Search all of my sites with Google

Friday, May 29, 2009

So Much For That Idea

After being unable to find the answer (because I was looking for it the wrong way) via searches, I asked the question at DIY Electric Car forums, and got my answer:

Quoted from that forum, the basic problem I'd face in the conversion would be this:

2 pole armature has the coil sides (wires in the steel slots) spaced approx 180 degrees. A 4 pole armature needs coils side spaces at approx 90 degrees. So a complete wire tear out and rewind is required. And then there are the magnets. You'd need to change to N-S-N-S instead of the N-S 2 pole arrangement.

I'd actually have to remove the magnets, physically cut them in half, etc., in addition to the total rewind of the motor. I'm not sure if I managed to do it I would find enough gain to be worth the effort. :-(


So the quest for easy power continues. :-)

Thursday, May 28, 2009

Four-pole Motor Conversion

At the suggestion of a blog reader, I may try converting the damaged powerchair motor into a four-pole version, by adding a second set of brushes. First I have to determine if the motors are wired in a way that allows this to work. I am not sure if they are, or if they are not if I can fix it without actually rewinding them.

I have to make a new brush mounting plate for it anyway, so this would not be that much more work. I have spare parts now for brushes and brush holders, from more motors that have the same softer-plastic type of brush mounting plate (though neither of them looks particularly problematic, compared to this one, excepting the large amount of gearbox grease that has oozed into the motor housing of one of them).

I've got to dig out the pieces of material I have from old component-mountings and whatnot that are sturdy enough to withstand the heat without deforming. I think I have a few large enough to make the plate from. I might also have to modify the motor housing a little bit around the plate-attachment area, to ensure clearances for the extra new pair of brushes (90° around from the original pair), so they can't short against the metal motor casing.

I've been doing some web-based research on the topic, but an hour or so of googling has not found anyone else that's attempted to do this. I must not be using the right keywords, because I am sure that someone has tried, even if it failed.

Friday, May 22, 2009

Wheelchair-ComputerPanel Flatbed Trailer

Today I put together the basic version of the flatbed trailer, using the wheelchair main wheels and quickrelease mounts for them at the back corners of the aluminum panel, along with the wheelchair caster wheels (about 5" diameter) for the front corners. Those wheels could be replaced with larger diameter ones, and even maybe 1/2" wider, if I find a pair that are I'll probably change them out (these are pretty hard rubber, with urethane-foam "tubeless tubes" inside), so it will ride better over bumps.

They're mounted to the old mainframe-style computer front panel, which is made from aluminum a little more than 1/8" thick, folded and welded into a semi-box shape, with lips folded back around the sides about 1/2", rather than making a complete back to the "box".

So it's fairly strong. I can stand in the middle of it with the wheels attached, and wiggle around, and it rolls across the floor a bit (as it should) but it doesn't significantly bend or distort under my ~150-pound weight. I expect that with a bit of reinforcement, such as a 2x2 wooden X-frame underneath it from corner to corner, bolted to the top and sides, I'll be able to carry significant amounts of cargo on it without worry. Possibly 300+ pounds, which should be more than I will need to move at one time for the forseeable future.

Above, you can see the caster wheel mounting axle bolted to the side front corner of the panel. It is very slightly rearward-angled at the moment, and needs to be straightened. The bolt holes in the caster mount are not round, though, they're elongated. So I need to find some larger diameter bolts than the ones I have at the moment, which I can torque down harder so it can't angle like that under load. Otherwise it isn't going to keep the load on the mount/etc properly, and might damage the corner of the panel, even with the reinforcement I'm planning.

The above pic shows the rear wheel mounts, which have only half of the width of the bracket bolted to the panel's sides. The bolt holes are wider-spaced than the actual side of the panel, so I can't directly bolt both top and bottom holes to the panel. I have not yet made them, but there will be an angle-iron (aluminum, actually) bracket holding the top part of the wheel mount to the top of the panel. I still need to drill the very large hole in the bracket for the wheel's hollow quickrelease-axle-mounting bolt , so the angle-iron can be flush-mounted against the wheel mount and the top of the panel, plus find more short bolts with nuts that I can use to fasten it together securely (I have a number of long ones, but I don't want to use them on this unless I have no others). Somewhere in my salvaged stuff I'm sure I have what I need.

Because the wheels are quickrelease types, I can take them off easily to store the trailer, so it doesn't take up all that much space, unlike my other trailer which essentially takes up the same room as a very large armchair or a wheelchair.

Now, I still have to put together the hitch and towbar, which will go thru the hole in the front of the panel that was originally used for a wiring bundle, and be bolted to the top of the panel along it's length. I may use a crutch for this, as I have a number of them that are missing various pieces essential for use as crutches, but are otherwise fine, and have fairly strong tubing.

The towbar itself isn't going to need to support any vertical or lateral weight, because the front and rear wheels will support that, so the tubing only needs to be strong in tension and compression along it's length, which a crutch is great at.

The hitch itself I'm still working on, but it will go directly behind the rear wheel on the bike(s). The recumbent Crazybike2 has it's cargopod frame extending that far anyway, so I only need to install the crossmember between the ends of the two sides of the frame to mount the bike end of the hitch on.

I'm probably going to put the whole joint system on the bike, as a simple ball joint. If I use a plate with a hole in it large enough for 2/3 of a ball to stick thru, I can drill a hole in a ball of some hard material (nylon would be nice) and run a bolt thru that with a hole in the trailer end of the bolt. Then I can put that into the end of the crutch, and pin it with a cotterless pin (one of the pair I got off the baby stroller). A bungee cord around the assembly, firmly attached to the trailer and wrapped around the crossmember should make a good security mount so that even if the pin comes loose I won't lose the trailer in traffic or anything, and can just stop and re-pin it (if I carry the other pin with me in the toolkit). I can of course also just use nut-and-bolt to secure them together, but I'd rather have the quickrelease.

Then I can build a wooden frame inside the bottom of the panel, working around whatever tubing/etc is now there, to help strengthen the whole panel as a load-sharing frame.


I will put reflectors bolted to the back end and sides of the panel, and at some point I'll add lights to it. For now I can simply move my regular taillight/signals assembly to it when needed, as I used to do with the other trailer. Since this one is a lot lower to the ground, I expect it will be harder to see, especially when not loaded. When I do grocery runs, I always have a couple of large white styrofoam coolers tied down to the trailer, so on this one they should provide a lot of visibility from rear or sides. I might paint orange stripes on them just to make them a little brighter in the daytime, as I did the white stripes on the back of my recumbent cargo pods.


Since it does come apart so quickly and easily, I'll likely take it apart when I'm stopped at a destination to lock it up with the bike. If I have nothing on it, I'll run the regular bike lock cables thru it's detached wheels and panel (via one of the larger holes already in it for knobs and such, if it fits), plus around the bike. Then it will be as secure as it could be short of taking it in with me (which typically isn't allowed). If it's already loaded up, I'll just take the wheels off and cable lock them to the bike, and cable lock the trailer separately to the bike.

