Search all of my sites with Google

Sunday, August 31, 2008

Fork In The Design Road, And New Sketch

I was too worn out to remember to post the main sketch of the ideas in the previous post until now, but here it is:

As usual, click the image to bring up a full-size version. You might still not be able to read the notes on it, but they're better read in the previous post, anyway.

Still no pics of the actual parts to be used, as we've had more storms and rain keeping me from doing it when I get home at night (most of the parts I need to take pics of are too big to keep in the house right now), and I forgot to do it yesterday during the day before the storms came in, while I was actually working on them.

I did get to sorting thru what I have, and found I'll probably need to use the shock-fork off the 24" bike instead of the 20" one I'd rather use, because the 20" does not operate--the springs are either broken, stuck, or not mounted correctly inside. Possibly even missing--I can't tell because I can't open them up without first making a really long (12" or so) allen-head wrench to undo the bolt inside. Since I can't tell what size that is for sure (probably 5mm), I may have to make more than one, out of some old metal rods that used to be laptop-display-locking-bars, plus some cut-off allen-wrenches, welded together, with some sort of t-handle to turn it by. Easy enough to do, once I have a couple extra hours to dig up all the pieces and make them.

Using the longer fork means either the front end is going to be several inches higher than I wanted it, or else is going to be at a more severe angle than is desirable for correct steering angle/trail. I am not sure, but I don't think I could successfully modify the 24" fork to be shorter and still work right, and even if I do, I'd have to make the same tools to take that one apart as I would the 20", *and* I'd have to weld on new brake mounts a couple of inches lower, in order to use them with the 20" wheel. The 20" also has a bit of offset to it's wheel mounts, where the 24" is perfectly straight, requiring near-vertical installation to get the originally-designed trail from it (which will mean a very tall frontend for the bike, either blocking part of my forward central vision or forcing me to make a higher seat, which I also don't want for mostly COG reasons).

Yet, the 24" fork is almost exactly 1 pound lighter, at 4 vs 5 (for the 20" fork). Part of it is materials, as I think the steel in the 24" is a lighter but stronger one, and part of it is construction--the 24" is a pair of single shaped-pinched tube for the wheel-mount part, with standard welded-on U-bracket with brake mounts, and the 20" a thicker, heavier material that is basically just tubes of different sizes welded together, with flat endplates inside the bottom largest ones, and plates cut to shape for wheel mounts welded onto the bottom largest ones. If I could just make the 24" shorter, it'd be ideal.

Now, there *is* one other way to fix this, which I thought of while typing this post up: Mount the 24" fork *backwards*, so the brake mounts are on the back, and weld on new mounting points for the wheel, as slotted tabs, on the front edge of the 24" fork, in such a way that they are higher up and far enough away from the tube to both allow the brake mounts to be used on the 20" wheel, and also to put that wheel's contact patch at the correct trail point for the angle of the fork I will have to use to compensate for it's length. That compensation will be less *because* of this modification, but it's still a taller fork above that new wheel-center, and will still need a greater forward angle to keep the front end of the bike as low as I'd like. This is going to take some careful measuring and sketching before I actually try anything, as I will only get one chance to weld it right--screw it up, and the fork will probably be useless.

Oh, and that magnetic shock idea is probably worthless--I realized there will be nothing to limit the rebound speed, so it'll likely be bouncy as all heck. :( It'd probably need a spring or something to limit the rebound, and if I have to use a spring, I might as well just use only a spring, for weight. We'll see what I can find in my scrap piles and do this all as simply as I can. Right now, I think the rear shock is going to end up being an adapted front fork shock, or rather, one side of one, as I have a damaged fork with one trashed bent-up side and one alright one that could be cut apart and used for this.

You know what? I didn't really think it would be all that much fun to try designing a *bike* around the limitations inherent in the parts I have laying around, though it certainly has been fun doing that for the motor on my Columbia. It's more of a challenge than I thought it would be, but having already read about the construction and design of many other homebuilt recumbents (almost none of the ones I've seen made of scrapped bike parts, and none *entirely* from them, though I'm sure someone *has* done that that I haven't yet seen), I find modifications of ideas others have used for their projects popping readily into my head whenever I run into problems with something--sometimes too many possible solutions, and I have to narrow down to just one!

Some ideas I thought were original (like my chain/sprocket/cable steering idea) actually aren't, and I probably thought of them because of having seen something vaguely similar on others' projects--I'll never really know if anything I'm coming up with is actually original thought, after all I've already read about, but at least I know in most cases the ideas are already road-tested and work. :)

I'm hoping to have a testable bike by Saturday. It'll need lots of improvements, I'm sure, but this ought to be interesting!

