Updated / Issue resolved itself - New MP5 Vectors - knocking inside hub

[Holeshot]

***EDIT*** See my posts below after 2 years / 10,000km of riding with the motor. All is well and the MP5 has been very reliable for me.

Hello All,

I purchased the MP5 Vector 700c rear kit from GM Canada for my urban style raleigh bike. The bike is fitted with standard equipment and aside from the battery mounted on the downtube, and the rider (between 170-200lbs) there is no unusual additional weight or haulage involved.

Upon installation and a single afternoon of riding I noticed a very light knocking sound which seems to echo out of the motor case.

Video: (this is the second motor, after the sound has grown louder over a weeks time)
https://youtu.be/eNyXze2GE3Q

I checked spoke tension and tolerances between bike components. The spokes had worked in and needed to be tensioned lightly but none were severely out of spec. The brake caliper had no issues with clearance and there are no observed markings on the screws for the MP5 cooling cover nor the brake caliper.  As the MP5 is new to me and I'm unfamiliar with it's characteristics I wondered if it was the spokes, even though under load, with the heavy motor .. still working themselves in. Over the next week, with careful riding the sound got worse. It will get louder if the riders center of gravity is leaned to either side of the bike, for example when turning.

I removed the wheel from the bike and removed the freewheel, disc brake rotor, cooling fin cover and the controller leaving just the hub intact and installed it into the frames dropouts. I removed the bikes rear derailleur and the freewheel so that the hub could spin freely in the frame with no drivetrain or braking components able to come into contact. As the motor was spun freely in the air by hand, each rotation there was a dull clunk and at the opposite 180 degree position, a loud knock.

https://youtu.be/ulo8g-bsmOY This is the first motor with no controller, brakes or drivetrain on the bike. The bike is upside down with the motor free spinning in the air. This was taken indoors and the sound is much louder in this recording than in person.

Being unfamiliar with the warranty aspects of GM and the MP5 I opened a support ticket with GM Canada and didn't want to open the motor covers. As the sound was coming from inside the hub and there were no bike components involved I could only expect this was a manufacturing issue. I was really enjoying the MP5 and as the kit was only a week and a half old, I asked GM canada if we could cross ship the motor and they could assess for warranty refund and the exchange went well. I kept everything from the first kit except the MP5 motor and wheel assembly. I retained the original vector controller and screws, cooling fin cover and screws, freewheel and all associated wiring accessories from the first kit. The new hub worked well but sadly experienced the same decline. The new hub suffers the same knock and bump sound as the first motor after only about a week. I've been the only rider of the bike and neither of the 2 hubs experienced any potholes in the time on the road. The tires are 700x32c and I kept them between 30-40psi to keep the shock of the road off the motor and myself. After finally deciding to break down and pop the covers off of a new product, edit: I removed only the cover from the controller side of the housing. The large bearing remained on the controller side cover and slipped easily off the stator/controller housing. I removed the screws from the freewheel side cover but it seemed like it wouldn't budge with light tapping and I didn't want to put unwarranted force on the new parts without prior experience. I found that the stator, winding, magnets and wiring all look to be in expected condition and are not causing any issues. There is no visible scoring on any. The side covers, bearings and axle align to hold everything together in spec. I believe that either the side covers, which don't appear to have any keying to keep them in place other than torque from the screws are allowing one of the bearings to sit loose on the axle or in the cover or the stator is finding itself free movement on the axle. It sounds like the magnets are pulling the stator around within some tight tolerances that the bearings or covers are allowing through the rotation and one of the three are knocking. The motor still felt "tight" on the axle with respect to free movement from side to side and in any unintended direction. It is important to note that there has not been a discernible change to this with either of the 2 wheels and motors. The one I currently have still feels true and tight which leads me to believe maybe the stator is moving on the axle.I have read numerous posts regarding MP and Smart Pie hubs with bearing preload issues but have not seen a post regarding new hubs and the old posts no longer have video or audio attached so I can't compare. The first hub motor I did not open myself and it was taken to GM Canada. I included the video of the hub knocking in my support ticket correspondence with them but they determined it had "no issues" and that I needed to keep my spoke tension in check. They issued a full refund. As the bearing preload issue has been discussed many times, I feel very discouraged having 2 hubs experience severe knocking within a 4 week period on new hubs.


https://youtu.be/eNyXze2GE3Q
This is the second motor, today, as fitted on the bike. I removed the brake caliper to eliminate question of it's spacing with the hub. I have checked and tested with no caliper or cooling cover installed and still, the same noise. The noise is present whether under human power, electric power, or no power just rolling, even in free air as shown in the last video.


Installation information and other details:
Two torque arms were installed, one on each side of the frame. The frame would not accommodate the correct orientation of the torquearm on one side so it had to be installed in the opposite orientation, while the other torque arm is installed as descibed in grin's installation diagram. Neither of the torquearms come within less than a centimeter of any moving parts and have no contact with anything unintended during operation.

Equivalent washers (one inside and one outside of the dropout) and torque arms are installed on both sides of the axle and the wheel appears centered and true upon installation in this manner. There were no fitment issues regarding the drivetrain components and the freewheel was supplied by the kit.

The rear was fitted with a 160mm tekro novela brake disc and the associated caliper. The spacing between the caliper housing and the MP5 cooling fin cover is tight (between 3-6mm) but even with the frame loaded with the rider the caliper remained the same distance from the MP5.  Also, the disc brake rotor screws are not making contact with the vector controller and both hubs have been tested with the rotor, cooling cover and controller removed entirely.


The kit installed with
LCD display,
PAS sensor (not installed)
Thumb throttle
MP5 with Vector controller
7 speed shimano freewheel
2x grin torqarm v4
52V 11.6AH Shark Pack

Has anyone experienced issues like this with new hubs?

Thanks for your help!
-BCC

[Bikemad]

Hi andto the forum,

I have been carefully studying both videos, and I would say that neither of the noises are being caused by something binding or catching either inside or outside of the motor.

It is also unlikely to be a loose stator, as this would only produce a single clunk each time either power or regen was applied as any play in the keyway slot was initially taken up as the torque direction changed.
A loose stator would not clunk every single turn of the wheel, especially when there was no torsional load being applied to it in either direction.

If there is no discernible end float movement between the hub and the axle, then this leaves only two moving parts, the two wheel bearings.
If one of the bearings is not fully seated into the side cover correctly, this could cause the bearing to be slightly out of true, resulting in slight flexing of the side cover.
The slight flexing of the cover could cause the inner part of the bearing to move slightly on the axle (or stator/controller housing), which could produce the noise that I was hearing, bearing in mind that any noise produced by any movement of the bearing on the axle/stator would be noticeably amplified by the large side cover (no pun intended).

Are you able to determine which side of the motor the noise is coming from?

Alan
 

[Holeshot]

Thanks for the very fast response, Alan.   

With much cooler temperatures tonight the motor was only producing about one knock roughly every 15 feet so I couldn't tell which side the sound was coming from. The video recorded today was in 73f weather and knocking was consistent. Similarly, the motor consistently reproduced the sound every day ( ~9/10 rotations knock) for the last 6 consecutive days. Each time was with consistent temperatures of 70-75f. My attempt tonight was in 59f weather. On previous nights, the sound was consistent but would grow louder as time went on, or maybe as heat was sinking into the motor and covers. The weather calls for rain and lower temperatures for the next few days but it is warmer inside. I'll have to test inside tonight or tomorrow morning and see which side is making the noise.

However, the previous motor began making this sound at the end of April / beginning of May with temperatures only reaching as high as 50-60f. With the bike stored in that temperature it would knock from the first rotation, so the motor and casing wouldn't have any heat soak from the coils in it at that point. Not to say that some variance wouldn't exist between two motors with the same sound, but it's worth noting.

