I took R2D2 down to San Francisco last night for an event at the DNA Lounge called Barbots. Dave Caulkins put together a bunch of robot builders with entertaining drink delivery contraptions. R2 spent the night carousing with patrons. He was very popular. Chris James went with me and provided much needed droid wrangling. Thanks Chris. He also brought along his R2 drink tray (Return of the Jedi--Jabba's barge.) But we couldn't quite get it to work. Here are a couple of Chris' pictures:
The new drive system worked great. It's very zippy, maybe too much. After talking to Chris, some tweaks might be in order. He suggests that Dimension Engineering might be able to provide some ideas on how to curve out the performance specs on the Saber Tooth speed controller so he slows to stop and accelerates a b it more slowly. I think I'll pursue that. But I can also try some different gearing to bring down his top speed. I've got an 18 tooth gear on the wheel and a 14 on the motor now. If I dropped that down to a 13 or a twelve on the motor, I'd go from 3.3 mph, to 3.1 mph, to 2.8 mph respectively. I don't know enough about performance curves and engineering to know if this would actually change the way he starts and stops, or whether it would just cap his top speed. Seems like if there's a smaller gear on there, he won't jump off the line so quick. It will be pretty easy to check. I think I bought a couple of these other gears from The Surplus Center before, or I can order them for $2-3 a piece. The chains would need to be shortened a bit, but it wouldn't take more than an hour to do the switch given the way the drive is set up.
Several cool developments today. First, I finally got a technique figured out for bending expanded cell PVC board to the diameter curve I want. With a bunch of trial and error, I was able to heat up some squares of the board with a heat gun.
This is tricky because the heat gun doesn't have a wide spread, so you've got to keep turning the piece over and trying to get it all heated up evenly. It will warp and buckle as the front and backside expand at different rates, but once it's all hot, the piece will lay flat and it will be pretty floppy and soft. This is 1/8" pvc board. Thicker pieces would be harder to manage, I think. A lot of the Boba Fett and Jango Fett guys use this same technique to make armor pieces for their costumes. Once it's evenly heated up, then I quickly wrapped it around a piece of PVC pipe that's 2.85" O.D. or so.
I pulled the edges down around the pipe and let it cool off. The resulting piece has a nice even curve to it and it's 3 3/16" in diameter on the outside, which is exactly what I want.
The official size of the battery boxes is 3" wide. I want to go a little bit bigger than that to make the fit over the NPC motors easy, but not so big that it is noticeable. This should be just right. So with some cuts and cleaning up, I got these:
I cut a the side wall pieces and did some gluing and got these:
I should be able finish these up in a few days. Rather than tear my old hoses out of my original battery boxes and the knurled hose fittings, I order some new fittings from a guy on Ebay, and some hoses from Mike Velcheck. Both very cheap. That way I'll have another complete set of battery boxes for later.
I also worked on the drive system to tune it up today. I found the spec sheet on the Sabertooth 2 x 25 speed controller here at Dimension Engineering. And I set the dip switches for a linear throttle after following their guidelines. That seems to have smoothed out and improved the performance of my motors.
Then I took another look at the front foot shell wheels. I've always thought that having two caster wheels in there would be better. They would distribute the weight and keep the caster from crabbing if the shell gets out of level. And I want to make the ride as even and smooth as possible. I had two caster wheels that I picked up in the hardware store near Mike Senna's house from R2LA. But there's not enough room in there for both of them to swing freely all the way around. They intersect this much with the top one in the down position and the bottom one in the up position:
.
Now I've always assumed that this kind of configuration would be a no go. I assumed that if the two wheel intersected, then when you drive it around and made a turn, you could have them bind up on each other, or one would interfere with the turning of the other and they'd start sliding sideways. But I thought I'd just give it a try to see what happened since it was just a matter of screwing a few screws in to test. Much to my surprise, they worked fine. I ran the R2 forward and back, side to side, and tried everything I could to get them to bind up or interfere and there was no problem at all. First, whenever you change direction, both casters tend to pivot in unison immediately. You never seem to have one going one direction with the other turning the other way. Second, if they are colliding, they seem to be aligning and straightening out from it so fast it's not detectable. So I'm all set with a double wheel in front now.
