Aluminipede, a small walking machine

[rodiponer]

In the spirit of The Contraption thread, or the Playa-Piano, I am creating this thread to ask anyone with interest in my vehicle for advice and thoughts on details while constructing it.

I am building a twelve legged walking machine. If I am lucky it will work. And if I am even luckier I will finish it in time for Burning Man this year.

The machine uses the Theo Jansen 'Wind Beast' style of legs, which have an entirely different gait from the Walking Beast: http://www.moltensteelman.mechanicalspider.com/
The Theo Jansen style of leg does not require the drive system to lift the weight of the machine, so small electric motors can be used instead of a large engine.

Here is an example of the legs I am using:


This will be a small vehicle, 6' long, 8' wide, and 6' tall, constructed of aluminum tubing and sheets, and it should weigh less than 300 pounds. The legs will be four feet tall and each one will be four inches wide. The passenger compartment will be about 4' by 4' and 4' tall, with pillows or bean bags to sit on.

I have been working on this vehicle since October and, in the process, have learned basic machine shop work, CAD, and a whole lot about mechanical engineering. I am currently machining a couple hundred joints for the legs and drive system, and have enough parts to make the first couple legs.

I just traded my MIG welder for a TIG welder, and have about three days of practice TIG welding. I have figured out how to reliably weld thin aluminum tubing. But welding the clevis joints to the thin walled tubing is much more difficult. The clevis joints are a larger solid piece of aluminum that heats up much more slowly than the thin tubing, and so it's hard to get both pieces heated up to the right temperature at exactly the right time without blowing a hole in the tubing. I have solicited help on specialist welding forums, and know how an expert would do this, but it may just be beyond my skill level with the TIG at this point.

I am thinking I may have to fall back on mechanically fastening the clevis joints into the tubing. Each joint should see, at a maximum, about 100 pounds of compression, so even very small rivets are strong enough.

Do you guys have any advice on solid rivets (not pop rivets)? I like rivets more than bolts because they have a more finished look and will sit more flush with the surface than bolts. The legs will only have one inch of clearance between them, so it's better to not have bolts protruding from their sides.

From my Internet reading, it looks like there are four ways of attaching solid rivets:

- Bamming with a hammer (free, or $80 tool for the special rounded punch if I do not use flat rivets)
- Bamming with a pneumatic riveter hammer ($100 tool)
- Squeezing with a river squeezer hand tool ($150 tool, if I can find one that will squeeze a 1.25" long rivet)
- Squeezing with a pneumatic squeezer ($500 tool)

I have to fasten about 200 of these joints to the tubing. These would be aluminum rivets that are 3/16 or 1/4 in diameter, depending on what I can find with a 1" finished length (1.25 to 1.375" starting length).

I've never bammed a solid rivet before. Does anyone have any advice or experience on this? How hard is it to bam an aluminum rivet? Is it nuts to try and do a couple hundred with a hammer and punch thing-- is the pneumatic hammer really the way to go? Is the pneumatic hammer frigging loud? Is the cheap one Harbor Freight sells good enough-- I mean, all it has to do is bam, right?

Here is a photo of the leg joints I am making, with the bottom side slipped into the tubing. On the right is another failed welding attempt.

[Ugly Dougly]

From my measily experience in riveting, I know that you gotta have some sort of heavy and solid surface on the other side of the rivet you are peening.
And ear protection.

[epic_elite]

welding aluminum is an artform.

i went to school for welding many years ago, and never sought out a career as the economy was bad and industry was worse.

but aluminum is tough to weld because the oxidized surface has a higher melting point than steel. so when you finally burn through the surface you tend to blow through the non-oxidized metal with a lower melting point. that's why aluminum welds are usually really wide and ugly.

judging by the picture, you've brushed the metal. but i can't tell if your using the whip and pause method of your GTA welder. (the American Welding Society doesn't call it TIG any more because you can use carbon dioxide which is not Inert).

pause until a pool develops , whip the arc away and dump in your electrode, return arc and pause until a pool develops, whip the arc away and dump in your electrode.

it takes a lot of technique to acomplish succefully.

maybe cut all of your pieces, and use your scrap for practice. these things are a lot easier if you standardize your work, and get your settings fine tuned on the machine.

