Archive for November, 2008


Project RazEr: It’s (99%) Legit.

Nov 30, 2008 in Project Build Reports, Project RazEr

Update:Test video! Now it officially exists. Low speed only for now, since I b0rked the controller and stuck it in Sissy Modeâ„¢

After eight months of development, five of which were spent sitting on a shelf at MITERS, Snuffles Reloaded finally moved under its own power. And boy did it move… I have already managed to nearly faceplant once on it.

With the project having entered a state of semi-functionality, and with it standing a decent chance of becoming something worthwhile, I’d like to move onto a less whimsical name. While “Snuffles” was cute, it just doesn’t have that aura of coolness.

Okay, so “RazEr” doesn’t either, and the coolness of a Razor scooter maxed out some time in the year 2000. But see what I did there? It’s an E! Instead of an o! For “ELECTRIC”! And it’s a Razor scooter! Get it?? INNOVATION!

Without further ado, here’s the 99% finished vehicle.

Yes, that IS a small, amorphous blob of wire and circuitry duct-taped to the steering column with a knob on an equally duct-taped stick acting as the speed governor!

No, it’s not spring returned. It’s cruise control for cool. Without caps-lock, and potentially more blood.

I finally mustered up the effort to pull it off the shelf and hook up the last four wires remaining on the thing which had kept it from moving for the past few months. With the help of some friends, we got everything bunched together as a test rig – then decided to call it good and wrapped everything in tape. And good it was. This time, functional cameras are immediately accessible, so expect some “action shots” when the sun comes up.

So why is it “99%” done? Well…. right now, it will probably get me arrested if I take it to the airport (then again, when you’re an MIT student, anything will). Obviously there needs to be some wire cleaning and rerouting. I don’t have a correctly sized box to house the onboard voltage regulator and R/C signal interface yet. I have a proper E-bike throttle that fits over the handlebars. I also need to make a charge-balance plug for the battery pack.

It also needs underglows.

However, it is at least mechanically sound, and I was able to ride it back from MITERS. Therefore, I know the range is at least 3000 feet minus one railroad crossing.

The business end. The relevant power transmission implement – there is only 1 – is housed completely in the rear wheel. It is a 80mm diameter custom-built 3-phase brushless DC motor, conveniently hidden within the confines of a 125mm scooter wheel. Maximum power on it is probably about 1000 watts. I have yet to properly meter it.

Even with no torque advantage (as a direct drive motor), the acceleration is pretty absurd. It’s not quite the neck-snapping and rider-launching takeoff of Snuffles 1, but I do need to hang on pretty hard. It is, however, a controllable launch, and will be even more so when a proper spring-loaded thumb throttle is installed (you know, so I don’t have to hang on with one hand and one leg while twiddling a knob on a stick with the other hand)

The whole vehicle is very low profile. Yes, I managed to drop the height of a Razor scooter – with an add-on battery pack. You think pebbles were bad before…

This is the “larger” flavor from Razor, which has 125mm wheels. With the belly pack, I still have about 1.5 inches of clearance. Not going to be bounding over any curbs or potholes, but it’s enough to get around on bike lanes and sidewalks.

There are four 4.2AH lithium polymer flat cells inside, and four outside. They are run in series for 29.6 volts nominal. The packs are rated for 20C discharge (80 amps) and can burst 30C, so I have plenty of reserve battery power. As for capacity, calculation showed that my optimal range at constant power on flat, frictionless ground in a universe with a constant gravitational field at a constant temperature in air with zero viscosity and me approximated as a point mass… is about 4 miles, no entropy generated. Realistically I expect far less, and need to do a real rangecheck some time soon. Anything above 2 miles is really enough to putz me around campus and into town and back.

With an extension of the belly box, and some custom control electronics, I could conceivably hide all the control stuff inside the tube frame, and only have a throttle cable coming out.

Yeah… that’s about how it is.

And it weighs this much. With better mounting facilities, it should weigh around 12 pounds. That’s a hair under 6 kilograms for you people in metricworld.

Anyways, work will now continue on dressing up the control system. I’ve officially moved past the “90% Zone of Project-Related Self-Loathing” which occurs with everything I build – I would furiously build it to 90% of the way, then just get sick of it and forget about it for a while. Now that I know it works, I should be a little motivated to push it through.

