Introduction: Electric Longboard Build & Clever CIM Motor Drivetrain
This Instructable documents the improving of a Proline 600 Altered Electric Longboard. It also serves as a useful resource for anyone attempting their own electric skateboard/longboard build. It ALSO serves as documentation for a mini-build: a new drive-train for the ever-so-popular CIM motors, which are often used in FIRST robotics competitions.
Altered/Exkate electric skateboards/longboards are not known for their turning radius (it's HUGE), their weight (the one we used was 48 pounds stock!), nor their performance (1, yes ONE, wheel drive). In fact, about the only thing they are known for is being one of the first companies to produce electric skateboards/longboards. So we decided to make it better. The further we went, the more we realized we could have started with a deck and built up; thus, I will endeavor to make this Instructable useful to both types of builders.
CIM motors have been without a cheap, easy, and compact gear-reduction drive-train since their beginning. Now the wait is over. Read on!
The team was part of MIT's Edgerton Center Summer 2010 Engineering and Design Class, a month-long class where high school students learn and apply real-world engineering skills to various projects.
Note: I'll apologize in advance for any bad quality photos.
Also Note: I am in no way affiliated with any of the companies mentioned in this Instructable.
Step 1: Parts and Tools
Note: All parts listed are for/from the Proline 600 model. Other models' parts may be able to be used with modifications.
Option A: Rebuild an Exkate.
1) Proline 600 Exkate (It's not worth buying one for $600. Go to option B if you don't already have one)
2) Two CIM motors
3) Two of these planetary gearboxes .
4) Some ¼" aluminum plate. At least 12" x 12".
5) Some 2" Aluminum round. At least 2" in length.
6) Two timing belt gears. A good source for these is: http://www.sdp-si.com/
7) Two timing belts . (We used 400-5M)
8) Two ball bearings that have OD’s small enough to fit inside the timing belt gear hubs and have a 5/16" ID. McMaster has cheap ones.
9) Two MBS mountain-board trucks .
10) Four ¼" rubber shock absorption (soft) risers
11) Nine lithium polymer batteries. We used these wired in 3S3P for a total of 9S3P (~36V and 6600mAh). Read up on safety information before using LiPo's!!! They can explode if used improperly!
12) Nine electronic project boxes (used as battery boxes). Preferably plastic and as close to the size of your batteries as possible.
13) About 6 feet of 14+AWG wire.
14) PTC fuses~1Amp. At least 24
15) Deans connectors or other low resistance battery connector. 1 pair
16) Three of these charger s or equivalent. (why you need three will become clear)
17) Three of these power supplies or equivalent.
18) Low voltage detectors or equivalent.
19) electrical tape/ assorted heat shrink tubing
20) Blue RTV silicon sealant or equivalent.
21) LED strips. We used two 12V 16" Red strips and one 12V 8" white strip.
22) Lots of various machine screws.
Optional: 3 balloons (for waterproofing LVD's), some extra hard bushings (rider's preference, but due to the extra long trucks, harder seems to be better), a small switch for the LED circuit
Option B: Build from ground up.
All of Option A minus the Exkate and plus the following:
1) A deck
2) An Altered electronics module .
3) Exkate Wheels: Two regular (http://www.alteredexkate.com/servlet/Detail?no=104 ) and two with drive gears (http://www.alteredexkate.com/servlet/Detail?no=103 )
Dremel with cut off, sanding, grinding bits
Hot glue gun
Hex keys, screw drivers, wrenches etc.
Tap and die set
Optional: Metal Lathe (very helpful), Cold saw
Step 2: Defining Problems and Goals
Note: if you are building an electric board from scratch, you can skip this step.
Problems and goals:
The Altered longboard presented an interesting challenge. On one hand, it sorta kinda worked. It ran, accelerated reasonably fast, and had a reasonably fast top speed. You could get from point A to point B. On the other hand, it had some major flaws:
1. Turning radius.
The minimum turning radius was measured to be about 10 feet. This is simply unacceptable for a longboard. It wasn't error on our part either. When the turning radius was measured, we were leaning so far out that the motor contacted the deck and the outside wheels started coming off the ground; it really was the most the board could be turned. But why is 10 feet too large for a turning radius? If you were riding on a sidewalk and needed to go around a corner, 10 feet would put you in the street. This means that you had to stop, pick up the board, turn it, and then keep going in order to make turns. Now that's a pain, but it was compounded by the 2nd major flaw: the weight.
The board weighed approximately 48 pounds stock. That is ridiculous. The lead acid batteries were the major contributor. The next was the 10 pound motor. The weight made it hard to pick up and carry (say at stairs or when turning).