[Dustdevil]

Cherry Max rivets come in various sizes. They are installed like a pop rivet, but they are structurally effective. When I replaced the windshield of my airplane I used them to re attach the cowl area. They are really not that expensive either. The only down side is that you may have to purchase/borrow an air powered rivet gun. They take a tremendous amount of force to compress them. A normal hand pop rivet tool will not work.

Jack

[LeChatNoir]

Whoot, rodiponer!!

I like this project already. And Theo Jansen’s stuff is very very cool.

Regarding welding aluminum. As has been touched upon already, clean both surfaces well before you weld them. Scotchbrite works well. Clean your filler rod with it as well. Try preheating the solid part with a small propane or MAPP torch. Be careful not get it too hot, but preheating will certainly help. TIG is indeed something that takes practice, but it is certainly doable. Having a foot control will make your lie easier, but is not an absolute necessity.

If I remember correctly, you need to run your machine in AC for aluminum. This has the bonus of helping to clean the base metal as you go due to the back and forth action of the AC current.

Will definitely be keeping up with this thread.

PS: Regarding rivets... you can use a cheap pneumatic air hammer ($25 at harbor freight?) to peen the heads over. Round the point of one of the punches you get with these type hammers until it looks like a small ballpeen. Works very well. Practice on some first. Blind rivets this way, however would be very tough since, like Dougly said, you do need to back them up with a solid surface or something like a sledge hammer.

[LeChatNoir]

Also, try starting your puddle on the larger piece and rolling it over onto the tubing. Stay bias to the solid piece about 60/40 with the center of your puddle until takes good hold. You may have to keep this position as you run the bead, or you might try weaving your torch in a figure eight pattern, dropping your filler rod in as you pull the torch back each time.

[Sail Man]

I cant wait to see this lumbering through Kidsville

[Sail Man]

I cant wait to see this lumbering through Kidsville

[Sail Man]

Well, eplaya seems to be full of interesting buggy surprises tonight

[Ugly Dougly]

It's working.

But on its own sweet goddamn schedule.

[rodiponer]

I am excited to be getting closer to building the first leg. I am waiting for the 0.25" rivets, pneumatic rivet hammer, rivet set, and specially sized rivet hole drill bit to arrive later this week. Harbor Freight has mini drill presses on sale, so I picked up one of those.



In the mean time I have been machining more parts and practicing TIG welding. I still don't have the skill to reliably weld solid pieces to thin tubing. I still usually end up melting the tubing away or getting close to collapsing the solid piece into a liquified gob. I'm getting better at fixing this before it gets too bad, at filling in the melted holes or gaps, but I end up with a huge pile of aluminum where there should be a neat little bead. So I will rivet. The rivets should be good for at least 400 pounds, which is way more than these things will ever see (that would be 7 adults trying to ride the machine at once). If they fail I can always go back and weld, and maybe then I will be better at welding.

I still can't weld the inside corners of the tubing, where it is joined into a T shape. It seems like the pretty blue arc just wants to stick to the sides of the T and not go into the corner. So I melt a little above the corner or a little below on the other tube, without being able to get the puddle into the crick of the corner. I am using a needle sharp 1/16" tungsten electrode. Do you folks who are good at welding have any ideas? This was easy with MIG, since the wire would feed right into the corner. Or do I have the wrong technique-- should I be melting above and below the inside corner and then trying to get that magic surface tension that melted aluminum has to somehow bridge the gap? It's started to feel like playing with water. Or make a big blob with filler and try to blow it in the corner? That can't be right. This isn't critical, and construction will proceed without it, since the tubing is strong enough just welded on the top and bottom faces, and I don't really need the strength of also welding on the corners of a T joint. But I should know how to do this...