(read: expect another update in 8 months)

I have kept a detailed build log of this project and its predecessor on this site.

Bot on??

Übercløcker RЭmiχ

Nov 29, 2008 in Bots, Project Build Reports, Überclocker Remix

Can I add more Unicode symbols to the name? Yes.

But I won’t.

It’s that time of year again – time to prep for the winter competition season. The event of interest is Motorama 2009, only 2.25 months away (again…). Last year, I prepped Test Bot 4.5MCE-SP1-WTF-BBQ for the event, which sort of ended in semi-epic fail. A few months ago, I built Überclocker for Dragon*Con 2008, which ended in epic fail.

The mission this time is to bring Überclocker for the 30lb Sportsman Class. The 30SC division is a recently created weight class which attempts to shift the focus of lolbot combat back to innovative weaponry and designs. I originally designed ÜC for this class.

But of course the bot can’t just go in its current form, since nothing really works. Therefore, over the past few weeks, I have began to refresh the design to address the shortcomings of ÜC1, which pretty much meant starting from a new file.

Let’s begin. A few hours of diddling got me the basic shape of the frame, which incidentally is sort of like the frame now. The big difference is the method of construction. Since proving to myself that I can build a bot in two weeks, I’m going to try to “Chinese Puzzle” the frame like I did on NK, with a variety of tabs, slots, nutstrips, and other elements cut into quarter and eighth inch aluminum plate to be magically assembled with the help of some retaining screws.

With prudent design, I should have to do minimal machining (e.g. tapping, enlargening, or boring out holes) on the superstructure of the bot.

A few more hours of diddling, and the basic shape of the bot is defined. The extended legs will return in a different form. The old fr0k assembly is in place to gauge the dimensions of the bot, and will actually be replaced with a whole new assembly.

Comparing side-by-side with the current bot.

This version will be 2″ narrower (since I will not have a problem with lateral stability) and actually a quarter inch thicker. Did I just make a bot *TALLER*? Yes, it is so I can fit the electronics in more efficiently. I should be able to double-deck the Victor 883 controllers inside. Victors, while extremely reliable and properly rated, are not exactly compact. By not requiring as much electronic floor space, I can shorten the middle of the bot and move the center of gravity further back.

The wheelbase is also 2″ longer than the current build, and the fork drive assembly (the single heaviest part in the bot) is further back. Ultimately, the goal of all this is to move the center of gravity as far away from the front as possible, in an attempt at mitigating faceplants.

Let’s move on to the drivetrain. A major change from the current build will be the ditching of timing belts in favor of #25 chain. While there’s fundamentally nothing wrong with belts (TB and NK both use belt drive), I found that a set of timing belts and pulleys was just too wide for the new rear drive solution.

What is it? Direct gearing! The drive motors will be 1:1 geared to the rear wheels. This solves the problem of getting power to at least one wheel on each side reliably. On TB, the drive motor cranks the rear wheel directly, but doing so in Ãœberclocker would require much larger wheels than I feel like using.

These gears will be 12 pitch and waterjet cut from 5/16″ steel. I’ve cut gears on the WJ before (hit refresh a few times to spot the site banner featuring some of that work), and figured this was a better (read:cheaper) solution to getting a handful of gears than buying the gears (which do not come in the face width and with the hub features I need) and machining them down. All of them.

Furthermore, cutting my own gears allows the use of custom bore features – like the double-D shape seen on the motor shaft. This allows me to cut a flat on the motor shaft and have the gear sit at a known location while transmitting torque. It will be retained by a screw and washer.

The sprocket on the other side of the wheel will also be a custom cut piece (thank goodness for geometry generators) with a large DD to match the one machined on the hub. A big retaining ring keeps everything in place. This allows the transmission of torque from wheel to sprocket with mininal width (just wide enough to pass a chain).

The whole assembly is 2″ wide.

Oh hey, it’s a new fork. This one is much lighter, weight and build, than the current. Sections of threaded rod with nuts will bind all the plates and whatnot together, hopefully forming a stiff structure. Only the outer forks, which will be taking the first hits, are half inch metal, with the inner forks 1/4″ thick.

