Aluminipede, a small walking machine

[rodiponer]

Hi Le Chat,
Thanks for checking in. I really enjoy reading your updates about The Contraption.

I am currently assembling the chassis cube. It's neat to see something come to life out of a computer rendering. The math said the chassis would be light, but it's a lot different to lift it and see.

I am almost finish riveting the top of the cube, and will then work on the bottom, and then the leg hanging structure. Even with the kids 'helping' me, I should be finished with the chassis structure next week. I am trying to track down some perforated aluminum sheets locally, for the skin. I found it online but it is unreasonably expensive. I have a lead in LA, only a two hour drive from here.

I am thinking of sliding some kind of light tubes into each of the 12 edges of the cube. I have to order these soon to have them on before I add the skin. I think it would look cool to have the machine glow from within and have the edges lit like that.

It is easier to drill straight than I thought, so I have switched to clamping the tubes on and then drilling through the water jet cut plates. I need to find a drill with a body that is as narrow as the chuck. The one I have is fatter than the chuck and won't allow me to drill the inner rivet holes straight.

Is there a trick to keep the drill bit from slipping in the chuck? Mine tends to slip no matter how hard I crank down the chuck. I am only drilling through 1/16th inch thick aluminum so I don't feel like I am reaching the limit of what the chuck can do.

[masho]

loving this. Thanks for keeping us posted.

[Ugly Dougly]

2 cute.

[jella]

It's ALIVE...ok well almost

[LeChatNoir]

Too cool, rodiponer.

The progress is looking good and the help seems to be putting their back into it. Or maybe putting it into your back. Regarding slipping chucks, I’ve never had one that didn’t slip at some point, even on small holes. You might try to find a hex shank drill in the size you need. The hex shank is usually soldered onto the drill. It will positively engage with the chuck jaws and not slip. Most hardware stores have them in the drill section. Beyond that, it could be that the drill is getting too aggressive of a bite. If you know how to sharpen them on a grinder you could try making the back angle more slight, but beyond that maybe just lighter pressure. As to the width of the drill body, some right angle drills are narrower up at the chuck, but they’re pricey.

[Token]

Try IMS for metal supplies.

The Los Angeles warehouse is always well stocked.

They have a SD location as well.

http://www.imsmetals.com/

[Snow]

Try to find a drill with a REAL chuck, one with a key. Fuck them keyless chucks, they are complete crap and ALWAYS loosen up and slip. Air drills typically have a body not much larger than the chuck and they have keys.

[rodiponer]

The chassis frame is mostly finished. I ran out of argon and have to wait until Tuesday to finish a few more welds.

Thanks for the tips, guys, on drills and metal suppliers. I purchased an air drill, which would have been perfect, except that my little air compressor could only power it for about 10 seconds every three minutes. So I found a cheap electric drill and then ground off the 'extra' bits on the plastic case so that it would fit in the tight spaces.

The frame is light and solid. I can lift it with one hand. There are only a few places that bend if you step on it, and it will be completely stiff once the perforated aluminum skin is attached. That should arrive next week-- Token, you are right, IMS has perforated aluminum. The first sales guy I called there said they don't stock it, but then I saw it in their retail store when I was picking up bits to finish the chassis...

I am now researching motors. I want to order the motors, controller, and remote control system this week. I've been reading BattleBot and Robot sites. They have nice motor controllers that interface directly with off the shelf RC transmitters, and have safeties built in-- like gradual accelerations and cutting off the motors if the radio is unreliable.

Do you guys know anything about DC motors? At the 4mph target speed I will need the crankshaft to rotate at 175 RPM. I think 500 watts is more than enough power to make this thing go. Should I size the reduction ratio so that the motor is turning at maximum RPM at the target speed, or is it better to have an xx% reserve in there? I don't want to overheat the motor by running it at it's maximum RPM if it's better to have it a little derated, but then I think I might lose torque unnecessarily if I decrease the reduction ratio of the gear box.

I want this to be energy efficient since I hope to power it with about 100 pounds of AGM batteries, and to recharge it with 270 - 405 watts of solar panels. Some of the most energy efficient motors I've found are a little oversized-- 750 watts-- do you guys know if they are less efficient to use a larger motor at a lower output (though still spinning it at it's maximum RPM)?


