Search all of my sites with Google

Friday, October 31, 2008

Swivel Joint Works; Cable Steering and Drivetrain Thoughts

Well, I still haven't found another of those nylon rings or anything I can make one from, but I did get the bracket-mounting plate attached to the front frame, and test-fitted it together.

Oh, and I painted it all. :-) Only a temporary paintjob, as I already know the crappy RustOleum would come off no matter how I paint it, so I didnt' bother with primer/etc, and just did this to protect the many areas of bare metal on the bike, where I've had to remove large areas of the original paint to weld the parts together. Should at least keep it from rusting up. Be my luck that the only areas the paint won't come off of are the ones already a little rusty. ;-)

I dont' know that this will be the final paint scheme; I think there is too much orange and not enough yellow up front. The red on the rear triangle also looks just like the orange in these images (perhaps a shade darker). Hopefully once I get the Krylon version of these colors and repaint it all after it's a usable bike (or it's MarkII if this one has too many flaws to fix), the Krylon red will be very red compared to the orange. If not, perhaps I'll dust it with pink.

Here's some pics of the joint itself, without the chainrings in the way:

The first pic shows it from slightly below, behind, and to the right. The second is from the bottom. The third is a simple side view from the right. You can clearly see the black nylon ring and how it fits flush to the bottom bracket tube, and the gap the U-bolt has on the left side without one. The only modification I might make to the ring is to notch the inner part where it meets the chainstay, to allow it to fit a little straighter on both sides.

The reason the metal plate is not symmetrical side-to-side is because the chainring side needs more clearance, so the chain itself doesn't catch the corners of the plate. Even as it is, it will need a bit of rounding at the corners themselves to prevent any touching of the inner edge of the chain on the plate when shifted onto the grannyring--it's a VERY close fit.

The plate itself is actually two identical 1/8" plates welded together along the edges. I still need to trim the angled piece off, and weld that edge together, assuming I can't think of anything to bolt to it (left it on just in case of that). The plates came off a washing machine, for the tub and motor supports. They appear to have been designed to take more mass (given a full tub of water's weight) than I will be putting on that pivot. I tested one of them (there were four) by clamping it to the bench vise and trying to bend it, and could not. I even hit it with a largish hammer, and only bent it a little at the edge where the hammer struck. So it should work for at least the testing stages, and if it shows signs of stress I'll find something else to replace it with.

Drilling the holes was pretty tough, as I could not press hard enough with the hand drill while keeping the drill straight to even begin the holes. After an hour of trying that, I was about ready to go find a phaser and shoot the holes in it. Then I realized we don't have those yet. :-)

So the lathe came to the rescue, by placing the drillpress chuck in the lathe's chuck, and clamping the plate to the toolholder, so I could use the lathe's own cranks and adjustments to press the metal against the bit much more precisely and powerfully than I could possibly do by hand. Since I ran the lathe at it's slowest speed so as not to overheat the bit or the plate, it did take a while, perhaps 10 minutes per hole, but it worked perfectly.

Here's the problem this whole joint was designed to solve:

The motor's chainline would change length if I had to have the swivel point anywhere other than the center of the bottom bracket's axle, which is the center of one end of the motor's chainline. The chainline of the rear drivetrain (where all the power from both motor and pedals goes to the rear wheel) also has to have this point as the center of one end.

Of course, I could always run that motor to a point earlier in the front chainline than this, but I really didn't want to, partly because it would add more complexity and partly because it would likely force me to put the motor farther forward--I'd rather have it in the center with me (basically it's directly under the line between my head and the road). I'm trying to distribute the weight as evenly as possible about the center of the bike, and also keep it all as low to the road as I can.

