Crystalyte UFO acting up at end of long rides

Joined
Jun 8, 2019
Messages
12
After several years of use under high load, my hub motor starts making an "angry duck" noise at high rpm and under load, after about 15 miles into my rides. When the noise comes up, the motor stops providing output. Luckily I can still coast at low rpm's. The next day everything's fine until I rode it again for some time.

The motor doesn't seem overheated, I don't measure more than 140F (60C) at the stator.

The motor is rated at 3KW continuous, has ferro fluids inserted since about a year, and is hooked up to a 72v battery and an 80A controller (KLS7218S).

Here's an Donald Duck video (ignore the oversized mudguard, it's... work in progress): https://youtube.com/shorts/y1D31ifSqag?feature=share

Any ideas? Tired hall sensor, magnet, or dust?
 
Sounds like it is losing sync with the controller.

The most common reason is a connection problem on a hall signal or the hall power or hall ground. This could happen at any time, but usually does so under vibration or in cold temperatures. It's probably not this, simply because it is happening only after a specific amount of riding time (or that time is creating specific conditions).

If it were a bad hall, it would generally just be failed, and not work at all, unless it only happens when the motor gets really hot inside. That could be happening as your ride heats stuff up.

If it were mechanical contamination, it would probably make the sound all the time.

If it only happens when you push current hard, it could be a magnet unglued, that then rasps on the stator. But that shouldn't happen under any condition other than enough magnetic force from the stator to actually pull the magnet off the rotor, which is a lot of force, and takes a lot of current. It could still have glue on the magnet that works until it gets hot, and then is unsticky enough to let go under high load.

A slim possibility is damaged phase windings on the stator laminations; until things heat up enough to push things against each other hard enough to short it may have no symptoms.

Phase wires could be damaged at the axle exit or elsewhere, but that kind of short usually happens randomly rather than under specific conditions (heating, time of ride, etc), except for the condition of vibration, etc. Or the short is a hard short and just stays nonfunctional.

But on both of those kinds of issues, it's common enough to blow up FETs on the controller, too, and that failure would not be intermittent.
 
I recently diagnosed an intermittent "blip" in the power on my 29er e-bike. I eventually found that it was a combination of deteriorated electrical tape and the crank brushing across phase wires that caused quick cuts in power.
 
E-HP said:
Maybe you have a phase wire connector failing and is heating up. Did you inspect them already?

Inspecting the connectors was how I discovered the problem. Taped them off and lashed them underneath the down tube and the problem went away.
 
Chalo said:
Inspecting the connectors was how I discovered the problem. Taped them off and lashed them underneath the down tube and the problem went away.

I had the duck noise a couple of times, and luckily it was a loose connector both times.
 
Thanks all for the suggestions. Though there are XT150 connectors on two different places along the path between motor and controller, and I have found those to be treacherous at times (especially when shrink wraps covers their butts, hiding they've half slid out of their enclosure instead of plugging into the other end), everything seems tight outside of the motor.

Given the specific conditions wherein the quack takes place, I'm putting my money on Amberwolf's hint about the loose magnet. If this were the case, I suppose this would be repairable by applying epoxy while keeping the magnet pressed tightly in place?

Also, I'm a bit concerned about "my precious ferrofluids". I imagine these are slammed against the magnets, and could interfere with such an operation in a number of ways? Or maybe will it drip out when the hub is opened, and the rotor must be held down as a pan to collect it?
 
Grinning_Ernst said:
Given the specific conditions wherein the quack takes place, I'm putting my money on Amberwolf's hint about the loose magnet. If this were the case, I suppose this would be repairable by applying epoxy while keeping the magnet pressed tightly in place?

The magnet should hold itself in place against the rotor, kept spaced from it's neighbors by all of their fields. If you find that's not the case, you can make spacers of plastic or wood that equal the gap of still-glued magnets, for either side of the magnet(s) that are loose. If necessary, you can clamp the magnet in place using well-padded clamps (if you don't use padding you may crack the magnet...)


Also, I'm a bit concerned about "my precious ferrofluids". I imagine these are slammed against the magnets, and could interfere with such an operation in a number of ways? Or maybe will it drip out when the hub is opened, and the rotor must be held down as a pan to collect it?

The FF simply acts as a "bridge" from the stator teeth to the rotor magnets, kind of like little "waves". Somewhere in Justin_LE's posts about Statorade / Ferrofluid, there are pictures and probably a video (of a motor with a window installed into the side cover) showing what actually happens, if you're curious. May also be on the ebikes.ca "blog" pages, or the statorade product info pages.


The FF will stay stuck to the magnets and the stator teeth; it's very oily and the magnetic particles in it should stay within the magnetic fields they're in.