Unfortunately this trailer may look spiffy enough to want to steal, unlike my other one, even though it too is made of junk. Sometimes just thinking about it makes me hate people. :-(

Sunday, May 17, 2009

Alternate Trike ideas

I thought of a possible alternate "trike" that would really be a kind of diamond-quad, with the trailer-wheels/hubs using 26" tires on the rear sides, at the same spacing as the WC ones in the pics I have in prior posts, but with a bike's rear wheel and mountings just behind that, as the actual driving wheel. This way there'd be no worrying about how to hook up the drivetrain to the wheels.

It'd be about 10-15 pounds heavier, though, and it would be even harder to turn, because I'd have three rear wheels all in a single plane that would need to stay in ground contact. A side-wheel suspension that allowed pivoting around the center wheel during turns could cause problems with rocking back and forth on anything but perfectly smooth pavement (which is fairly rare). If I coupled that with the steering so that it only tilted during turns it'd be ok, but it would be more complex and even heavier.


Now that I have those shocks and springs, I am tempted to build the tadpole trike instead. A potentially serious disadvantage of the tadpole is that the cargo section in the rear would have to be split into two halves, on either side of the wheel, just as it is now on every other bike I have.

This can be solved by either putting the cargo section up front between the front wheels, which limits how high I can stack things due to visibility concerns, or to add length to the bike and place the cargo section between the back of the seat and the rear wheel.

It should be possible to do the latter in such a way as to be extensible, so I can keep the bike short when not carrying cargo, but that would require a foldable or soft-sided cargo "pod". I'd rather have one that is fully lockable so it's more secure, so I think that I am much more likely to leave it as a solid longer trike (longer by as much as 2 feet!), with even more cargo space on either side of the rear wheel as the pannier-style cargo boxes I have now.


If I do the tadpole, several problems go away:
--trailer rear wheels can be used for the front wheels as they are, no changes
--WC wheels can be used on the trailer, no changes
--drivetrain connection to rear wheel can be exactly like a normal bike, or like my current CrazyBike2
--Front of bike gets shorter, so easier for me to see around corners before I get there

The front end and steering gets more complicated, but it would also have better suspension *and* I think I could even do the rear suspension I want (still using springs rather than airshocks, though).

I can use either regular steering or under-seat steering (USS). The regular steering is as pictured above, and the USS is below:


The versions sketched above (in MSPaint) are 20" front and 24" rear (might not really be proportional). Medium blue is the main frame, black is the front double-A-frame suspension and the tires. Dark blue is steering. Gold is the rear triangle. Orange with black lines is the shocks. Teal is the seat. Light green is the motor. Red is chainlines, purple the pedals/cranks, pink the bottom brackets and chainrings. Gray is the cargo box (only pictured on the long version, but present on the other as well as just the split-box rear half, without the full-width box just behind the seat).

I'll need to work up a parts list for the tadpole version and see how many things I don't have already, to see if I can build it faster than I could come up with solutions to the other problems involved in the wheelchair trike's drivetrain and such.

Wheel Dilemma

I had thought at first glance that I would be able to simply swap the wheels/hubs from the wheelchair to the ex-stroller/nee'-trailer's axles, but unfortunately they're different diameters.

The wheelchair axles are around 7/8", and the trailer axles are around 3/8". The sealed bearing units on both of them are, however, the same apparent outside diameter, for one side of the wheelchair hubs, but not the other. The inside of each WC hub appears to be the same as the trailer hub, but the outside of each WC hub is just very slightly smaller (perhaps 1/2mm), so the bearings from the one won't fit the other.

I haven't gotten the bearings out of either WC hub's inside holder yet to verify, but since the outside ones are recessed for the quick-release button, it's easy to see the problem, by attempting to insert one of the trailer bearings, which *almost* fits into the hole, but not quite. It's possible the inside ones are the same, but eyeballing them with the trailer bearing sitting on top of the WC bearing, it *looks* the same.

This means that unless I use the whole WC bearing, axle, hub, wheel, and mounting plate at the same time on one vehicle or the other, I can't use them at all. That's disappointing, as it makes my decisions harder.

If I use the WC wheels/hubs on the trailer, it'd make for a no-flats-causing-me-to-unload-cargo-to-fix-flat advantage. Also, would keep the trailer from bouncing around as much from the pneumatic tires, as the WC tires are solid, but can only be used on the WC rims. I don't think that in the condition they're in, that I could get them off the rims in one piece, or if I did that I could get them back on in one piece. I'd have to take them off to easily unlace the rim to use it with the trailer hubs, or to change the rims on the WC hubs for bike-sized ones, with pneumatic tires (probably would use 26" ones, right now).

The trailer rims/hubs are steel, 20". I can easily change the rims on these, so that makes them a better candidate for the trike wheels. I can also weld to the hubs if needed, and more safely machine them in various ways, so that I can attach whatever drive mechanism is necessary to them.

However, the trailer hubs are narrower, by nearly an inch, meaning they'll have less side-strength against stresses during turns, etc. This won't be a fast trike, but it will be heavy. Even with three wheels supporting it instead of two, it'll still need as much wheel strength as I can give it, while still keeping them light.

I think the trailer axles are smaller diameter mostly because they are solid, where the WC axles are hollow so the QR pin can move inside them. The QR pin itself could be bearing some of the load, too, I guess, but I doubt that it is--if it was, it'd be very difficult if not impossible to move it in and out with a load in the chair, and it's just as easy to move the pin in and out with the chair full (of big lead-acid batteries and other heavy objects) as empty.

I suspect they can both handle the same weight of load, but I *know* the WC was designed for at least 250lbs on it, while the former stroller only 75. Primarily I think that low limit is because of the plastic frame interconnects, and that the axles could bear much more, but I have no evidence for it, and can't find out without testing, which may destroy the axles if I exceed their limits.

Got more thinking to do.

Saturday, May 16, 2009

New Trailer From Baby Stroller

Today I was out with a friend and we stopped at a few thrift stores. I don't like buying things new for this project, but buying used stuff isn't really out of the scope of the recycling ethic as applied to the project, even though I'd rather be using stuff other people would just throw away (or already *have* thrown away).

I found two very useful items, the first of which is this baby stroller for joggers, with an additional plus that it was made in the U.S.A. :-) It's from Racing Strollers Inc. (The Baby Jogger Company), and a quick search finds it was rated for about 75 pounds (more than I expected, less than I need).

It's fairly sturdy, and very light, as it's mostly aluminum, with basically bike wheels and a bike style caliper brake. They still sell a similar model: http://www.babyjogger.com/perf_jogger_lp.aspx which is capable of holding up to 100 pounds, and is designed to fold up, with quick-release wheels.

The cloth baby seat snaps on and off.

The top frame/handle has two pins that undo it from the forward bottom end joint with the bottom frame, and those pins are similar to the wheelchair quick-release wheel pins--they have little inset spring-loaded ball bearings on the sides that keep them from just sliding out, but can be pulled out with a good tug.

