Search all of my sites with Google

Tuesday, April 28, 2009

New Bars, "Power" Steering, Chainline Fix

The chainline I fixed on Saturday, after being incredibly frustrated with putting the chain back on the pedals several times a ride, whenever I didnt' throttle down fast enough when I had to slow down, or I habitually paused pedalling as I checked the road situation or something. I'd rather have a freewheel or a throttle controlled by pedal-chain tension, but for now this will have to do.

For now all I did was remove a few links so it's at the minimum necessary to go around the sprockets. Because of the spacing between the sprockets, and no adjustment possible for them, there's almost exactly 1 link (1/2") extra, which results in a lot of droop, and the chain bouncing off if there's no tensioner, whenever I hit even a little bitty bump.

So the above is how I rearranged the derailer so it is simply taking up that little slack--it's VERY VERY tight, and I expect it will wear out the derailer wheel more quickly than I would like, but it should survive until I build the real tensioner that will include the throttle control.

The whole chainline as it now exists:

Looks a lot more normal, and it has ZERO problems going in reverse (such as when backing up the bike) or being driven forward by the motor, neither of which was true before. It's worked even over some potholes that would have previously derailed the chain.


So now the bike looks like this:

including the other work I did today, such as changing out the handlebars for some old Schwinn "urban comfort" bike bars from the pile. I'd never thought I'd use them, because their grip angle doesn't suit riding upright, but they seem to fit a 'bent position pretty well.

I like their grips a lot better than the round grips made today for most bikes. They're sort of an ergonomic cross-section, and have between-finger-ridges that seem to fit my hands ok. Won't know exactly how good they are until I get a long ride on them, which should happen Friday, though I may test it Thursday if I have time, by riding around the neighborhood a lot instead of a long-distance destination trip.

The rider's view isn't much different, though positions of things changed, and I'll have to get used to them again. A few things to fix still, like the throttle control is only ziptied on, until I find it's best position, after which I'll make a screw-mounting point for the bracket.

While trying to fix the slack in the motor chain, caused by vibration causing the motor to creep backwards on the mounting plate over time (because I didn't file down the edges of the boltholes in the plate well enough, I think), I managed to break off the only really long bolt I had holding it on:

The frontmost hole on the side towards you used to have that bolt. Now it's still got half of the shaft in the hole, blocking all the threads, so I can't put another one in there unless I can take it out of the motor's gearbox mounting plate. That plate is removable, so I should be able to do that, but it might be all the way down into the gearbox housing rather than the plate itself, and if it doesn't stick up at all I may not be able to remove it with the tools I have available. I *might* be able to drill a hole in the remains of the bolt, and tap threads into that, then screw something else into it with locktite on it's threads so that I can then remove the bolt remains, but I'm not confident that will work.

I also don't have any more bolts that are long enough with the right threads, so I'm going to have to dig into more of my salvaged items to see if there are things with 2"+ long bolts with the same threading and size as the brake-stud bolts on bikes (also the larger size accessory mount bolts are the same, but neither is long enough to fully thread in on the side row, only in the middle row).


This is a rear-looking oblique shot of the midships chassis area, showing the motor controller ziptied for now to the leftside seatstay of the Magna frame, with it's MOSFETs/heatsink ziptied to the downtube of the Schwinn frame.

The motor's hub and chainring are just visible at the bottom center of the pic, with the temporary lighting battery ziptied down (with one broken tie) at the left, just behind the Magna's seattube. Looks like a lot of stuff in there, but it's not. Just junky. :)


A full shot of the new handlebars, including the new position of the mirror. As before, I don't have the mirror above the bars because if I have to turn the bike over for any reason (like wheel maintenance) I don't want to have to remove or loosen the mirror.

Of course, the shifters are still in the way of flipping the bike over on it's bars, but one slight turn of the allen bolt on each one's mount and they can be rotated out of the way easily, and put right back when done. I'd rather move them totally out of the way, but there's not really a place to do that yet.

I'm considering a small clamped-on bar down from the inside top part of the bars that would let me parallel the shifters *inside* the grips, toward the center of the bike, with levers still sticking out. Like an L-bar, top of an L at each side of the bars before they join to the center mount, and bottom end of an L near the top of each grip, to hold the shifters. Not sure, so will work on the idea in sketches or something, then maybe a mockup, before altering the bars at all.

The new bars in side-view, from the left.


The whole bike from the front:

It's a pretty dark pic because the flash keeps getting tricked by the front reflector's brightness, no matter how I set the camera. :( There will be better pics in daylight in a future post.

While I was doing the other stuff, I also finally got around to both flipping the fork around and changing out the steering head to make steering more stable, by forcing me to use more handlebar input angle to get less actual steering angle (it's almost 1.7:1 now, I'd guess, where it used to be about 1:1).

The turn signals had to be flipped too, since they'd point backwards otherwise.

As I had welded their brackets to the fork to make things easier and quicker at the time, I got the grinder out to cut the very small welds I'd used, but the cheap HarborFreight grinder chose that moment to strip all the teeth of it's internal right-angle gearing, so now the motor works great, but spins the disc not at all. :( I don't think I can fix it.

Since the way the brackets and signals are made won't allow reverse mounting of the signals on the brackets, I just bent them (they're flimsy) down and twisted them 90° to make flat bottom-supports for the signals, and put a self-tapping screw thru the former mounting hole, up into the plastic bottom of the signal casings. Works fine, and now they're closer in by a little bit, too, making it maybe a bit easier to get around a doorframe as I take it in and out of the house and such.

