halls.

but the halls have to be supplied with voltage and the ground, red and black, and when the halls switch, they sink current coming down the hall sensor wires from the controller, and the voltage drops when the hall sensor is turned on by the magnet moving by. then passing the magnet again should switch it off and the voltage will climb. that's why it is easier before opening because then the hall sensor wires toggle up and down as the wheel is rotated and the magnets pass by the hall sensors.

also, the halls are chinese knockoffs of the SS41 but they cannot use that on the case or they violate trademark laws, so they just print a 41 on there so every one really knows, but not in violation, and the F and other stuff is just a factory designation known only to the chinese. they do the same thing with knockoff transistors. there has to be a secret code somewhere, but i bet it is a guvment secret.

 

[Oakwright]

I reassembled the motor, just because I couldn't find a strong enough magnet laying around the house. Gave it a few spins while plugged in, and found the source of the problem. One of the halls is constantly reporting 5v. In the picture I posted it would be the one on the right (the blue wire). Which is funny, because it's actually the one that visually appears to be in the best shape. I'll have to tear it apart again tomorrow and inspect it closely to try and verify that there isn't a short in the wiring. Fun!

 

[dnmun]

you can see the other two hall sensors toggling? just the blue does not toggle? i have some halls laying around here. pm where you are you located, should add your real loco in your profile too.

don't tear totally apart again yet if the blue is stuck high and the others toggle. the ground wire to the blue hall could be open on the pcb so see if you tell when you take the cover off again, just the side opposite the wires, because that is where the halls are. check continuity of the blue ground wire back to the black wire in the 5 pin plug. you can do that first.

 

[dogman dan]

Once fixed, it will still be a problem, since the motor will suck water or at least moist air inside every time it shuts off.

I like the idea of just one hole drilled in each cover. Then tape over the holes. Once in a while, when the weather gets nice, do a run with the tape removed, and let the motor cook the water out.

I don't have a link to the product, but others have painted thier stators with waterproofing material. You'll have to dig for the threads that talk about it. Likely in technical. My guess is that regular woodstove or engine paint would work fine though.

 

[miro13car]

It seems there is a need for hole in hub motor cover with removable plug for ventilating moisture out of motor.
He said he rode in couple of light rains.
So clearly rain water got into the hub and evaporated there=rust.
I opened my Tidal Force motor/controller at about 10,000 km and no speck of rust was foud inside. That's after rides in many rains including very heavy deluges when I could not see through my glasses.
Riding TF in even heavy rain is no problem at all . However mech. construction of TF motor/contr. is so different.

 

[Oakwright]

So, I tore it back open, and cleaned it up a bit. First I used an old toothbrush, and a mixture of water and baking soda. That took the majority of it off. I ran over it for a while with a dremel tool, but it was a pretty crappy bit (plastic bristles, not metal) so before long that bit was toast. So then I went back to the toothbrush method. Next I polished it up with vinegar. I scrubbed it up with the toothbrush, and used some ultra fine sandpaper. Then I rinsed it real good. Then I dried it off with a heat gun. And I think it polished up pretty good actually!



There were a few spots that simply refused to let go of the rust though. One spot I polished several times and I seriously could watch the rust reform right in front of my eyes. That stuff can be stubborn sometimes!


I am now fairly certain that the water came in through the axle, following the wires. That rubbery epoxy was not attached very well, and it slowly chipped off while I was cleaning everything, and it revealed some much richer looking rust under it. Usually whenever I was in the rain I didn't run the motor (since my connections were not completely sealed). But I think it was still enough for it to get in there.


After examining the "bad" hall sensor closer, I am not entirely sure it is bad. The wires were not insulated very well, and instead were covered in the same crappy epoxy as the axle. It flaked off while I was working, revealing corrosion and rust working it's way up the wires. It cleaned off nice and easy, but I'm thinking the rust was causing a short between the red and blue wire (+ and signal). So I'm thinking tomorrow I'm going to redo the wiring to that sensor and get it insulated proper and retest everything. I might just get lucky and not have to replace the sensor!



Still, I haven't decided what I want to do for a long term permanent solution. Just from looking at the wires I can tell water can be sucked through it, there is plenty of gap between the individual wires, and I was even able to blow air through it. So I know water can get back in there through the wires. The inside of the axle where the wires go through obviously can leak water. So, I'm thinking first order of business is to treat the whole thing with something to make it more resistant to rust. I'm trying to decide between Boeshield T-9 or ACF-50, but will probably go with T-9 since I can probably just get some at REI. And then I'm thinking of drilling some holes in the plates. Since I do sometimes ride in the rain, I want them to be something that I can plug to avoid mud and water pooling in while riding, but then I can unplug it to let it dry out. Perhaps something I can get my air compressor to blow air through to help dry out. Also, I'm going to very seriously reconsider my cable routing to ensure that all connection points are kept dry.