[Holeshot]

I took my bike out for a test ride (a little warmer at 69f weather) and the knocking is consistent again with almost every rotation. It seems to be coming from the freewheel side of the motor.

[Bikemad]

When you removed the side covers to look inside the motor, did the side covers come away complete with the bearings, or did the bearings remain on the axle/stator?

Alan
 

[Holeshot]

I mis-spoke in my original post. Sorry.

I had decided to open the covers earlier this week but when I inspected the hubs internals I removed the cover from the controller side of the housing. The large bearing remained on the controller side cover and slipped easily off the stator/controller housing. I removed the screws from the freewheel side cover but it seemed like it wouldn't budge with light tapping and I didn't want to put unwarranted force on the new parts without prior experience.

edit--

At this point, I don't think much can be gained from further testing while fully assembled and installed on the bike. I'm going to ask for direction from GM Canada via the support ticket and probably take the MP5 wheel off my bike tonight or tomorrow for further disassembly and inspection. I will post here with updates and photos as soon as I can.

[Bikemad]

I think it might be difficult to see any physical signs of the cause inside the motor.

I suspect that the noise is being produced by a tiny amount of bearing movement because the inner or outer race of one of the two bearings is probably not 100% perpendicular to the axle.
If a bearing was misaligned with the axle/stator, I don't suppose it would have to be very far out of true to cause a slight flexing of the cover, ultimately resulting in a small movement of the bearing within the cover or on the axle/stator.

If the bearing has not been pressed into the cover 100% true, the cover would not rotate uniformly. If the cover is removed slightly leaving a ~3mm gap, the gap should remain constant at a fixed point on the hub as the cover is rotated on its bearing.
If the gap increases and decreases when the cover is rotated gently by hand, the bearing is either not properly seated into the housing or the machined recess is slightly out of true (but this would be most unlikely as the flange edge of the cover and the bearing recess should have been machined while the cover was held in the same fixture to avoid any misalignment between the two).

However, if a bearing is slightly misaligned on the axle/stator,but correctly seated in the cover, the cover would rotate evenly relative to the bearing's inner race, but the gap between the cover's flange and the hub would vary at different points around the hub.
Unfortunately, the gap between the cover flange and the hub face will also vary slightly as the stator is unlikely to be perfectly centred within the motor ring magnets.
With one cover removed, the axle/stator will usually move sideways due to the pull of the magnets and the clearance in the cold bearings.

If the machined bearing surfaces on the axle and stator are not 100% parallel and concentric with each other, this would be difficult to measure without access to a lathe and a dial gauge (DTI) to check for any run-out of the bearing surfaces and the machined edge that the inner races of the bearings press up against.

From what you have already said, the problem only seems to occur when the temperature rises. This could be due to the bearing clearances reducing when warm and making them more rigid and no longer able to cope with a slight misalignment without physically moving either on the axle/stator or within the covers.
If the bearing was worn and not new, I don't suppose the noise would occur at all.
My theory is that as the bearings rotate slowly, the strain of the misalignment builds up until it is sufficient to cause the bearing to suddenly move (flip sideways from one position to another) and it is this sudden movement that produces the noise which is significantly amplified by the large side cover acting like a speaker cone.

If the noise was being produced as a result of movement between the spokes/nipple and the rim, or the spoke's elbows and the flanges on the hub, the noise could only occur if sufficient external forces were being applied to the wheel rim. But we can clearly see by your first video that this was not the case.

Alan
 

[Holeshot]

 

[Holeshot]

OP here. Update***

I never found the source of the knocking sound inside the hub motor, but GoldenMotor support insisted that it would go away on it's own. It's been about 2 years since I installed the MagicPie 5. I've run it at 52 volts and 1500-1650 watts on road bike tires with very little sidewall or give (700c x 32 - 35 tires) and NO suspension. I've put 10,300 kilometers! (6400 miles) on the motor / bike.