3. One wheel drive.
Only one wheel (rear left) was powered. This meant that whenever you turned right, the drive wheel would lose downward force, thus losing traction. It would often come off the ground completely, and you would just stop moving.
The giant plastic battery box underneath the deck really hurt board flex. This made for a very bumpy, uncomfortable ride.
1. Decrease turning radius by at least half (5ft).
2. Decrease weight to something more carry-able.
3. Make it 2 wheel drive.
4. Increase board flex.
4. Increase board flex and shock absorption.
Step 3: Turning
The first step was to understand why the turning radius was so bad. Part of the problem was that giant motor attached to the rear truck; its can would contact the deck, limiting how far the rear truck could turn. Trying to relocate the big motor was pointless, since the one large motor was going to be replaced with two smaller ones anyways. However, this wasn't the major problem. Exkate/Altered (same company fyi) trucks are custom- they have a unique footprint and use torsion bushings instead of compression bushings. They try to make their new type of truck sound awesome on their website, but the fact is they don't work as well as regular compression trucks. The front truck on our board would only turn so far before stopping. So we took it apart and found that it was stopping due to an irreparable metal design feature (it's complicated, and the only way to fully explain it is in person with one taken apart). The rubber torsion bushing is held in place by three metal pins and showed some serious strain at those points. In summary, we didn't think it was a very good design, and our opinion is backed up by the poor performance of the board with respect to turning.
So we replaced the trucks. Regular longboard trucks weren't long enough for the giant wheels (even 180mm trucks). More specifically, the length of the axle wasn't long enough for the super wide wheels. Mountainboard trucks are designed to accept very wide pneumatic wheels, so we went with those. We chose to use the bushing-style trucks to save money, but shock absorber mountain board trucks would probably be fine. The MBS trucks came with small metal brackets (used to mount brakes) on both sides. The drive gears on the wheels would interfere with these brackets, so both brackets were cut/ground/sanded off of one of the trucks (making it the rear truck). The front truck was unmodified. Oh, and it turned out that the axle diameters were so similar that new bearings didn't need to be purchased (it was 9.5mm for the new ones vs. 3/8").
We found that using an extra hard bushing in the back and a hard one in the front gave the best ride, but it's of course up to the riders taste.
These trucks cut the turning radius from 10ft down to our goal of 5ft!
Step 4: Losing Weight
48 pounds is heavy. Our main strategy here was to replace the 20 pounds of stock lead acid batteries with lighter lithium polymer batteries, while maintaining or improving speed, acceleration, and range characteristics.
The original batteries were three 12V lead acid batteries wired in series. We chose to use nine 11.1V, 2200mAh LiPo packs for various reasons. 9 was a nice number for wiring 3 sets of 3 packs in series and then wiring those sets in parallel. This gave us a total of approximately 36V (37.8V at full charge and 27V dead) and 6600mAh. We didn't want to go too far over 36V because the Exkate speed controller was designed for 36V, and we didn't want to increase capacity much because the stock range was fine and increased capacitance meant unnecessary battery weight and expense. LiPo's were chosen for their high energy density and (now) relatively low expense thanks to Chinese exporters like HobbyKing. It also turned out that 9 would fit under the deck nicely. We could have gone with six or even three higher capacity packs, but having larger packs would have decreased board flex (larger, stiff battery boxes) and modularity.
Modularity was important. Lipo's are infamous for their violent explosions when handled improperly or punctured. Individual, waterproof, hard-plastic boxes(actually electronic project boxes-these come in almost any size fyi) were used so that if one pack failed, it wouldn't take out the other packs with it. The one issue with the boxes was that if a pack failed violently, the tight enclosure would pressurize and go off like a bomb. This necessitated the intentional weakening of a side of the boxes (aka. taking a dremel to a side and grinding down a thin spot) for a controlled blow-out point.
The batteries were also wired such that if one failed, it wouldn't cause any others to short across it. Battery connectors were not used between packs; they were all hard wired together. However, not every pack was directly connected to every other. Wiring 3 packs in a series set, and then wiring those sets in parallel at the end meant that if one of the batteries shorted, the series set would just cut out. See circuit picture below. If the packs were wired in parallel sets (and those sets wired later in series), and a pack shorted, the rest of the parallel set would short across it and likely blow up, too.
The packs were then wired in parallel sets. Yes, I just contradicted myself, but let me explain. The charging/balancing taps between corresponding packs in the series sets were wired in parallel (3S3P) to allow for easier charging. This way, there were only 3 balancing taps instead of 9. The main power leads were not wired in parallel between sets (they were wired in parallel at the end). However, we didn't want large amounts of current flowing in the small gauge tap wires (which would happen in the case of, for example, a shorted pack). Thus, small 1.2A PTC fuses were placed inline with the tap wires between the packs. PTC fuses act like little circuit breakers, and were small enough for our application. See picture for how the taps were physically wired.