I also took out some of the spacers that the wheels attach to--I think about 3/8". This had and interesting effect. First it lowered the front end a bit and helped to hide the front wheels, so that's good. But it looks like the drive motor shells in the back legs are a bit out of level--the front wheel is engaged and the rear wheel is engaged and touching, but not as heavily. This unweighting a bit of the rear wheel made the skittering problem stop in tight turns. But the wheels are still engaged and when they need to dig in, they do. Combine that improvement with the dip switch change on the speed controller and the handling now is just great. It's exactly what I wanted. He's responsive and smooth with lots of torque. So I'm psyched about the drive system.
Don't know if you can read it, but what I've tried to do here is import a spread sheet from Excel (as a jpeg) that gives the full list of all the parts I used on the new drive train and the costs. The total cost is around $535. That's compared to $1k or so for the JAG drives, or about the same for the Senna drive system. I was able to cut a lot of the costs by getting all the gears, chains, and other stuff for a fraction of the price at The Surplus Center. A scooter motor set up will cost you less--you could probably get it all done for under $200 if you had the tools and the determination to do it all yourself. Email me for more details if you need them.
I've also been looking at some of the documentation for the Sabertooth 2 x 25 speed controller I'm using from Dimension Engineering. It looks like you can get different sorts of response curves with different dip switch settings. So I'm going to tinker with those a bit and see if I can get those high torque motors to start up and stop a bit more gently. Anyone have any experience setting that up with NPC motors and the Sabertooth?
I've done a lot over the weekend. After the bench test of the first motor, I worked for a while on getting it mounted into the shell and onto the leg. That went pretty well--there were some hassles taking things apart and putting them back together to get everything just right. But it went in fine. I've got the R2 up on a custom dolly I built for him and laying down on the work bench. So this is looking up into the foot shells. You can see that the clearances are really tight. I had to trim down the overall length of the side plates a bit, probably because I used .25" thick stock to make the shells.
The next day I went to work assembling the other drive frame. Since I had worked out all the details on the first one, this one went to together much quicker.
You can see that I've modified the sidewall on the top plate here (the whole thing is upside down) so that the bracket for the motor will sit down flush to the side plate. This helped too because I needed to move the motor in tighter to the shell to get it to fit inside a battery box. I did some work on this one, put keys into the keyways, cut chains to fit on this side and the other side, and so on. Tricky work, but rewarding. I've really enjoyed this part of the build.
So this motor went into the shell and onto the leg pretty quickly. There was a lot of trimming and adjusting of the shells to get it all to sit right, but they went in.
All that crud on the foot shell is where the weathering ran down behind the battery boxes when I painted it on. Had a brief scare when I hooked these up at first. I didn't get one of these wire nuts on tight, and I had one of the motors wired to work the other direction from the other one, so the connection between the wires in the wire nut got really hot and burned up the wire nut. I haven't seen that before. But no harm was done. I replaced the nut and wired the motor in reverse of this one and they run in tandem now. You can see how fat and tall this thing is to fit inside the battery box. Here too:
I also discovered when I got these in that they lift the R2 a bit higher than my old ones. I think my old clearance was about .5" I did that deliberately, not wanting a lot of wheel exposed. But these show quite a bit:
In fact, I had to jack up my front caster with a piece of 3/4" plywood to get it flush. That's too much I think, but fixing it will be hard. There is room in the drive frames for the wheel to run further up. That is, there's .5" of clearance or so between the top of the wheel and the bottom of the frame top plate. So the wheel could be raised in the frame effectively lowering the R2 by the same amount. That would mean cutting 4 new drive frame side plates and redrilling all the holes in them. Then I'd have to take these completely apart and reassemble them with the new side plates. But making the new plates wouldn't be too hard.
