Archive for May, 2008


Pop Quiz 2 Update 4

May 29, 2008 in Bots, Pop Quiz 2, Project Build Reports

Now that the lathe is better equipped to produce shiny, round objects, it was time to get back to work on Pop Quiz. Having a better tooling stup dramatically improved efficiency, and I was able to complete the motor mechanics today. Now I just await my SDP-SI order for the bearings and some shoulder screws.

Oh, and I also discovered that the lathe does indeed have back gears. They were rather well hidden, but now the operations which need <180 RPM at the spindle can happen

Boring out the 1/4″-walled steel tube to form the motor can, using the not-boring-bar.

A stiff toolpost does wonders for surface finish and accuracy. Here’s the finished can.

Next up was the motor mounting base and the hub for the can. I dug up the micrometer depth stop which helped stop the carriage at a repeatable, set distance. This meant I could manually feed using the carriage handwheel and not the compound slide, which eliminates the risk of accidentally cutting a taper. A DRO would make this unnecessary, but…

The mounting base begins to take shape… The same shape as the first try, but with provisions for mounting two stators stacked on eachother.

After cutting that part off (with a real cutoff blade!), I started working on the can hub. This involved two precarious internal boring operations. The larger one I could take care of using the not-boring-bar, but it couldn’t fit into the 1/4″ starting hole for making the circular bearing pockets (which are only 3/8″ in diameter anyway)

I should have bored these small holes on the milling machine, but instead I took the risk and used a 3/8″ endmill in the tailstock chuck. Oddly enough, it came out way *UNDERSIZED*, at .365 or so. This is weird. Very, very weird. What the deuce?

The three completed pieces after processing. The can presses onto the hub, and is retained by a healthy dabble of green Loctite. The whole thing sits over the mounting base, and a shoulder screw threads through the hub and into the mounting base to keep the whole thing together. I didn’t make the exterior mounting ring as seen in the 3d drawing yet.

Next steps: make the exterior mounting ring thing, as seen in the 3D rendering. Make a fixture for the UHMW slab chassis, then machine it. Get some carbon fiber and make the cover plates. Wait on Banebots and SDP-SI for the load of internal parts.

Now that I have at least part of the machine tools I used the most before, things should go much smoother. Dragon*Con 08 will be here in no time…

Ghettopost II.5

May 28, 2008 in MIT & Boston, Project Build Reports

The second half of that aluminum billet was just sitting on the mill with the “make something using me to end your pains” look.

And so I did. Now introducing the other Ghettopost: a cutoff tool holder!

For the uninformed, the MITERS lathe has no parting /cutoff tool holder. As addressed in my Intro to Ghettomachining post, the only way to get your part off the stock you started was by hacksaw. Some would dismount the part and hacksaw it manually. I usually just slammed the hacksaw on the stock, with the spindle running, and slice it using power.

Either way, it wasn’t fun, nor particularly safe. And so with the relative success of my Ghettopost II from two days ago, I decided to complete the square blocks of aluminum and make a cutoff tool holder.

Using lessons learned from the turning tool holder, this build went much smoother and with more accuracy.

Sketching the outlines of the finished part in lieu of a full drawing helped alot. This was the other half of the giant aluminum billet that I bandsawed in half. First step was to shorten the height to 1.5″. For this, I used the giant shell mill, doing it all in one .400″ deep pass.

After the block was planed down to height, it was time to mill top features. The “stairstep” shape is to hold the tool as well as a clamping bar thing to lock it firmly into the holder. The large hole was started with a 13/16″ drill and then finished off to size with a boring head, just like last time.

After the top features were machined, I turned the block around to drill the horizontal clamping hole. Then around to another side to mill the Slit of Clamping. Finally, I flipped it over and cut 5/16″ out of most of the underside.

The reason for doing this is because the cutoff tool is a very tall and narrow tool. I sized this post according to the remains of a 3/4″ cutoff blade that I found. Because its cutting edge is correspondingly 3/8″ higher than the highest I accounted for. As a result, the tool has to sit lower than the flange on the 1″ post I turned.