The color changing LED strips look great. The light reflects off of the inside of the structure in a cool way, even though the LEDs only face one direction. My father in law is going to chop the 15 meter spools into 48" segments and soldier lead wires onto them. Each edge of the cube will then glow and pulse.

I am now looking for LED tail light/head lights. The machine is symmetrical, and so the front is whatever direction it happens to be going in. If it reverses then the front is on the other side. So I want to find or make little light fixtures that will change from headlights to tail lights, depending on which direction the machine is going. I'm thinking of cutting 2" square tubing into 1" lengths, placing some kind of bright RGB LEDs or red and white LEDs in them, then covering the open end with a little 2" square piece of lexan. Do you guys have any suggestions for LEDs that are bright enough to work as headlights like this? There would be four on the front (one on each corner of the square), and four on the back, so they would need to be as bright as a flashlight, not a car headlight.

[jella]

http://www.goldengadgets.com/led-lights ... light.html

Very cool...I can't wait to see it walk across the Playa

This might be what your looking for

[Token]

DC Motors 101:

Go read these fine pages at MIT. They will give you the basics:

http://lancet.mit.edu/motors/index.html

It is a quick read and a good foundation for devising a plan for your gizmo.

Here are the two key relevant graphs:



The torque vs. RPM is the graph that tells us where the sweet spot will be in terms of the power output on the motor shaft.

Key to note that maximum shaft torque is at zero RPM.

Also of interest is that at maximum RPM, the shaft is essentially zero torque, i.e. power is near zero and all torque is used to overcome internal motor friction.

You do not want to run a DC motor at maximum RPM! No power there whatsoever.



This picture tells you where the real power is produced. Power is essentially the product of the torque and rpm. Note that the "power band" is in the middle of the RPM range.

For your application, the motor will run in the 0 to X RPM range, where X is likely somewhere in the 50% to 75% of the RPM range.

The big question will be the estimate of how much power you need in order to meet your vehicle top speed requirements.

Given that at this point in our build you do not have hard data on what the friction losses will be in the considerable number of mechanical links in the Aluminiped, going big on the motor makes sense.

With the power curve in mind, if you make your gearing field adjustable, you can gear down a big motor just fine at the expense of battery life. If however you have a motor that cant handle the load, there is nothing that can make it better.

It would be nice if you can find a motor manufacturer that makes suitable motors with identical interface mechanical parameters in multiple sizes. This way you can upgrade or downgrade motors as needed once the beast is built.

[Token]

High Power LED:

For the headlight/taillight question, there are two manufacturers that are at the cutting edge of the technology:

Phillips Luxeon

Which is used by Audi for their automobile DRL application.

and

Cree

Cree usually provides a more affordable alternative.

There are of course cheap Chinese knock-offs of the above designs which are just as good.

The problem with high power LEDs is that they are often near 180 degree light pattern. Collimation lenses are often recommended for headlight applications.

I have personally worked with the older 5W Luxeon Star design and it is very bright. A single LED provides all the light we need a camp. I do overdrive it at 9V but I keep the current within the design limit which is critical for LED lights. Current kills LED lights. Keep that in mind.

[rodiponer]

Thanks Token, that was a good start for my four day bender of reading about motors, motor controllers, sprockets, chains, gear ratios, machining shaft keys, and etc… And combining all of that into a huge spreadsheet to find The Motor. There are a lot of very nice motors out there with some kind of odd-ball shaft that I can't find sprockets for. Or motors that were perfect except for being 12" long when I have 6" of space, so I wasted a day trying to figure out how to make a 90 degree angle transmission, or find one for less than $100 (that didn't happen).

The good news is that when I got sick of that I could go out to the garage and weld. Welding is so relaxing, if I ever make something again I want it to involve more welding and less thinking. I am just starting to get that 'stack of dimes' look that people think is pretty, and the inside corner penetration is much better than it was in the past. I am starting to really like aluminum.


And, after all that learning and reading, the current plan is now:
Motor: Ampflow E-150, 0.5hp 4700 RPM, 2.2 pound/foot torque motor ($79, but has a bigger brother that will fit in the same bracket for $299 that is 5x more powerful, in case it doesn't work)
Motor Sprocket: 9 tooth, 1.3" diameter, #35 chain
Crankshaft Sprocket: 114 tooth, 13.8" diameter
Gear ratio: 12.6:1
Torque at crankshaft: 28 foot-pounds ("big brother" would be ~130 foot-pounds)
Motor RPM at 4mph: 2200
Chain speed at 4mph: 700 FPM (within the 600-800 guideline)
Center distance between sprockets: 13", chain length: 53" (within the 30-50 times pitch length guideline)
Wrap of chain around small sprocket: 122 degrees (120 is the minimum)
Motor Controller: Roboteq AX2550 80amp continuous dual channel motor controller
Remote Control: Spektrum DX3E 3 channel RC car controller
….

Success will be if the machine crawls forward at 1/2 mph. The 4mph target would be icing. I just don't want the machine to need a push start or to not budge an inch while the motors hum.

The metal supplier doesn't cut perforated aluminum. Does it mess up their shears? So I am thinking of using electric shears from Harbor Freight, the kind that look like an electric drill with snips on the front:
http://www.harborfreight.com/14-gauge-s ... 92115.html

I've never used those things before, but saw a Youtube video of them cutting and think they might work well going through all the holes. Do you think it will work? I am thinking of those things over a circular saw because I am out of room in the garage, and that looks like something I can lay a 4x12' long sheet on the grass in the yard and cut up without making a huge mess.

And now I am looking at trailers. I want to find one in the next two weeks, so that I can bring this machine to my collaborators house in Los Angeles. He is helping me with the electrical parts of this machine, and I have a mean plan to park this thing in his driveway for two weeks while my wife and I head out of the country for a wedding. This has become another huge page in a spreadsheet, since the machine is 62x96x92", which is a pretty wide to get through an enclosed trailer's door, even if I put it through sideways. So we are looking at making the legs fold in somehow for transport, so that the machine is 73" long instead, to fit sideways in a 7' wide trailer, as long as the door is wide enough… If I can't figure this out I am half tempted to just go with a small flat bed trailer, though then this thing will sit outside under a tarp for 11 months of the year on my friends farm, instead of completely inside a little trailer. The 8' wide trailers look like they might be harder to drive, and mostly come in very long sizes (14' and above). I've never towed a trailer, but the 7'x12' trailers I've seen in the neighborhood seem more human sized than an 8x14 or 8x16. Do you guys think there is much of a difference in handling the larger trailers?

My bicycle chain is horribly rusted from the Playa. On McMaster I see chain that is 'Corrosion Resistant Treated Carbon Steel' -- is that good enough or should I look at going up to stainless steel chain?

[rodiponer]

I'm very happy with the electric metal shears. They are easy to control and make clean edges. Not bad at all for a $40 tool.




This weekend I will be cutting the sheet metal to skin the machine. I am really happy with how it looks in the test fits.



I refilled my bottle of argon four days ago and I am almost all the way through it again. I keep running out on weekends, so I guess I need a second bottle.

The motors and electronic bits will arrive next week. I want to make some boxes out of the left over aluminum sheets to keep the dust off of the terminals and etc. The electronics enclosures on McMaster Carr seem way over priced, and I can't use the waterproof plastic boxes at Fry's since the motor controller has to dissipate heat. Instead of using sealed connectors for each wire, I am thinking of welding a ~2" long tube to the bottom of each box. The wires will then go through that, and on the playa I can then stuff something into the tube to make it reasonable dust tight. Think that'll work?

[LeChatNoir]

I love the progress. Just zipping through tonight, but will give more feedback later. I like the spirals that those types of shears make. Nice and consistent.

[masho]

looking good

[LeChatNoir]

Re: Trailers

We haul our camp infrastructure and The Contraption in a 7’ x 16’ enclosed trailer. I’ve pulled multiple types and have found the easiest to pull are ones with tandem axles. They seem to track better. The single axles, especially if loaded too heavy, can really sway around. One year we pulled a small single axle and it would really get buffed about just after passing semi as the crosswind hit it again. You just learned to really brace for it each time and it was nerve wracking at some points. Then pulling the enclosed tandem axle some years later, it didn’t really have much of that problem at all. Maybe slightly on I-80 in Wyoming where it can really blow… more like you felt the pressure but it didn’t sway. It tracks well and feels like it is a rigid part of the vehicle.