This lets the frame swing up at the front end, though there is no similar quick release on the back, nor does it seem to pivot far enough in any direction to make it easy to fold up. Makes me wonder why they used the QR pins at all.

However, since they came with it, I now have them, and I'm probably going to use them as locking pins to hold the trailer hitch to the bike/trike/etc. I haven't designed the hitch yet, but there are number of simple ones on the web, including several at Instructables.com that look sturdy.

Another nice feature on this stroller is it's bike-style brake handle, with an extra feature.

It is also a parking brake, which can lock to the engaged position by pressing the little button you see above at the juncture of the silver handle and the black mounting frame, while squeezing the handle to the engaged position.

Simply squeezing the lever a bit after it's engaged will disengage the locking button with a click.

I'm going to remove it from the stroller and put it on the bike as my main brake lever, so I can have the parking brake on the bike's front wheel. (currently the only parking brake lever I have is on the old scooter's brake/turn-signal control, which is on the left handlebar and thus for the rear wheel, which on my current bike doesn't have a use as there are no rear brakes yet).

So it will go on the right side, for my front wheel. I'll need to make a different kind of brake light switch, as it doesn't have one built in to this lever, but it should be easily possible with a reed switch and a magnet, which would even be weatherproof, unlike the existing switch.

The button would also then be on the "bottom" of the lever, which puts it on the inside of the downturned bars I'm using now, which would be helpful as I can then push it with my thumb while squeezing, only needing one hand to engage the parking brake.


The bike-style caliper brakes are identical to many I have from old scrapped bikes.

Mounted underneath the stroller, bolted thru two aluminum plates riveted to the frame, they seem sturdy enough. I will probably run the activating cable from them up thru the trailer lighting wiring harness to the bike's brakes. I'll need to fit a splitter or other mechanism that will let me pull this cable at the same time as my bike's brakes.

The wheel mounting is similar to the wheelchair's:

It's not a quick-release, but it does use sealed bearings (which feel very nice, not worn, and still very low-friction).

The design of the hub mounting point is not as strong as it could be, but that can be fixed.

Most likley, since I will be taking the top frame off entirely, and moving the back wheel brace (the sideways L-bracket in the picture, which actually goes all the way across as a short-riser U) to the rear of the aluminum ex-computer-faceplate that will be the trailer bed, I will have the wheel bracket flipped across the bed so that the bed's facing will be opposing the axle-bolt's pull tension while the then-upside-down bracket opposes it's push tension. Then it should be stronger against a load pushing down between the wheels. It still won't hold as much as a wheelchair-based trailer might, but I think I need those parts more on the trike.



The second good find was this pair of shocks and springs:

They're not the best shape, but they still work, and even if I don't use the worn shocks, the springs themselves are separate, and may work for part of the suspension on a heavy power-assisted bike or trike.

I am still looking for info about them, but the only part markings are "MX 14C1 Made in Mexico". I expect the MX is for MotorCross, but I haven't found the company that made them yet.

I'd guesstimate each weighs about 5 or 6 pounds, but I haven't actually used a scale on them yet.
All I have is ideas at the moment, and no absolute plan for them, but the price I couldn't pass up (nothing on the parts car is small and relatively light like these, and I haven't found springs and shocks just laying around, or on Freecycle, etc.).

Tuesday, May 12, 2009

Wheelchair Trike Layout Ideas

After some pondering, I've decided to build the front frame "from scratch", rather than modifying a complete existing bike frame. It should be lighter this way, and being purpose-built it will allow me to have all the bits I need in all the right places without quite so much hacking and working around, and probably end up being completed faster as a result.

It should also be shorter by perhaps a foot, since not so much clearance between the front wheel and the pedals is actually needed as is provided by the frame itself, with the pedalling position I am in recumbent-style.

One possible layout for it, with the cargopod atop the frame:

It's all just laid out on the ground flat, on it's side, to get an idea of possible placements of things.

The frame would not be quite as simple as it looks here, as it would have a triangular cross-section from the front as well. The downtube will split Y-style somewhere around the front edge of the seat, so that it can be then bent upwards and be bolted inside of the chair frame tubing where it used to go to footrests. It won't actually go as far back as it looks in the pic.

The top tube will end just at the center of the crosstube at the top of the chair frame, and be bolted/clamped to it.

Those three connections should be enough to make a stable and sturdy frame out of the chair and tubing. If they're not, well, I'll find out soon enough after it's built. :-)

The actual length is not yet determined, I have to measure out where the seat will go, then the angles I'll be sitting at, how long a leg-extension I'll need and thus where the pedals will go, and how much clearance I need from pedals to tire and thus where the headtube will go. I'm guessing it'll be just under 8 feet long, about like the CrazyBike2. Even though it's front tire is smaller and the cranks are closer to the tire, all that extra space goes behind the seat so I can have more cargo in *front* of the wheel's axle than I can with the CrazyBike2. I might make it 9 feet long *just* to get more cargo space in front of the wheels but still behind the seat.

There are several elements not shown at all, such as the steering tie-rod from the front steering tube back to the handlebars, which are also not shown. I will probably use underseat steering, because I should have clearance to make it work on this bike, as long as I give up the idea of under-seat cargo pods.


Currently it has no planned suspension, but it is possible I will at least make a front suspension fork similar to a number of freakbikes I've seen on the FBN forum, etc. I have a number of partial concepts for a rear suspension, but all of them require modifying the chair frame in significant ways, so until I can implement one without doing that, they'll just remain concepts.

The "simplest" would be to bolt the wheel support plate not to the frame itself, but to a pair of tubes clamped around the vertical frame risers. Springs would be installed just above those tubes, mounted around the frame risers, so the tubes holding the wheel would press against the springs and give some amount of vertical damping (probably not much, with such short springs, given that I have no springs I could use there yet, either).

The next best would be to bolt the wheel support plate to a car leaf spring's center, then support that leaf spring at both ends from the frame. This would not be very easy to do and still provide enough space for cargo *in front* of the wheel, as it would move the wheel forward by almost half the frame's length.

Another variation on that would put the wheel support plate at the far end of the leaf spring, with the spring's middle and center mounted to the frame itself. This would actually allow the wheel to be extended out behind the frame, assuming the leaf spring can bear the bike's weight (estimated at 100-120 pounds due to the 3rd battery and complex drivetrain) when suspended at it's end that way, which is an unknown to me.

Any shock system I come up with for the rear needs to be independent for each wheel, partly so that I can turn easier using the tilt of the trike to turn with, just as with a bike.

I have another idea for a linkage from the steering that during a turn would proportionately force the appropriate rear wheel upward along it's suspension, tilting the trike toward that direction (the direction being steered toward). It may be completely impractical, and might not even be possible for me to make using parts I have available, but I may try it.



The major thing I needed to establish was how the drivetrain might work. Assuming I will also motor-assist this one, then the drivetrain will be a little complex, marginally more so than it currently is on the Crazybike2.

The primary reason for that is there is no provision on the wheelchair hub for power input from anything other than hands on the tires or handrail. The probable complexity of a rugged solution including a freewheel and cassette necessitates a separate stage in the drivetrain for that part, and simply bolting a chainring (the orange one) to the hub, or more likely to the spokes themselves, the way it was done on Eric Peltzer's ebike.