I am also having a problem with white cloudy stuff forming on the aluminum when I try to form a puddle on the solid blocks. It looks and feels like dried toothpaste. I've played with the argon flow rate and do not think that is the problem. I've also tried freaking out and cleaning the aluminum as if it was going to be a surgical instrument. So I don't think it's contamination of the aluminum or tungsten. I can eventually blow through it and get a puddle to form, but I think it is causing me to heat up the solid blocks much hotter than they need to be before I get the puddle. What could this be?

Also, I am really hitting the limit of my machinist knowledge. My CNC mill is making oblong holes. They should be perfectly round. The computer thinks they are perfectly round. At first I thought the bit was deflecting, or that it's some kind of inertia effect of the mill since I have ramped up the speed as fast as my machine will go (I have to make so many of these, that even a few minutes less for each part is days less). But I added a very slow 0.005" finishing pass to these holes, and they are still oblong. So it's not a bending bit or a over speeded machine shooting past it's coordinates with too much inertia. I think it may be something called 'backlash', which I understand only in the way that someone understands something who was read about it in the dictionary. So, wonderful machinist folks, would backlash cause an oblong hole? I've searched the Internet on how to measure and compensate for backlash in the software, or adjust it with some kind of screws on the mill, I'd like to confirm that this is the problem before messing with something I don't understand.

Until I figure out the oblong holes, I have programmed the holes to be cut slightly larger, so that the bearings still fit. The joints will work fine this way.

I'm about to start putting the aluminum stock in the freezer before I machine it. I'll have to ask the wife for permission. But after a few hours of machining parts, the vice and the bed of the mill heat up and stop keeping the aluminum cool as it is cut. So the parts gets way too hot and the aluminum turns into that gummy mode where it sticks to the bit and clogs it up. I broke a bit today, and chip welded another-- I am not sure if there is a way to rescue the bit and get the aluminum globs off of it. I do not have a fancy flooded coolant system or automatic lubrication system, so I think I just need to start feeding it frozen aluminum.

[motskyroonmatick]

When I weld aluminum I ball the tip of the tungsten.
I just did some research and apparently this(balling the tip of the tungsten) may not be necessary in some cases. This page touches on that and also states that Argon/Helium mix gas might be best for welding on thicker material.

http://www.millerwelds.com/resources/ar ... -beginners

Inside corners on aluminum have been the hardest for me to make. I think if my machine was more powerful I might be able to muscle my way through it.

I learned TIG on stainless Wine and Brew tanks. That is still where my best skills lie.

[mdmf007]

Your getting sage advice on the welding, but ill tell you about rivets.

Rivets are not used on parts that are subject to a lot of movement. Rovets are typically used to hold static things together, airplane skins, huge beams on old bridges etc.

If there is any force along the axis of your legs they rivet holes will simply wear egg shaped holes and eventually fail. This will happen very quickly with movement

To affix a rivet you will need access to both sides, or have to use pop rivet style rivets. Without access to both sides to place a buck you cant squash out the tail.

or you could simply switch to steel for these parts.

[ygmir]

it seems, somewhere, I read that properly riveted lap type joints are stronger than welds?

[rodiponer]

I have access to both sides of the tubing and so am using solid rivets. If you look at one of the photos above, you can see 1/8" holes completely through the tubing (and the solid clevis inside) about half an inch down from the edge. This is where I will place the rivets, though I will use 1/4" rivets since I was not able to find 1/8" rivets of the right length with a strong enough aluminum alloy.

I don't want to use steel since I am worried about corrosion with the aluminum tubing and clevis joint. Maybe I am overly paranoid about this from sailing, and things on land don't corrode that easily.