Notice the little trusses and holes starting to appear in the frame. This is my attempts at keeping weight down while maintaining some semblance of structural integrity.

Also, the fr0kdrive assembly has been compactified. I realized that with some tweaking of the geometry, I could stick the lift motors right under the pivot and drive the fork directly with gears….

And gears there will be. The two lift motors (the über-ghettofrakenb0xen) have been melded together into one case as the Integrated Dual Frakenb0xen. Each gearbox is 216:1.

On the other side is a 3.2:1 arrangement of 12 pitch spur gears. The pinions are stock commercial parts from McMaster-Carr, and the upper assembly will be waterjet-cut. Pinions ride in their own enormous ball bearings, and so does the entire fork assembly. Overall reduction from motor to fork is 691:1.

Such absurd gearing can easily destroy itself if there’s no limit on the torque produced by the motors. Ãœberclocker currently uses shaft collars clamping on a split shaft to limit the gearbox output torque. This works well, but there’s no intermediate shaft this time.

I have decided to implement a cone clutch assembly on the drive gear. The blue and cyan parts are rigidly attached to the fork structure, with the blue part movable axially with adjustment screws. It is tapered externally and mates with an internal taper in the gear. The gear will slip at different torque levels depending on how hard I crank the screws. Or so I hope.

This will be a venture into precision secondary operations on a waterjet-cut part. Up until now I’ve been doing pretty mundane stuff to my WJ’d parts, like threading, enlargening holes, and the occasional boring. However, this would require some pretty precise lathework on the bore, and keeping the less-than-consistent edge of the WJ’d part in line will be interesting.

The gear, blue cone, and cyan mount will all be roughly profiled on the waterjet (straight sides). Then they will be externally and internally machined as needed.

Also  being revamped is the clamp arm drive. The bot currently uses a really dumb leadscrew actuator mounted to a weak-ass insect-class motor. For some reason I expected this to stay together in battle.

It didn’t, so I’m upgrading. The new actuator is a 540(or 550) drill motor run through its own 150:1 gearing. I figured that as long as I was making a weird gear for the main fork drive, I’d make one for the clamp drive. The drill motor is mounted next to the gearbox, and power makes a U-turn from the motor into it, through a 1.75:1 belt drive (not really visible). This feeds into a drill gearbox assembly (36:1) and outputs onto another custom WJ’d spur gear set (2.5:1).  The large output gear is actually part of the main fork structure, so the clamp arm can move relatively to the arm, carring the actuator with it.

This gearset is clutched at the drill gearbox ring gear (like a normal cordless drill), mostly to prevent it from exploding if another bot tries to force its way out.

This is the number one “weird complicated mechanism” that might get revised before I get there in the actual build process.

And so the culmination of everything is…

The upper grabber structure is assembled the same way as the fork – truss plates and main side beams bound together with threaded rod. I actually planned ahead and am implementing squishy bumper feet things for industrial furniture (from McMaster) into the tips. This should make bots less likely to just slip out.

Top and bottom plate attachment is not yet designed, nor are internal electronics mounting implements. Now that I have discovered that McM sells carbon fiber panels in normal thicknesses and sizes, I’m going to try and CF the whole bot.

If all goes according to plan, I’ll have the large Ãœberclocker logo on top backed in white plastic with an EL panel, so it glows with a sinister light. Or something.

So this would be Update #1, I suppose, for the new bot. The design is not yet final, and things will probably change some (like that clamp drive assembly). I do need to get stuff cut out soon, though, if I want to make Moto. I need to factor in the fact that there will be a roughly 1 month gap between now and then in which I will not have 24/7 access to some form of machine tool…

LOLriokart Update 6: Criticality Edition

Nov 08, 2008 in LOLrio Kart, Project Build Reports

In nuclear physics, criticality is when a chain reaction becomes self-sustaining. In MITERS physics, criticality is when your project can actually support its own weight without being propped up by 2x4s. It’s only taken me six months to put 4 wheels back on the kart (which can now function as a shopping cart again!). The driveline mechanics and steering linkage are still not complete, but at the very least, it can be rolled around at will to clear space in the lab.  So, naturally, it’s picture update time.