The swivel point I chose is the most elegant solution to the chainline problem and the rear shock problem I could come up with. I've looked at quite a lot of rear-shocked bikes, and none have anything that would allow the chainlines to not have to change length as the shock operates. Most of them actually cause the rear chainline to change length during operation, which seems silly to me, as there is no need for that to happen. But most have the swivel point behind the bottom bracket, instead of in front of it, and none have it *at* the bottom bracket's axis. I guess for most bikes it doesn't really affect the operation significantly, as the rear derailer already has a tensioner that takes up the slack easily enough. But it prevents any bike that doesn't have a tensioner in the rear chainline from having a rear shock that pivots (or at least, it runs the risk of jumping the chain off a sprocket if something were to jolt the chain in the wrong direction during a shock that loosened the chain because of the pivoting).

I wonder if something like this idea is patentable. :-) Probably not. Just in case it really is a new idea, and someone tries to patent it for themselves, I'd like to offer the idea for free use, under the license of share and share alike, where improvements to the idea must be made freely available under the same license. :-)


The bottom bracket's left side is going to be a bit complex to make, as it is the point where both the pedal drivetrain and the motor drivetrain come together to drive the bottom bracket's axle, which then drives the rear chain/drivetrain on the right side. The bottom bracket axle is of the square mounting point type with a threaded hole. The hope is that I can use a long bolt I already have that fits the threads in that hole, secure it to the axle (by welding if need be) so that counter-clockwise rotation isnt' going to unscrew it. Then I will need to mount two freewheels to it, one in line with the motor's sprocket, and one in line with the pedal's second sprocket.

The motor has to be geared down as far as possible so that I can run it at a higher speed and still keep the bottom bracket's rotation at about the same speed that pedalling it will give. I don't know exactly what that speed is yet, as I have not determined the gearing from the front pedal sprocket to the rightside second pedal sprocket, and then from the leftside of that to this bottom bracket assembly. What I will most likely have to do (because of the severe limitations in which parts I have available to fit the motor's shaft that also have a very high ratio between them and their matching larger sprockets for the bottom bracket axle) is to calculate the rate at which the bottom bracket will need to spin if the wheel is to go a particular speed, for the range of speeds the bike is designed to operate at, for each of the combinations of chainrings on that drivetrain.

Then I will have to calculate how fast the motor will need to spin (based on whatever ratio I can get in the motor's chainrings) to get that speed at the bottom bracket, and setup the controller so it delivers that voltage as needed.

I then have to calculate out appropriate ratios for the chainrings in the pedal drivetrain so that I will be pedalling around 80-90RPM or so (my comfortable base cadence) and still get the appropriate speed at the bottom bracket. I'm sure I won't be able to find exact parts to do this in the stuff I have laying around, but I'll get as close as I can. :-)



Now here's where the steering is going to complicate things for me.

As you can see, a simple chain all the way won't work, because the angle of the front fork I had to use is too steep to allow a straight line back to the USS crossbar. That means the chain will not properly engage the sprocket on the steering column of the fork, and will not turn it very far (about 20-something degrees) before the chain begins to just come off of it. I already knew this would be a problem, and had planned on a cable attached to each end of just enough chain to be able to fully rotate the front fork for full steering angle, which is less than 180°.

The cable bits I was going to use were brake or shifter cables, but I found it is very difficult to attach and bind them together properly around the ends of the chain's links, in a way that will not come loose. They don't bend easily enough to simply run a full loop from bar to chainlink and back to the same bar, without crimping and then sometimes breaking some of the individual strands in the cable at the bend point (at the chainlink end).

Then I was taking apart what someone had left behind of a large copier, and found these pulleys and cables in the document scanning section on top:

The pulleys are better than the ones I had originally intended to use, although their "steel" mounting brackets are of so soft a metal I can bend it in one hand. And I'm a wimp, so.... I'll have to move them to a better bracket if I use them. Anyway, the cables are teflon-coated, and a little thinner than brake cable, but quite flexible. I'm sure they're nowhere near as strong as brake/shifter cable, though. But they'll make it easier to play around with various cable steering systems without messing up and wasting a bunch of otherwise-usable bike-type cables. Once I have a working solution, I can then replace them with tougher cables. The pulleys will handle either kind.