Friday, August 29, 2008

Recumbent Thoughts

Today I took a trip to pickup some frames and a bag of assorted parts, basically. Since it was so far away, I'd already planned on stopping at some food place for an hour or two to cool off and stuff after I'd picked them up, but I had no idea my day was going to be quite as bad as it was. :(

The good stuff happened while I was at the food stop halfway thru the trip, for about two hours (only finally feeling cooled off over an hour and a half of being there!), I had my sketchpad out and doodled up thoughts about the new bike I am making.

I have had trouble figuring out a simple way to do the steering, because there is not enough room under my legs for a regular handlebar/etc, without hitting my legs as i steer, no matter how I shape the handlebar around the other things it has to clear. I can't put it *over* my legs kinda like a normal bike because then it would be like what is called an "ape-hanger" which for me is really really uncomfortable after a few minutes, and impossible to use on any normal ride.

So instead I thought that if I put a sprocket for a chain welded to the steering tube just above the actual front fork (but below the place the fork is mounted into), and a short length of chain (just enough to allow full movement range of the front wheel steering by pulling that chain to one side or the other), and regular bike cables (like brakes or shifters use) running from each end of that chain thru guides to a handlebar that just sticks out from under the seat on each side, which pivots under the center of the seat, I can get usable steering without interfering with anything else on the bike. It's more complex than I wanted, because it adds a bunch of parts but it is less complex than linkage bars and stuff. I can adjust it's tension and whatnot just by changing the length of the cables to the chain. I just have to figure out where the best points to mount the cable ends are on the handlebar, to give the easiest steering, and pick the right size sprocket to give the right turning sensitivity so I don't have to move the handlebars 1:1 with the wheel like I would if the handlebars were directly connected to it like on a regular bike, but aren't' so sensitive that I can't easily learn to work it safely. That's some thing I wish I could do on my regular bike, but never figured out a way to do before (I could do it now, that I've figured this out, but it's significantly more complex on the upright bike to do, so I won't).

Another thing I want to do is have the pedals (which stick way out front over the wheel) be adjustable in distance from the seat (since I can't easily make the *seat* adjustable), which means basically using two different sizes of frame tube so that the one holding the pedals is small enough to fit inside the one connected to the bike frame itself, and have some sort of clamp (a seat tube clamp, probably) that holds them at the desired inserted length. That way I don't' have to worry about rewelding the thing over and over to get the length right for my legs, because since I've never done one of these before I have very little idea of how exactly I will be pedalling it, and the amount of extension my legs will find ideal while producing the torque I need to drive the bike.

The problem with this adjustable length is that there will then be potential slack in the chain running from that pedal sprocket (#1) to the mid-front sprocket where the front gearshifting will take place (so that I don't have complex delicate stuff hanging off the front end of the bike where it can easily be damaged by even a minor scrape/bump/etc). The normal front shifter will not do anything about slack n a chain, but the *rear* shifter, in the form of a derailer, *does* do this as part of it's normal functions. Since I wanted to use a freewheel on this mid-front sprocket anyway, to keep the pedals from being moved by the motor if the motor is running faster than they are, that ends up being perfect, because I can use a complete rear derailer and cassette setup off a rear wheel, instead of the typical 3-gear setup usually found on the pedal end of the chain on most bikes. Yeah, the sprockets are significantly smaller on that, which presents an issue with gear ratios probably not being enough, but I may be able to partially solve that with one translation sprocket set at the mid-rear sprocket on the motor shaft, which also needs a freewheel (to keep the pedals from turning the motor shaft if they're running faster than the motor). If I use a larger sprocket for (I think) the chain on the rear wheel end, and a smaller sprocket for the chain on the mid-end, it should compensate for the difference in total number of teeth that using the smaller rearwheel cassette in place of the regular front pedal chainrings (sprockets) creates. I'll have to do some drawings and calculations to see if that really works that way, but my melted brain says it ought to be fine. I should trust it, right? :)

Mounting the rear derailer will require some extra welding to put a tab in the right place to bolt it to, but that isn't' much extra work. Getting the spacing right so that it neither hits the ground (or speed bumps) nor interferes with the front wheel during ride or steering might need more figuring out, since by my estimate there's not a lot of space up there for it, and it'll have to stick out more to the front than downward, unlike a normal rearwheel derailer. I have to make sure there's also room for it to extend during shifting, etc, without hitting anything. More work than either of those will be inventing something out of scrap bike parts that will let me mount the cassette/freewheel there in a usable way, since I won't be able to just use the normal pedal-bearing/bottom-bracket combo as I'd planned, which would have been very easy. Most likely, it will be done simply using the entire hub and axle off of a rear wheel, mounted on just the tabs from a rear triangle frame from a bike, welded in the right place on the underframe. Gotta draw that up more to figure out details.

The part that is causing me a bit of trouble is that I need to feed the drive from one chain to the next via the various sprocket transfer points (2 transfer points, mid front and mid rear, from the pedals in the very front to the rear wheel at the very back). I *think* that I will need to run the pedal-to-mid-front chain on the right side, then the mid-front-to-mid-rear on the left side, then the mid-rear-to-rearwheel on the right side again. That's mostly because the only derailers I have (or have ever seen in person) are for the righthand side, and will not work on the left side on the rear wheel, which forces my hand on the other two chains for their order. The motor I need to use is a treadmill motor, which thankfully has a shaft extending out on both ends, which helps tremendously in that I can use that shaft to transfer from one side to the other...except that if I do that, I will *always* be turning the motor when pedalling, which makes me pedal harder when no power is applied--something I need seriously to avoid.

So I have to come up with a way to probably just run all the chains on the same side and mount the single sprocket that I need to not be on the mid-front freewheel in front to the hub directly just behind the freewheel, and something similar on the motor (mid-rear) freewheel. It's the most straightforward approach mechanically, as the other approach I can think of (at least, that's possible for me to build) is to put the motor only *indirectly* into the chain path, by having it's freewheel engage the chain on one side, and not be an end-part of the actual chain loop--instead having the chain loop end at another pedal-bearing bottom bracket with a pair of sprockets (possibly of different sizes for compensation of the previously noted problem of the front chainring). That adds more weight (not a lot, but still), and more complexity, in that the motor needs to then have some sort of way to adjust it's tension against the chain (well, I might have to do that anyway).

There is one more problem...the freewheels are designed to turn in a way that prevents something in back rotating from transferring that rotation to the chain attached to that freewheel. That idea works *perfectly* for the rearwheel and mid-front freewheels, as I only want to transfer power from front to back, never the reverse. But for the motor, I have (I think--I'm having trouble wrapping my brain around it just now) a problem, because if it is rotating forward, it will not engage the chain at all, because the freewheel will do the same thing it would if a rear wheel was turning it--it will just spin and keep the chain from moving. If it were mounted backwards, it *will* work the way I want it to, but I don't have a normal freewheel that can be mounted backwards. I *do* have a couple of those single-gear freewheels with a sprocket that's actually cast into the metal of it, but cannot get them off the hubs they're attached to (I'm not strong enough to turn them in the correct direction with any tools or setups I've been able to make, and do not have the *real* tool to do it right--I don't want to have to buy the tool just to do this one thing, especially as I do not know if all this stuff will even work well enough to use as a bike when I'm done (I have to build and ride it to find that out). But those single-sprocket freewheels are actually threaded evenly all the way thru from front to back, and thus should be able to be mounted backwards on the hub. Then I can just take a hub-section and bolt it to the motor shaft (which I'd already figured out how to do for a previous idea on how to use the treadmill motor for my regular bike, mostly based on Eric Peltzer's belt-drive--but for that idea, the motor would have been already backwards-facing, on the left side, and thus the *normal* freewheel would be backwards and rotate the way I need it to.

I may be overthinking stuff again, especially as much as my head hurts right now. :( Hopefully that's all it is, and it's really not as ahrd as it seems to ensure chains only go the way they need to to get energy only from front to back, never the other way around, and never into the motor, only out of it. I don't intend to use regen to brake or to charge batteries, so freewheeling is exactly what I want.

Another thought I had was about the shock/suspension I need between the top of the rear triangle and the back of the seatframe, to give me rear suspension. I've got springs and stuff like that, but they don't work that well for what I want to do without a lot of added weight for long or large diameter springs, along with the tubing to keep them from warping like an inchworm under compression. :) Now, you may have seen "gauss guns" that use a magnet facing one way to repel another magnet facing the other way to repel yet another magnet, etc., all in a straight line. What would happen if you put all those magnets in a tube with a limited physical space between all the magnets, and the tube were compressible (like a shock absorber or collapsible-ring cup, in segments that enclose each other)? As you compress it more, the magnets continue to repel each other, but don't have anywhere to go, being kept in line by the tube, and pushed toward each other by the magnetic fields of the other magnets next to them. As you get them closer it gets harder and harder to push them together, rather like the way a spring compresses.

Now, I thought I had just had my lucky Eureka moment of a lifetime, and invented something to change the world...but no, it's already been done (there is even an entry right now in one of NASA's contests, according to a Google search on the subject of Magnetic Shock Absorber or Magnetic Suspension--even Wikipedia mentions the idea in the Shock Absorbers article. There's even research going on to use the action to generate electricity!). :(

But still, it's obviously a workable idea, and shouldn't be that hard to build, if I can find the right magnets for it. Oddly enough, cylindrical ones seem to be cheaper than rectangular and other shapes, possibly because there isn't as much demand for them? That means I might be able to buy enough to build a workable one on this bike. An adjustable airshock would be the best option, but I don't have that kind of money for just one of them right now. (actually, I don't have *any* money to spend right now, which is a part why this bike is being built of scrap--the other reasons are just to see if I can do it, and also on the principle of recycling this entire bike project is being done by).

Wednesday, August 27, 2008

Recumbent Bike From Scrapped Bikes

The treadmill motor idea is going to be implemented on a new bike I thought of only a few days ago, and have been working out in sketches since then at breaks and lunch at work (no other spare time so far). This bike will be made as much as possible from only old bike frames, with a seat frame made of the remains of an old bedside toilet seat holder frame whose plastic bits were all sunrotted off. It's a recumbent type, bike instead of trike, mostly because I really can't make the complex steering mechanism for a trike properly right now--I just don't have the tools or knowledge yet.

I actually want to build an airshock-based recumbent trike, which I've already drawn up, but I simply can't afford the tools and materials needed to build it.

Today I am cutting up the old scrap frames I already have for a few parts, and will hopefully get some more scrap frames for the rest of the parts I need later this week from a fellow Freecyler. There'll be pics once I have them ready for pre-assembly fitting and filing, hopefully of all stages from "whole" frame to just the parts I will use. Some of the frames I got as pieces and such, so they're not all whole to start with. At least one was pretty trashed and bent up and I'd already cut those parts off months ago, so it's also not whole.

In the meantime I'll put the basic sketches up here:

The steering idea is still in progress, as currently both drawn methods conflict with my legs pedalling at anything but mild steering angles, and I think the trail is probably totally wrong for this setup. I had planned to use a 24" shock front fork with a 20" tire or smaller, with some serious modifications to the fork's headstock mounting end, but I discovered one of the tiny kid's bikes I've got a frame for actually has shocks in front, and it uses (I think) a 20" tire or smaller. Have to actually measure it (doesnt' have a wheel in it of course). Anyway, it's definitely short enough to not have to use such an extreme angle to keep the front end lower than the pedals, which will both help the shocks work better (more vertically) and let the steering work more easily.

I'll probably have to cut and weld a set of handlebars for the steering to make it work out right, because I don't think anything I have will either reach far enough back or be bent in the right shape to be out of the way of my legs, the bike frame, etc.

Not sure exactly what I'll do for the rear spring between the seatback frame and the rear triangle, but possibly the intact side of a bent-up 24" front shock fork could be used, if mounted correctly.

The seat will just be a lawn-chair style tube frame with material strung across it that won't hold water in, and allows as much airflow as possible (to keep from getting sweatsoaked back and thighs).

Since the pedals won't be in the way, I can have battery mounts along the bottom edge of the rear frame triangle, and my cargo pods or baskets (haven't decided) can be mounted much lower, with only the rear derailer to worry about clearing.

The seatpost from the rear triangle will be sticking up as far as I can get it (probably nearly a foot above where a seat would normally be), with a lighting module for visibility in place of the seat. There'll also be lights elsewhere on the bike, to ensure as full a visibility as possible, and it'll be painted the same DayGlo Avenger style as my current bike (which will remain intact for use when this one undergoes alterations, as is nearly certain to be required).

There will be 3 chains, the first runs from the pedals (above the front wheel, with a single chainring) to the first bearing set you see hanging down just behind the front wheel. That is where the front derailer and chainrings will be, mounted on a freewheel. That freewheel will prevent the pedals from being moved if the motor is running faster than the pedals, since they'll all be on the same drivetrain. The second chain runs from a fixed chainring on there to the bearing set (actually the motor) just in front of the rear wheel. That motor shaft will have a freewheel on it mounted so that the pedals don't turn the motor if pedalling only with no motor power, but instead only pass power to the 3rd chain that runs the rear wheel cassette (which will be a standard one with typical derailer and freewheel, etc).

I'd originally thought of putting the front derailer/chainrings on the motor so that it could also benefit from those additional ratios, but the inital sketch-scribbles showed me it was too complex and would be difficult to adjust and maintain, especially if the motor had any issues requiring it's removal. It will be difficult enough to setup the freewheel on it.