The main reason to flip the fork around was to fix braking, which it did marvelously. Now I can actually lockup the wheel, which while undesirable in actual practice is required by Arizona law (brakes must be good enough to cause the bike to skid on a dry surface). Now I just have to get them adjusted for no squealing and the best grip without rubbing during a ride. (the wheel needs a bit of truing, too).


The change in steering heads lifted it up a bit higher in front, as the mounting point is above the head instead of below it, so I also flipped the rear mounting point:

Additionally, I had to use the Dremel to very slowly remove the built-up welded steering stops I'd put on this area of the frame, so that the front wheel can still be turned at least most of the way it used to be.

The side view of the new head:

I didn't have to do anything to this MTB steering head (handlebar clamp) besides leave off the top of the clamp, so I'd be able to thread in the bolt I already had in the steering tie rod from the old head (it's the same size and threading, apparently, despite different companies and perhaps 3 decades between them).

From the top it's more apparent what I did:

The four clamp bolt holes are spaced differently than the single clamp hole was on the old steering head, so that one set is closer to the pivot point in the steering tube, and one set is farther away. I'm using the farther set, which gives me a different ratio than the ~1:1 I had before, making it easier to steer, but less twitchy, since every little motion I make with the bars is not directly translated into the same motion of the front wheel, but reduced instead to not quite half. Much more stable now, and easier to steer.

One unfortunate side-effect is that due to the design of my rear steering rod pivot tab, since it's right and left limits are defined by the frame around it, I can't get the same amount of steering as I had before, meaning I will have to make wider turns than before, by a little bit.

You can see the full range above.

I can improve it by making the steering tab itself narrower, but it will also reduce it's strength unless I can make it thicker at the same time. For now I'll just wait and see how it rides over time, to see if it's even necessary to change it.

That's all for now, until the fixes for the fixes start. :)

Saturday, April 25, 2009

Motor Design Flaw

There is a serious design flaw in this powerchair motor, in the form of a plastic disc used to hold the brushes, insulating them from the motor casing and each other. The "clever" designers used a soft plastic to make that disc, and it's not heat resistant. Not a problem in a small light-duty motor, but this one is a heavy-duty motor intended to be run hard and get warm to hot--remember, in a powerchair like this came from, it (along with it's mirror twin on the other wheel) would have directly driven the small diameter wheel at around 4MPH with up to (I think) about 250 pounds of person *plus* the chair/batteries/etc.


It seemed like it might be working a little less well than it ought to even when I started with it, but Friday on a several-mile ride, whenever I was in an area safe to do so in, I tested out how well it would work on motor-only, so I can find out how dependable it is on that if I need it in the future. That of course heated up the motor even more than previously, plus it was a pretty warm day (90-something °F), so the already-warped soft plastic got even more pliable, and the brushes were pulled farther away from the motor until it stopped working at all.

As I have not run the motor much yet, I can tell what wear on the brushes is from prior usage, and what is from me. The brushes were already at an angle away from the motor due to their spring-loading pushing the whole assembly away on that plastic disc, and that left an angled wear on them, curving away from the motor as it slowly pushed farther and farther away, pressing more and more on the forward edge and less and less on the rear edge. When it got even hotter during my tests today, it moved a little further and hit the stop of the front wall of the motor, and stayed there, wearing a little angled notch on the front edge of the brushes.

The plastic disc is simply far too soft for this. It actually melted a little bit right at the point it contacts the brush-mounting plates! They could have used any number of plastics or non-conducting materials that are MUCH harder than this, including the old bakelite-in-fiberglass that most motors I've taken apart have in them. But this Chinese-made motor was done the cheap and dirty way, so now I will have to make my own replacement for the disc, and disassemble the motor to put it in.

Then I have to polish the armature and the brushes so they match again, and properly conduct power to the motor so it runs right and doesn't get damaged by arcing. :(

For the moment, all I could do was replace the thin brass rivets that were no longer holding the plate to the plastic (because of the melting) with tiny screws threaded into the brush holders where the rivets used to be, but those screws will melt into the plastic, too, if I run the motor hard at all--difficult not to do because of that throttle peculiarity I wrote of previously, where it suddenly jumps from around 2/3 power to full power, and usually I need something just above 2/3, so I am stuck with full or nothing most of the time.

This is going to take some serious time to really fix, and I haven't yet had the time to do it, and probably won't until this coming Thursday at the earliest. The risk is that I won't be *able* to fix it, and might leave the motor in an unusable state (right now, even a good bump will pull the brushes away from the motor far enough to keep it from working correctly, and sometimes at all--tapping lightly on the end-casing (the plastic part that formerly held the electromechanical brake) will usually get it started again, but sometimes I have to open that casing and reposition the brush holders carefully.

I am probably not going to fix it this week because there is a group I just discovered called PhoenixPhreaks, a part of FreakBike Nation where others that also build "freaky" bikes gather. PhoenixPhreaks will be having a ride/get together on Friday night, which I'm going to take the bike to even if the motor is not working (I'll leave the batteries and motor home, if it's not, because that's a lot of mass to move by pedal power alone!), as long as I don't get my work schedule changed between now and then.

Friday, April 24, 2009

Cargo Pods Remounted, Rack Reinstalled

V1.1 of the seat, tightened up a lot, is working great so far. Only gotcha is that I need to bend the front tube end on each side down, so I don't keep hitting it with my legs when I am getting on and off, or putting a foot down to balance at a stop. :)

It's really nice that it lets air flow thru a bit, so my back doesn't feel like it's roasting anymore. Also, it's stronger and stiffer, so it doesn't flex around while pedalling, making it a bit easier to steer the bike, and much more efficient at getting my legs' energy into the pedals rather than into twisting the wooden seat around.