There's one more problem that needs addressing. If three chargers are used in tandem, and the chargers are relatively cheap (like the ones listed in the parts list), then you will need three separate power supplies for them all, even if the power supply can handle the wattage. The problem with using a single power supply isn't related to power or the supply at all, but to the charging circuitry. Refer to the third battery diagram. Cheap chargers likely have the negative balancing tap lead output (the black one) connected directly to the ground (black/negative) input. This means that if all the chargers are wired in parallel to a single power supply, all of the chargers will be at the same potential and not allowed to “float” up to the higher voltages, e.g. 33.3V and 22.2V, necessary for charging the upper parallel stacks. If a separate power supply is used for every charger, the chargers will be able to float up to the necessary voltages. When in doubt, use separate power supplies for chargers. Alternatively, you could wire your batteries with connectors at all the series connections and disconnect them for charging.
It turns out that we also saved weight by changing motors. Each of the smaller CIM motors we used weighs about 3 pounds. By using two of them instead of the one larger 10lb motor, we saved about 4 pounds. Switching trucks also saved us a few pounds.
This battery scheme, or some variation, would work great for anyone building their own electric longboard from scratch.
Step 5: CIM Motor Drive Part1
The first step is figuring out how/where you are going to mount the motor. After figuring that out, you can cut (band saw) the aluminum plate into the proper bracket shape. For example, we were mounting the motor parallel to the truck, so our mounting brackets were shaped as in the pictures below. There is a small raised ring near the shaft of the motor that is approximately 1/8" high and 1" in diameter (see: http://www2.usfirst.org/2005comp/Specs/CIM.pdf ) . A large (~1") hole has to be drilled in the bracket to allow it to sit flush against the motor face. The next step is to drill the two clearance holes for 10-32 screws that will secure the bracket to the face of the motor. Then countersink the holes so that 10-32 flat head screws may sit flush with the face of the plate.
The next step is securing the planetary. The great thing about this particular planetary is that it is built for a 5/16", which happens to be the shaft size of the motor. All that you have to do is grind/file a flat spot on the shaft so that it can fit in the D-slot planetary hole. The next step is actually securing the planetary to the motor/bracket assembly. This is done by carefully transferring the 8 hole locations of the outer ring of the planetary to the bracket. Then those holes need to be drilled and tapped for #8-32s. The outer ring of the planetary cannot be drilled out or threaded due to the very brittle steel used in its construction (it WILL crack if this is attempted). The #8 screws will be a pretty loose fit, but this is ok and actually good because it is VERY hard to get the 8 hole locations exact with a drill press (if you have access to a mill, that might make it better); having the loose fit builds in error protection.
The next step is to locate the four shallow holes on one side of the planetary's inner disk (see pics). Drill out and tap (using a *bottoming* 10-32 tap) these holes; make sure you don't drill the holes any deeper because you run the risk of drilling into the planetary gears. These holes will be the ones holding the timing belt gear. The best method we found for transferring these hole locations to the gear was cutting the heads off of some 10-32 screws, grinding them down to points, screwing them into the tapped planetary holes with the points facing out, and simply pressing the gear onto these points. Doing this would leave marks for clearance hole locations in the gear. If the gear you are using has support struts, then you will likely need to cut out (probably with a dremel) some of them.
The next step is to modify the gear (this step can be done before the others). The hub of the gear needs to accept a bearing that has a 5/16" ID. The OD will be dependent on your gear. For example, if you are using a small-ish gear, you might need a smaller bearing. The whole point of having the bearing is to transfer forces (from the belt) from the gear to the shaft of the motor. The gear cannot be directly attached to the motor shaft because the shaft will be spinning at a different rpm than the gear (because the gear is attached to the planetary). Thus, a bearing is needed. I tried to draw a cut-away of the assembly showing the force transfer (see pics). Anyways, the best way to do this is on a lathe because bearings require very precise mounting surfaces. I first bored out the hole in the gear to a clearance diameter of 3/8" so that the motor shaft wouldn't contact the gear hub. You can see in one of the pics that the gears we used had a hub that extend past the front and rear planes of the gear. I cut this extension off. Then using a boring bar, a recess was cut in the hub to accept a raised portion of the planetary, allowing the gear to sit flush against the inner disk of the planetary. Then, again using the boring bar, the bearing surface was cut into the hub. I didn't put the recess and the bearing surface right next to each other; afterwards, I wish I had- see the pics for why, but basically, the motor shaft wasn't long enough to reach the bearing, so we had to flip the gear over, meaning that it was no longer flush with the planetary's inner disk.
After machining the bearing surface, the whole thing can be assembled. First, screw the bracket to the motor. Next, slide the planetary on, making sure that the holes for mounting the gear face out. You'll need to find at least 3 out of 8 good holes to bolt the planetary onto the bracket with. "Good" here is defined as putting the least amount of stress on the planetary and making the planetary as concentric with the motor shaft as possible. If the planetary is off-center, the gears will grind and make a lot of noise. The inner disk/motor shaft should be somewhat easy to spin by hand. If it is not, or if it locks up, then you'll probably need to find another set of holes. There is no good method for doing this…just try lots of combinations until one seems to work best. The planetary will likely need to be spaced out from the bracket slightly to prevent rubbing. We found that #8 nuts with the threads drilled out and small washers worked best for spacers between the planetary and bracket. Another thing that helped was not tightening the screws down all the way; this lets the planetary find the spot it likes. Next, slide the gear on and screw it to the planetary. Now undo all the screws one by one, put Loctite on them, and screw them back in. Loctite will keep them from vibrating off. Finally, mount this assembly on your robot/longboard/invention/etc, put on the timing belt, and your good to go!
Mounting the planetary in this fashion gives a 1:4.5 gear reduction (not including your timing belt gear ratio, of course). If you mount the planetary in a different way, I believe you can also achieve a 1:4 reduction. You can get some pretty serious final gear ratios if you use the right timing belt gears. Another great thing about this drive is that multiple planetaries can be stacked to achieve multi-stage gear reduction for a lot of torque. For example, instead of attaching a gear to the 4 inner disk holes, you could attach an aluminum disk that would also bolt to the 8 outer holes of another planetary. Then the timing gear would be attached to this second planetary.
One flaw with this method is that you have to use a relatively large timing belt gear (for the gear attached to the planetary) because the 4 mounting holes are pretty far out. Another flaw is that getting the 8 holes in the bracket drilled precisely enough is very very hard; that being said, we did it twice without having to re-do anything.
Step 6: CIM Motor Drive Part2 - Application to the Longboard
The CIM motor drive was originally developed to be compact enough to fit under a longboard deck. We decided that only one bracket was sufficient to hold the motor in position. We could have fixed other side of motor, too, but they're so light that ¼" aluminum on one side was fine.
The CIM motor spins at a much higher rpm/V than the stock motor. In order to keep stock torque and top speed characteristics, we needed to gear them down- thus the need for the planetary gearboxes. The CIM motors @ 18V + the planetary + 1:1 timing belt gear ratio ~= stock motor @36V + 19:44 timing belt gear ratio in terms of rpm and torque. Note: we know that CIM motors are meant to be run on 12V and that we are running them at 18V (2 motors in series on a 36V circuit = 18V per motor). This is fine; they handle the higher voltage and rpm without any issues.
The Exkate drive wheels have a 44T gear permanently attached to them. We did the gear ratio calculations and it turned out that a 1:1 timing belt gear ratio was fine, so we bought 44T gears to attach to the planetaries. If you use a smaller gear, you'll get more torque (and thus acceleration), but I can tell you from experience that there is PLENTY of torque. A larger gear will give you a higher top speed (but less acceleration).
But why did we go with 2 motors instead of 1 bigger motor with a solid rear axle or differential? (You can buy differentials for tricycle-style bicycles that are small enough to fit.) One reason is that both would require a complete chop-and-rebuild of the rear truck in order to get them to fit in the proper place; we'd basically have to design our own truck. But besides that, there are other problems with both. A solid rear axle would mean that both rear wheels would be spinning at the same rate. This is bad for turning. When a car turns, the outer wheels have to spin faster than the inner ones. If you have a solid axle on a longboard, the outer wheels have to slip in order to make the turn. This would seriously hurt the turning radius. A differential would alleviate this problem; however, in the case of longboards, it causes a different problem. A longboard is turned by leaning in the direction you with to go, causing more force to be on the inner wheels than the outer wheels. This means that the outer wheels have less traction. With a differential, the side with less resisting force (traction) gets more torque and vice versa. This means that, in very sharp turns where the outer wheel comes off the ground, the outer wheel will get all of the power and the inner wheel will stop spinning. This is the same problem we had with 1 wheel drive, but now on either side of the board! In summary, a solid rear axle or differential were not good ideas.
So we used two motors. However, one thing we didn't foresee until it was too late was that we had created an electronic differential by wiring the motors in series. If one motor has less load than another, it will steal power from the other motor. Ideally, the two motors should be wired in parallel. However, that can't be done with CIM motors because, while 18V is fine, 36V would probably cause them to explode. The second best possible solution is to find two relatively low rpm/V (so that you don't need the planetaries), light weight, compact, 36V motors and wire them in parallel…I couldn't find any such commercial motors. The best possible solution would be to completely overhaul the power system by using two motor controllers (one for each motor), finding motors that match our needs exactly (and don't need the planetaries), and creating a new radio scheme (because the radio receiver is integrated into the current electronics module).
Note: when I say “motor”, I mean brushed motor. We decided to go with brushed motors over brushless because of their simplicity, being cheap, and ability to wire them in series or parallel. Brushless motors are more efficient and more powerful, so if anyone wants to undertake a brushless version of this project, that would be cool! (I'm building one with in-wheel brushless hub motors, check it out here: http://www.mitrocketscience.blogspot.com / ) .
The last thing to mention is how we mounted the brackets to the trucks. A hole was drilled in the bracket plate large enough to fit the truck through. Two small two piece clamps were machined out of 2” aluminum round stock. First, we cut 1/2 inch pieces of the 2” round. Then those had a big hole drilled in the center smaller than the diameter of the trucks at the point we were going to clamp it (if these holes end up being too small, filing can fix them). Then they were notched (see pic) and holes drilled (for a #10-32 tap) perpendicular to those notches. Two clearance holes for 10-32 were drilled through the face of each (to be used to screw the bracket to). Then they were cut in half (see pics). Then the bottom sections were threaded for 10-32 and the top sections were drilled out for clearance. The end result was two, 2-piece clamps that would fit snugly onto the trucks when screwed together. After fixing them to the trucks, the two face hole locations were transferred to the brackets and drilled and tapped for 10-32. Then everything was shimmed (to get the motors straight because I can almost guarantee that the clamps won't sit perfectly straight on the trucks) and bolted (with loctite!) together. Note: a better way to do this than clamps would be to weld the 2” round disks onto the truck.
Step 7: Flexibility and Shock Absorption
The deck is one of the two things Exkate really got right (the other is the electronics module). It is a VERY nice deck. Each of the 9 layers that make up the plywood deck was stained prior to assembly. It also flexes like a dream when it doesn’t have a giant battery box screwed to it. Having individual battery boxes really helped flex, and thus, shock absorption.
The other thing that helped shock absorption was a 1/2" of soft rubber risers between the trucks and the deck.
Step 8: Miscellaneous Notes
The low voltage detectors (LVD) plug into the taps and monitor cell voltage. If any cell drops below 3V (the lower threshold for LiPo's), then the LED turns red and an annoying alarm sounds. A neat trick: You can put them in balloons and fill the neck of the balloons with hot glue to waterproof them.
Hot glue can also be used to hold wires in place.
While the longboard we built is technically water proof, I wouldn't run it through anything more than a light drizzle.
Adding LED's: Three waterproof, 12VDC LED light strips were purchased for the board: two 16” red strips, and one 8” white strip. These were wired together in series (with a switch inline so that they could be turned on and off) directly to the speed controller input power leads. They are held in place with hot glue.
Note: the batteries were not hardwired to the controller. A deans connector is used to connect and disconnect the batteries from the controller.
Step 9: Cost and Feasibility
The following are estimates:
Electronics module: $170
CIM motors: $60
Chargers/power supplies: $45
Other parts (wire, aluminum, hardware, gears, etc): $100
Total: ~$795 (not bad and totally worth it)
Proline 600: $600
The extra $200 makes all the difference. It turns something barely rideable into something amazing to ride.
The Altered board we altered was donated by a private party (thanks Stephen!) to the Edgerton Center.
Step 10: Performance and Pictures
It is truly an amazing machine. It accelerates smooth, has ridiculous amounts of power, and is fast (I'd say around 15mph top speed). It also has a turning radius of about 5ft, which isn't bad for long boards. It weighs a little over 30 pounds, which is significantly better than 50, but still rather heavy to lug around. One flaw with it is that it only has 1/2" of ground clearance, which is fine for sidewalk bumps, but bad for large cracks, pot holes, etc. It's also somewhat noisy due to the planetary gear chatter.
There is extra, lip-like ¼" aluminum on the bracket extending above the planetary (see pic). During turning, that hits the deck before the belt does, sparing the belt from damage. On your deck, make sure the belt doesn't contact the deck. If it does, you can try cut outs or more aluminum.
Now enjoy some pictures of the longboard that has been affectionately codenamed "eXKate.C.D."
For more, including links to videos, check out: http://www.mitrocketscience.blogspot.com/