Working forward from the wheel, the chain will go to a hub (represented here by the small white chainring/hub part just forward of the wheel, below the chair frame). That hub will be mounted on a dropout (including derailer hanger) that will hang from the chair frame, bolted/clamped into place. The left side of that hub will have a chainring the same size as the one on the wheel itself bolted to it. The right side of the hub will remain a normal freewheel/cassette unit, which will have it's own chain going up to the next part of the drivetrain. There will be a derailer on this cassette to shift gears just as if it were still on a rear wheel in a normal bike drivetrain.

This next part will be the most complex piece, as it will take the power from both pedals and motor and transfer it to the wheel's drivetrain. On it's right side, as normal, will be a triple chainring with a derailer, so that it can be shifted as normal. On the left side will be a device I've wished for previously, but will redescribe in detail below. Basically all it does is allow two independent freewheels to be installed, without fear of them unthreading themselves from the crankshaft, on the left side of a square-taper shaft.

It's a 3" long cylinder the same diameter as the threaded portion of a rear hub, and threaded like that, so freewheels can be screwed onto it. There is a thick lip left unmachined on the outer left edge, so that anything threaded onto it (from the right side, before attaching it to the crankshaft) can't unscrew itself.

To set it up, just thread the freewheels on, put the device on the crankshaft, line up your chains by threading the freewheels back and forth until at the desired point, then measure their positions, take it and them off, and insert spacer rings of the desired thickness of whatever material is available (large steel washers would be good, but old rear cassette spacer rings should work, too), then thread the freewheels back on, and reinstall the device on the crankshaft.

Hook the chains back up, and voila! there is now a freewheel for keeping both of the power sources isolated from each other, and still able to input power into the rear drivetrain (which now does not need it's own freewheel, either).

Now we come to the power input to the drivetrain. On one of the freewheels of the device above, a chain goes up to the pedals, which have their chainring on the left side rather than the right, for this reason. The pedals would only need one chainring, but if great hill-climbing ability turns out to be needed, two or more could be used, with the derailer as they would normally have, or stopping and manually moving the chain from one to another can be done. A chain tensioner plus throttle will be installed on the top of the chainline as described a couple of posts back.


The other freewheel of the above device would have it's chain running back to the motor's output chainring. The motor is represented by the rusty-chrome sprocket within the chair frame, and is currently planned to be the treadmill motor, which has not yet been used on a bike yet (discounting stand-tests on the previous one). It's chainring would be small on the motor, and as large as possible on the freewheel. With the 36V setup this bike will have, most likely a double-reduction would be required, and so the device I made for that using the belt from the motor with 3:1 reduction, then small chain with about 10:1 reduction would take care of that.

The motor chain tension will be adjusted by sliding the motor (and it's intermediate reducer, if needed) back and forth in slots within it's mounting plate, which will be bolted to the chair frame.

The motor will be just in front of the batteries, which will be in a box directly between the wheels along the line of their axles. The box should be able to contain the controller, charger, and batteries. Since it will be aluminum, I can simply bolt the controller and charger to it's walls, and the box itself should help dissipate the heat. The controller's MOSFETs will if possible be mounted so they are on the inside of the box at a point where a heatsink can be bolted to the outside surface exposed to maximum airflow, but at as little risk of impact or tampering as possible.

The battery box won't be accessible without unbolting the cargo pod from the chair frame, partly for security of the contents.

In the space leftover around the motor and it's drivetrain components, I'll have a toolbox bolted in place with whatever I might need to fix the bike on the road at least enough to get me home. Again, it wont' be accessible without removing the cargo pod, but hopefully I won't need it very often. Because of this, the tools to unbolt the pod itself need to be something small I can carry with me, so I'll probably use allen-head bolts.


The front fork I've decided on for now is the one that came with the Magna frame that's now on the CrazyBike2. I chose it primarily because it's lighter than any of the others I have in that size, yet still strong and apparently well-made. It also has brake-pivot studs already on it, so I can use that type of brake instead of caliper brakes in front if I choose (and if I have enough working parts to do it with).

The seat is just the plywood version of my other seat, and will not be used on this bike--I'll probably build a v2.0 of the tubing/mesh seat I now use on the CrazyBike2, just for this bike.

That's the basic overview of how this one is currently being planned.


This is just a pic without the cargo pod, but otherwise the same.

Flat Tire Again, 26" Tube In 24" Wheel = Bad Idea

Remember my bright idea of using the 26" tube in the 24" tire? Not such a bright idea after all. Only a week and a couple of days in, and it began to fail, though as it turns out not from the same thing causing the other tubes to fail (still undetermined).

When it started to fail, I was on my way home from work, and could feel the tire slowly getting softer. I stopped and reinflated it, hoping the Slime would fix it, and it did help, but not enough--a half-mile later, it was soft enough to feel problematic again. Having little choice at 10pm on a Friday night, I continued to reinflate it two more times before reaching home. I estimate it would have lasted a mile and a half before being totally flat each time, but I didn't want to damage anything with the massive weight of the bike on a flat tire if I could avoid it. I expect the rim could actually cut thru the tire itself, and certainly the tube, with me plus the bike pressing on it for any length of time.


The failure this time was because of the folds themselves, and the thick rubber. As it was stressed in a direction not designed for, it actual *tore* holes at the corner of the fold on one side.
This hole is long, but is pinched together by the rubber around it for the most part, and the chunky tubeless-style Slime is able to seal it, amazingly enough, when I tested it in a 26" tire.

This hole, however, looks rather like a hole chewed in something by a mouse.

It's not actually just a tear, there's material missing; presumably rubbed away by friction with the inside of the tire sidewall. It's not fully sealable by the Slime, and is probably where the majority of the air leak is from.

I taped it with packing tape to hold a temporary patch of rubber just placed over it, then placed it inside a 26" tire (not on a rim) and inflated it, and the first hole could be sealed by Slime, and so could this with the patch over it, but not without it. I dont' think the Slime could actually *stay* in the hole as the tire rotates against the ground, though, so it probably wouldn't stay inflated long during a ride without a good permanent patch on it.

The other is intact, though you can see creasing on it.

It does not bulge out or appear misshapen when inflated, so it is likely the only real damage on this side is the scuffing from rubbing against the sidewall, which is only visible on this side.


This is the problematic wheel, which I have again closely examined and found nothing that could cause a problem in the rim, spoke head/tensioners, rim weld, tire bead/sidewall/tread, Slime protector strips, or spoke cover strip.

The red ellipses are the approximate locations of the two folds in the 26" tube to fit it into this 24" wheel.

I'll be taping up my patched-up 24" tubes and trying them again, since I can't afford new ones at this time. I have to tape the patches because the holes are not where the patches will press evenly against the inside of the tire, and thus aren't held down against the tube. As my patch kit consists simply of rubber cement and patch pieces and/or old innertube slices, it doesn't seem to hold as permanently to the tube as a really good patch kit would, unless it's helped by the pressure of something external holding it on, even after several days of drying/hardening. Wrapping tape around the tube after slightly inflating it seems to help tremendously for holes in places where it does not press against the tire.