It's even possible I may not use front shifting at all, depending on the complexity and clearances of working out a tensioning scheme for that chain loop.

3 separate chains instead of one long chain is partly to make derailing the chain accidentally harder to do, as well as keeping the chains tensioned properly. It's also so I don't have to go buy a bunch of thin chain, and can instead use for the front two chains some of the single-chainring-width chain I already have off these scrapped bikes, and then just the single regular thin multi-chainring-width chain I have off one of them.

Should be interesting to see how different the actual bike is from those idea sketches above. :-)

Friction Drive 2.0.2 Still Works, Except In Rain

I've been working on the treadmill motor idea off and on since the last post, but haven't made final designs yet, mostly because to my surprise, the Friction Drive using radiator fan motors and roller skate wheels has worked well enough for quite a lot longer than I expected, with little maintenance.

So far the only big problem was when I skidded on the lip of a shopping center driveway entrance as I was turning in (it is over 1.5" high--practically a curb for a bike tire, though probably not noticeable in a car or truck). I ended up too far sideways to recover and skidded out--I stepped off the bike with no problem but the bike hit the ground hard enough to knock a battery loose (had to re-zip-tie it on using the spares I always carry) and break one bolt and one rivet holding the motor plate on. A zip tie in a strategic location enabled me to keep going and get home after shopping, but it definitely wouldn't've held too long that way.

When I got home, it turned out a second bolt had been stripped a bit, so I took the whole plate off, redid the rivets holding the mounting brackets on the bike frame, and used the next size up for bolts. Took a bit of searching thru my junk boxes to find a set of matching nuts and bolts that also fit one of the wrenches I already carry with me (since I don't want to add another tool to it just to road-fix any motor issues). These bolt heads and nuts are a bit thicker than the originals, and thus on occasion if I am pedalling hard with little motor input (or during accelleration from stop when the motor is least effective) the left crank scrapes a tiny bit against one of the nuts, even though the nut's filed down as far as it safely can be. Can't really do anything to fix that, without using thinner-headed bolts I don't have (and would have to buy).

The rollers themselves have worn some, requiring occasional (2-3 weeks apart) retensioning of them against the tire, but no other issues have developed than those previously posted about. The tire itself doesn't show any excessive wear from it either.

The only major problem the system has is that when it's even a little wet, it really doesn't work. If I just *barely* throttle up once I'm already going, I can add a bit of motor speed to the bike, so I'm not completely pedalling, as long as it is not actually raining or I'm not riding thru puddles. During anything more than a light sprinkle, then even at low motor speeds the slickness is too much and the motor speed overcomes the friction, rather than transferring any torque at all to the tire, and the motor just spins at it's free-run speed at whatever throttle input level it's getting at the time.

Since this *is* Phoenix, Arizona, it generally doesn't affect my rides, since even when it's cloudy or rainy, it doesn't stay that way for long, and roads dry up enough to work with pretty quickly. During this season, it's a little tougher, and the other night my ride home from work was all pedal-power due to the nearly-torrential downpour the entire way.

It's other major disadvantage is the huge amount of power being wasted in the fan motors, since they're designed for low voltage (12V), high speed, and low torque (they drove small plastic fans to cool off a car engine in their previous life). Since I'm using them at 36V (actually about 28-30V since I can't get more than roughly 75-80% duty cycle out of the controller I've got on there, either because it's damaged or because it's designed to only do that), and at much lower speeds at those voltages than they'd free-run at, to drive a pretty heavy load, they get very very hot, perhaps 150F or more by the end of a less-than-2-mile run at their axle cores. Given that they're bolted to relatively large aluminum plate that gets significant airflow across it's entire surface during almost all of a ride, that's a lot of wasted power in heat. That's why I would like to go to the treadmill motor, as it was designed much more for this type of purpose.

Since I have only old SLA batteries to test with, I don't really know what the range would be if I had new batteries, but the most I get out of it when being conservative with it's usage is roughly 5 miles before having to stop for at least 30 minutes to let the system cool off and the batteries self-revive, which gives me another 2-3 miles after that before it *has* to be recharged for up to 14 hours to get that same range again. If I had new batteries that allowed higher C-charge rates, I could probably get a lot more range with a quicker turn-around at the end.

All in all, it's a definite success. For those reading and looking for a good method to drive a bike via motor, I can seriously recommend this particular one for it's simplicity of design and build, and low maintenance, as long as you're not in a wet area. If you are, you're gonna need different tires and rollers to generate higher friction when wet, which will probably wear your tire out faster. Other than that, and the motor waste heat/energy problem, it works pretty well, old batteries notwithstanding. :-)