SmartPie Restoration

[Lollandster]

About a year ago I sold my much cared for Smart Pie powered bike. I bought the motor back in 2012 when it was first introduced and it has worked fine since. I have had to changes ball bearings and controller to keep it working though.

The new owner, which is a friend of mine, decided to use the bike as an all season vehicle as he no longer has a driver license (DUI). While driving in heavily salted slush, early this winter, the bike stopped working. Instead of dealing with the problem at the time, he put the bike in a heated storage without doing anything with the bike.

Now a few months later he asks me for help getting the bike going again. I was quite impressed with the sight when he got the bike out of the storage. As you can imagine the salty slush that was not washed off had done a number on the bike. I regrettably didn't take any photos of the bike in that stage (all pictures I have are from after washing). Except for the frame (aluminium) I think the whole bike is now completely broken. Since the motor was still moving freely I removed the wheel from the bike and took it with me to work for investigation.

The rim had cracks at every spoke hole, so that was not salvageable. The spokes were rusted and could not be removed without cutting them. I also had to drill out the screws on the brake disc.


Note that the cable is not the original. The motor was converted to external controller some time ago after the internal one died.

First thing you notice when looking at the motor is that the paint is flaked off and the aluminium below is quite oxidized from the salt. The motor still turns without problems so I don't think any salt water got inside.

Using an air gun, I removed all the loose paint revealing more oxidized aluminium.


There was nothing else to do than start disassembling the motor and try to save it.
Getting the gear side off was pretty easy and I could quickly confirm that the inside looked OK with only minor corrosion on the magnets (I think it's chrome coated neodymium magnets?)


Getting the other side off was much more difficult. With insistent persuasion with a hammer I got the first cover off.


The next cover (the controller) was not that easily persuaded. I had to cut the cable flush with the cover and use a propane burner to heat the cover before it finally popped off. By that time I have broken a couple of fins.

The electrical connections inside the controller housing was not corroded at all.

Removing the controller housing and then the left motor cover went smoothly. Separating the stator from the magnet ring was as hard as expected.


There was some corrosion on the iron core, but nothing indication salt water intrusion. Outer layer showed some signs of de-lamination, but not enough to make me worried. 


Now the restoration starts.

[Lollandster]

I am lucky enough to have access to a glass bead blasting at work. This is the perfect tool for removing corrosion from aluminium.
Here are some before and after shots of some of the parts:

(I did that one better before painting)
















Next up is painting and assembly. The parts are already painted and is baking in the oven at work as I write. I will take some pictures tomorrow.

I also pressed out the old bearings as they had lots of play and need replacing. Getting the freewheel off was not easy and required a big pneumatic impact driver.

[Bikemad]

It's scary what salt water can do in such a short period of time.

I'm guessing that the masking of the "Golden Motor Smart Pie" lettering was not intentional:



Presumably the vinyl stickers absorb the impact force of the glass beads, preventing the stickers from being removed, along with the paint underneath them.

The end result will be worth all the hard work. 

Alan
 

[Lollandster]

It's scary what salt water can do in such a short period of time.

I'm guessing that the masking of the "Golden Motor Smart Pie" lettering was not intentional:
...
Presumably the vinyl stickers absorb the impact force of the glass beads, preventing the stickers from being removed, along with the paint underneath them.

I came to the same conclusion when I saw the effect. I'm very pleased with the quality of that vinyl print, I'll definitely you that company again in the future.

[Lollandster]

My painting skills might not be world class, but the result is good enough and the preparation is good so the paint should stick well.




I pressed in new ball bearings. Used Molycote P40 as lubricant, I think that is a good grease.


I tested the hall sensors since the motor wasn't in a known working condition when I started. It turns out the sensors are a latching type, so I learned something new today. All of them was working fine.

I used Dow Corning 744 to seal the side plates. I may regret that in the future if I have to open the motor, but until then it should be tight.





i noticed that the controller housing isn't a perfect circle. The cover only fits one way. This means I have to position the housing correctly so the cable exit corresponds with the axle groove. As you can see in the picture above I did not do that on first try.
If you notice the missing O-ring, I can reassure you that I have a new one, but it makes the cover hard to get off so I won't install it yet.

I have a new internal controller somewhere that I will install. I got it as a replacement when my original controller died, but I never used it as I upgraded to an external one at that time.
Except for that I feel the restoration is finished.

When the silicon sealant is dry I will test the motor to see if it still works.

[NR]

Damn nice job, it actually looks better than new, the flashy dragon claws originally isnt as cool as the flat black. 

[Lollandster]

Damn nice job, it actually looks better than new, the flashy dragon claws originally isnt as cool as the flat black. 

Thanks.
The gear side came out more glossy than then rest. I'm not sure why, but I suspect it's just me not being consistent enough when painting. It is no longer my motor so I'm not going to repaint it. The painted gear that is on there now is just for show. I'm not sure what bike this motor will end up on so I'll hold off buying a new freewheel until I know. With a new freewheel I suspect the contrast will hide the inconsistency of the paint job somewhat.

[Lollandster]

I wonder if should do the same to the other project I'm working on?

[Pwd]

Fantastic restoration! I hate how road salt corrodes aluminum and metal that bad. I know the pain; I've got a couple of vehicles that I drive all year in similar conditions. I'd recommend spraying the outer casing with a spray like FluidFilm or Corrosion Free 3000.

You did such a great job on the Smart Pie; you should tackle that Magic Pie and post some pictures of course.

[Lollandster]

I finally got around to installing the controller. I hocked up a throttle and a power supply. From what I understand, the controller should be OK with anything between 20-60V. My power supply only goes to 30V so I set it to max.

When I turned the throttle the motor just wiggled back and forth very fast for a second before it stopped. To me that indicated that the controller was putting current to the windings in the wrong sequence. I had to solder the hall sensor wires to the controller PCB, since I didn't have the correct plug, so I had therefore double and triple checked that I did that correct. GM has even been so nice as to silkscreen the color coding on the PCB under the socket. The phase wires on the other hand are spade connectors so I didn't bother double checking them. It turns out I got those in the wrong order.

When I tried the motor with the correct wiring, it turned smoothly with barely any current draw. But I could clearly hear the square wave pulses. If it was my motor I would upgrade to Vector controller.

Just for fun I decided to hook the motor up to a oscilloscope. I did kind of know what it was going to look like, but I didn't know how the back emf would look like before I tried it.





The first picture is zoomed out so you can see the phases going in and out. The second and third pictures are zoomed in some to show the PWM.

I don't fully understand the EMF part, but you can see the EMF is building as the the magnet crosses from one phase to the next. The peak at the start of each wave is very normal with EMI from fast rising waves (like PWM), so my guess is that induction from the active phase is changing the magnetic force of the magnets and thus causing a feedback to the next phase. But I'm speculating.

The PWM duty, on this controller, is dependent on throttle only; making this sort of a PWM modulated square wave controller? I would like to compare it with a vector controller. I would think it's still PWM, but it will ramp up and down the duty instead of just turning it on and off.

Picture of the test set-up: