December 31, 2012

PocketBoard Part 2: Now it Need Wheels

To clean up the rough surface left by the angle grinder, and to make the board flatter, I planed the both sides by hand.

The small block plane left the surface smooth enough that it barely needed sanding:

I jigsawed out the rough form of the deck, and cleaned up the edges with a plane and some coarse sandpaper.

I did not want the bolts that hold the trucks on to split the wood, so I added aluminum plates to the top of the deck, to spread the load from the screw's caps.  To make the plates level with the surface of the deck, I routed depressions for them.  The edges were finished by hand with some tiny chisels.

The plates were epoxied to the deck, and drilled for the bolts.

To prep the wood for finishing, I sanded the wood progressively to up to 1500 grit sandpaper, leaving the surface extremely smooth and almost polished.

Rather than applying some sort of varnish or polyurethane, I finished the sufrace of the wood with tung oil.  Tung oil gives the wood a lovely color, waterproofs the surface, and (in my opinion) looks and feels nicer than a clear coat that adds material to the sufrace of the wood.

Here's a comparison of the oiled vs unfinished wood:

And here it is with the trucks bolted on.  As usual, I used countersunk stainless steel socket cap screws.

Once I get some wheels, it will be ready to ride.

December 28, 2012

PocketBoard: The Tiny Oak Skateboard

While my scooter is great for medium to long distance trips (when it is working...), especially with a load of cargo, it is too large and heavy for short excursions.  In order to improve my response time on Free Food and Reuse posts, I needed a very small wheeled vehicle that I could easily stuff into my backpack once I reached my destination.  A couple weeks ago I snagged a pair of skateboard trucks from a reuse pile in CSAIL, which decided that this vehicle was going to be a miniature skateboard.  

I brought the trucks back to Atlanta with me, because it has been a long time since I made anything out of wood.

About 10 years ago, a large oak tree fell across our back yard.  This past year, my dad got a chainsaw so he could cut up and remove the massive trunk, so we have a number of oak trunk sections in our back yard.  Despite sitting outside for a decade, only the outer couple of inches of wood are rotten, so there is still plenty of good wood left, if you are willing to work for it.  My dad split off a 1.5" thick board for me from one of these stumps.

Here's the board.  I passed it through my granddad's planer a few times to get it flat on both sides.  I originally planned on thinning it down completely on the planer, but for some reason the wood was extremely hard to push through.

The board was thinned down to just over 1/2" thick using a 36 grit grinding wheel on an angle grinder. This method proved to remove material remarkably quickly: I removed a half inch of solid oak in 15 or 20 minutes.  The surface of the board will be planed by hand for a nicer finish.

December 17, 2012

Improper Use of Machinery

With the arrival of some parts from Surplus Center and Monster Scooter Parts, I was able to finish the differential and assemble most of the drivetrain.  

Since I don't know where to find a proper broach on campus (maybe the Edgerton shop has a set?) I had to horribly abuse the poor MITERS tools use some unconventional machining techniques to make keyways in the differential's output gears.  I basically followed this guide, but used the mill instead of the lathe, and using a tool I  ground out of some tool steel blank.  I tried using the lathe at first, but found that the tool holder would twist if I accidentally tried to make too deep a pass, so I switched to the mill, locking the output so the cutting tool wouldn't turn.  The keyway isn't perfectly the same depth all the way down the gear, since the tool flexed some, but it works well enough.  I imagine this method would work much better on aluminum than steel.  

Yes, I chucked a piece of square metal in a 3 jaw drill chuck...

I widened out the bore of this 15 tooth sprocket, and bolted it to the output adapter I made for the hub gear.  All the hubless sprockets I got had teeth numbers that were multiples of three, so I could easily widen their bores on the lathe.  I totally thought of that beforehand.

And here is the assembled drivetrain.  I was not as lucky with the left side as the right side, so I had to make a chain tensioner to take up the slack in the chain.

Here's where the trike lives when I'm not working on it.  ALL The Bricks below it, Hat Coil to its right,  and parts from MITERS plane behind it.

When I started building the fork, my plan was to attach the front wheel via two blocks of aluminum that would clamp to the legs of the fork.  I made these two clamps, but I managed to sheer off one of the tightening screws in the process of clamping the block to the fork, making the piece useless.  I decided to go a different route rather than remake the broken part, and just weld on some steel supports for the front axle.  I cut the supports out of some scrap steel bar, and bolted them together at the correct spacing for welding.  I'm not sure what alloy of steel it was, but it made almost gold colored shavings on the mill, along with lots of unhappy noises.

Tacked in place:

This weld made me really happy.  The others weren't quite as nice, but they still didn't need much grinding.  I seem to be improving at this MIG welding thing.

The axle is a random stainless rod I found.

I may replace the rod later with a threaded rod, so that I can turn the axle holes into proper dropouts like on a bicycle, so that the wheel can be easily removed.  A threaded axle would also provide a convenient way to attach foot pegs to the fork.  The foot pegs could even double as the nuts that clamp the axle in place.