I built a flatbed trailer years ago and bought the enclosed one in ‘08. I’ve not used the flatbed much since. Stuff stays dry and secure and I can unload the next day after getting home late from a job. I love it.

I did not buy the 8’ wide model since the 7’ was wide enough for what I needed. Besides, the width seemed like a lot to tow to me, even though the axle width may well be the same. It just seemed like a lot to watch in my mirrors. You’ll have to weigh it all out compared to your needs and resources, but that’s my two cents.

Re: bike chain. My opinion is that only the stainless will hold up without rust. I’m amazed at how quickly the playa can rust uncoated steel. That being said, I’d still use the regular bike chain and just replace it every year if it gets too nasty. Regardless of what it’s made of, it’s gonna accumulate playa dust.

Check this link here for 90 degree gear boxes.

[Token]

The electical bits you have chosen are very nice. The Controller is a great piece of engineering.

The Roboteq looks like a sealed unit with the powerFETs on side heatsinks. You should be able to seal the edges where the I/O plugs in without killing the cooling.

I've used liquid masking film available in most hobby stores to seal electronics stuff in my RC car days. It applies like nail polish but cures to a thin rubbery film that seals well but can be removed mechanically with a fingernail and some pulling. Hit the seams on the box with this stuff and around the I/O port once you work out the harness.

Since you are using RC for control you need to register your freequency with the Borg.

RC Hobbyists

If you are planning to bring any sort of radio controlled model, vehicle, or art installation to BRC this year, please contact rchobbyist (at) burningman (dot) com with the radio frequency and channel number under which you will be operating. Radio channel conflicts must be avoided to prevent loss of control of an RC project. An out of control model/vehicle/art piece can cause serious damage to people or property, so it is important that RC'ers help us organize regarding radio use. If you have an RC art piece, please inform the Art Team, art (at) burningman (dot) com, or contact a team member at the Artery in Center Camp.

Since the Roboteq supports it, might be good to configure analog joysticks as input. You could use two sticks like old school tanks or figure out the channel mixing for single stick control.

I would love to see your project when it gets to Los Angeles and assist if needed.

[rodiponer]

Cool, Token, I'll shoot you a PM when we move this thing to LA for the wiring.

I am just talking with my father-in-law (who is doing the electrical part of this machine) about heat on the controller box. He also thinks it should be externally mounted. I prefer putting it inside a diecast aluminum waterproof/dustproof enclosure, like this:
http://www.budind.com/view.php?part=cn

And using thermal grease between the controller and the aluminum box, then mounting the box to a thick aluminum plate on the chassis of the vehicle. If that's not enough, welding little cooling fins onto the diecast box. I don't know if that will be enough. But the controller box is the single most expensive part of the whole machine and I hate to get it crapped up.

I was checking my RPM and gearbox math and learned that I was totally wrong. I need the shaft output speed to be 60 RPM at 4mph, or 70:1 gear ratio instead of 12:1. So I need to either return the motors I have and exchange them for wheelchair motors, or add a gearbox, or add another chain and two sprockets to further reduce the RPM.

LeChat-- the gearboxes you linked are great. They look perfect and are the right price. But they are rated for 1750 RPM-- how much do you think I would kill them by running them at 4000 RPM? Granted, this is a mutant vehicle, so it's not 4000 continuous-- probably more like half an hour here or there with cool down periods between runs. Do you have any experience with this? I am very happy if I am cutting a 20,000 hour service life down to 1,000 hours. That's still more than I will use the vehicle. But I don't want to do something that would break it in the first 20 hours of use.

Also, thanks for the advice about dual axle trailers. I was favoring a single axle since this machine will only weigh a few hundred pounds, and thought that a lighter trailer would be easier to handle than a heavier two axle one, but we already brace ourselves in the pop-top-camper-van when we pass a truck, and I don't want to add even more wind sway to the vehicle. Our every day car is a little compact car, so the van already seems a bit unstable to us.