I'm tempted to try out the method at this Instructable:
http://www.instructables.com/id/Bicycle-dual-inner-tube-system/
which essentially consists of two innertubes placed in one wheel, with the stem of the second exiting the rim via a second hole drilled partway around the rim from the first.

It would at the least let me quickly get back on the road after a flat bad enough for Slime to be unable to fix, but before I do it I will need good tubes to start from that don't have problems on their inner circumferences.

If I have good tubes, I suspect I won't have flats in the first place, as I have not had any in well over a year on the upright bike using Slimed tubes plus Slime protector strips, as I now have in the rear tire of the recumbent.


The problem might be rendered partially moot if I can quickly design and build the wheelchair-based trike, especially if I use the "solid" tires that came with it until I can afford better ones, at which point I'd be replacing the wheelchair rims and tires with bike rims/tires/tubes.

That will require unlacing the wheelchair rims from the hubs and relacing the bike rims to them, and a decision of whether to use 26" or 24" rims. I think 26" rims might be better mostly because it seems easier to find road-tires with slicker treads in 26" size than 24", especially with recycled stuff.

Saturday, May 9, 2009

Tension-based Throttle Improvements

I've decided to change the tensioner itself to an idea borrowed from AussieJester's work on this thread on the FreakBike Nation forum:
http://www.forum.freakbikenation.com/phpBB3/viewtopic.php?f=17&t=1252&start=15&st=0&sk=t&sd=a
specifically this post:
http://www.forum.freakbikenation.com/phpBB3/viewtopic.php?p=10488#p10488
where he's used a brake lever (wheel end) from a bike, along with it's spring and pivot point, with a derailer tensioner wheel bolted in place of the cable retainer, to create a chain tensioner.

(pic used with permission)

A quick GIMP mockup pic:

This would greatly simplify my tensioner's construction and reduce it's size and weight, compared to the one I originally planned in this post:
http://electricle.blogspot.com/2009/02/minor-improvements-still-in-progress_16.html

So now the tension-based throttle is simply this:

Mounted on the lever is a magnet, which passes by a frame-mounted hall-effect sensor wired to the throttle input on the controller.

The harder one has to pedal, the more pull/tension there is on the top of the chain, and thus the higher the lever is pushed, closer to the sensor, causing more motor power output, which helps ease the tension on the chain by reducing the power needed from the pedals.

Since the amount of tension desired might be different for different people or bikes, there can be an adjustment of where the magnet is along the lever (and thus how much power the motor outputs and when).

The magnet itself was originally intended to be moveable on the steel lever simply by sliding it around, to calibrate the throttle. Using an aluminum brake lever necessitates some changes to that, but can still work the same way. Probably the magnet will be mounted to the brakepad-adjust slot, if I do have to move it.

Alternately, this adjustment can be done via a scaling control in the controller electronics, but that would require extra changes outside the scope of most people's capability, for those who are using this blog to make their own DIY e-bike.


For my original design, that's what I would have been doing, though, by using an op-amp just after the Hall sensor output, with an adjustment knob either on the throttle mounting box, or on the handlebars.

Since my first design was expected to have (and has had!) breakdowns, I also wanted a manual throttle override so I could use the motor itself to get home if I found myself unable to pedal due to a system failure (which has happened, unfortunately before I had the motor on there!). This would be done by using the cable-control throttle off an old Honda Spree scooter, and the cable would pull the tensioner up as if the chain were pushing it up, causing the motor to receive power as a constant based on throttle position, rather than the feedback-based control provided when the chain tension moves the lever.

Monday, May 4, 2009

Wheelchair-Based Trike

A donation intended to provide essential parts for my flatbed trailer inspired a trike almost as soon as I started working with it. It's aluminum, so I can't weld it, but I can bolt things to it, and it's capable of supporting at least 250 pounds on it's two tiny front wheels plus it's rear spoked wheels.

Since I'll be removing the tiny front wheels regardless of what I do with it, and almost certainly re-lacing the hubs onto bike wheel rims, most likely using this whole cage as the rear of a trike, it might be able to support even more weight (but I won't push it unless I have to).

I don't have any really lightweight small frames to use for the front end except for the Magna that's already part of Crazybike2, but I do have a nice-looking copper-colored Mongoose frame that will probably fit the project well enough for at least the first try.

A quick first idea, just setting the parts together in a first-guess position:

The bike frame has a 20" front wheel, with the 24" tire-size wheelchair wheels on the back. Unfortunately they're not the same as 24" bike wheels, as the tires are much smaller thickness, and so the rims are larger diameter than 24" bike wheels. Currently they have the urethane foam flat-proof "tubes" in them, and very very badly worn tires (past the tread, into the fabric).

A few problems arise, of course. First, since there's no place to put a normal bike chainring set on the wheels, I have to come up with an alternate way to get the power to the wheels. Front wheel drive is out, as I don't want to deal with those kinds of complications. I could move the handrails to the inside of the wheels, and either cut gear notches in them or replace them with smaller bike rims (or wheelchair rims if I can find any) and use those as pulleys.

In the pics, I've already moved the left handrail to the inside of the wheel just to see if there was clearance, which there is plenty of. I need to take at least one of the tires off the rim anyway, as there is one broken spoke tensioner head (the tensioner is still on the bent-up spoke, but is missing it's head; I assume that's still in the wheel). Once I do that, I have the option of unlacing the wheelchair rim and lacing on a bike rim instead. Then I need to decide if I want 24" or 26" tires, since I have more 26" tires with road-style vs MTB-style tread, and I'm more likely to run across 26" slick tires than I am 24", anyway.

If rim-as-pulley is used to replace the handrails, a belt would then drive the wheels, and be attached to a device at the other end that takes the bike chain drive from the front pedals/etc and converts it to the pulley drive, along with a freewheel and shiftable chainrings (pretty much like the device I made for the treadmill motor setup, but with the pulley on the left side and the chainrings on the freewheel as normal).

Alternately, I can clamp a chainring to the spokes on the wheel(s), driving it from a chain fed from that device, in place of the belt/pulley. (this would be better for traction and torque, and is like the way Eric Peltzer did his bike at http://peltzer.net/ebike ).


Since the wheels are on separate axles already, they'll freely turn separately when I steer, unless I use a drive system that drives *both* rear wheels (which would be better to do, but more complicated).

I'll probably need to alter the small bike frame so that it's rear triangle is replaced with a triangle of small tubing spread far apart at the rear, so it can be inserted into the former foot-rest tubing post holes (where the black posts go in from the bottom front of the chair right now), and secured in place there with the bolt pins. Probably it will also need an upper set of tubing to connect from the seattube area to the top of the wheelchair frame, to stiffen it against folding up with weight in the middle (me).

The second set of pics below show the bike frame upside down, with the pedals up higher. I'm not sure it'd work as well as the other way, but I'd have more ground clearance, and a teensy bit more fork rake.