They would be mounted on a post that intersects the line the chain would make if it continued rearward and downward at the same angle the sprocket on the fork is at.

Imagine in the pic above that the chain only goes about 1/3 of the way back from the sprocket to the post, with the rest of the distance covered by cable. The cable loops from there around the pulley (one on each side of the post, or stacked in the center on a single pivot axis, whichever is easier and/or works better), and on to the adjusting clamps on the USS crossbar. Thus the flexible cable will make the angle change at the pulley, and should not cause a problem of jumping tracks like the chain would, and the chain is all at the straight line angle of the fork's sprocket.


I'd like to use a simpler steering mechanism, but I haven't got the parts needed to make either a U-jointed shaft-style or a rod-pushed/pulled system, and buying either of those sets of components would be too much of a break in the goal of using only recycled parts on this bike. I still hope to find some recyclable U-bolts I can use for the swivel joint, instead of the cheap purchased ones I have now.


Tomorrow I am going to try to get the steering working, and if there is time, the pedal portion of the drivetrain. Might not be possible on the latter, as I still have to make the axis extension and freewheel system on the left side of the bottom bracket. Maybe I'll just do the steering and the seat's covering (which I just have to cut, sew edges together on, pop-in the lacing rings, and then lace to the seat frame). Then I can at least do downhill test runs of the steering and shock systems.

Oh, I guess I should at least cable up the brakes, too. :-)

Monday, October 20, 2008

Swivel Joint Design

I've been working on a way to get a swivel joint between the front and rear sections of the recumbent project for at least 3 weeks now, and so far all the ones I could come up with either would not fit in the space between the sides of the bottom bracket, or they would require parts I'd have to go out and buy--not allowed since this is a recycled-parts-only project, if at all possible.

Another problem was that all of them except one (which would need me to machine some sleeve parts, probably possible with the lathe, if I figured out how) would put the joint in such a place within the chainline as to cause shortening and lengthening of the chains to that bottom bracket when the shock and joint were in operation (during bumps/dips/etc). That would require more complications to the drivetrain, requiring tensioners (like the one in most common rear derailers) on each chain going to that point from the front half, which would be both the pedal's rear chain and the motor's chain. Don't want to deal with that, so kept thinking, and came up with another idea in a sudden "eureka" moment.

If the swivel point for the frame centers on the shaft of one end of the chains (the bottom bracket shaft), then the chainlines will never change length. So if the swivel hardware actually mounts *around* the bottom bracket, using the bottom bracket tube as the actual swivel joint in back (the "axle"), and the front frame mounts to something that swivels around that tube, as a pair of rings of some type that pinion the seat tube and rear chainstay tubes between them in such a way as to prevent the sideways wiggle of the joint, it will allow the whole bike to swivel vertically as I need it to. It doesn't have to swivel very far, just a few degrees during bumps/dips, so it won't get a lot of rotational friction or wear.

I spent the last week or so looking thru hardware in my boxes and bins for something that would do this without much modification, and found some U-clamps. They were too small for the purpose, but they gave me the idea to go find the right size. Nothing I had was nearly the right size, so I had to start checking thrift stores and other places that might have odd hardware bits to re-use. I finally found a big bag of assorted fasteners, clamps, and other hardware stuff at Deseret Industries for 50 cents that had something similar, but not quite right:

I don't know what it's called, though it appears to be something to use in buildings to hang pipe, conduit, or wires in, and would be screwed to or bolted to wooden beams. The black ring is a nylon insert that happens to be almost exactly the same inside diameter as the outside diameter of the bottom bracket tube-it's about a half millimeter bigger, which is acceptable.

The problem with the galvanized bracket it comes with is that it's too thin to support the entire wieght of the bike around that swivel point--it would just bend totally out of shape in any torsional stresses, *and* it has an open top which means I'd have to weld something across that to prevent the bike from just falling apart at that point.

