Re-gluing the magnets on a motor with oil lubrication
This week I had a new type of electric bike hub motor breakdown. The Xiongda 2-speed rear hub motor on my 26" wheeled bike started to make a harsh scraping sound - similar to a disc brake with pads worn to the metal. The sound only came when when power was applied against resistance (like going uphill), and the bike felt like it was losing power. Stopping at the roadside, I found I could stop the rear hub-motor wheel completely with the brake, run the motor with the throttle, and hear the motor turning inside the stationary wheel, making the awful scraping sound.
Here's a short video of what happened when I powered the rear wheel, off the ground. First it runs in low speed and I apply the brake while leaving the throttle on, then I do the same in high speed:
The sound could be made in high or low motor speed. Limping the bike home (up the mountain…) I found I could run the motor on 3/5 power without it slipping and scraping.
I dismantled the motor, expecting a sheared key on the freewheel plate, or some other sort of freewheel problem.
After checking it all carefully, and trying a new freewheel plate which didn’t stop the sound, I noticed some of the magnets in the magnet drum were a fraction out of alignment with the drum edge. I found I could slip the individual magnets out with gentle levering. Evidently the magnets had come unglued and were spinning inside the drum, not turning the drum.
This motor is one of 2 XD 2-speed hubs I’ve put Automatic Transmission Fluid (ATF - a light oil) into, in the hope of improving cooling of the copper coils on steep climbs. It also makes the motor sound rather smoother - the XD 2-speed is a pretty noisy motor, with its complex gears and freewheels. As I’ve discussed elsewhere in my blog, the coils of a hub motor have trouble losing heat when labouring. To escape from the motor windings, heat must travel either by conduction through the laminated core to the axle, or by radiation from the coils and core to the magnet drum, then radiating again to the hub case - either way is very inefficient. Adding some ATF provides good lubrication throughout the hub, and the flow of fluid can be expected to carry some heat from the coils to the case. The drawback is that some leakage of oil can be expected through the axle bearings and elsewhere, especially if the bike is laid on its side. I haven’t yet concluded whether the ATF is making a significant difference to motor temperature but the motors certainly run quieter.
I’m not sure what effect the ATF has had on the magnets coming loose in this case. Clearly however, once they came unglued, the oil would have made it much easier for the magnets to slide around inside the drum. Perhaps without ATF the magnets would stay in place without glue. But did the ATF have a role in ungluing them? This motor has had ATF in it for a few hundred kilometres and a few months (it's my less-used bike), but I have another XD motor which has had ATF for 1000 or 2000km and a few months, without trouble.
The magnet drum has 20 curved high-strength magnets arranged around inside the drum cylinder, that rotate around the 18 poles of the armature. Before removing them, I marked the magnet ends with alternating patterns of file grooves: 1 groove and then 2 grooves. This represented the alternating polarities of the magnets, so I could replace the magnets with the right orientation, polarity and pattern.
To repair the motor, I needed to glue the magnets back into the drum. The biggest challenge was cleaning the oil off the magnets and the drum. I used turps, metho and degreaser in turn.
I left the 2 sealed bearings in the drum, and took care not to get solvent or degreaser into them. After cleaning, I sanded both surfaces, and ground a bit of roughness into the drum surface with a die grinder and disc. My main concern is that there be enough friction to prevent sliding.
I used runny epoxy resin (like is used for fibreglassing) to glue the magnets in. The magnets hold themselves into the drum very firmly with their magnetism - no need for clamps. I left all magnets a little proud of the drum, and aligned them together by pressing the drum down onto a flat surface.
I recognise that the epoxy will weaken when the motor is hot, but the glue's job is to resist shear forces - stopping the magnets sliding around inside the drum again - so even if it was softened by the heat it should do its job.
Installed back on the bike, the motor is working perfectly again. Of course I can't be sure until it's done 5000km like this...
Here's a short video of what happened when I powered the rear wheel, off the ground. First it runs in low speed and I apply the brake while leaving the throttle on, then I do the same in high speed:
The sound could be made in high or low motor speed. Limping the bike home (up the mountain…) I found I could run the motor on 3/5 power without it slipping and scraping.
I dismantled the motor, expecting a sheared key on the freewheel plate, or some other sort of freewheel problem.
Here's the tool I made to unscrew the motor side plate |
Here are the magnets slipped out of their places, but still holding onto the drum with magnetism |
I’m not sure what effect the ATF has had on the magnets coming loose in this case. Clearly however, once they came unglued, the oil would have made it much easier for the magnets to slide around inside the drum. Perhaps without ATF the magnets would stay in place without glue. But did the ATF have a role in ungluing them? This motor has had ATF in it for a few hundred kilometres and a few months (it's my less-used bike), but I have another XD motor which has had ATF for 1000 or 2000km and a few months, without trouble.
The magnet drum has 20 curved high-strength magnets arranged around inside the drum cylinder, that rotate around the 18 poles of the armature. Before removing them, I marked the magnet ends with alternating patterns of file grooves: 1 groove and then 2 grooves. This represented the alternating polarities of the magnets, so I could replace the magnets with the right orientation, polarity and pattern.
To repair the motor, I needed to glue the magnets back into the drum. The biggest challenge was cleaning the oil off the magnets and the drum. I used turps, metho and degreaser in turn.