The seat would be similar to the one on my current bike, and placed in front of where the wheelchair frame's seat would have been, with the base of the seat about the same height the pedals are at in the pic above, or a tad lower.

I should be able to do the same remote-steering and forward-located pedals I have now, although since I have no other square-taper threaded-in bottom brackets I'll be using the old one-piece cranks that came on the frame I'm considering using. Remote steering because I don't like tiller-style steering, and I'd rather have the bars right back where I'm actually sitting. More convienent for mounting things on with switches and stuff, too.

The three U1 batteries will fit very well along with their charger and the motor controller in back of the chair, between the axle points, on the double-barred part of the frame (in a box made for the purpose, probably from aluminum).



The motor(s) would be wherever they need to go to fit into the drivetrain before the shifting part of it, just as now on the current bike. I might be able to use the treadmill motor easily on this one (with a double-reduction setup), as I could enclose it in a fan-cooled/filtered-air box to keep road dirt out of it, as there's plenty of space inside the wheelchair frame. If you can think of a
better motor to use, I'll try anything out.

I *will* need to get or make a longer cable for my bike lock for it, since the wheels can be taken off so easily--I can imagine someone stealing *those* off the bike (or trailer) just for wheelchair use. :-(


As I don't have a front shock fork for it, I will probably make one like those seen on several of the freakbikes, where it has a lower fork that supports the wheels for turning, and an upper fork that
contains the springs or whatever, which should allow me much more freedom in the steering design. (I can't believe I never imagined any kind of shock fork in front other than the kind that comes on MTBs and common road bikes, until I saw the freakbike forum pics--I apparently never noticed or never saw anything like them in my google searches of bike designs previously!).

All in all, this looks like a workable idea, mostly dependent on making a way to easily and reliably transfer power to the rear wheel(s). I've already come up with several possibilities in the few hours I've been thinking about it, so it should be possible to do this with stuff I already have laying around.

With the big heavy U1 batteries, the trike would be MUCH better as a ride than the bike is. Then I can maybe find some small batteries to assist the Crazybike2 for short jaunts, and use the trike for longer trips, since I can easily put the third U1 on there for either spare power or for more overall power the entire trip (a more likely-to-be-used solution).

Phirst Phriday Phoenix Phreaks Ride

After fixing the motor, I had one other problem to fix--the rear tire was sunrotted (remember, this is all salvaged junk!) and the cracks bulged enough under the load of the bike, batteries, and me to make me worry I'd get a sidewall blowout of the tire itself, which would then let the tube shred itself against the frame, most likely.

I *had* pics of all this stuff, but something happened to the camera card on insertion into the computer to read it, and I ended up having to format it to use it at all. :-( So I took some pics of the things I could, to at least give an idea of what I'm talking about.

The tire itself was easy to replace, as I have a number of 24" tires. None are road tires like I'd prefer, but some are in better shape than the ones that came off the Roadmaster bike, and one is less knobby, and not quite as wide, with a completely solid center 1/8" or so for a bit less rolling resistance.

These are a couple of the ones I have, all are fairly similar, except others have even less continuity between center treads. One looks more like a waterfall scoop than a wheel. :-(

The one on the left is the actual rear tire I was using.

This is the tread on the one I now have installed.

As you can see, it's more continuous in the center, which hopefully reduces rolling resistance a bit. I couldn't really tell a difference, at the same 50PSI I keep them at, but the bike is really heavy with the batteries and motor, so it's possible there is a difference I could feel if I had just the bike and no motor/batteries/pods on there.

Changing the tire should have been quick and easy, but as things tend to be on this project, it turned into a sub-project all it's own. After successfully putting the new tire on, keeping the same perfectly-good innertube, and adding the Slime protective strip between the tube and the new tire's tread, I installed it on the bike and left it on the test stand.

The test stand is just a pair of Mac Plus cases (without any computer/etc in them) I put under the cargo pods to hold the bike completely off the ground, kickstand and all, to make it easier to work on the back end and to run motor and drivetrain tests, which I can't readily do with the kickstand down (it interferes with the chainline when down, but not when stowed). So there was no pressure on the tire other than the 50PSI I keep them aired up at.

Not long after I put it up there, I heard a hiss, then a quick rush of air. Upon taking the wheel and tire back off, I found a very large split (about 3/16" long) in the *inner* side of the tube, nowhere near a seam or anything else. I VERY carefully inspected the whole tire, rim, etc., just to be sure I had not missed a sharp edge or thorn on my first pre-install inspection. Nothing there, visible or tactile. The 3/4" wide spoke-cover strip was installed correctly, covering the spoke tensioners, no issues I could find. I even rechecked my tire-installing tools, which are heavy-duty plastic with very smooth ends, and found no damage or anything else to explain it.

Must be rotting rubber (remember, the tires were sunrotted), I guess. Digging around I found two other 24" tubes I could use, a couple of 20" that wouldn't by any stretch of themselves or the imagination fit on the 24" rim, and a small pile of assorted 26" tubes, including a new-in-box Slime Thorn Resistant tube I used to carry on the upright bike as a spare, until I'd gone a whole *year* without a flat.


These are the three 24" tubes I was working with.

The yellow patches were added by me, as were a few black ones you can't readily see (some made from other unrepairable totally-blown-out tubes, such as the batch of them I had more than a year ago that had valve stem failures). One patch on the right outer part of the rightmost tube, with a reddish outline, was already on that tube when I got it.

What was most frustrating about this is that the original tube had not had any problems before this, and that since the hole was on the *inner* edge, the slime could not seal the hole! I patched it but reinflating it found yet another hole, and holding a finger on that, another, and another, most of these on the sidewall or inner edge, really making me question if the tire tool was doing this (since the new tire seems to have a much tighter bead than the previous one, and is much harder to get on and off the rim).

However, the same was true of the other tubes, which I tested before even attempting install; I inflated the first and heard it hissing right away, without even 30PSI in there. The leak was on the inner edge. I patched it and tried again, only to find another and so on and so on. No inner-edge leak on the last one, so I tried it out, and only had to patch one other place on it besides the existing patch. Now I theoretically had a tube to try out.

Installed, tested, ok. Put on the bike, ok. Bike off-stand, resting on tire, ok. Rode around the block a few times, ok. I had the batteries and motor off the bike for the motor repair at the time, so it was lighter than usual by a lot. At this point, I was getting tired and frustrated, and hot at maybe 2pm or so, so I went to the store for some fake Dr. Pepper; in this case Safeway's brand of Dr. Skipper, as it was closer than any other store and they had a sale I could almost afford a 2-liter of, for 89 cents I think.

Almost halfway there, about 1/4 mile, and PSSSSSSHHHHhhhhh the tube pops. Ugh. I decided to walk it home and fix the tube first, so I wouldn't trash the tire I just put on there.

Another hole, about halfway around the outside part between the old patch and the new one. I gave up on it and started thinking about the 26" tubes.

Finally I decided that I could fold up the extra bit of tube in two places at opposite sides of the wheel, 90° either side of the stem and reflector.