The solution was to raise the contact surface of this holder to account for the increased height of the tool. The cutout is 2 inches square, which overlaps the 2 inch diameter flange and allows the little step at the bottom to hang off the edge of the compound slide.

After the holder body was done, it was time to make the clamping bar. This would have taken a few minutes if I had some 1/2″ steel square rod. Unfortunately at MITERS, everything is around when you’re not specifically seeking it out, but I couldn’t find a single square rod near half an inch.

I had to settle for a very odd-looking .5″ thick, 1.25″ wide bar. After taking a few test cuts, I realized it was stainless steel. Oh boy, work hardening. Slow and steady was the answer here, and I somehow trimmed this chunk down to 2.5″ long x .625″ wide x .375″ thick.

The complete clamping bar, with holes drilled. The first hole was accidentally drilled for a 5/16″ clearance because for some reason I thought I was using 5/16″ screws. Fortunately, it does not affect anything, as the tool will point out the other way.

The clamping bar reaches over the tool bit and has a little lip on the edge to prevent it from tweaking sideways.

And thus, with some added hardware, the cutoff tool holder was complete. Here’s a video of the inaugural cut, through nothing less than quarter-inch walled, 2″ diameter steel pipe. It was great.

This tool holder uses 1 5/16″ clamping screw instead of two 1/4″ ones (okay, one 1/4″ one.) I might rebuild the other holder, or at least rethread it, for a 5/16″ screw.

Now that I’m done making tools, I can start BUILDING ROBOTS! To complete the full complement, I’d have to have a dedicated solution for a boring bar. However, most of those fit in normal tool holders, so it’s not a priority.

Oh, here’s a picture of the ass end of a liquid nitrogen tanker truck. They go around every week refilling massive (20+ foot tall) tanks of LN2 around the labs. Shortly after I took this, the driver came around and chased me away.

Ghettopost II

May 26, 2008 in MIT & Boston, Project Build Reports, Stuff

I finally cracked.

I got absolutely fed up with the toolpost on the MITERS lathe tonight when it, despite my best attempts to prevent it, loosened on a cut in some steel pipe (for Pop Quiz’s motor can) and chipped the head off one of the new carbide-tipped bits.

So I spent the next 8 hours after that designing and building a new toolpost to replace the 80-year-old-steel-on-a-stick business current on there – it’s called a “lantern” toolpost, and is just about the biggest pain in the ass ever contrived by man.

After browsing through some designs for DIY toolposts, I settled on a simple one-tool, height-and-angle adjustable design. I decided to not go totally balls-out and try to make a real quick-change one. Rather, I just wanted something that didn’t have 96,105 degrees of freedom, and was alot stiffer. With the amount of time I spend messing around at MITERS, it was worth the effort, but not too much. We also have no dovetail cutters.

Off we go.

The first step was to make a post. Luckily, a strange plumb-bob-like thing made of steel, with a 2″ body and a 1″ stem, was hanging around in the metal bin, saving me from having to turn an entire 2″ steel round down to 1″. Some quick passes to cut off the threads on the stem and I was in business.

Here’s a great picture of blue steel curls. This is probably bad, but ignorance is bliss and so are cool colored chips.

But now you can laugh at the amazingly shitty finish on the post. Also notice the T-nut machined from some handy steel bar, and the 1/2″ bolt that holds everything together.

After this picture, I put the post back onto the lathe and smoothed it down with some sandpaper. The OD on the post is one inch.

Next was the tool holder itself. I found a giant aluminum billet which someone apparently tried to bore through and stopped midway, and cut a chunk off on the bandsaw.

I finally got to play with the big shell mill. Quite a change from the little endmills, as it was consuming aluminum at quarter-inch depth & full width as fast as I could crank the handle. It also left a brilliantly smooth finish.

I carved the block down to 2.5″ x 2.5″ x 1.5″ thick.

In another episode of new tooling experimentation, I decided to try out the boring head to make sure the holder fits snugly over the post. This 1.01″ hole goes all the way through, enabling the holder to pivot on the post and be clamped down anywhere needed.

