Sunday, December 20, 2009

Holding

I'm waiting on parts from The Surplus Center to get here for the upgraded drive system. Then I'll document the full process. I've been looking at Grant Imahara's (of Mythbuster fame)book about building combat robots: Kickin Bot. Dumb name, but the book is a great resource for an R2 builder. Very well written and researched. He gives all the necessary background and lots of clear explanations of the basics. You can pick it up cheap on Amazon.

Saturday, December 12, 2009

Drive System Upgrade

I've been researching and planning a new drive system for the R2. Right now, I've got some E150 Razor scooter motors in there, with 5" caster wheels. I used the chains and gears from the scooter. That system has gotten me around pretty well, but I think I pushed one of the motors too hard early on and its performance is pretty poor now. I can't make quick tight turns, especially with the left motor, and he won't spin in a small radius. Overall the motors just aren't that responsive and don't have a lot of torque. So it's a bit like driving an old worn out car around.

The motors of choice for builders in recent years have become the NPC 2212s. These are in the JAG drive systems, and they are what Mike Senna uses in his drive system (See the Senna files under Drive System in the Yahoo Groups page.)
The Surplus Center has some versions of these motors much cheaper. NPC wants $150, The Surplus Center has them for $50.

Several club members have investigated and here's the short story: the SC motors have a very short shaft on them, maybe 1/2". Apparently, however, it is possible to get this short shaft to work on a Senna style drive. That could save you another $200 overall. It is possible to use a press and some machine shop tools to make a new longer drive shaft, press out the old one, and put the new one in. But the consensus seems to be that this is difficult to do yourself without the right tools, and expensive if you get a shop to do it.

I bought a set of NPC motors. They are hard to get ahold of, and their website doesn't work. But if you call them persistently, they'll help. And they will give R2 Builders a discount to $135 if you mention the club, although this hasn't worked for everyone who tries. I got it and the motors look great.

I've seriously considered just building one of those two systems. Mike's has four wheel drive so that drive system is really rugged and doesn't run into any trouble getting across door thresholds or even going on grass. That's really appealing. I also found that if you get as many matching parts as you can from The Surplus Center you can save $200-$300 on the price that Mike projects in his Bill of Materials. I like the idea of building it up out of parts I collect or build more than buying a ready made system. Even if I had the cash to get the system already made, I wouldn't prefer that.

The only draw back that I can see, and this isn't much of one, for the 4 wheel system is that when you do tight turns, there's no differential gear to proportion the power to the different wheels, so one set of wheels will turn at a slightly wrong rate since they are tracking through a different radius circle, and they will chatter or slide laterally across the ground. I don't think it's a big problem, and as Victor Franco said to me it's a perfectly acceptable trade off for the killer drive system. I find that pretty compelling. And I may still go to a Senna type drive.

The JAG system has been very reliable, lots of builders use it, and it only has one wheel, so there's no chatter problem. And it has a belt drive. My experience with scooter motors is that the chains and sprockets are noisy. You can hear the chains spinning on the gear. So having a belt drive is pretty appealing too. I have spent a lot of time lately trying to learn about gear, sprockets, and timing pulleys and trying to decide which sort of system would be the best to build. Some of the good websites for parts for these things, besides The Surplus Center, are McMaster-Carr, and sdp-si.com. I have been searching through these sites and others looking for just the right timing gear pulley to do what I want, but I can't find them. A timing gear pulley will have grooves that accept a grooved drive belt. I'd like to use them on a single wheel drive system and get the best of both systems.

That search led me to the question of just what gears do I need to run for the R2. I don't know much about this aspect of engineering, so I've been trying to educate myself.

The NPC motors turn at 285 rpm at top speed. That's pretty fast. If you hooked those straight up to the wheels, it would be 4.2 mph. Here's how I figured that out, and I hope it's right. The standard wheel is 5" in diameter. Nothing much bigger will fit in the shells. So that times pi gives you 15.7" for the circumference. Turning at 285 rpms, means 4474 inches per minute. Multiplied by 60 gives you 268470 inches per hour. Divided by 12 inches per foot, equals 22372 feet per hour. And divide that by 5,280 feet in a mile, equals 4.2 mph. I figure a very fast walk is about 4 mph.