So I did quite a bit of drive testing here. The motors are a dramatic difference. First, they have amazing amounts of torque. Nothing will stop them. I even was able to just drive him up the curb--this is ill advised, but I wanted to see if he could pull the load. No problem.
That was pretty cool. And the responsiveness is just great. You touch the throttle and he goes, pull to the left or right and he turns instantly and proportionally to the stick. With the scooter motors the response was awful and I just couldn't control what he did that well. I'd have to throw the stick over all the way to get him to turn and then over compensate to straighten him out. He also stops immediately now. Chris James had commented on this with his NPC motors. When you take the power off of these, the motor stop completely with no coasting. So his stops are kind of abrupt. It can be dealt with by driving him more gently with the radio. I'll need to get used to it. He also can drive backwards with great control now and turn on a dime.
I am seeing the skittering problem in really tight turns that people had warned about. With the four wheel drive system, when you do a tight turn, one of the wheels will not be tracking or turning at the right speed because the radius of the circle it is driving isn't the same as the radius for the other wheel. So it will break loose and jump sideways a bit causing some jerking. That's annoying. Not sure what I'd do about it. One possibility would be to cut up some new side plates for a one wheel system instead of the two wheel ones, and maybe put a small caster up front for stability, and raise the wheel in the frame to lower the R2. That would take care of a couple of problems. The result would be close to a JAG system, but with chains instead of belts. BTW, these motors and the chain/gear system are much quieter than the old scooter motors. The noise isn't an issue at all. I'm very happy about that. And with the excellent level of control I've got now, I won't worry about taking him into a crowded environment.
The battery box problem is next. I think I'll try making some slightly oversized boxes to save me all the trouble of hacking and sanding and modifying my old ones or the motors. My old boxes are a bit narrow. So I got a piece of 3" PVC pipe yesterday for a measly $2.50. It's 3.5" O.D. I think I'll cut it in half down the length and then take .25" of an inch off so that the overall width of the half piece is 3.25" or so. Then I'll build the battery boxes around that. I'll post pictures when I do it. I think I'll make the boxes a little taller than called for too so that the tops of the motors clear on the inside. They will be big, but I think it won't be noticeable to anyone, probably even builders. Should take me too long to whip those up, although I will have to paint them and then get the battery harness and the braided hoses and knurled hose fittings off of my old ones and transfer them.
Here a picture of Sam (my three year old) checking out the new rig:
And finally here's some video of the new motors in action. It's pretty hard to drive with the RC radio and take the movie at the same time, and I'm not used to the motors yet, but you get the idea.
Did a bunch of work on the drive system today. I finally had all the parts in hand, so I could move forward without stalling. Most of the day was spent modifying, cutting, trimming, drilling, and filing all the pieces to get them all to fit together right. I got the foot shells off of the R2 and checked the fit. It turns out that there's not much width in there, so I needed to reduce the width of the frames. That meant moving the side plates to the inside of the lip instead of the outside of the lip on the top plate. The wheels have wider hubs than the treads, so I used the band saw and sander to trim those down to have a narrower profile:
Once the wheels were trimmed down, I could get the axles and bushing in there and start cutting the chains and getting all the gears set up. This all went pretty well. It looks like I won't need an idle gear on this side. The chain is just right for snugness here.
Getting the other side with the motor bracket, gear, chain and motor all cut and sized right was tricky. I used some carriage head bolts on the inside there for a low profile. That way the bolt heads holding the bracket on don't interfere with the motor gear.
There are several little things to do here. The inserts need to be tightened up and secured with a small screw. The axles need to be cut to a miminal length. And some of the bolts haven't been put in there. It took me a little while to get everything lined up and get smooth rolling action on the wheels. But it seems to work well now.
Very beefy. Notice that I cut the lip off of the top plate on the frame so that the motor bracket would sit in there flush to the side plate. To get the motor to line up where it should go up inside the battery boxes, I needed to get it shifted in closer to the frame. More on that in a minute.