I decided to accomodate up to a 3/8″ bit for now. The slot was positioned such that the top surface of a 3/8″ square bit would be aligned with the rotational axis of the chuck. Since we mostly have 5/16″ and 1/4″ square bits, it means the tool should never “bottom out” in normal use.

Five set screw holes will eventually hold some big cap screws to lock down the tool bit. For some reason, I drilled them all the way through when halfway would have been enough. Does not affect anything.

Center test. It passes quality control.

Here, the “slit of clamping” and clamp screw holes have been machined. The top half is threaded, the bottom half is a loose clearance, both for 1/4-20 cap screws. I could have used a single large screw, but couldn’t dig up a 5/16″ or 3/8″ cap screw of the needed length. All threaded holes are 1/4-20.

The robot gods had to intervene somewhere, of course, and it was on the very last hole tapped for the night. There wasn’t enough to wrench it back out, so I ground the broken tip down flush. One clamp screw *should* be enough. There is space between the two screws to add a third if necessary.

Here’s the height adjustment in action. Another cap screw threads vertically through the holder and pushes against the post, allowing precise levelling of the bit. Once that’s set, the clamp screws are tightened down and the whole thing locks up solid.

It allows clean height and rotational angle adjustment, plus the ability to “hold” settings between tool changes. One twist doesn’t affect the entire setup, unlike the lantern post.

The bottom of the holder can be raised up to 1/2″ off the post. If it is up this high during operation, something has gone wrong. With a 1/4″ bit, it should be around 1/8″ up in order to cut on the centerline. Adjustable to taste, material, feed rate, bit style, etc…

The finished product. It really should be steel too, but I couldn’t find a big ingot of steel lying around. Big aluminum works also.

Notice the slot cut at a right angle to the original clamping slot. This was done after I realized that about .75″ worth of aluminum was not going to give way to a single screw’s clamping pressure. So I had to weaken my own part on purpose with the slot to enable the clamping to be more effective. The tool holder tightened down easily on the post after this quick mod.

Time for a test run, in a steel round. This was a really ballsy cut – .03″ depth, but at a somewhat slow power feed (196 TPI?). About half this depth was the “danger zone” for the lantern post. I wanted to take it further than I would normally do so just to see what would happen. It took the cut fine. By the way, the sudden shake near the end was when one of those flaming hot steel curls hit me in the nose.

In summary, 3 tools to change a bit (original lantern post wrench, a vise grip, and a 7/16″ box wrench) down to 1 (one 3/16″ hex wrench), over 9000 degrees of freedom down to 2 (height, angle), 99% more predictability, independent DOFs.

So what’s next? Hmm, now that I have a grasp at how this works, maybe it’s time for a cutoff tool holder. I’m seriously getting fed up with wielding a hacksaw on my pretty finished parts, too.

Pop Quiz 2 Update 3, Ãœberclocker Update 2

May 24, 2008 in Bots, Pop Quiz 2, Project Build Reports, Überclocker

Due to finals week, I haven’t made much physical progress on either design since the last update. I did receive parts that I had ordered over the week and the previous week, however, and messed around with some of those. Past that, I’m in the process of filling out the internals for Ãœberclocker. Electronics are stupidly expensive, and who knows, I might get crazy and delusional enough to make my own controllers.

Also, the MITERS lathe has been under a rewiring job for the past few days. Nobody really notices it because it’s hiding behind the machine, but the 220v 3-phase outlet has been hanging by a loop of duct tape stuck around the conduit box channel, with exposed conductors and everything. Occasionally, a metal curl will fall between two live conductors and there would be a bright spark as it vaporizes. If you swept under the machine and jiggled the wires wrong, it would lose contact and the lathe wouldn’t run.

Not only was that a pain in the ass, but it was one hell of a fire hazard (along with everything else at MITERS). MIT Facilities has been working on rewiring that entire conduit as well as “updating” the ancient wiring of the lathe itself. It’ll probably be a little while.

In the last Pop Quiz episode, I discover that hard drive motors are all the same, and worse, all slightly short. So I took the opportunity to redesign the motor, not around the small stator, but…

…around two stacked stators. I also took the opportunity to simplify the motor design significantly, reducing the amount of weird internal boring I have to do with the not boring bar.