So that suggests that the speed needs to be geared down some. If there is a gear on the motor shaft and a gear on the wheel and a chain or belt connecting them, then the way to figure out the speed is a function of the ratio of the sizes of those gears. That is, if there is an 18 tooth gear on the motor and a 10 tooth gear on the wheel, then for every full revolution of the motor, the wheel will make 1.8 revolutions. This set up would speed up the wheel faster than the motor. What I need is to slow it down.

The JAG system has a 20/32 gearing. 20 on the motor and 32 on the wheel. So the wheel turns .625 of a full revolution for a turn of the motor. That makes 285 rpm into 178 rpm on the wheels at about 2.6 mph. The Senna drive runs at 18/11 for 2.6 mph, or 18/14 for 3.3 mph.

So I figure a speed somewhere in the range of 2.6 to 3.3 mph will be good. The problem for the timing gears and belts idea is that I can't find just the right size timing gears to give me that gearing. The limitation is the .5" axle. There's a .5" shaft on the NPC drive and I want to have one on my wheels too. But there's only a certain range of gears that are typically produced for a .5" bore. That range starts up high, so if I got them I'd have these huge honking gears on the wheel and motor.

And I think that as gear size goes up, the necessary torque to turn the gear increases. So stalls are more likely for the motor if you get close to its stall torque. But I'm no engineer. No doubt it would really help to talk to one at this point.

Long story short, I'd like to get the 30/22 or 18/14 gear ratio with gears and belts, but I can't find the stock parts for that. I could have some custom made up, but that would be a lot more money and time to wait. And I don't have enough confidence at this point to launch into that.

But there are cheap readily available chains and sprockets for these gear ratios at The Surplus Center. They have sprockets with keyed .5" shafts for $2 and $3 for example. That's a deal. I can afford to experiment a bit at those kinds of costs. They also have different size chains: #25, #35, #40, and so on. I think #35 will be suitable. That's cheap from them too.

So I think I can put a whole drive system including NPC motors together using chains and sprockets from SC for less than $500. I'll have an itemized list up here shortly. Compare that to $1000 or so for for the Senna or JAG systems.

I'm going to try to reuse the 5" caster wheels I have, and use the aluminum channel and sheet that I picked up here locally. The wheels were $13 a piece, and I have spent maybe $40 on aluminum. And then I'm going to buy parts from the Surplus Center and maybe McMaster Carr, and construct these myself. I'll document the build and give a parts list, prices and sources as I settle on final options.

Sunday, December 6, 2009

Legs Finished!

I did the final things on the list today to the legs. Yesterday I had dabbed a bit of touch up body filler--the Bondo kind in a tube, it's petroleum based, so it dries with air contact-into a lot of little flaws revealed by the first coat of primer. Today it just took a few minutes to sand those down flush with some 320 and 400 grit paper. Then I vacuumed the legs thoroughly and took them outside for another coat of primer. The temp was probably around 50 degrees, which is just at the bottom of the operating range for that spray paint. But primer is pretty forgiving. The second coat went on without any real problems and these are ready to go now:


This is the whole package for the builder who commissioned these from me. I'd probably put a coat of black paint on the hubs if they were going in my droid. I'm happy with the results:







The only remaining flaws at this point are very small and will be covered by paint. The next steps would be to add a few more coats of primer with some very light sanding--maybe with 00 or 0000 steel wool in between--then several light coats of a gloss white with some light sanding in between the early ones.