I still haven't cut the outside axle off here. And I used my old drive frames to map the holes on top where the threaded rods from the legs come down. So far it looks like all of this fits pretty well into the foot shells and onto the legs. Some other views:
I had to do some work on the foot shell for the bracket to fit through the hole into the battery box. Cut the holes much bigger.
The battery boxes are going to be a problem. With the bracket, the motors stick out a lot. The outside width of the motors is about 2.5 inches. My battery boxes are made with PVC pipe that's about 2.9 inches wide on the outside. So the interior space with .25" side walls is 2.4. That is, the motors won't really fit inside the battery boxes. For now I'm going to get these motors all assembled and installed without the battery boxes. I can think of a couple of possible fixes for the battery boxes. 1) I could cut out the whole back side of these, and then add a pieces of 1/8" PVC cell board across the back to increase the interior volume. Then I could build up the curved sections to blend in with the box with body putty or something. That would take quite a bit of work. 2) I could make some new battery boxes with the next size up of PVC pipe. The problem is that the next size up is 3.5" O.D. The plans call for the battery boxes to be 3" wide exactly, so that would be a pretty big fudge. You'd be able to see how fat those were right away. But the inside diameter would then be 3.042" which would be plenty to house the NPC motors. I'll think about it some more. I don't think this will be too hard to solve, but I will probable have to effectively rebuild my battery boxes. Fortunately, since I have been selling those, I have several of them around and I can build them quickly.
And here's a video of the bench test of the motor all mounting into the drive system. Works well. I am going to try to take the half link out of the chain on the motor gear.
I snatched a bit of time today to do some work. Almost all of the my supplies have arrived now. On the Senna design, the wheel is fixed to the axle and the gears are fixed to the axle to drive the wheel. So the connection between the axle and the frame needs to tolerate some metal on metal friction or have a bearing or something. I have these two options:
Those are bearings on the left with a .5" inside bore and an outside diameter of 1.25" or so (Can't remember.) I picked them up at a local surplus yard (Blue Collar Supply) for $.50 each. On the right are some impregnated bushings from McMaster-Carr. I think these are the same ones Mike Senna recommends. The inside bore is .5" and the outside diameter is .75" I think I'd prefer to have the axles riding in a bearing that could turn. But I decided that I would start with the smaller ones. That way I can bore out the frame plates to fit, and if I don't like them, I can go bigger to the bearings. Plus, I don't have an effective way to cut/drill an accurate hole for the bearings to mount in.
So I set up a jig on my drill press that would let me drill all eight holes in the side plates as accurately and consistently as possible. It's important to get the holes aligned across the frames exactly right, or the wheels will sit crooked and not track right. Earlier, after I rough cut all the trapezoidal side plates, I clamped them all together and sanded all the edges on the belt sander. That got all four side clean, and it got the size off all of them as close as I could get them. That helps for cuts and drilling later. It's always worth it when you're creating multiple pieces of something to set up a jig or fence or some other method to make all the cuts as uniform as possible. That makes fitting and every step thereafter much easier and it will produce more accurate results. Here's the jig on the drill press. Just a couple of blocks clamped down to hold the plates in exactly the right place:
I'm using the step drill bit here because it's so effective at drilling these larger holes. This way, the steps each just have to take out 1/8" of material or so at a time instead of cutting the hole thing at once. I just discovered these recently and I use them all the time now. Don't buy one of the outrageously expensive ones at a tool store--$50-$75!! Harbor Freight has always got some for a few bucks. Get a couple and when it wears out, throw it out.
Drilling here gave me some trouble. The bit was binding and heating up when I got down deep. I have heard people say that you need to lubricate metal when you are cutting it so that the blade or bit slides. It reduces friction, and the cutting edge does all the work. I was skeptical about this, but I tried it here on the bit and it worked great. The bit just slide right through these pieces once I had a good shot of WD-40 on the bit and the piece. And it didn't catch on fire, like I thought it would. Valuable lesson about working with metal.
That gave me some clean, uniform holes. I think they were no more than a millimeter off across the different pieces so the jig worked pretty well:
A bit more grinding on the inside of these holes with the dremel tool, and they were ready to take the bushings for the axles.