 

[dnmun]

rust doesn't look that bad. the hall sensors can short between the red lead on the sensor to the blue lead if there is water on it and it shorts through the water. not sure how big a risk that is, usually the phase wire shorts to the hall sensor wire when the wires get twisted off from lack of a torque arm so the axle rotates in the fork and ties the wires in a knot and rips them out. that's usually what shorts the hall sensors.

otherwise they are super reliable and withstand a lot of heat and vibration.

if that is the factory epoxy on those halls, you need to leave it there for now. make sure it is dead before tearing it out.

check the continuity of the ground of the hall sensor back to the black wire from the controller, the ground. if there is ground, and the red is 5V then the blue should turn off if you move a strong magnet past, and then turns back on when the magnet is past. it switches state when it sees the north pole, then switches when it sees the south pole. that's why the magnets in the hub are arranged side by side with the magnets adjacent each other reversed in polarity, and an even number around the hub so the totally number of cycles is always a multiple of 360 degrees

 

[auraslip]

If you watch that video, it's clear that the only fool proof method of preventing water damage is to ventilate the hub. Water comes in, and is spun out right away. Or if your running your motor anywhere near it's limits, the water is instantly turned to steam. I ride my ventilated motor in the rain, and it stays dry.

see: http://www.shelbyelectro.com/motors/mods/ventilating-a-hub-motor

 

[dnmun]

if you think the water is entering through the sheath with the cables inside, then you can put your drip loop below the axle and then make sure there is drain hole in that sheath, at the bottom of your drip loop and the water that gets into the sheath would leave it before entering the axle.

i think we decided on the KingFish's thread that the phosphoric acid was pretty easy to use, and cheap, but you don't have much rust and just keeping the motor hot by using it more when it is raining will reduce the chance of condensation. so maybe do the opposite of what you said, use the motor a lot when raining and see if that helps. storing outside in the rain has gotta be the worst. get a tarp if needed.

but first to isolate the hall sensor problem.

 

[Oakwright]

Great link there about the hole drilling! That gives me a good idea of how to do it right.

I have a bit of an update on my latest batch of testing. Now that everything is all cleaned up and dry I was able to do some more testing and such. The "blue wire" hall sensor was always reporting 4v, and sometimes it would twitch down to 2v, but very sporadically (about once every third turn of the wheel). The "yellow wire" hall sensor was toggling on and off as it should, between 4v and 0v. However, the "green wire" hall sensor was erratic, sometimes toggling on and off properly, and sometimes staying at 0v, and sometimes staying at 4v. I find it interesting that they were all reporting at 4v, not 5v (testing between red and green I had 5v). If the red wire had been shorted directly to the green or blue wire, then I should have gotten a 5v signal, not a 4v. So I'm thinking that the sensors are essentially sticking on.

So I tore it apart again, and started peeling up insulation on the blue and green, and their leads were even worse off than I had thought. Two of the leads ended up just coming apart. They are toast. The yellow sensor supposedly is good, but I'm not sure how much I trust it considering the other two died. So I think I might go ahead and just buy a handful of sensors, they aren't that expensive. I also went ahead and ordered a Lyen tester, since it will make it way easier.

 

[ohzee]

Nice job jumping feet first.. looks like you are well on the road to recovery.

 

[dnmun]

yes, i guess it never gets up to 5V, just toggles is what you want to see.

 

[Oakwright]

Update: I drilled some holes in the cable side plate to help it all drain. I used much smaller holes than was suggested in the previous link. My thinking is that I'm not overheating this thing, so it doesn't need the ventilation for cooling, just for keeping dry. So a few small holes should be sufficient for that. I did put them in pairs though, to create the high pressure / low pressure exchange. I put the outer holes on the edge of the curve to let a bit of the spinning action fling the water out a little easier too.

I only put 6 total holes in it, partly because that's all the more battery my drill had, and it takes forever to recharge (my bike has way better batteries than my drill).

I did not put any holes on the chain side. My chain and gears get gummed up with road gunk on a pretty regular basis, and I didn't really want it to fling gunk into the motor.

I figure if these holes don't cut it, it's pretty easy to either widen them or drill more. But it's not so easy to fill holes in!





I also cleaned out a ton of really nasty rust and grit filled sludgy grease.

I tore apart the bearings and cleaned them up, and repacked them with grease. One of them had a very gritty feeling. The good bearing hardly had any grease in it, but what little was in it was kind of a clearish-whiteish color. The gritty one however was gross. It had so much rust and gunk in it that the grease had turned to a sort of russet amber color. But once cleaned up they felt good. I repacked them with grease (used too much at first) and they feel very nice. Although, I used white lithium grease, which is probably a little more thick than needed. However, they still spin faster than they did before I cleaned them, so it's still a win. Plus the white lithium should last a very long time, and I don't really want to have to tear them apart and regrease them constantly.

 

[Oakwright]

Big update!

The hall sensors came in, and I installed them. I went ahead and trashed all three of the old hall sensors. It took forever to solder them on, since my soldering iron kind of sucks. I didn't end up with any serious burns though, so I consider that a success.

I found some really pretty colored shrink tubing, so that was awesome:


For comparison, this is what it looked like before:


And in the end, the bike is up and running! I rode 8 miles, without a problem at all (besides needing to tweak my rear brake and readjust my seat):


The wheel actually spins better than it ever did. So I would say cleaning out and repacking the bearings was a good thing. Additionally I fixed up one of the bullet connectors that was flaky. I didn't bother taking any pictures of that.

Here is a shot of the motor up close:


I'm using a pouch I got from REI wrapped over the connectors with bungee cords to protect the connectors. I should come up with a better solution:


All in all, it was a total success. Now if only it would get warm enough to ride more! I almost froze my fingers off while riding, and I was wearing TWO pairs of gloves!