That's not the tube itself, as its' in the tire and I don't want to unravel the other NIB one I used for the pic above, either. (they never go back in the box, do they?)

But basically like that, stuffed into the 24" tire, with the Slime protector strip around it. I inflated it a little bit, then bounced it till it stopped creaking and shifting inside, then fully inflated it, installed the wheel, and test rode around the block a while.

When I felt safe enough, I went after that soda with no problems.

Now, one thing I do figure is that the slime inside the tube wouldn't be able to circulate freely, due to the folds, so I used an extra dose of slime in it before folding it:

This is the chunky slime, meant for tubeless tires, and it will seal a much larger hole than the regular stuff, in theory. Works so far in my other tires on other bikes, and for a couple hundred miles on this one (until this tube problem).

I just had to squeeze it around in the tube till about half of it felt like it was in each half of the tube, fold and install.


I did a few other things before heading out to the Phirst Phriday ride downtown (about 10 miles away from my house), which included changing the chainring on the motor from a 24-tooth to a 21-tooth, while leaving the crankshaft end of that chain at 24-tooth.

This means that now the 120RPM of the motor at full speed, which currently turns the crankshaft at 120RPM as well, and thus is about 1/3 higher speed than what I would pedal at normally (90RPM), would be roughly 105RPM now. Pedalling along with the motor's full speed would now be much easier, without feet being knocked off of the pedals by any sudden burst of speed (caused by a slight downhill grade, for instance, like going out of a driveway, etc.). Also gives more torque to the motor for startup from a stop.

This chainring is of course smaller diameter, which means it will not be able to bolt directly to the face of the hub on the motor, or else the chain won't engage the teeth. I used a stack of washers on each of 4 bolts to keep it just far enough away from the hub to let the chain fully ride on the ring and not touch the hub. The hub itself had to be flipped over and then spaced a little away from the motor, so it would still line up with the chainring on the drivetrain crankshaft, and not rub on the next chainring there.

None of that would be much of an issue, except that since it *is* a smaller diameter, it also can't use bolts the size of the ones holding the other chainring on, and actually must be pretty small due to the tiny 3/16" overlap of the holes in the chainring vs the holes in the hub. I knew that this arrangement couldn't sustain resistance against the torque needed to drive the bike from the motor, even with my pedalling, but I didnt' really have any idea how long it would last.

I set it up, tightened the chain tension, and tested it just on the stand. No problems free-running. I setup a friction grip (leather welding glove) for the rear wheel, and put as much strain in jerky and smooth ways as I could on it, and still it held fine. Rode it around the block a few times, and still fine. Guess it was better than I thought, as I didnt' see any wear on the bolts yet.

Well, it did last a while--almost 5 miles. Then as I was riding south in the bike lane on 15th Avenue, I heard what sounded like a bucket of coins on the road, and suddenly had no motor assist, so I stopped. The sound was the washers spilling off the bolts, which had finally been cut thru by the hub vs chainring torque. Since I was expecting failure, I'd brought the tools and old chainring and bolts with me, and changed them out, setting off again after about 20 minutes or so.

My next failure on the way there was completely unexpected. The bottom wheel on the rear derailer just fell off.

The roller pin it's bolted in with had somehow unscrewed itself (I still can't figure out how), and the roller wheel was just yanked right out with the pull of the chain vs the springs in the derailer tensioner. I was still pedalling at that point, trying to figure out what the noise had been, but couldn't stop due to traffic, as I was coming up on a green traffic light and the bike lane had ended for the right-turners to get to the edge of the road, so I kept going thru the intersection and then found I couldn't shift up or down.

I braked to a stop and got off, looking at things, thinking a cable might've broken (but it didn't feel like that, it felt jammed), then I saw the derailer's empty bottom spot, and my jaw dropped.

I left the bike on the sidewalk with it's lights on, key pulled from the motor switch, on it's kickstand (no one is running off with *this* bike that fast without motor assist!), bike lock around the wheels and frame, and walked back along the path to see if I could find the roller. I did, but not the center-spacer or one of the two side washers.

Great. Halfway from nowhere to home, and I've got a two-speed bike. :(

But wait--I have a SPARE DERAILER!

Well, sort off. It's a Suntour, not a Shimano, so of course the parts don't match well enough to just put the unused bottom roller wheel's washer and center spacer on the Shimano, but the whole wheel, washers, and spacer did fit, being a tad larger in center-hole size than the Shimano.

Wow. I'm really glad I didn't actually take the derailer off of that chainline now, and instead just left it as the tensioner, even though it isn't the best way to do that. :-)

It'll hopefully never happen again, so now I should be safe to replace the derailer-tensioner with a purpose-built one with my throttle in it. :)

After seeing a few of the bikes on the FreakBike Nation forums, I think I might scrap my idea of a box with all the throttle stuff in it, and instead just use a brake-arm with the roller mounted on the end where the bolt would have been to hold the cable tight, and the pivot point one of the brake-studs welded onto the frame along the chainline somewhere. It's simpler, and the parts are already pre-made for torque in those directions.


The Phirst Phriday ride itself was interesting, though I found the bar-hopping part of it at the bars themselves not fun, as I'm not really a bar-type person. I did like the gallery we stopped at, and I really liked the riding around in a group, and talking with the few members I managed to for any length of time. Wished I'd been able to talk with more people, but I'm still new to them, and they already have friends and people they know in the group, so it'll take a while before I can really become part of their conversations. I'm also not a very good in-person person; I do much better via text formats, like blogs, forums, and emails. Gives me more time to think about what is being said, and what I'm going to say, so I don't sound quite so much like an idiot. :-)

The Crazybike2 was a pretty big hit, apparently--it's not quite like any of the other freakbikes there. Lots of people interested in it, at least in passing. The lighting was an oft-touched on topic, as was the reuse and repurposing of so many odd components.

A pic by Stephon from the forum, from outside the first bar:



I liked quite a few of the other bikes I saw there, and some of them gave me specific ideas I will have to use myself eventually. One in particular that I liked was the black delta trike by "lostideas" with the airtank seat and airshocks on the front fork, which can even be bounced up in the air. Even has a rack for a big jambox on the back. It's too small a trike for my own purposes (hauling cargo, etc.), but it's definitely interesting, and I will likely research more on it's design when making my own.

The chopper-style ones had interesting points, but I really wouldn't want to ride one. :) Looked cool, but not my style.

