Cold Arbor Update 7

Cold Arbor is reaching that point in a build where another day of work will suddenly make a pile of parts appear cohesive and robot-like. Most of the fine details of the two linear actuators have been addressed, and I’m almost ready to move onto making the saw assembly itself.

Additionally, I’ve had the chance to cut out more parts and put the entire frame together.

Pictured here are most of the parts for the swinging saw assembly and the two clamping fingers. These almost go together as-is.

Back frame rail temporarily slipped together. Drive motors mount to the two projections, and the clamp actuator is (mostly) integral to the frame itself.

Here’s the front frame assembly in the “glob on as much brazing alloy as you possibly can” stage of fabrication. I heat up all the requisite areas of the metal and liberally distribute and wet the surface of the aluminum at the tabs with the zinc-aluminum braze.

The interior fillets on this piece are some of the best joints I’ve done so far. I was able to properly fill these joints because of…

…my +1 Frayed-Aircraft-Cable-And-Chunk-Of-Copper-Tubing Brush of Oxide Breaking!

Instead of ordering $9 stainless steel pencil brushes from McMaster, I chopped one up out of some stainless steel aircraft cable and the nearest small tubing I could find. In retrospect, making the handle out of copper, a highly heat-conductive material, was probably a bad move.

The long and thin pseudobristles allowed me to get the brush into narrow corners where my stainless steel toothbrushes could never hope to go.

The stage after globbing is sanding the proverbial daylights out of the part. Using a large and wide belt sander helps establish a new flat surface that somewhat resembles the old. This actually makes the whole part look really nice, almost like it was meant to look like that!

Here’s the backside in the globbing stage.

… and everything put together. Well, kind of squished together for the shot, that is. It looks great, but how well will it perform…

T-nuts installed into the thicker frame components. The waterjet cuts accurate enough such that these things either press in with thumb pressure, or, in the worst case, require a tap from a rubber mallet.

Mocking up the rear actuator. The bottom plate hole pattern mounts a Banebots 20:1 28mm gearmotor.

I bored, drilled and tapped one of the Surplus Center sprockets to fit the 1/2″ ACME leadscrew. The screw itself has two flats to let the set screws grip properly, and an end-threaded hole to act as a physical stop for the clamp.

I ran into a foreseen-but-ignored problem when making the output shaft for the saw actuator. 11 tooth #25 sprockets have a hub that is barely over 0.5″ diameter. The output shaft of a drill gearbox is typically 12mm, or 0.472″. I already sized the output bearings for 12mm.

So that means there’s no way to actually attach the sprocket to the shaft. Too little thickness to set screw or cross-pin, at least that I was comfortable with. While I could have turned the drill shaft down, this required either changing bearings in the actuator body or making some kind of adapting sleeve. I thought the 11 tooth D-bore motor sprocket (from a scooter motor) that I found would save the day, but alas, it was 10mm.

Naturally, I take the solution that would allow me to abuse machine tools: Make the sprocket hub bigger.

Uh oh.

I turned a steel ring that was fitted over the existing sprocket hub. This increased its diameter to around 1″.

Then I welded the ring to the sprocket on the exposed end. While all this was fixtured on the lathe chuck, of course. The sensitive machine surfaces were covered with a welding blanket first.

I could have done this on a less expensive or important fixture, but the steel ring bore was a hair too big to align properly without wobble. I used the machine spindle to correct the wobble, and decided to just weld it right there while everything was still squeezed together.

Mmm… porosity. After depositing the weld, I turned the surfaces clean.

You can tell I didn’t focus very hard on cleaning the sprocket surfaces beforehand. Oh well – this isn’t going into space.

And all ends well.

Not really a pretend-o-bot, but more parts are assembled. I got some 1.5″ long cap screws to close up the gearboxes, so now they are actually complete.  I still need to find a replacement 400-size motor that can stand 18 to 20 volts, however – the stock BB motors are only safe to run up to 12 volts or so.

Things left to do:

  • Finish the saw assembly, including all the random pins that attach things together.
  • Give Deathrunner some windings!
  • Front wheel hubs
  • Waterjet the last of the components – electronics mounting provisions in particular, and the top & bottom plates.
    • Design this stuff first.
  • Panic
  • Panic
  • Panic
  • Panic
    • Panic
  • Panic

Cold Arbor Update 6: More Drivetrain and the Saw Actuator

Hey, where the heck did January go? There’s only a week left!

That means there’s only 3 weeks left until Motorama 2010. Time to pump up the volume on Arbor work. While awaiting more waterjetting time, I finished most of the more complex machined parts for the robot.