It doesn't make as good a shock absorber, but I can live with that.


I also need to angle the handlebars back a bit, which I should be able to do by replacing them with some I found in my junkpile I'd forgotten about, that are the old-style raked-back type Schwinn used to use on some 10-speeds and such.

Since the seat was working ok, I decided to put the rear rack and the cargo pods on:

The rack is at an angle mostly because there isn't yet a convenient place to bolt it to in front. I was not intending to put it back on, but changed my mind when I realized that I should have somehting there to keep anything (like backpacks) slung over the back of the seat from pushing on the fender. If I redo something later to give me a front anchor point, I'll probably put it flat.

I'm still working with the cargo pods, only held on with three bolts right now (will be six once I am sure exactly what position they'll be in). I already raised them up a couple of inches to give better clearance during hard turns, as sometimes they'd touch ground. However, seeing them up higher made me think of raising them a bit more, so they'd clear the top of the rear tire/fender entirely, and then I could put a small rack bolted between their top edges, to let me stack things across their flat tops without worrying about support in the middle, etc.

I'd also like to move them forward some, to have more of the cargo's mass in front of the rear axle, but until I rebuild the lid so it hinges from the front edge, I'm stuck with having it farther back so it will clear the seat's back edge to open.

From the back it's hard to tell it's angled so much, as it just looks like I took the pic from higher up, but the camera is about level with the top of the handlebars. I need to also put a duplicate reflector on the rightside pod, and paint it to match the leftside stripes (and neaten those up a bit, perhaps). I'd actually like to use reflective red/white striped stickon material, but that's not something I've ever run across as a recyclable item, just already on things, and been there long enough that any removal just destroys the reflectivity.

More work tomorrow, though exactly what I end up doing might change as I think of new ideas.

Thursday, April 23, 2009

Tube-and-Mesh Seat V1.0 and 1.1

After the last post, I worked some on the seat, cargo pod supports, etc, but had too many other things to do to really finish them up. Today I didn't have a lot of time either, but had to at least finish the seat a little bit so I could ride it to work.

The frame for the seat isn't anything special.

This was the way I test mounted it at first, but as I found during my ride to work, it didn't work very well as it was too far forward by a couple of inches, preventing full extension of my legs (one of the problems with the plywood seat, too).

Because the tube intersects *under* the top tube above, in order to move it back any, it had to go back at least a couple of inches to clear the top tube. I don't have a pic of it yet (the camera's card was full just as I was trying to finish stuff, and I need to back it up before erasing anything), but it really doesn't look much different from the above.

V1.0 had it laced with separate laces on each side, between the edge of the mesh and the tube.

This didn't work very well, as I simply couldn't get it tight enough to not slip down, and it wouldn't keep me from banging on the tubing below for more than a few seconds.

I also cannot find the brass eyelets I had to make this seat with, so I ended up having to use an old worn-out soldering iron to *melt* holes in it, such that the edges of the holes in the nylon mesh are solid, to keep the cords from ripping thru the mesh.

That part actually worked fairly well, for now. As soon as I find the eyelets I'll put them in, though, because I'm sure this won't last long like this.

V1.1 came pretty quick, where I unfolded one layer of each edge, so it would be wide enough to go beyond the edge of the 19" wide seat tubing, and be able to lace shoe-style around the tubing instead.

This works MUCH better, and aside from not yet having it tight enough, lasted almost the whole ride to work before beginning to sag.

After I got home, I took the seat off and sat there for a while tensioning the cord (parachute cord from a box of assorted stuff I got from Freecycle.org's mailing lists last year). Starting from the front/bottom edge, I pulled each diagonal as if drawing a bowstring, then pulled the next diagonal to take up that slack, and so on all the way to the top end of the seat. I can get a semi-musical tone from the bottom half of the cords, they're so tight now.

Works very well so far. We'll see how the next few days' rides go.

I didn't have time to put the cargo pods back on, so I rode it with a backpack hanging off the top rear tubes of the seat. Since I didn't have the pods, I also couldn't carry the batteries. Without that weight, the bike is still pretty easy to ride, as long as the motor gearbox clutch is disengaged.

I also did end up using that heavier square tubing on the bike--I needed *something* that would make a good seat mountpoint that was parallel, flat-surfaced, etc., and the only good way to do it was with either L-shaped bar, or the square tubing. The latter was lighter by a fair margin, and stronger in torsion, so I just used it as I would have for the bottom ones, on the shorter top bars--that's the rusty bars across the top of the rear wheel.

They stick out in back just about right to use for a trailer hitch mount, too, made from that black fork I have on the upright Columbia bike right now. Might even be able to make something lighter and just as stiff; I'll get to that after I build the flatbed trailer for this bike, which is probably weeks away unless I come up with a sudden real need for it.


More pics in a later post, showing how the seat now looks, as well as the permanent lighting setup--the one above is just zip-tied into place, as I only had a few minutes left before I had to get to work. :-( The way it is now is much more interesting (read: strange) but more functional and more visible. The way it was in the pic above created a problem hanging the backpack on there, because it blocked the taillight and brake light, though not the two scooter turn signals.