To stop the front half of the trike from falling over without a person on it, I attached some air springs from SEGFAULT.  They don't provide nearly enough force to lift up the frame when it has been tilted, but they do stop it from falling over.

To interface the rear wheels with the differential and drive shafts, I first pressed out all the bearings, and removed the stock sprockets and band brakes.  I milled one side of each of the wheels flat, so I could bolt on these hubs with 1/2" bores and 1/8" keyways, also from the now disassembled SEGFAULT.  Later on, disk brakes will also be bolted to these hubs.

I made some little aluminum inserts for the opposite side of the wheel, to guide and support the axle through the entire wheel.

And now it rolls!  One possible issue I've noticed now that it's assembled is the questionable steering geometry (rather like the first edition of my scooter's steering, but not nearly as bad).  Because of the alost 90 degree head tube angle, the steering has negative trail, which may make leaning difficult.  I guess I'll have to ride it to find out.  If it is a problem, I could either simply flip the fork around 180 degrees, or cut and re-weld the head tube at more of an angle.

Two sets of parts I forgot to take any good pictures of the making of are the two bearing blocks and the two outer vertical plates.  Each of the bearing blocks has a pair of bearings pressed into it.  They are held in place by two screws from below, as well as two screws each through the vertical plates.

Cutout for the end of the motor:

The protruding corners of the bearing blocks wil be removed:

This is roughly what the top plate will look like.

I split the top plate into two parts.  The smaller bit will be the cover to the battery compartment.  I plan on making two battery packs for the trike, so that I can quickly swap them out when one dies, and I did not want to have to take off the entire top plate each time.

Since I had to oversize the chain between the hub gear and the differential, I made a chain tensioner to take up the slack.  I started with a rod of some sort of round plastic stock, nylon, I think, cut it to the proper diameter on the lathe, and bored out a hole to press a small bearing.  I then used the indexing head on the mill to drill out ten holes around the perimeter of the plastic cylinder, corresponding to the low spots in the sprocket.  Since plastic is soft, I just used a coarse file to sharpen the teeth.  It rolls quite smoothly, and since it is on the low tension side of the chain, it should not wear too much.

Battery pack time.  I pulled out one of MITERS's many boxes of A123 28650 cells, to see exactly how many I could cram into the battery compartment.

I can fit about 36 cells, with room for wiring, which will give me a 12S3P pack. I also did some test fitting with some 18650 cells (the size commonly found in laptop batteries) and found that I could fit 96 cells, for a 12S8P pack.  In A123 land, this means around 28% more capacity, at the cost of much more labor to make the packs.  Also, there are way more of the larger cells freely available, so I'll probably just use those.

This is the last update until early January, when IAP starts, and I will have as close to unlimited time as I ever will here.

December 2, 2012

Building the Differential

I started out with a foot of 12 tooth, 20 pitch spur gear rod, two 20 tooth pinions, a pile of bushings and bearings, some 5/16 precision steel rod, a 2.5" diameter solid aluminum cylinder, and some miscellaneous other hardware.

I cut the spur gear rod roughly to length on the horizontal bandsaw, and faced the ends on the mill.  I bored out the centers, and pressed in pairs of bronze bushings.

The configuration of the gears inside the differential looks something like this, with the two 20 tooth pinions connected to the two output shafts:

I made the end caps from some 2.5" aluminum round found in the MITERS scrap pile.  I faced the end to make it easier to align properly in the lathe.

So I wouldn't have to constantly refer back to my computer, I printed out some engineering drawings with all the dimensions I needed.  Some are penned in, because I forgot to double check if I had all the dimensions before printing them.

This is the finished end cap that goes opposite the drive sprocket.  Things I learned from this:  Go very slowly when making press fits.  A tiny bit of material can make a huge difference.  I botched one of them, and made the bearing hole a hair too wide.  Fortunately, it was nothing some precision shim couldn't fix.

Six steel rods span the differential, serving as both axles for the planet spur gears to spin around, and as the structure that transfers torque from the sprocket side of the differential to the non-sprocket side.  I wanted these rods to all be the exact same length, and I didn't want to rely on the sketchy MITERS calipers to make them all the same.  To make sure they were identical, I decided to face their ends simultaneously.  To hold all six rods in the mill vice, I made a fixture out of some aluminum square rod.  I drilled 6 holes in the rod, and cut alternating slits through to the holes.  When the square rod is clamped form its ends, the slits compress, and clamp the rods together.

It also came in handy for tapping the ends of the rods without grunging their surfaces in the vice.

The rods are countersunk 1/2" into the end plate, and are held by stainless steel socket cap screws from the other side.

Here's the mostly assembled differential.  I still need to add keyways and set screws to the output pinions, and make spacers to keep the planet gears axially aligned 

And a rendering, for reference:

I found some 1/2" aluminum tubing lying around, so I used to temporarily assemble the drivetrain.

Wheels are on their way, so I should have a rolling frame relatively soon, assuming the last couple weeks of classes and finals don't eat all my time.