I’ve grown into the habit of making my own linear actuators for the robots. Acme nuts and threaded rod are cheap on the surplus market, and I can tune the characteristics of the actuator to suit. I prefer linear actuators and linkages for moving robot assemblies around over direct torque couplings on the end of a gearbox, because it’s a bit easier to make a part strong in tension or compression rather than shear, and the leadscrew isolates the actuator motor from torque shock. Überclocker features a prominent exception to this because the clamp arm needs almost 150 degrees of travel, which is much more difficult to accomplish with a linkage and linear actuators.

Anyway, the saw arm of the robot only needs to travel about 60 degrees, which a single 3 link planar linkage can easily do. A clear image of the actuator is visible in this early design rendering of Arbor. It’s essentially a scaled up version of Clocker’s top clamp arm actuator – a leadscrew nut trapped between two thrust bearings.

To make the actuator body, I needed two big chunks of 2″ x 2.5″ rectangular aluminum bar. I had the bar, but both bandsaws at MITERS were simultaneously down.

Well that sucks. People need to stop accidentally cutting hardened steel on them or something.

The time of day I usually work on these things coincides with absolutely nothing else open on campus that houses a machine tool capable of cutting metals. So, I had to create a hackaround…

Bridgeport lovers and shop instructors avert thine eyes.

Wow, what the hell is that?

It’s a 6 inch milling cutter (with R8 arbor!) that a fellow MITERer picked up to cut a bunch of deep slits in steel. So, I’m not fundamentally doing anything worse, but it’s still one of those exercises that has the potential to destroy property and cause personal injury.

So, with the machine in low gear, a constant stream of Tap Magic, and everything cranked down as tight as I could manage, I plowed the cutter through just under 2 inches of aluminum leaving a roughly .02″ thick edge uncut. This was done mostly to keep the block from being pitched through a window as soon as it fell off the saw.  The whole process took a minute, and…

…left a brilliantly clean finish, almost fly-cutter-like in appearance. To remove the block, I just ripped it off with some vise grips.

Alright, so that was the fun part. Here’s a leap of faith and some finished actuator housings. No other special machining hacks were involved in the making of these parts.

Well, maybe one. Bridgeports have a Z-axis knee handle that can detach from the machine and be stored elsewhere. This is so you don’t accidentally run into it and get OSHA on your case, or move the Z-axis setting. Unfortunately, they have a bad habit of detaching themselves from their drive splines, especially when you’re trying to crank the Z as hard as you can. This has resulted in me clocking myself with the cast aluminum handle in the forehead at least once.

After having the handle fly off too many times, I finally got pissed off enough and put a shaft collar on the handle shaft so the stupid thing doesn’t come off. Ever. I don’t care if Bridgeport made them this way for a reason.

The first tool casualty in a long time comes in the form of me dropping a fully loaded boring head after finishing one of the actuator halves. It landed on the point of the tool, and so the entire tip of the boring bar cracked off.

Sad face. Time to start checking Ebay again?!

Fortunately, I was about to continue with a spare.

There’s something weird about the output gear in this actuator. It’s not a gear. It’s actually a #25 sprocket!

I found out that surplus chain and sprockets can literally be an order of magnitude cheaper than using spur gears. You can’t seem to find industrial 24 or 20 pitch metal spur gears for under $20 to $25 a pop. I was going to have to pay out almost $90 in spur gears alone for the two actuators in the robot.

But short runs of chain can perform the same duties. Surplus Center’s chain and sprocket selection was just too cheap to not explore these options. So, I decided to make the motor-to-leadscrew connection using chain instead. Enough sprockets and chain to build both actuators ran just under $11.

The 14 tooth sprocket shown here is squeeze-fitted onto a 1/2″-10 Acme nut, which, incidentally, is also Surplus Center hardware.There will be an 11 tooth sprocket mounted on the drill motor output shaft.

I heart Surplus Center.

The sprocket was actually once an independent power transmission component. To remove the “sprocket” part, I bored it to death on the the Old Mercedes. One cut at the diameter of the hub, and the outer ring with the teeth just pops off and lands on the tool.

While I had the machines still set up, I popped off these protoforms of the rear drive hubs. The flanges will have a 3 point bolt circle drilled into them later, and two flats will be machined near the retaining ring groove in order to make room for custom D-bore sprockets, just like on Überclocker.

Mounted on their respective gearboxen. The smallnubs will fit into bronze bushings in the side of the robot.

The Scene™ as of yesterday. Lots of things “almost, kind of, sort of” done, but not really.

Time to go catch up on waterjetting so I can continue building…