Monday, April 20, 2009

New Seat Started, Building CargoPod Supports

Today I was going to just go out for a power-only jaunt to completely drain the batteries the rest of the way, and see how long it took and how fast it might go. I'd've just disconnected the pedal chain, and tucked it up out of the way. Unfortunately I didn't even get out on the street before I could detect a pretty big problem (which was not really unexpected). So my priority on what needs fixing and when got changed around for me.

The weight of the batteries is around 23 pounds each, or 46 pounds total. The way the cargo pods are mounted depends on the load they have on the bike being distributed along these lines:

  • The over-wheel rack's supports, and along the rack itself into the back of the seat's supports (which are plywood).
  • The seat's rear supports, made from an ex-front-fork off a 10-speed, which go to the same rear-dropout hardpoint at the bottom end as the over-wheel rack, and at the top end go to a large bolt thru the seat's supports, a few inches above where the rack attaches.
  • The chainstay across the bottom, which feeds both into the rear axle and the rest of the bike frame.
  • The rear axle itself, in that the side of the pods press against the axle nuts.
As you can see, the weak point is the seat supports, which I could hear groaning occasionally before, especially when I was pedalling hard trying to see how fast it would go on that canal test trip (of which there are two more videos at the end of the post). When I heard quite a creak as I backed the bike out the front door over the doorsill, I stopped and took a good look at it.

None of the photos I could take of it show the problem, as it just looks like woodgrain, but I can see cracks forming in both the supports and the seat itself, primarily around the areas near the mounting points the cargo pods are stressing. I knew I would need to change the mounting method eventually, probably soon, but had hoped it would be safe longer than this.

Since it is not, I decided that while I'm taking the back of the bike apart, I might as well also replace the seat itself with the tubing-and-mesh version I wanted to test out. I can't bend the tubing for the really good version that would be contoured like my seat is, but I can use parts of the seat I had started to build for the original recumbent idea I never finished.

Since that seat was never quite wide enough to keep me from banging shoulders on the frame anyway, I cut both top and bottom ends off, with the curved sections, so I can use them as the crossbar supports for it underneath. As they are not long enough, they will then be cut in half and the white tubes you see with multiple holes in them will be sleeved over them in the middle to extend them.

All this tubing was originally part of one of those invalid's potty chairs, which was rated for up to a 300lb person, so the tubing is pretty strong. The short white tubes were two of the adjustable feet (there should be two more but I can't find them). The silver (aluminum frame you see with them is something I was considering taking the rails from to use as seat mount slides on the bike itself, but later decided to use for cargo pod mounting instead, as you'll see. That frame was originally a medical chart rack, holding two rows of them side by side. I had altered it using the rails on there from a rackmount tower (something from Data General, I believe, found as scrap being tossed out by some repair company years and years ago) to be my music-equipment rack for taking to sci-fi conventions, where I used to play live and computer-rendered music. Two more identical rail sections were in an ex-camera flight case (huge, at least 3 feet long by two wide by a foot and a half deep) aqcuired at an auction again many years back, which I had also altered into a music-equipment rack for conventions, which also held the tower computer I used at that time.

I am still working on the seat parts, ensuring they're all fit properly before welding together, so there aren't any pics of the completed seat yet, as it probably won't be done till tomorrow at the earliest. Then I still have to make the mesh covering, and lace it on.

The new seat should weigh, with it's support brackets, about 6 pounds or so. The old one is at least 15, including it's cushions and covering. So that should help make up for the weight about to be added by the cargo pod supports.

Next up was altering some of the frame parts that needed rewelding anyway, if they were to really support the battery and cargo weight indefinitely. That white ex-fork I had used as a rear-to-front frame bottom support (running from the Schwinn's seat tube down to the dropouts on the back of the Magna frame) had only been lightly welded onto the Schwinn frame, as I had intended at some point to do this modification but had not yet gotten around to it.

This is a before pic from the front angle; you can see a lot of paint rubbed away on top of it by where the seat's supports were. The only weld to it is two small points at the seatstays on the back side, which you can't see at all in this pic.

The idea will be to cut out the entire curved section, and weld just the straight section onto the side of the seatpost itself.

Below I have cut and welded the left side, leaving the right side as-is to keep the bike's shape supported, so I don't have to brace it all to keep it straight. Without that, the only thing holding the two halves of the bike together is the seatpost joint on the Magna frame, which could swivel, if on it's own.

I cut it with the hacksaw, then shaped the end of it with a grinder and then rasps to make as close a fit to the seat tube and seatstay joins as possible, then clamped it down and welded it. It is difficult to tell, but I also moved the support's bolt-on point at the Magna's dropouts from the outside to the inside, to make room for the cargo-pod support frame.

After completing the left side, the right side was done and the barely-welded-on section removed.

It's ugly, but it is now a lot stronger than it was before, and a bit lighter (that curved section was just under a pound).

From the side there's not a lot of difference in shape, except for the part that is no longer there.

All the angles from the side remain the same. Without the seat in place, the horn is a lot easier to see. It's just clamped on the frame with a radiator hose clamp for now, until I decide if I'm keeping it there or not, at which point I will probalby weld it on to keep anyone from being stupid and stealing it.

On to the cargo pod mounts. They will also hold up separate battery boxes at some point, once I have a chance to find something suitable, or make it out of something else. Until then, the cargo pods will carry the batteries as they have so far.

First up on the pile of stuff to test-fit was a piece of bedframe I got on that windy trip to Arrowhead, and found via the Freecycle.org lists.