I found a video of people who made a walking machine with legs that appear to be about as big as mine. It's great to see it working, I just love the motion of this leg mechanism, but I am also a little concerned that they have a wheel underneath it. There is a longer video that shows the construction of the video and more of it walking, and it looks like they had some kind of binding issue-- they had to lift the machine up and rotate the legs to a specific point and then set them back onto the ground, or kick the legs to get them to start. Maybe they just lacked torque, or I can't even tell if they used bearings in the joints...
[youtube][/youtube]

So my plan is still, if the legs don't work, to flip this thing upside down and let the legs wave in the air like some kind of jelly fish monster.

[rodiponer]

Since the Roboteq supports it, might be good to configure analog joysticks as input. You could use two sticks like old school tanks or figure out the channel mixing for single stick control.

Thanks, I wrote to the BORG rchobbyist email address. I got the futuristic spread spectrum frequency hopping remote control thinking it would be very resistant to interference. I even watched a YouTube video with some RC airplane guy who had a signal generator making RF noise on it's frequencies and he was not able to interfere with it's link. He was able to confuse other brands of radios. Not complete proof that it will work, but it gives me hope that the machine won't freak out if a Tesla coil or worn out alternator is around. It is, after all, a 250 pound machine. We are also wiring up e-stop buttons to it.

I would like to wire up two analog joysticks. The controller does the differential steering thing, so a single joystick would also work. I added the remote control so that the driver can also be the side walker. And because I am not sure how much weight this thing can take, so it may just be a weird sculpture that follows us around rather than a conveyance. Or something just for my 60 pound daughter (who has, happily, recently gotten much better at walking and is almost completely out of her wheelchair. I think she could walk to Center Camp from Kidsville, but getting as far as The Man is still way too far, for this year, at least). But yes, if this thing can actually take weight, then onboard controls would be nice.

[LeChatNoir]

LeChat-- the gearboxes you linked are great. They look perfect and are the right price. But they are rated for 1750 RPM-- how much do you think I would kill them by running them at 4000 RPM?

I don't know, really. I suspect they're over built somewhat, so they could probably take more than they're rated for, both in Hp and rpm, but I've no idea how much. If you have the space to do it, I think I"d try for stepping it down closer to 1750 with a sprocket. Better to make it easier on the machine than get all the way out there to have one of the gearboxes crap out on you after moving a few feet.

I've become a reluctant pro at the stop/backup/reconfigure thing.

[unjonharley]

Any of you see the full sized metal elephant in 08.. It walk about one block then locked up.. It took the DPW's largest fork lift to take it back to camp..Did it make it to 09?

[unjonharley]

A few years ago a ant machine showed up at the Or. kenetic races.. It broke down at the parade staging.. The welding wouldn't hold. The following year it showed up in two units. This time with wheels and legs..

[rodiponer]

I need sheet metal advice. Do you guys know how to attach it to something so that it is completely flat and not warbly or like a pop-bead, where it will pop back and forth if you push on it?

For my first attempt I clamped the sheet aluminum, tack welded it every few inches all around the outside edge, and then welded it in short 3-6" lengths. Then I'd let it cool down while welding some other random short part, until all of the short welds were joined together. I ended up with a bad case of the warbly pop-beads.

I am thinking of just pop-riveting the sheets on instead of welding them.


I made a major welding improvement-- I got those funny leather sleeves and thicker TIG gloves. Now that I don't have to worry about being shocked through the thin Tillman gloves and cotton shirt, I can rest my body more on the work, be extra steady to make prettier welds, and the whole experience is so much more relaxing.

I exchanged a few emails with the gentleman who made the Cortege Walker (the video above). He says the wheel was for steering in the narrow streets, and not necessary for the legs to work. He also says that the legs would only bind when he went around tight corners, since they would twist and overlap each other. So that is good news.

[LeChatNoir]

Rivets or machine screws. Or...

You could try clamping the entire edge you are welding with a piece of angle iron, sandwiching the sheet between the tubing and the angle. Tack each end, then the middle, then in the middle of those, then in the middle of those, until you get the spacing you desire between your welds.

[unjonharley]

Rivets or machine screws. Or...

You could try clamping the entire edge you are welding with a piece of angle iron, sandwiching the sheet between the tubing and the angle. Tack each end, then the middle, then in the middle of those, then in the middle of those, until you get the spacing you desire between your welds.