What did I learn this time through:
I got the high end baltic birch 12 ply plywood. It's very good, stable, and straight. But the outermost layer, the layer that will end up showing to the eye, is fragile. Home Depot and Lowe's usually have some good 12 ply cabinet grade plywood, and I have found that the outer layer on that stuff is a bit thicker and a bit more sturdy. I'd go with that next time. I'd also cut a 3" circle out of the insider layer of the legs at the top where the shoulder hub is going to sit, although the router method I used yesterday wasn't too bad. The router method for cutting the pockets out under the shoulders is a very clever idea I got from Mike Senna through Victor Franco. The best way was to have a couple, but not all, of the layers glued, and then rout out the pockets with a template from the inside. Each leg weighs several pounds. I think that next time, I could rout out a lot of material on the insides of the pieces without compromising the strength at all. The ankles could have a lot of interior material removed, the channel for the wiring could be wider, and the shoulders could be more hollow. That could save a pound or two per leg, and that would add up. The modification to the center leg tip to make a better connection to the foot shell is an improvement. Next time, if I was setting up a fixed three legged droid, I think I'll work out the plans to do that on all three legs--building the ankle locks into the design.

Saturday, December 5, 2009

Legs 9: Finishing touches

Several small things finished up in the last few days on the legs. First, I needed to cut the second 2.5+" pocket in the bottom of the shoulder hub pocket. After fooling around a bit, I remembered my technique from before. The easiest way to do this is to set a router bit down to the right depth plus a bit, and then just freehand the inner circle out. There's not much you can mess up, and the pocket won't show, so the freehand line won't be a problem.



I'm all about fast and efficient, and this was both. The results:



I think the plans call for 1.34" depth on the shoulder hub. I cut this to 1.4" or so. There will be some variation on the resin or the aluminum shoulder hubs that the club produces. I couldn't get mine off my droid to check. But I think that the resin hubs will fit, or if they don't, the backs can be trimmed and sanded a bit. And there should be clearance here for the aluminum hubs too. If not, a bit of dremeling could bring the levels down for a good fit. The hubs should fit flush to the face of the body of the leg.

Then I spent some time trimming, sanding and cleaning up the under shoulder detail pockets:



After painting, the wood grain and some chiseling will show inside of these if someone was really looking for it, but it would be hard to spot. I think if a builder wants the inside behind the greeblie to have a really clean look, a piece of styrene, maybe .020, cut to fit and glued back there would polish it just fine. That would take 5 minutes.

And now after some finish sanding down to 320 grit, and some filling in a few odd spots, these are ready for a coat of primer:



The primer makes any chips, dings, gouges, or gaps that I have missed so far immediately obvious. So I'll let this dry over night, fill some remaining spots, do a light finish sanding, and then one more coat of primer and these are ready to ship out.

I've completed that process on the center leg and they are ready to go out.





Any remaining wood grain will be covered by paint. At this point on my droid's legs, I layered on several coats of primer, sanding lightly between each coat with 320 or 400 grit, and then I built up several layers of gloss white, sanding between the early ones. The layers and sanding will progressively fill and cover any remaining micro wood grain texture to the point that the finished surface will be complete glassy and smooth. But I'm leaving that to the guys who bought these from me. Final steps tomorrow, and then shipping early next week.

Friday, December 4, 2009

Davis Holiday Parade

R2 got a place in the Davis Holiday parade this year. Lots of kids and floats and characters walk down through downtown to a tree lighting ceremony. I was foolish enough to just drive R2D2 in it instead of getting a ride on a trailer or float or something. That turned out to be a bad idea because for one thing it was dark and I couldn't see the street surface very well. So any potholes or cracks were real hazards. He hit one at full speed and really jammed the center leg wheels into it. The force was enough to torque the center foot shell and pop some paint off of it, but otherwise no real damage. It sounded awful when it happened. I was PAINFULLY aware that R2 has no shock absorbers or suspension system of any kind. So all that weight resting on the center leg creates an enormous blow to the frame when that front wheel jams. He made it, but I'm going to devise some kind of shock absorber for the front. Maybe a layer of some synthetic material between where the center leg mounts to the frame or something. But the parade was fun and a lot of kids were totally psyched to see R2D2 up close and personal:





The rest are too blurry to use. I guess the night street lighting screwed the camera up. One of the most entertaining things is the reaction from the adults. The first question everyone asks me is where or how I got the droid. I tell them that I built it (mostly) from scratch and they just won't believe me. Some of the just reject that as obviously impossible. I guess it would look that way from the outside. I have to say that there's no way I could have done it if it hadn't been for all of the other builders who figured this out before me in the club and created the plans and all. And I have to say that building something this complicated is really just a matter of breaking it all down into tiny digestible pieces and working on each of them one at a time. This time a year ago, I was completely absorbed in thinking about the cutting order for the angles on the booster covers I was building, or something like that. Little pieces start adding up to bigger pieces over time. If you think in the really long term, some really complicated projects are possible. That's just one of the valuable lessons that this project has taught me. Another one is: put the tools down, go away, and think about it for a while.