It was exactly long enough, but it's quite heavy, and it's also an L-shape, making it difficult to bolt the pods to, as they ride too low to seat the bottom edge into, and thus would need their own L-shape lip to mate to this one, and bolt them together. That would make them even heavier.

Next up was some durable thick-walled square tubing picked up with those big posts and lamp pole, also via the Freecycle.org lists.

Alas, it too is heavy, though I can bolt the pods directly to it's vertical outer surface. It is also a bit longer, and thus would provide more support for the part of the cargo pods that stick out beyond the rear axle.

Third and what I finally chose was the rackmount rails shown in the pic of the seat parts. They are not L- or square, but rather a...hmm. Hard-edged S, perhaps. Imagine a lowercase "h", with the leftmost lower vertical part missing, so that it doglegs down from upper left, across, then down to lower right. I should have taken a pic of the cross-section. :-)

Two of them bolted end to end, with five holes of overlap, and three bolts in those holes, is long enough to fully support the entire cargo pod length, and still reach the Magna dropout accessory hardpoint as it's front mounting point. They're a LOT lighter than either of the other options, even though they appear to be stronger, and they are definitely longer. I'm estimating each one weighs about two to three pounds; I'll weigh them before they're finally bolted on.

Because of their shape, one vertical edge (the top inside one) is bolted to the bike frame at the Magna dropout hardpoint at the front and at the Schwinn hardpoint at the rear, and the other vertical edge (the bottom outside one with all the rackmount-spaced holes) is available to bolt anything else I need to it, such as the middle-back side of the cargo pods, the battery boxes, etc.

Once I finish the seat, I will know more exactly how much space and where the space is for a top rail similar to it, if necessary, for mounting the top part of the cargo pods to.


As threatened, a couple more videos of the test ride to/on the canal. The first one is on 29th Avenue heading towards Metrocenter.

The second is on the CaveCreek canal bike path, just north of where Hatcher would be, I think, if it went thru there.


More to come as progress continues.

Saturday, April 18, 2009

Test Ride Video, Some Fixes Completed

Well, I didn't get a whole lot done today. Somehow I bruised my rightside ribs on Thursday, but didn't feel it until Friday when I got up for work. It got worse all day long; a couple hours in I could barely lift a 35 pound bag of dog food, and after that I couldnt' do it at all--even sliding the 20-pounders hurt, and lifting them was difficult. Sleep took a long time to come after I got home, and I was still sore and tired until midafternoon.

I did do some minor work, mostly fixing the rubbing the motor chain had on the plate above it. Part of it was caused by the sprocket bolted to the motor hub not being perfectly centered, and thus not being able to fully clear the plate without extra shims between the motor and plate. The shims are just some old broken hacksaw blade and utility knife blade (good steel, never throw anything away you might still be able to use!).

Another place the motor chain was rubbing was some bolt-heads I had used to secure the largest sprocket back onto the 3-chainring set, back when I had originally gotten it on one of the junk bikes from the "Bicycyle Bonanza" post early in this blog (2007, I think). I filed the bolt heads a little flatter for better clearance, but one or two are still rubbing a bit and need a hair more filing.

Since the derailer being used as a chain tensioner on the pedal chain was a bit too far leftward for proper lineup, but had no further adjustment range to fix that, I moved it temporarily rightward by adding spacer nuts between it and the bedframe corner bracket being used as a hanger for it. Since the bracket has long slots it's screwed thru to the main mounting plate, I can adjust the whole plate farther rightward, but at the moment it's butted up against the motor already, so I need to cut away some of that edge of the bracket so it can move farther rightward. Not much, perhaps 1/4" at most--just enough to let the derailer fully align with both pedal drivetrain chainrings. I also need to use slightly different spacers on the two bolt-points for the derailer, so that it does not sit at an angle, or else slightly increase the hole size for the axle-hole point, which will do the same thing.

The biggest problem with this tensioner is that unless it really is lined up exactly perfect, then the chain tends to catch on the front chainring's tooth tips instead of engaging them, when I'm backing the bike up or when the motor speed exceeds the pedal speed (or I hesitate even for an instant during pedalling without the same hesitation in throttle control, a common thing so far in the 3 days I've been getting used to how it works and handles.

If I make the tension-based motor throttle, it will fix this problem because it will in theory never go faster than the pedals, or run when I'm not pedalling. Any hesitation in pedalling will stop the motor, too. This I'm still working on the design for.

If I add a freewheel for the pedals so they can't be driven by the motor, that will fix it too. This is difficult at best, with the limited parts I have to use for it (and no means to readily make a square-taper freewheel holder that would hold two freewheels plus chainrings on the freewheels), as it means I must make the pedal chain go not directly to the receiver chainring, but rather first to a freewheeled ring on a separate axle, then have another chain from a secondary sprocket on that axle go to the receiver chainring. More weight, more complication, etc.

Ideally I'll have both freewheels and tension-controlled throttle, but they each will take time and resources to make, so I have to do just one at a time. I think the throttle really needs to be first, as it will have more effect than the other on my whole riding style with the motor, including removing the distraction of a hand-operated throttle completely.


After I finished what I could of the work today, I took it out for another test ride, finding that while the chain problem is less than it was before the spacers, it is still a pretty bad problem, requiring perhaps a dozen stops to put the chain back on the pedal chainrings after it was jolted off by bumps or by my hesitations or other pausing without throttling down fast enough at the same time.

A helpful camera-wielding bike rider captured this video as I was testing it out today:

There will be more videos of today's tests as I receive them, on a separate new post.

This one includes part of a 22.3MPH test run on a paved canal path that's long and fairly straight, with no other bike or foot traffic.