Part of the reason I like rivets is because they expand to fill the hole they are in. The rivets are 0.250", and the drill bit is 0.257", so it only has to expand 0.007" when I bam on it with the pneumatic hammer to completely fill the hole. So I feel like these joints will have less play in them than with bolts. I will read about cyclic loading on rivets. I suspect that they can cycle to some fraction of their strength OK, since the skin of an aircraft cycles every time it climbs and is pressurized. I thought rivets were best in shear, which they will only be in here. But I am not an engineer and am just reading, learning, as things come up, so thanks for the heads up.

With the intended loads, these rivets should only see about 80 pounds in static compression. I don't know about dynamic loads, which is why I am way over sizing them.

On the welding, yes, I've read a lot that says to ball the tip of the tungsten. And a lot that says to make the tip pointy. My understanding is that it's better to have a pointy tip unless the tip is so fine that it melts and drips. At that point you should have a ball. I haven't seen any tungsten melting yet, so I have been using a pointy tip.

I will look into using a helium/argon mix. Thanks for the tip. It looks like a mix of helium gives you a more focused arc, which might be what I need to melt the inside of the corner.

[ygmir]

can aluminum be forge welded?
I've done it with steel, and, it's super strong........

[LeChatNoir]

I am not aware of aluminum being forge weldable under normal conditions. It has such a small widow of plasticity that I suspect it would not be possible outside of a controlled industrial setting, if even then.

I would advise trying to weld with your tungsten balled and see how that does. We used to use a trick where we would use the torch somehow and it would heat the tungsten up and ball it. It’s been so many years since I’ve TIG’ed aluminum that I can’ t remember. motskyroonmatick might know.

Regarding the oblong holes, without knowing more about your mill, my first guess is backlash in one axis. This is the amount of play or wear in the lead screws of the machine. In manual machines, you can detect this by the wiggle or slop in the handle before the bed actually moves. In every control I’ve run on a machine without glass scales, there is a backlash compensation setting for each axis. You’ll have to find where to enter this number in your control and set up a dial indicator to check things out. Move your machine in say… in X- for a bit. Set up your dial indicator on something other than the machine bed and zero it against your vise or something similar bolted to the bed. Zero out the readouts and now begin to manually jog the machine in the X+ direction in the smallest increments your machine will move in. Once you see that dial indicator move the least little bit, you can compare that reading with how many thousandths your readout says you moved. Subtract what the indicator say from the readout number and you have your backlash.

ETA: Can you tell me about your machine? Many smaller machines have perforated disks on the stepper motors that have a little optical eye that counts how many perforations go by and it uses this to count how far the bed moves. If there is backlash, the disk and lead screw will move, but the bed won’t, making the control think that things have moved when they do not.

Glass Scales tied to the machine beds eliminate the backlash issue. It either moves or doesn’t.

[rodiponer]

I will try a balled tungsten electrode. I have read that to make the ball you run the TIG with DC current until the tip melts into the ball of the size that you want.

Thank you for confirming my backlash hypothesis. You are right, now that you say it, I see that the holes are larger in the direction of the positive X axis. I will study this and learn how to fix it. The CNC software I am using has a backlash setting to compensate for it, and the mill also has something called gib screws, which I believe can be adjusted in some way to reduce the backlash. I have been running the mill quite hard, and so it might be that I've knocked it out of perfection with all of the work I've asked it to do.

I have an inexpensive desktop (~80 pound) mill with stepper motors. There are no optical encoder wheels to measure the movement-- the computer just trusts the stepper motors to not lose steps. It seems to work well, at least for the precision that I want.

Here is a photo of an oblong hole (notice the space between the top of the bearing and the hole):
http://www.rodiponer.com/2010-02-24-oblong_hole.jpg

And a movie of making the part, with the two hole cutting operations (the rough and then the finish pass) starting at 6 minutes 25 seconds in:

[LeChatNoir]

The gib screws are for tightening the gibs as the name implies. Those are the long tapered plates that adjust play in your ways (the v-notch thingies that the bed slides back and forth on). Unless you have slop in your ways indicated by the table moving side to side a little bit (perpendicular to the axis of travel), I’d leave these screws alone.