Legs 8

I'm working on the final list of things to do on the legs. Today I got the circular pockets cut for the shoulder hubs. The diameter of the hubs is 3.35" or so (I can't remember--check the plans), so I need a way to cut a clean perfect circle into the center of the shoulder. The best method I've found for small diameter circles like this is the router table. I've made an extra top for the table and after putting in a relatively sharp groove cutting bit, I've drilled out a pin at the exact radius away from the FAR SIDE of the router bit. If you measure to the near side of the bit, your hole will come out too big. Here's what I mean:


I only bring up the mistake because I did this once and ruined a pair of horseshoes that I had put a lot of work into.

I have a plunge attachment for this router but I don't like it. I just drill a small pilot hole into the center of the shoulder and then I work that hole down onto the pin sticking up in the picture. To start, I set the router height very low and turn on the router and ease the piece down onto it. Then I gently rotate the whole leg around that axis point at the pin and the bit does all of the work. The danger here is that a dull bit, or a mistake will rip out a nasty section in the beautiful perfect legs that I've worked on so hard. Routers operate at very high speeds so the slightest mistakes invariably spin into big gouges that tear across the face of your work piece. I did several test runs with this set up before actually putting one of the legs on here. Once the groove is started, I change the depth on the bit until it has cut to the 1/2" or so depth I'm shooting. This is a tricky operation, and the details of how you do it will vary depending on your tools. There are some very nice circle jigs for routers out there, but I don't think many of them will scale all the way down to a 3.35" diameter. Another option would be to use a collar and a circle template. Here's the results of the first circle that I get:

The cut is clean and perfectly round. So I'm happy. Once I got it set up, I ran all four legs through.

For the shoulder hubs to sit down into the legs, there needs to be another smaller circular pocket down inside of this one. Check the plans for the shoulder hub. You'll see that it has a stepped profile. But I've got an empty chamber in there that I left for the wiring. You can see that I already made a round hole in that chamber, but I wasn't planning ahead very well. The bit I used to round out the top of the hollow chamber in the legs was a 2 1/8" door knob hole cutter. But I need a 2.5" diameter hole here. So I've got a problem. Now that the plug is popped out, there's no material for the pin to ride in for the router jig:



That is, I set my router circle cutter to the 2.5" diameter and started cutting, but when I got down to this depth, the piece that the router pin was sticking into popped loose because there's nothing holding it in. In the picture you can see the diameter of the hole that I need. It's just a bit bigger than the one that is in there. I hope that makes sense. So what I'm going to do is put a 2.5" hole saw bit on the drill press. Then I'll get this piece centered just right and clamped down under the drill press bit and I'll cut the large inner circle down to the right depth that way. Then I think I'll be able to pop that plug out with a chisel. I'll show more pictures when I get to it. But here's a really valuable tip. When you are doing the initial cutting out and you get to the middle layer of plywood that has the channel down the middle for wiring:



Make this circle that's at the top of the channel 3" in diameter. That way you won't have to do this extra work when you go to cut the pockets for the shoulder hub. Make sense?

The one other thing I did yesterday was glue and clamp the curved ankle details to the outer legs:



This is a pretty easy step. I put a couple of pins into the leg and then drill the curved pieces to accept them. And when I say "pins," all I mean is that I am taking a small finish nail and cutting it off with some pliers, then I'm sticking that down into the work piece. The pins keep the work pieces from drifting out of alignment when they are under the clamping pressure. As before I put enough glue evenly spread around both gluing surfaces to ooze out of the sides all the way around. That was all the gaps will be filled for finishing and painting work and I'll have a good bond. Resist the temptation to clamp these too hard. The MDF is prone to get dents and it doesn't take much pressure to make a really strong bond.