I might have acheived faster speeds if I had fully-charged batteries, but I was still doing the rundown test, with around an hour of runtime so far, with about ten miles on them. By the end of this trip, I had another 5 miles or so, and about 30 minutes (lots of stops to fix that chain, unfortunately slowing me down, as it actually took at least 45 minutes to do the test ride, even though only 2/3 of that was spent riding).

If I finish readjusting the front chainring set's derailer (on the wheel drivetrain) for the 2-chainring set that replaced the 3-ring set, I will be able to shift onto the larger ring and the motor could drive the bike faster than the approximately 15MPH that it can currently do by itself with no pedalling, but I'm not sure it's really necessary. I can't use the motor faster than 20MPH on the roads anyway, whether I'm pedalling or not, so trying to go faster with it (to keep up with traffic, for instance) is not a good idea, regardless of how much more that would make motorists happy (if I could keep up with them so they didn't have to go around me).

Going faster has safety issues as well, since it makes it very much harder to stop quickly, since the bike's brake system was originally designed to stop a lightweight 25-30 pound bike and it's rider, not a 120-pound (dry weight with just motor/batteries/bike!) behemoth cargo bike plus 150-pound rider, plus any cargo I might carry (theoretically another 100 pounds max, 20-30 pounds typical). Even more once I redesign the cargo pod mountings so they are very stiffly mounted to the frame, such that the loading is not all on the rear triangle of the bike, and the batteries are in separate boxes forward under the seat. Less if I get some smaller batteries (like 17Ah) to use for short-range stuff, saving the U1s for the long trips I make roughly every two to three weeks nowadays, and for trips I will be hauling a trailer full of heavy stuff.

Even after I install the rear brakes and flip the front fork around to fix the front brake issues, it's still going to be harder to stop than a typical bike, with just these rim brakes. I may put two sets of rear brakes on, one on the seat stays, and one on the chainstays, to give me more frictional area to stop with. If I had any, I might install disc brakes. I've considered trying to design some from recycled materials, as an additional system to the rim brakes, but it's more complicated than it seems at first. :-)

Friday, April 17, 2009

Motor Controller Wierdness, Battery Life, Throttle Issues.

I need to build a new motor controller. The one I repaired and upgraded off that old dead electric scooter has a (new) wierd issue with the throttle input voltage. It outputs more or less linearly to around 60-odd percent of full power, tracking along with the throttle input voltage up to around 3V (can't remember the exact voltage), then at a specific point just beyond that, it suddenly jumps up to full power.

It's not the actual throttle input device itself, as I've used different kinds, mostly pots of various types and a few Hall effect sensors, and any of these setups that outputs up to that point and beyond will give that sudden change. Anything below that and I can't even get full power.

As bad off as the original controller electronics were when I got them (multiple fried components), I'm just going to start from scratch, and wire one up based on one of the motor controller chips designed for this sort of thing, and wire that up to the 100A/100V MOSFETs and drivers I had upgraded the old controller with, which are all on their own heatsink.

Even with that heatsink just sitting there on the plate at the bottom of the Magna frame's rear triangle (the one the motor is bolted to underneath), those MOSFETs aren't getting hot, just fairly warm, even when I used the motor by itself to drive the bike at around 15MPH for about 5 minutes or so. The ambient temperature was around 75°F, so they'll probably get significantly hotter during summer. To prevent that, I'll be putting the whole controller in a box with a good heatsink bolted to the bike (or the outside of a cargo pod, by the rear wheel so the wheel's motion will also help move air past it, which worked pretty well on the upright bike's controller). I may also put a small PC-style thermally-controlled fan on it to ensure there is no overheating, with the sensor mounted on the heatsink near the MOSFETs.


Battery life seems good so far. I've run them about an hour today, in four trips--one to the grocery store on the way to work, then a second to get the rest of the way to work. A third was after work back to the same store for something I forgot, then back home the same way I came before work. Almost 10 miles in all, with no noticeable difference in power provided from start to finish (although I'm sure there is a difference). That's much better than I could get with the old 12v-12Ah batteries and the frictiondrive setup.

Life test will continue tomorrow/etc, until they're run down enough that there's not really useful power from them, so I can see how long they might last on longer trips (before I try taking one with them and being stuck without power).


The throttle knob I have is difficult to use in traffic, partly due to the lack of freewheels on the pedals/motor interface. Having to hold it in place all the time against the spring (even as light a tension as it is) is tiring, and because it's only really reachable by my thumb, just like my rear shifter, I have to let it go to shift. If I'm shifting gears for slower speeds, that's ok, but if I'm trying to shift into faster gears it's difficult to keep pedalling at the faster speed when the motor is now being a drag on the pedals, so it starts to take my attention away from the road and put it on the shifting/throttling which I need to be able to do automatically.

The only good solution I can do anytime soon is to build that tension-controlled throttle. I have some ideas that should be reasonably quick to implement and test, though they're not ideal and won't be a chain-tensioner in themselves, which is what I'd like to use so I can get rid of the derailer being used as a chain-tensioner.

The first idea is simply to use a lightly-sprung pivot arm with a rubber roller against the top of the chain, right in the center of it's top chainline where the difference between tense and not is greatest. That arm would have a magnet on it that sweeps near a hall-effect sensor, wired into the throttle. It's basically like the one I want to make, except it's not mechanically strong enough to actually be used as a tensioner. I have an arm setup out of some old high-end Sony VTR machines I was given as boxes of salvaged parts, but I almost certainly have to change the sensor on it which is some form of CCD optical sensor, as far as I can tell, and I have no idea what it's pinout or ratings/etc are (I didn't see a p/n on it). That arm was probably used in the VTR to do something very similar, to track tape tension and control motor speed, or something similar, to prevent damage to the tape.