After making a mirrored set of bearing blocks, it was time to test the axle alignment. Either my parts aren’t square with respect to the cart frame, or the cart frame is just not square at all, but it takes a bit of malletwork to push the axle through. I have to keep the retaining bolts pretty loose during axle service also.

I thought of a few workarounds like remaking the double blocks from one single block, or using self-aligning mounted bearings, but put it off to work on the rest of the parts.

Temporarily mounting a drive hub to test fits and clearances. As expected, I’d need a shaft collar on the outside of the bearings to keep the tires away from the mounting bolts in addition to the collars on the inside. I may end up making this one large shaft-block-spacer assembly to reduce partcount.

With the rear tires bolted on, it was time to drop the thing off the milk crates which have acted as a work stand. Here’s the first “on-the ground” test. Not much a test, since it had no steering linkage or drive components, but it was a good chance to test the rigidity of the entire system by jumping in the basket and rocking around. Verdict: I approve.

Let’s get to work on the steering linkages. Out of the same mysterious machine that supplied my brake-o-hub original material come these trunion-like things. Each one is a one-piece machined aluminum block with a bronze bushing in it. How convenient – they seem like just the thing to turn into steering arms.

The left part is the original part, and the right the finished steering arm. I had to cut down much of the material to get to that stage, but it was the most compatible with the months-old steering knuckle design, and also an excuse to make the space smell like Tap Magic.

Test mounting the finished part. The chamfer allows a farther range of motion in that direction. Normally, the edges of the knuckle blocks themselves act as the hard travel limiters.

The combination of hard machined lines and angles give the front end a rather distinct look that I rather like.

So remember how the bearing blocks weren’t exactly square with the frame? They actually not square with themselves either. Small variations in how the steel tubing cart frame was welded made aligning the bearings to pass the shaft through extremely difficult.   To alleviate this, I made an inter-block spacer with pilot circles that snugly fit the bearing bores. At the very least, this forces the bearings to be square with eachother. The two sides may still be out of square, but retaining bolt holes can be fiddled with.

After properly spacing out both sides, it was much easier to remove and install the driveshaft. Speaking of which, I milled double flats on the ends to give the hub screws something to bite onto. The shaft is also internally threaded on both ends, so I can at least have something physical interlock (like a bolt with washer) between set screws and certain death.

With both linkage arms made, it was time to try and link them together to check out the steering geometry. The spirit of the mysterious machine supplied these dual-bushing aluminum spars which, by some chance alignment of Cybertron and Earth, was the exact length to provide perfect straight-ahead wheel alignment (Any toe angle is too small for me to notice).

And so the first inaugural test run of the kart was completed, with three people involved : One to ride, one to push, and one to flick the steering.

Steering linkage in (mostly) straight-ahead position.

….and at maximum departure from center, turning right. Notice the difference in angle of the wheels – this is an attempted implementation of the Ackermann steering geometry, which is what modern car-type steering is derived from. That’s the reason I took the extra step to make the steering arms offset to one side.

So that’s all for now. Even though the big aluminum link was fun, it will probably be replaced with some big tie rods to allow for fine adjustment. I have not planned out how to actually attach the steering wheel to this linkage yet, nor have I designed the motor mounts or rear powertrain. Or anything for that matter – the entire kart build so far has been design-and-build on-the-fly. I don’t know what parts I’m making until 5 minutes before I actually make them.

The primary goal is to get the steering and front brake system working, since that allows the kart to be used as a rolling chassis for any number of different propulsion methods.

Hope will be the only thing you have.

Nov 05, 2008 in Stuff

The European Union and Canada welcome the United States in the formation of the new Dark Age, where our oppressors are not nomadic raiders and manor lords, but bureaucrats and lawyers.

Of all tyrannies, a tyranny sincerely exercised for the good of its victims may be the most oppressive. It would be better to live under robber barons than under omnipotent moral busybodies. The robber baron’s cruelty may sometimes sleep, his cupidity may at some point be satiated; but those who torment us for our own good will torment us without end for they do so with the approval of their own conscience.

- C.S. Lewis