I’m betting that after you determine and compensate for this backlash you’ll find it works much better. I suspect this is just is just a part of the leadscrews and not due to any hardship you’ve placed on the machine. In the video, you look pretty sane in the passes that your taking.

[motskyroonmatick]

I think you ball tungsten by putting the machine in DC electrode positive. I set out a piece of copper on my welding table and put the tungsten directly above it vertical. Other metal seems to contaminate the tungsten.
I then very carefully ease on the power with the foot control. Creation of the ball will happen fast so be ready to let off of the pedal. The ball should not be lopsided. If it comes out lopsided give it another shot. It you try it too many times the ball will become large and you will need to grind it or snap it off and start over.

[dragonfly Jafe]

this is an incredible sounding project! How will you turn (different walking speeds on each side?)

I hope you get this down to the Playa - I want to see this in 2010!

[LeChatNoir]

I think you ball tungsten by putting the machine in DC electrode positive. I set out a piece of copper...

Yep… now that you say a piece of copper, that’s what I remember is that and DC-Electrode Positive.

I hope you get this down to the Playa - I want to see this in 2010!

Me Too, dragonfly. And I’m curious about turning as well.

[LeChatNoir]

rodiponer,

Do you have a foot control for your TIG?

[rodiponer]

Yep, I really want to see this thing, too! That's the main reason I am making it, to see it.

My TIG welder has a foot pedal. I have the least expensive Miller, the Diversion 165, which I traded with someone on Craigslist for my MIG. It is an inverter machine for hobbyists and so I can't adjust the AC wave form, balance, or the other things that more advanced machines can do. It just has an amp knob and an AC/DC switch. Which is fine since just the amp knob is pushing the state of my knowledge.

I like rivets. The rivet gun, 0.257" drill bit, and two pounds of rivets arrived today. I bammed a few rivets. It's cool, it takes about half a second and they seem way strong. The pieces don't move or shift at all, so I think I can assume that both edges of the rivet will be under shear equally, and thus the 1/4" rivets should be able to handle about 1000 pounds. Way more than necessary.

I am getting close to finally assembling a test leg. Can't wait! If it works I think I will use it to build some kind of jig for the next 12, or actually, 24 legs that are joined in pairs (2" apart) to make 12 legs that are 4" wide box structures.



Do you guys have an opinion on how to cut aluminum sheets? These would be 0.0625" or thinner, probably perforated (with little holes), and I would need to cut these reasonably straight (+/- 1/8" over a 12" length) in 4" widths for the legs. My local metal supplier charges about $3 to $5 per cut, which adds up to a lot over 8 pieces per leg (96 strips of various lengths). Can you cut perforated aluminum in a shear, or does that hurt the shear in some way? Maybe I will call a water jet cutting place and see what they would charge to cut up a single 4x6' sheet into 96 pieces, if I supply a CAD file with a bunch of boxes.

[motskyroonmatick]

Years ago I had a bunch of aluminum diamond plate sheared in to strips that were eventually sides and legs for a coffee table I made. I would not recommend the shear. It twisted the metal slightly as it was cut and I ended up fighting those slight twists during the entire fabrication process. I had to brace the bottom of the table portion to remove the sprung nature that the strips gave the top. It has been a great coffee table and a good conversation piece. I plan on bringing it to the burn minus the smoked glass this year. All of my best constructions/projects end up on the playa.

I had to have the legs for the table welded on professionally. I was unable to make those inside corner welds that you are having a hard time with. I think I was unable to do it because my welder did not have the raw power to quickly melt the flat surface to start a puddle that could then be worked on to the edge piece. Butt welds and outside corners were not so much of a problem after a bit of practice. If I am thinking of it correctly I think your welder will do well on leg construction. It is probably the modern equivalent of my Econo-TIG.