One thing that I have seen on several bikes there and on the forums is a front shock arrangement that does not use the "shock absorber tube" style of fork that is common to all the MTB designs I've seen, and the roadbike I have now (and the fork I am using on the Crazybike). Instead they use a double fork--the bottom fork appears to be the load-bearing/steering fork, while the upper fork ties a pivot point on the dropouts to a shock-mounting plate near the headstock, with various kinds of shocks between that plate and the headstock swivel hardware. I never even imagined such an arrangement before, even though the idea itself appears quite simple, and I can't see why I didn't think of that. It's like not being able to figure out how to open a pop-top soda can, or something! :-(

Now I know how I will revise the look of my curvy version of the Crazybike2 frame, the one a couple dozen posts or so back, with the leaf-spring rear shock. Maybe I'll do a leaf spring shock front *and* rear. :-)

The ride home around midnight was nice, and very uneventful. The batteries lasted the whole way home, though they were pretty well dead by the time I got to my house. Since I hadn't charged them at all at or during the event, after having ridden the 10 miles or so to get there, plus at least a couple of miles or so during the event, I'd say that's a pretty good range, of about 22 miles. I'm sure if they were new batteries, and I had the bugs worked out of the drivetrain and throttle control, making it more efficient, I'd get a lot more than that.

Now I've got about a hundred miles on the motor/battery setup since first install, well over two hundred miles on the bike as a whole. Still working very well considering it's literally a ridable pile of junk. :-)

I rode up Central Ave, right thru downtown Phoenix, with barely a motor vehicle in sight, and more than a few other cyclists, in groups and alone, usually on sidewalks rather than the road, mostly without any helmets, lights, or other gear, dinging their bells when they saw me riding by. I didn't have a bell to ding back, just my car horn, which didnt' seem appropriate to use in response.

I think I will have to finish that teletype-bell idea I was working on a couple of months ago but gave up on after realizing no car would pay attention to it, if they could even hear it. Since I ride only on the road, and bells are more commonly used to warn pedestrians or other cyclists of a cyclists approach, I didnt' see the need for one until now. It would be useful for responding to other bells in a friendly fashion, though.

If there are future gatherings, I'll go if I am not working, and they're close enough to bike to.

Motor Repairs

With a second motor (from the other side of a chair, so the gearbox is mirrored) I was able to fix the motor brushplate problem.

This is the damaged motor taken off the bike; the wiring harness I left attached to the bike and is just unscrewed from the terminals on the motor.

There's just two philips screws holding that black plastic endcap on (right side of pic), so it's easy to remove.

The nuts holding the endcap/brushplate holder/bearing assembly on are a little more difficult. They're slotted, so a flatblade would work, except that they're threaded down far enough on the bolt that the protruding bolt end prevents access to the slot.

Dremel tool and an old wide flatblade screwdriver to the rescue. Comparison with similar screwdriver unmodified (bottom tool in pic).

The damaged motor's endcap interior, showing one side of the black disc of soft plastic used as a brushplate. The screw terminal for the brush's wire is visible, with the brush itself just below that.

The square metal bit on the end of the shaft is part of the electromechanical brake I'd removed; it would engage a steel plate that rides on springs whenever no power was applied to the brake assembly, preventing the motor from turning. With power applied, the brake plate would be pulled away from the shaft area, allowing it to spin. A type of cotter pin is used to hold the square bit on the shaft, and required removal with a thin metal tool and gentle tapping from a brass hammer before I could take the endcap unit off.

Once that is removed, the brushplate itself is held on only with a pair of small nuts and screws (one each to the semi-circular slots you see in the black ring), used to adjust it's angle around the commutator segments.

It's difficult to see the problem with the plastic in this pic, but you *can* see the irregular wear on the brush facing you, where it has been worn at an angle on the bottom (forward) end, due to the brush not contacting the commutator straight-on anymore.

In this pic, you can just see the slight melting of the plastic standoff area just between the black ring and the metal brush holder.

You'll have to click on the image to see it's full-size version to really see the problem.

In this shot it's a bit more readily visible in the small pic, but again clicking on the pic for the full size version makes it much easier to see.


This pic makes it obvious how the whole brush holder is at an angle due to the melting plastic.

Since the brush holder can't stay straight, the brush no longer fully contacts the commutator segments, and arcs, not only causing extra heat but wasting power and damaging the brushes *and* the commutator.

Below is the endcap assembly (including the coils from the electromechanical brake) off the undamaged motor.

Note that the ring is a different color. This is because it is made of a different plastic, one much more like I was describing before that it needed to be made from. It is much harder, and does not appear to be affected by heat (to test it, I poked a non-critical area with a hot soldering iron, which is MUCH hotter than the motor will get, and it did not damage the plastic).

It also has thicker standoffs, although the brush holders are identical.

Instead of using soft brass rivets, it appears to be using some other alloy for the fasteners (which appear to be a different type of pin, rather than a rivet). They may be cast into the plate, or they may be threaded in. I can't tell without taking it apart, and I don't want to risk damaging the only working brushplate I have. :-)

A comparison pic.

The leftmost unit is the better one, with the brownish plastic that is much more like thick circuitboard material, while the rightmost unit is the damaged one with the black soft plastic that melts easily. I have no way to know (since the s/n label on the second motor was unreadable) whether the company that made them found out the black soft plastic was a problem and *improved* their process, or if they created the problem by replacing a working (but more expensive to make) solution with the cheaper soft plastic. I'm really curious as to which it is, but I doubt I will ever know.

This is the core of the undamaged motor, with nice shiny commutator segments. it also has other better-designed features than the one with the soft plastic plate, such as tied-down armature wiring, preventing it from coming unglued and flying out into the magnets or whatever, which would destroy the motor.

The damaged motor does not have any tie-down of the wiring at all. It also has a cheaper way of connecting the wires to the commutator segments, by simply pinching them into contact.

A comarison of the two commutator ends, with the better design on the left, and the damaged motor on the right.

Both do have an anti-cogging feature of spiraled core segments, and both have the same shaft and bearings.

One thing not visible without taking the shaft out of the gearbox end (which requires removing all the grease to see what holds it in place, as the four screws holding that end's plate on are not all that secures it) is the gear teeth on the shaft itself.

This is the open gearbox on the damaged motor, which shows the gear that makes contact with the shaft's teeth on the right center end of the gearbox.

In this motor, there is grease left in the teeth as it turns. In the other motor, the fit is much more precise between teh two, and only shiny metal is visible after one pass, even if I first force grease into all the teeth as they go down into the gearbox--it's all forced out of the teeth by the very good fit between gears, meaning less slop and less wasted power, and also less wear on the teeth when I pedal harder than the motor is outputting power, which forces the output shaft to turn in a way that essentially acts to the teeth meshing as if the motor had reversed directions. Any slop is then turned into wear as the teeth bash against each other at that point, even if lightly.


This is the bolt whose head broke off when I was tightening the motor down, tensioning it's chain.

Fortunately it was easy to remove, by gripping with pliers and turning it, and it had not damaged the threads of the hole.


Because I could not change the entire motor between gearboxes (without further digging into the gearbox to find and remove whatever shaft retaining rings/etc there are), I simply ended up moving the brush rings between the two motors, so the good one is now on the motor I was already using. Since the motor itself has damaged commutator segments (scratched up and burnt on one end of all of them), I kept the original brushes in there to match them.

I did polish the commutators and brushes a bit with my Dremel's jewelry buffing wheels and rouge, and it made some difference, but still needs much better polishing than what I could do with that and my limited skills and shaky hands.

It now works much better than it did before, though, so it was worth the work, and now should not fail again the same way.