If a magnet came unglued, the FF will likely be on all sides of it, so to glue it back down will require thorough cleaning of it and the rotor surface to ensure the glue / epoxy sticks.
 
try puting wheel off ground, get someone to put power to wheel, and try just knocking wires with your hand, or slightly tugging on them, you might discover where there is bad connection that way.
I had an intermittent connection on hall wire, which was just inside the axle ( different motor than yours) but could wiggle wires and cause the fault while motor running
 
As fate would have it, shortly after posting this topic, a friend brought in his bike with the same hub motor as mine, for which I diagnosed and replaced a faulty HAL sensor. This helped me to compare his hub's innards with mine. (The only difference between his hub motor and mine is that mine's got ferrofluids inside.)

So when, after a lot of procrastination and parental duties, I pried open my hub on this blessed Sunday, the first thing that struck me was the stench escaping from it, which I would normally associate with a dairy farm. Is it the fluids? The organic decay of bamboo pieces that are oddly used to keep the stator laminations in place?

In any case, I see no obvious indication of loose magnets or stator laminations, in the sense that nothing feels loose. But I do see indications of contact between stator and magnets:

IMG_1118.jpeg

IMG_1120.jpeg

IMG_1123.jpeg

But then I saw this small piece of rope sticking out:

IMG_1125.jpeg

And so when I flipped over the stator I saw decayed pieces of rope all over the place:

IMG_1130.jpeg

Since the ropes are a bit stiff and seem enveloped in a gluey substance, maybe they get more flexible when the temperature rises, which would explain they jam up the space between the stator and magnets after about ten miles.

So now I guess I just hope to have confirmation from this forum before I clean up and slam this sucker shut:

— This is probably my look-no-further moment, and am I merely suffering from contamination by these decaying pieces of rope

— I can just preemptively snip off all these ropes, including those that haven't come loose, without fear of the windings getting out of place

— is the slightest scrape between stator and magnets problematic? I doubt those slight scratches I see would be caused by fiber debris?

Thanks!
 
While its open I would check the hall sensor wiring, make sure theres continuity on all the hall sensor wires, no pics of the sensor wires? I doubt the rope will matter much, the windings should be nicely settled in by now.
Bamboo has been used since chinese started making hub motors, cheap solution and works well.
Do a really thorough check of the hall sensors are working and wiring of hall sensors is ok, wiggle the wires when testing continuity. I'm suspecting hall wires.
 
Cleaning up the rope didn't solve the problem.

I took the bike with me on holiday to the South, and riding in the baking sun discovered that high rpm is not a requirement, as the issue always occurred under load and while the hub motor is hot. So I was mostly reduced to enduro "strolling", with strategic breaks in the shadow, which was a strangely relaxing adventure.

I have now bought a Crystalyte TC100 waiting for a 24" tyre. This will be more suitable for the kind of abuse normally thrown at it.

So now I'm a bit more confident to pry at those magnets on the H4080, and tonight I made an interesting observation when removing both flanges from the hub: there are air gaps between the magnets and the hub shell!

gaps.JPG

Furthermore, when flipping up one of the magnets, there are no glue traces at all -- turns out they are just held in place by their magnetic force:

flipped.JPG

I'm not only surprised there's no glue; my presumption was the ferrofluids would bridge any air gaps between the magnets and the shell. But while it seems smeared into the stator windings and snuck between the indentations on the magnets' edges, there's nothing the back of those magnets. Kind of makes sense if there are no pathways for the fluid to get in there when the hub is screwed tight shut and the magnets pressed into each other, but so much for the heat dissipation...

So the magnets are held in place by their own magnetic properties, and apparently now come loose under heat and load. Could one conclude that the rare earth magnets have degraded, and are more easily popped out of place?

Would it still make sense to try to glue them in place?

Can someone recommend a strong, heat resistant glue with good thermal dissipation properties?

Thanks for reading.
 
99t4 said:
Sorry I don't have answers to your questions, but those are great closeup photos! :thumb: What's your secret?
Well, thanks! I never guessed I'd get feedback on those :)
My aim was to capture the gaps under the magnets, and the white cardboard turned out a good choice for that. Instead of trying to get up close, I used the optical zoom from a foot or so distance and tried winging it to get a shot of the gaps like that. Other than that, I think I got lucky with the lighting, and the "neural engine" of the phone's camera cleaned up the mess.

To answer my own questions:

- the discussion of weakening and /or loose magnets is a rabbit hole that's extensively covered on this website, such as:

https://endless-sphere.com/forums/viewtopic.php?t=91625
https://endless-sphere.com/forums/viewtopic.php?t=97245

- JB weld seems a decent metal to metal glue for my purpose

- very similar adventure to mine, on reddit:

https://www.reddit.com/r/ebikes/comments/qn5rkz/success_i_was_able_to_repair_my_hub_motor_magnets/

The magnets in my hub are 40mm x 13,5 mm x 3mm, and there's basically no gap in between them. I can't even put the one back that I pried up, and damaged a neighbouring one in the process.

As it's impossible to find magnets of that size, I'm thinking to buy 40mmx10mmx3mm, and replace the 46 current magnets with 62 of those, thus changing the poles from 23 to 31.

I'd try to glue those, and bridge the gaps between the magnets and the hub shell, with JB weld.
 
Back
Top