[LeChatNoir]

This thread is good exercise for my brain. Stuff is starting to shake loose and float to the top!!

I remembered how I used to get down into those corners with the TIG. Try sticking your tungsten out about a half inch longer than you normally run. Turn the shielding gas up to really push it into the corner and cover the extended tungsten. You’ll have to be extra steady since your torch movement is effectively amplified. You may overcome some of this by purposefully weaving side to side just a bit.

Looking at some of your welds in the photos, something else came to mind. For this situation, a foot control was especially handy. I liked it way better than the thumb controls.

You may be doing this already, plus it looks like the rivets are going to work smartly, but you’ll run across a thick/thin weld situation again eventually so I’ll throw it out there. Start on your solid piece and really ramp the amps up to get a puddle started. As soon as the puddle forms, back your pedal down as low as you can and still hold the puddle as you simultaneously move your torch towards the thinner material. As soon as, and possibly just before, the puddle forms on your thin stuff, dip your filler rod in.

You may be able to find a happy middle ground with your amperage, but I’m suspecting that you will need to both weave your torch side to side while ramping your amps up and down with the pedal respective of each pieces thickness. This is the kind of TIG welding that I loved. It was like music, with the repeating patterns and rhythms. I could get lost in it… or maybe found.

Cutting the thin sheet:

Motskyroonmatick is right. Most shears will twist it. If you could find a sheet metal or HVAC shop, you might get it sheared in their equipment, which usually cuts the entire length in one shot, instead of on an angled lead-in like most shears. My local steel yard has a giant shear that works like this and I've had good results in 18ga. mild steel.

Your aluminum material might be thin enough to use a straight edge and clamps, then score it real good with a sharp utility knife and break it. It’s worth a try as east.


Automated choices are Plasma, Water Jet, and Laser, in order of cost (at least this has been my experience locally).


Plasma is the most cost effective but leaves a bit of a rough edge. Nothing that can't be dealt with in 90% of the cases, for me at least.


Water Jet cutting will leave a slight bevel on stuff 1/8â€

[Elliot]

About cutting of sheet aluminum.... Your pieces are awful small for this, but maybe you will be starting by cutting long strips from full sheets.
This is a very crude method, but....
When I skinned part of my bus with aluminum, I cut it with a Skil saw. To do this, you use a hard-tipped plywood blade and INSTALL IT BACKWARDS. You may also want to use a wax lubricant sold for this purpose. And ear muffs and goggles.
Far from neat, but cheap.

[mdmf007]

it seems, somewhere, I read that properly riveted lap type joints are stronger than welds?

It will get close if you ise epoxy with the joint.

Aircraft rivets are subject to forces that try to separate the skin from the body, and not in a linear motion.

Rivets will "egg" out a hole if the force is applied from the sides. Use epoxy as well to minimize this.

Bare aluminum will corrode very badly when exposed to playa. Check out the archives at www.landspeed.com We are replacing aluminum panels on the jetcar from the Blackrock runs we did.

sounds like a fun project - look forward to seeing it come to fruition,

[rodiponer]

Bare aluminum will corrode very badly when exposed to playa. Check out the archives at www.landspeed.com We are replacing aluminum panels on the jetcar from the Blackrock runs we did.

Oh no! I chose aluminum partly because I didn't want to deal with corrosion. The steel things I made for the playa last year rusted so much since I've returned. I've searched the landspeed.com archives and haven't found anything about the body panels corroding (yet, I've only asked Google to search the main domain-- they may have a blog sitting on some other domain). Do you know what alloy, and if it was a aggravated by dissimilar materials (I see they used a carbon fiber and aluminum together in some places, which is hard to get not to corrode)?

It will get close if you ise epoxy with the joint.

Thanks, I'm looking up how this is done.

[dragonfly Jafe]

just a data point - my 2006 BioTrak had aluminum track treads, and did not seem to suffer from corrosion issues. I washed it off when I got home after it spent a week+ on the Playa.