Wednesday, December 2, 2009

Legs 8: Shoulder Hubs

I need to make a set of shoulder hubs. These pieces will fit in between the main body uprights in the frame and the legs. They amount to a cylinder that's 6.60" in diameter and 2.6" thick. They will also need bolts holes going through them so that the legs can be attached to the frame and a hole in the middle for wiring from the motors to be fed through. Construction here is pretty simple. I scrounged some scraps--enough to cut several 7 x 7" squares. Then I stacked them, put some short screws in the corners to align them, then glue, and clamps:



Once the glue had set up on these, I laid them out with the various measurements:


The layout is pretty simple despite the appearances here. I need reference lines at a 90 degree angle so that these can be lined up on the frame. I want the top dead center of the hubs to line up to top dead center on the frame.

But the out legs are canted back from vertical by 36 degrees. So I drew one cross on these hubs, then I drew another one tilted off of it at 36 degrees. Then I can just put the square bolt pattern onto that second tilted square. Next step is to rough the circles out on the band saw:


This step won't be easy to duplicate for people without the equipment. I've got a 220 volt heavy duty band saw that is designed for doing resawing, and cutting these corners off was pretty hard work. I probably wore out one of my last good blades on it. Lots of heat. There's almost 3 inches of solid plywood here with lots of glue layers. That's hell on woodworking equipment. But I got them off. I'm trying to think about other options for builders who don't have a suitable band saw. There's always hand sawing, although I try to avoid that when I can. You might be able to set the blade on a decent circular saw down 1.5" or so and then do multiple corner cuts on the top and bottom and just carve them off. Be careful and secure the work piece. The reason I'm intent on doing the rough cut is to make less work and less dust creation at the belt sander next.
I realized that I've been talking about the belt sander a lot and it probably sounds like I mean one of these:



(On a side note: to my horror, I found out that a lot of the storm troopers who are carving blanks of MDF to make molds for their storm trooper armor and just going to work on the MDF with one of these and carving away. But MDF has a number of chemicals in it that are carcinogenic and the dust is really bad news. If you use a belt sander on it without some protective gear and a dust collection system you'll get a lot of it in your lungs. I've made my eyes and throat pretty irritated with it too. Back to the topic:)

But I mean the stationary belt sander when I say that. Here's my rig:



I know I have digressed about tools here a bit, but it's necessary. This is another 220 volt machine with a 6 x 48" belt and adjustable angle table. With that much power, nothing bogs it down. And it's precise enough for me to do a lot of finish work quickly with it. Having good tools pays for itself over and over. This, like the band saw, is a Grizzly. I've had really good luck with them--cheap and high quality. Even with shipping, their prices are better than the main line brands you see.

So I put a rough belt on there--100 grit--to make the rough work go fast, and rounded these hubs out. A word about sandpaper grits. Get the right grit for the job and be willing to pitch the old stuff when it gets worn out. If you're doing big rough stuff, use a low grit. There is no sense in wasting hours and hours scrubbing away at some big job with a worn out piece of 220 grit. The new 100 grit belt on here took these hubs down to the circle line in about 10 minutes total or less for both of them.



Now I can drill out the holes for the leg bolts to go through and the center hole for the wiring.


The bolt holes are 1/2". The bolts are 3/8", so that allows 1/16" of wiggle on either side. And I just used a 3/4" paddle bit for the center hole. That could be bigger to save some weight--2 inches would be nice. But that was more work that I wanted to do on this.

These should line up nicely with the bolt holes on the legs:


Speaking of the legs. You can see here that I've started laying on some bondo and filler in the gaps, holes, and tear outs on the legs. Here I've gone back after the bondo has set and sanded it down flush with an orbital sander and 220 grit disks. That should solve the tear out problem I've been having with this fancy plywood.

Only a few steps remain before these legs are ready for primer (and shipping): 1) rout out the recessed circles in the middle of the shoulder for the shoulder hubs, 2) drill the axle holes in the tips of the outer legs, 3) attach the curved outer ankle pieces, 4) add touch up filler and finish sand the whole leg down to 320 or so.