Since there is the possibility that something will go wrong with that setup, I'll wire it in using a diode on the input line, so that I can also keep the current throttle, also wired in with a diode.

Eventually I'll get the tensioner/throttle built, and then a cable from it to the cable-style throttle twist-grip control from the Honda Spree will be able to pull the tensioner up against the spring just as if I was pedalling, should I need an override for any reason.

Thursday, April 16, 2009

PowerChair Motor Installed And Working

After the battery fairy dropped off those U1 deep-cycle batteries plus a power-wheelchair motor, I was itching to get them on the bike and try them out.

It took a few days before I really had time to do it, and then it took two days to actually get it done, but now I have a motorized bike capable of about 15MPH under motor power alone, and up to 20MPH with pedalling. It could actually go a lot faster than that, perhaps 30MPH or more, but A) that's illegal here in AZ and B) it's not geared right to get that fast (on a flat road).

This is what it look like as of just before I went to bed last night (well, actually this morning around 6am, but I didn't realize it was that "late" until I got to bed). I forgot to take a side pic of it as it is right this minute, but the only visible differences are the thumb throttle being added and the pedal-chain tensioner (old derailer) being replaced with a non-bent one.

The front pic above is from just before starting this blog entry, although you can't see it very well because I needed to do these pics outside in daylight, several hours ago, but forgot.

Remember that trip I took in the extreme wind about two and a half weeks ago? I went after some Freecycle items, including bedframe parts.

This derailer hanger isn't made from one of those actual parts, but it is from an old bedframe I'd saved the hardware from, when I moved several years back and couldn't take the whole frame with me. I am not even sure what made me try it, but it fit almost perfectly the first time, and it was the first piece of hardware I grabbed out of the pile of brackets to try for this. Originally I was going to just weld a rear dropout from a scrapped bike onto the new bottom mounting plate, but then used the brackets when I ran across them looking for something else that day.

This is the same parts, from an oblique-ish forward view.

A closeup of the drivetrain section as it is now (though the chain shows slack that isn't really there--it was just caught on one of the bent derailer's edges).

On the left you can see part of a crank on the chrome chainring set. It was from a bike with an apparent crash on that side, which had bent the pedal and damaged the very end of the crank. The rest of the crank was intact, and since I might reuse these parts on another bike later on, especially if I get better parts for this one, I opted not to cut the whole crank off the chainrings (it's swaged on, like most cheap bike chainrings/square-taper cranks). However, since it won't clear the bike frame on the top part of the circle, I did cut off the bent/broken part, just at the edge of where the threads would have started. The crank is so long it can still be re-drilled and re-used as a shorter crank someday.

Before I put the controller and stuff on, it looked like this, with just the empty space above the mounting plate I welded onto the bottom of the Magna's rear triangle.

The drivetrain is different now, in that I now have two much larger chainrings on the right, at the front of the rearwheel drivetrain (the one I can shift gears with). The smallest of them is a couple of teeth larger than the biggest of the three chainrings that used to be there, meaning I can now get more speed, at the sacrifice of low-speed torque. With the motor on there, it hopefully won't be that bad, but really I didn't have a choice, as I had no other parts around I could do this all with.

I had to move the 3-ring set to the left side, replacing the two independent ones I had made for that purpose. I didn't take pics (need to) but the way I made them simply wasn't going to hold up to continued usage. I'm surprised they lasted as long as they did, given what they looked like when I took them off. The square-taper hole I had filed in the metal discs used inside the ex-freewheels the chainrings were mounted on had been notched by the torque on them, and would have eventually (at an unknown time and place) broken right thru, making them round and no longer able to drive the bike. The only reason I used them at all was that I needed the freewheels between the motor and the pedals so they wouldn't drive each other, and the freewheels themselves failed (one was already bad to start with and the other failed pretty quickly, forcing me to weld the freewheeling rings to the core just to be able to ride).

So since I expect that between the motor and the pedals to be increasing the torque applied at that point, I'd better use something with more surface area to transfer that torque. All I had without serious time making my own (possibly weeks of spare time filing) was two chainrings sets with square-taper holes (not counting the one on my upright bike, which has to stay functional). One 2-ring unit with the damaged crank above, and one much smaller 3-ring unit I'd already been using on the righthand (rearwheel) drivetrain of this bike--the orange one below.

Since the smallest chainring (granny ring) was the best match for 1:1 operation with the motor and it's magically-fitting chainring, that left me the middle ring and the largest ring for the pedal chain. Both of those being significantly larger than the pedal chainring at that time (the orange small one two pics up on the righthand side of the bike now), they'd leave me with virtually no speed when on pedals, and my feet flying off the pedals when the motor ran. Plenty of torque, though, I guess. ;)

That left me with needing to change the front chainring, on the pedals, to something larger, preferably something 1:1 with the smaller of those chainrings. I didn't have anything the right size for that that could be mounted on there easily. I had hoped I'd find something that I could just bolt onto the small already-welded-on chainring, but nothing had the same mounting holes, nor any place I could drill them without weakening it too far. There were some that were sort of close, but they'd've been loose and not stayed in position properly, and if it didn't cause the chain to come off it'd still wear thru the bolts holding it on, eventually, and cause a worse disaster out on the road someplace when I'd finally forgotten about the problem. :(

So I did the next best thing: I took the only chainring I had off a one-piece-cranks bike that would let me later change the chainring itself if I had to (it's held on with 5 screws), and welded that center spider to the *other* crank (the one originally *intended* to be on the left side of the bike) and flipped the cranks back over. So now I have an unused small chainring on the righthand crank, but it doesn't cause a problem nor add significant weight. It just looks odd.