When making up screen doors or recovering furniture I tack in the middle first, Then each end. A lite tac in the middle so it can be removed if needed to ajust slack..Then in the middle and middle..

[LeChatNoir]

Yes! I stand corrected. Tacking the middle first allows any slack to go to the ends, instead of being bunched up in the middle.

[Token]

SYN

SYN-ACK

ACK

ARP

ECHO

PING

Can't let this thread linger on page two for too long.

[Sail Man]

So my plan is still, if the legs don't work, to flip this thing upside down and let the legs wave in the air like some kind of jelly fish monster.

That is silly talk! I am giddy with anticipation at seeing this walk in Kidsville. Not to put any pressure on you

[bigbluedoggy]

Yes... very curious to see how it's progressing!!! I hope it's all going well! Such a cool project!

[rodiponer]

Thank you everyone, for your interest.

I have help! A good friend, my father in law, and my Mom (!) are now helping build this machine. I am hopeful that this machine will be finished in time for the playa this year, though it will be close-- it will probably take it's first steps the week before we leave, and even then it will need finish work to make it a ridable machine.

I gave up on welding the sheet skin to the legs. It was taking too long and did not look good. I found 3M VHB tape. I now love 3M VHB tape. It is double sided super glue tape with 100 pounds of shear strength per square inch, but unlike super glue it doesn't stick to your hands or make a mess. My Mom and my good friend spent two days taping the sheet metal skins to the sides of the legs, and then put in over 500 blind rivets to backup this tape near the stress points in the legs.


My father-in-law is currently wiring all of the electronics together in a waterproof box. The electrical system has turned out to be pretty complicated: a motor controller, remote control receiver, 12 channel radio controller relay box, an RGB LED controller, four LED dimmers, a solar charge controller, a 24-12 volt step down transformer, emergency stop switches, remote switches to power cycle the motor controller, a watts-up battery monitor, and dozens of fuses and little do-dads (diodes here and there) that I barely understand that will keep everything happy. I am so grateful to have his help. There is even talk of adding persistence of vision flickering light sticks to the machine, and a remote chance of a green laser scanner (with a salvaged two axis mirror galvanometer thing driving it).


I am using a new clevis design. The old ones require 1/128" tolerances between three welded leg parts, which is much too finicky. Everything will line up fine, then after I bam in the solid rivets it will shift slightly and the joints will start scraping. Leg #1 has had about three hours of careful filing and it's still not moving as smoothly as I want. So legs 2-12 will use a new clevis design that has much looser 1/8" tolerances. I plan to use nylon washers, which are easily sanded, to adjust the fit with these new cleviss. This also solves the galling problem, since I was worried that the 1/128" gaps between the joints would get playa dust stuck in there and the aluminum would start mucking up.

By making 8 of these clevises at a time I think I can machine 16 in about three hours, which is enough to make a leg. It takes about three hours to assemble a leg (there are a lot of holes to drill, then file, solid rivets to peen, then bearings to slip in and etc). So this change in direction may not set me back as much as it seems.


The next hard part is the crankshaft. I have barely started drawing it, and haven't quite figured out how to make it. I need two, and they need to be very strong, since the weight of the machine will be transmitted to them, in addition to whatever weird dynamic loads there are. I am hoping to make the crankshafts so that they can be disassembled-- something like three short keyed shafts going into four 1" steel bars that have keyways, set screws, and maybe shaft collars to hold the shafts in place. Each arm of the crankshaft needs to be 120 degrees apart from it's adjacent arm. I've never milled a keyway, and if I use 1" thick steel bars, the keyway will have to be milled in a way that makes it round on the top. I don't know if that is a problem. I also don't really know how strong it needs to be-- it seems like a crankshaft has the force going through it at all kinds of odd angles, where it has to turn from the shaft to the flat bars that connect the shafts. I am guessing I need to way over build this thing.


The rest of the work is relatively straightforward. Finish skinning the machine-- more cutting, VHB tape, and rivets. Weld brackets on to hold the batteries underneath the floor (half a day of measuring and cutting). Make brackets to hold the wheelchair motors. Weld brackets for the solar panels to bolt onto the roof. Add the floor. Attach the lights… None of that seems like a big deal, and is stuff that the people who are helping me can also do.

(fingers tightly crossed that this will work)