But the nylong ring is exactly perfect as a friction-preventing ring, because regular U-clamps like bigger versions of the small ones I already had would fit perfectly inside the groove around the ring. I looked and looked and could not find any used anywhere, so I did end up actually buying these new, from ACE Hardware, just so I could get the project moving again:

This is how the bracket without the ring would fit, because these clamps don't seem to come in anything between the 1-3/8" size and the next up 1-3/4" size, at least at any kind of hardware or automotive store in this area. The ones I already had were barely 1" inside diameter, so they definitely wouldn't fit. The bottom bracket is 1-1/2" inches outside diameter, so too big for the smaller and just too small for the big one.

Thus, that nylon ring from the other hardware set above, slipped into the U and then slipped over the bottom bracket tube will fit perfectly, both diameter-wise, and width-wise (fits just within the width of the space between the chainstay tubes/seat tube, and the outer edge of the bottom bracket tube).

You can see the basic idea of it above.

Now I have one more little problem. Since I don't know what that first bracket is actually called, just a vague idea of what it might be intended for, I have yet to locate one like it so I can get another nylon ring like the first. There's no name or anything on either part, just "123458-PS" molded in raised lettering on the black nylon ring. Searches of various kinds turn up nothing related using that or parts of it. I've already looked at a couple of dozen websites tonite, scrolling thru literally thousands of little images of parts they carry, but not found a single thing even resembling it.

If I knew what it was called, I could easily find another one, I'd bet.

So I will probably have to take it to Lowe's or Home Depot or some other building-supply store, and see if they can find me a similar item. Since those are all a few miles and more from me, I've first sent email inquiries to various places to see if they can tell me what it is, and if they can get me one just like it. All I need is the ring, not even the bracket it goes on. If I can't find out that way in the next couple of days, I'll just have to bike to the stores and see.

Tuesday, October 7, 2008

Video Comparison of Lights vs None

I decided I really should have some comparison up here for lights vs no lights vs just the typical bike-helmet light and taillight, for visibility in dark streets, so despite the problems with the camera not having settings to turn off autofocus or autoiris, and no one to pan it to follow me, the following three vids still show pretty well how on a typical poorly-lit city residential street (with one street light at each end) just how important good vehicle lighting is.

Pardon the clicking sound as I ride; the bottom rollerskate wheel got a chunk out of it due to a stick getting jammed in there while riding in some heavy winds, stopping it but not stopping my tire, which rubbed a flat spot on the skate wheel. :-( That's one of the problems with the whole friction-drive concept, as this sort of damage has happened 3 times now, though from different causes each time. Gotta put another one of my spares on there soon--it's still working fine, just noisy and annoying.



This one is with full lighting, and shows turn signals and brake lights as well. I really wish the camera didn't autoiris, even more than I wish it didn't autofocus. On still images I can turn those off, but I cannot do so on video mode, as far as I can tell. The background lights in each video are the porch lights on houses down the road.


Only helmet mounted 6-LED headlight here, with flashing 5-LED helmet-mounted taillight. Definitely harder to see, but much more visible than the next one.



The only way I even see I'm in this one is the little battery-meter light at the headstock, because there are no lights at all here. Just me in white shirt, black pants (my work uniform pants), reflective-striped safety vest, and the DayGlo Avenger-colored bike, on a pretty dark street.

Keep in mind the street is not nearly as dark as it looks in these vids (*everything* is dark, including the lights, probably because it is autoirising down because of the glare from the porchlights in every direction I could point it), but it is not well-lit at all.

So these vids are useful for contrast between the lit, partially lit, and unlit ways to ride the bike, but little else. I still need to get someone else to pan the camera as I ride, and to find a way to keep the autofocus and autoiris from happening on video, as well as get someone with a car with decent modern lighting to drive alongside me as I ride, to provide a lighting comparison for visibility of car vs bike, so all of you can see relatively how bright this lighting is.