The chainring just added and the largest of the 3 receivers are about the same number of teeth (I didn't count them yet, as I didn't really have any choices on parts for this section, so it doesn't yet matter). They also line up almost perfectly; well enough to not have the chain come off.

It didn't have a way to tension it, though, as I haven't built the tensioner/throttle for this drivetrain yet (still pondering it), so I put the derailer back on as a tensioner/guide for now, attached to that bedframe piece.

You can also more clearly see the plate welded to the bottom of the Magna's triangle, which has slot-shaped holes in it for the 6 mounting screws in the top of the motor's gearbox, and for the clutch exit/lever mounting point. The clearance for the motor and chainring on the motor's hub was exact, but I miscalculated the chain's thickness, so it rubs lightly on the bottom of the plate. I used shim-type spacers (made of a broken hacksaw blade) between the motor and the plate to try to fix that, but it is not enough, and I'll need to use more shim thickness.

A friend (Mark) was here at the time, and helped me file out some of those motor mounting holes in that plate, while the very stiff cold wind tried to blow us away.


Motor mounting screws I thought would be a problem, but it turns out that modern non-caliper brakes (with studs either side of the wheel for pivots) use screws with the same diameter and thread pitch, and I have a handful of unused ones that quickly filled the bill. They are not long enough for one part of the moutning, so they only grab a few threads there, but the other side is fine. Hopefully I'll run across longer screws in something I'm repurposing, before there are any problems.

This is the clutch for the motor's gearbox. It's that small lever pointed off to the left, in the bottom center of the pic.

If I have to use that for any length of time, without being able to figure out and install a freewheel, I'll be putting a thumb-style shifter lever on my handlebars for it, to pull it open when I need to disengage the motor (as that's only easily done at certain points in the gear revolution cycles inside the gearbox). I'd rather just use a freewheeling setup, but that is going to take time to figure out exactly how to do it with the parts I already have.

I do have some ideas, one of which is essentially the same as the belt-to-chain power transfer device I'd made for the treadmill motor, which included a freewheel on the belt pulley mount. That only gives me a freewheel for the motor, though, and still leaves me with pedals being pushed by the motor.



One more thing I needed to do for the batteries was add the rightside cargo pod, which I've been putting off because I really didn't want to cut a hole in it just to clear the derailer. I thought of this crazy temporary solution and tried it:

It's just a block of 1x1 scrap held to the chainstay by a radiator hose clamp. There are also two screws deeply in it from inside the cargo pod, to secure them together and keep the wood from rolling down under the stay, which might make it catch the rear wheel spokes or chainrings. That's generally very bad when movng, so I try to be sure and pass on any chances of that happening.
Before adding the pod itself, here's the wood.

One other thing it does is adds side clearance for the bike kickstand, the purple foot of which you can see right below the wood.

Attaching it is done the same as the other pod. Three radiator hose clamps around the tubing behind it, which are at different angles and thus prevent the pod from sliding around.

I'd rather have a frame just for holding these on, but that bike is still in planning stages, and may not be even started for months.

Hard to see, but this is the clamps on the bike frame itself.

The side panel of the box, with tape covering the vent holes.

Eventually I'll find a full sheet of aluminum to replace it with, thick enough to help with loads and anti-theft of pod contents, but till then they'll do.

Looking down into the new box, you can see the battery (covered for safety by it's original cardboard cover, so nothing can get in and short the terminals) My toolbag fits in there, too.

Unfortunatley I don't have a lid for it et, so no lock either.

Then the old box, with the other battery in it, and some assorted stuff, and the gray battery charger bolted onto the side.


I don't have another small pot of the right value range to try the same in-the-handlebars trick I did with the upright, to make my own grip-throttle again.

So instead I used a 5K pot of standard panel-mount pot. That fit with very little filing into the seat-post-clamp hole of an L-bracket for a seatpost-mounted-reflector. One of the reflector mounting holes is used to run the rear derailer lever's main bolt thru, which doesn't affect it's operation.

So now I have a good thumb throttle. By the way, this also fixed the problem I had with never getting more than roughtly 2/3 of the battery voltage at the motor (not full duty-cycle PWM).

WIth this,it goes from about 3.09 v (of 10V) at the throttle's output giving roughly 1/2 motor speed, to 3.1V, where it gives full motor speed. There is no fine control for those, and I think it is because this was originally on a hall effect sensor, with a magnet in a twist-grip throttle.

I decided to change out the little scooter horn (middle) for a car horn (from the Ford LTD).

It was quick and easy, though it uses the frame of the bike fro ground, so can' be powered easily from the main battery, and has me using a separate small SLA for the scooter-based lights, horn, etc. The LED lights are all on one of the motor batteries.

Here's the bike computer after a bit of riding around the neighborhood with the motor running, to test things out.


A closer shot of the controller area, from above. The rear bolts holding on the motor are visible at the left edge of the metal plate in the pic. I'm considering bolting the heatsink to the bike tube framing, as it will not fit on the flat areas down there.

More stuff and pics in a future post, after more test riding.