The knocking sound DID go away on it's own after about a month or so. It does come back each spring when outdoor temperatures vary between 10-15 degrees celsius but this only lasts for a moment and disappears again for the year. 10-15 seconds of knocking on each ride for a few days at most. It's definitely temperature dependent. Just to let anyone else out there know - if you have the same sound that I had and you're 100% sure it's not another bike component making the noise.. some knocking seems to be okay and it does go away. I was worried about it when it was new but it's absolutely a non-issue for me at this point.

The motor/controller (from GoldenMotor.bike, the Canadian distributor) have never had any issues. The only problems I've had with the kit are:

-The brake lever switches fail after about a year due to their cheap design and water ingress. Easy to replace.

-The 700c kit rim doesn't take well to potholes at speed when you have very little tire to soak up the forces. It bent once and cracked the next time. I replaced it with an Alexrim's part and haven't had any issues since. I have absolutely had the same kind of pothole hit on the new rim many times since then and it has caused no issues for the Alexrim.

-The spokes from the kit held up okay but eventually broke at the 90 degree bend by the motor due to corrosion. I'm in Canada and we salt like it's going out of style so this is kind of to be expected. I replaced the spokes with Sapim E-strong spokes and haven't seen corrosion over this past winter.

-The Hailong battery case had the old style 4 pin connector on the bottom. These failed after about 8 months and stopped making connection. Again, no suspension so it gets tossed around a lot and this was a known issue that I think was rectified with a new connector style.

All in all the motor is very strong and has been 100% reliable for me doing courier deliveries in the city at maximum stock controller power, despite the occasional knock in the spring weather. When the motor or controller eventually does fail I'll most likely buy the same MP5 again.

[Holeshot]

P.S I weigh between 200-230lbs so the bike has never really had a "light" load.

[mrgardiner]

Thank you so much for the follow up review and comments on the noise, and rim/ break lever change outs!  I am on the fence about ordering just a front motor, or the wheel kit, given the long term quality issues.  Same with the battery kit 48V vs. sourcing my own 52V kit.  Have you ever used the programming features on the Magic Pie 5?  It doesn't look like there have been many software upgrades, but I have not looked too deep.

[Holeshot]

Based on my experience I would say the wheel kit should be fine if you're running a front wheel kit (less weight on the rim) and don't hit big pot holes above 30km/h (~20 mph). Once you're moving too fast it's not easy to avoid potholes with car traffic around you. I've since reduced the maximum speed of the bike to our legal limit in Canada of 32km/h since 90% of my riding is on-road.

I'm using the software from about 2 years ago so I'm not sure about any new features or settings available. The parameters available with my old software seem like a good set of controls. Off the top of my head, I use:
maximum wheel rpm - speed
acceleration - percentage based, I have this left at 100%.
maximum phase current - control the amperage that the windings will take, again left at maximum. This does not directly correlate to amperage drawn from your battery.
maximum regenerative braking phase current  - adjust how much you'll push into your battery and how hard your EBS will brake at it's maximum force. Again not directly relative to the battery current.
regenerative braking deceleration in rpm/second - how fast the EBS will slow you down from riding at speed. I have this toned down a bit from the maximum because I want to use regen often and at maximum values you end up stopping pretty fast. I think maximum EBS throws off people riding behind me little bit
as well as some other similar controls

The throttle input seems to be more of a speed input for the controller. If you want to accelerate slowly you have to moderate your input as you ramp up your speed. I have used a Cycle Analyst V3 which adapts throttle input to more of a requested power or torque value but this isn't a feature built into the GoldenMotor controller. Opinions will vary but I actually like the stock throttle input speed control that the kit provides. Just roll your throttle on from a dead stop and the function feels natural. It helps to maintain your speed regardless of the terrain. With the throttle to requested power level of the CA3 you end up having to constantly adjust your throttle position as you go up and down hills.