Here are the magnets drying on a strip of sheet metal after cleaning - you can't just put them in a box or they'll smack together |
The magnets being glued back into place with epoxy. Note the file marks on the magnet edges, so I could replace them in the right orientation and order. |
I recognise that the epoxy will weaken when the motor is hot, but the glue's job is to resist shear forces - stopping the magnets sliding around inside the drum again - so even if it was softened by the heat it should do its job.
Installed back on the bike, the motor is working perfectly again. Of course I can't be sure until it's done 5000km like this...
Hi Bruce. It looks like you have been busy. Thanks for your very informative and detailed blog.
ReplyDeleteJust in case you don’t have time to look at my comments on the other blog.
Bonnie ask me to say Hi from her at Xiongda
Hi Bruce, just wondering if you have had this magnet slipping issue in your other Xiongda yet? I'm keen to do the ATF mod for better cooling but not sure it is worth the risk of the magnets coming unstuck!
ReplyDeleteHello Mike, this is still my only motor that's had slipping magnets. The XD 2-speed I've used the most (~8000km) has ATF but no problems with the magnets. The other couple of motors in use are fine too, but no ATF. I think ATF is unlikely to make trouble for you, except the usual issues of leaking out the axle if you lie the motor on its side.
DeleteHello Bruce, and many thanks. I'm pretty sure my situation is utterly identical. Went over a bad St. Louis bump which lost me afender bolt and the harsh started right away. Same feeling of stall, same dimunition of it with either less throttle or higher speed (centrifugal force landing the magnets?), same limp home. Got the motor mostly apart, and no damage to gears or clutches and all turns. I'm poised to pull the drum off today, but I already know that I have one cracked magnet that wiggles with a tiny screw drive. 2 question. 1. "Turps" and "degreaser I get, but what is metho short for? 2. What do you think about 75-90 weight Mobil 1 synthetic gear oil as a lube? FYI, I followed d8veh's lead and put on synthetic PTFE gear grease on the metal gear side and lighter PTFE lube on the nylon gear side. I'm happy with the lubing, per inspection, but lots of black flecks, which I assumed are charred teflon. Thanks again for sharing Bruce. I would have never have diagnosed this myself. I will take and send pix soon.
ReplyDeleteHello Bob, it's very pleasing to know my efforts have been useful!
Delete1. "Metho" is short for methylated spirits (in Australian), which means ethanol, with some added poison, from the hardware shop. Isopropyl alcohol would be a good option, but metho is cheaper.
2. Gear oil is worth a try, but I used ATF because of its lightness: I want the oil to splash from the windings to the motor case to carry heat. I can't see any harm in trying different oil. Perhaps there is some risk to the nylon gears from using oils, but I haven't seen it yet.
3. While you're fixing your motor, have a look at my recent edit to this post: https://bruceteakle.blogspot.com/p/it-seems-clear-xiongda-2-speed-motor-is.html . I recently had an XD breakdown that I thought was magnets, but it was broken TIG welds on the nylon sun gear. Just make sure this hasn't happened to yours.
Is your XD running on specified volts and amps, or have you over-powered it?
Good luck with your repairs, Bruce.
Qucik update, Bruce. A magnet id broken into 5 pieces and won't go back together correctly. Seems as if the internal magnetism repels the parts from each other, when fitted together. So I wrote to Bonnie for new ones, and a few sundry other parts. Until then, I'm riding my wife's 750 watt Evelo mid drive, with NuVinci Harmony auto trans. I'll also be riding my old mini bike rescue. It has the 48 volt, 750 watt, BLDC motor, geared down to 480 r/m, chained to my NuVinci 171 Developers Kit auto trans, used as a jack shaft. Auto shifting starts at ~350 r/m, and it is sprocketed for ~21 m/h. It will climb anything. St. Louis police are sanguine (bigger fish to fry) so I can grocery get with it, with impunity.
ReplyDeleteAs for lube, I'm convinced. The idea of it splashing and dissipating heat is compelling, so I'm going with your ATF. FYI, I see no evidnce of glue between the magnets and the drum. Might just have ground off (maybe that's the black specks I see in my lube), but do you know what they do at the factory? As for leaking, it's there, but minimal, even with PTFE lube that' probably less viscous than the ATF. I'm with you on that.
Here's a link Ii sent Evelo when they expressed interest in my Nuvinci based mini bike. Complete with halting narration and loss of words....
https://photos.google.com/share/AF1QipN0ROhDQnUCvUUxsw5bHGMQRKt8E1jwvUfjo3AXFIXY-IkYVpdndspvot9L6qlPyg?key=LXNhR29yMmo3S2g0ZmlzRVFNOUdMVWxib2xXVXBR
thanks for the articles. I am going to order this 2speed hub from china and put it on a trike. I assume one just takes off the lid and pours some ATF inside and then reaassembles right?
ReplyDeleteHello Rolland, you’re right: just unscrew the side plate and pour some ATF in.
DeleteThese XD 2-speeds can be good on trikes. I have a contact who has done this: an old man who put an XD on the front of his upright trike. He achieved his goal of improved hill climbing, and was happy with the high speed on the flat (but I was a bit worried about the high speed on the upright trike - they make me feel very unstable when they don’t tilt like a bike). Good luck!