Last time, I couldn’t find a steel pipe or tube with the proper dimensions to make the magnet ring. After my last final exam finished, I headed directly to Central Machine Shop (what priorities…) and got some 1″ ID x 1.5″ OD pipe, which should cover many small motors to come.

I also got some more prospective parts in the mail.

The chassis will be made from the slabs of UHMW, which are 1/2″ x 6″ square. I will actually mill them down to 3/8: while making the internal features. At this point I’m not too sure on how to clamp a thin bit of plastic in a milling vise, especially since I’ll be cutting the center out of it. I might screw it to an aluminum or MDF fixture plate.

The rubber cord is 1/2″ diameter neoprene rubber. Little chunks of it will form the bot’s wheels. This is another machining quandary, since I’ll need to drill a hole down the middle as well as slice 5/16″ wide chunks off.

Hmm, I wonder if I can stick a razor blade on the toolpost?

In Ãœberclocker news, I received a prospective drive motor candidate.

I designed the bot around a 900RPM drill gearbox, which is quite uncommon compared to the 36:1 gearbox that dominates the ‘cheap drill’ market. There was Harbor Freight, but for $40 a drive motor I might as well DIY a gearbox anyway. A tour through Google Products led me to a bunch of obscure, sketchy-looking online vendors which sold drills as part of their inventory. I found a (suspiciously similar-looking) drill package for $25.

After executing the Standard Cheap Drill Dissection Protocol, here are the usual suspects. A 500-sized motor, a gearbox, ball-detent clutch, and a variable speed controller. The battery pack reveals 15 1,300mAh sub-C NiCd cells, also standard-issue. Nothing too unusual.

However, this is a legit 24:1 gearbox with a 4:1 (instead of 6:1) first stage, and a legit 18v motor on the end of it. I had heard a few stories of some of these drills actually being factory-overvolted 600RPM units. I wouldn’t put it past Chinese manufacuring to do that. Fortunately, this one’s for real. Here’s the link if anyone else is seeking a faster drill motor for drive.

Slightly overvolted to 24 volts should give ÃœC a speed in the 10MPH range.

The same day, I got some luck at MITERS while sorting through the increasingly large bin of random motors to categorize them (“big”, “small”, “stepper”, “not motors but round and shiny”)

2 half-processed 12 volt DeWalt drills. Someone had cut the handle off and removed the chuck and clutch assembly to convert it to a drive motor, but looks like they stopped midway. Saves me the effort of having to gut it myself. They are the newer, less-explored 3-speed type.

I duct taped the gearbox parts together so they don’t crumble. I accidentally let the internals slip out once, and it took 20 minutes to figure out how to reassemble it.

These gearboxes appear to have no mounting provisions whatsoever. The best way to mount them might just be to cast them in epoxy or urethane, something formable which I can make a normal-dimensioned mold for, if I end up using them. They would be great for the drive, as the arm solution is taken care of. On the high speed range, the bot should go over 9000 MPH.

They are 2″ in diameter, which makes the mounting game slightly more problematic as the bot is 2″ tall at the frame.

I put these back in the pile, since it’s unlikely someone will randomly come and steal them both.

Stay tuned for more! The summer build season is on, and I should be getting a pile of materials for ÃœC after the Memorial Day weekend.

Black metal

May 23, 2008 in MIT & Boston, Stuff

Only a little bit of it, but new in the Guan Post-Academic Year Collection is this finely crafted piece of machinery.

It’s some sort of air-bearing spindle assembly, probably meant to spin obscenely fast. Two pneumatic bearings support a very smoothly polished stainless steel shaft, on the ends of which are perched what suspiciously appear to be former brushless motor parts. There is a 8-magnet assembly at the far end wrapped in Kevlar, next to a hub-like piece. On the close end is four very large rare-earth magnet assemblies, each with 8 magnets, and curiously, each is offset by a few degrees. I suspect this was a stepper of some sort. These are very powerful magnets.

I happen to have a small compressor that I might fire this thing up with, just to see it spin, I suppose.

Then I can build a motor around the magnets and….