In order to get the fat wheels into the frames, it looks like I am going to have to assemble the side plates on the outside of the lip on the top channel.
I'll have to countersink the holes so those bolt heads sit down flush. But you can see it come together here. That'll be a quick fix when I get some more time. And I'll need to cut a slot down the top of the top plate to receive the threaded rods coming down from the legs. More on that later.
Next I rough cut a couple of six inch pieces of axle. That should be enough to accommodate the gears, the bushings, and the axle collars. There's a bit of loose space inside between the wheels and the bushings here. But I think a single .5" washer on either side of the wheel will be just right.
Next up, I'll get the gears and motor brackets all figured out and start assembling the rest of the pieces. The plan is to make this a four wheel drive system. So the motors will drive one wheel, and then that wheel will drive the other wheel with a gear and chain. Check the pictures of his system in the Yahoo Group for comparison. More soon.
I've had several false starts on the drive system upgrade. At one point I thought I would be able to attach a gear to my old wheels and use my old drive frames and add the NPC motors in, but after I got the parts from The Surplus Center I could see that that wouldn't work. So I reconsidered.
I've been really impressed with Mike Senna's drive system ever since he developed it. So I've resolved to give it a go with some modifications. I searched around Sacramento a bit for wheels and found a place that
sold me these:
The one's that Mike lists in his plan are also Colson caster wheels, but his are 5" x 1.5". My distributor didn't have those, but he had these for about $10 a piece. They are 5" x 2". What that actually means is that the hubs are 2" across, but the treads are about 1.5". And I think the tread width on the 5" x 1.5" is 1.25". So mine are .25" wider. I've had a picture in my head since I started this project of wide fat wheels with lots of torque--or even tank treads--that would conquer any terrain. So I opted for the wide ones here. We'll see if that's realistic. The wide wheels may cause trouble in tight turns. The wide wheels won't work on Senna's plans, however, because he's using a piece of rectangular tubular aluminum to fashion the frames, and I think the width on those is 2" on the outside, and .125 or .187 on the side walls. So there's only about 1.75" clearance on the inside. But I have a solution for that in a minute.
I had some 1/8" aluminum plate and some 3" wide channel left over from before. So I have cut a couple of pieces of the channel. I think the lip on these is 1". Then I have drilled and counter sunk holes for bolts on both sides.
Next I have cut 4 trapezoidal side plates. I'm following the Senna planes on outside dimensions. (I've mentioned it before, but the Senna plans are part of the "Files" attachments to the R2 Builders Yahoo group. Go down to Senna files, then look in the Senna Drivetrain folder.) The side plates are going to bolt onto the channel top plates above to make the frame.
Even after building my first set of drives, I had apprehension about cutting all of this aluminum. I got tendonitis from the hacksaw before. So I talked to a friend at a machine shop and he said to use a fine toothed jigsaw blade in my jigsaw that is intended for soft metals. I was very skeptical. But h was right; the jigsaw cuts this 1/8" plate just like wood. Very easy to work with.
So I clamped these side plates into position on the U channel to mark and drill the holes for mounting in exactly the right spot, drilled them out, and then the bolts went through no problem. My bolts were a bit long and protruding into the space inside, so I trimmed them off with the Dremel tool.
Here's the result. I've put the plates on the inside but they can be easily be mounted on the outside to make a little more room. So there's a bit more versatility with the interior width here than working with the rectangular tubing. The drawback is that the bolts can work lose. So I'm going to add some locktite to them on final assembly or superglue. When the rest of my materials get shipped here next week, I'll start laying out the axle locations, add some bolts across the bottom for bracing, and so on. More details then. I'm excited to get those NPC motors in there.
Got it in my head to build an R2D2--I've loved Star Wars since it came out and I''ve got a D-I-Y maker ethic, so it was a pretty obvious project. This blog documents the whole build, and now that the R2 is done, it includes public appearances, parts construction, new projects, modifications, and tweaks. I learned a lot from the builders who did this before me. I hope this blog does the same for others who start the project.
