Robot construction ideas

I am starting construction on another robotics project, and wanted to get some ideas, and share some as well.

The limiting factor for me has been machinning capabilities. At work, I build with ABEC 5 and 7 bearings, AGMA 13 and 14 gears, and regularly tolerance parts to 0.0002". We regularly line bore bearing assemblies, and do whatever else is required. I use Maxon motors almost exclusively, and use other exotic components like particle clutches, linear encoders, etc.

Without a mill and lathe in my garage, I am religated to doing things in different ways for my personal projects. One constant that I have found, is that one way or another, you pay a price. Like the nomogram of the triangle that says "good, fast, cheap, pick any two". I have always found it easier to buy precision, rather than make it. For instance, building a ball bearing assembly from raw stock would be painful, and mostly un-necessary, but for less than $10.00 you can buy an excellent bearing. Another way you pay, is by building a robust assembly. You pay in the components you use, and the time you spend on design, but your reward is the ease with which everything else goes together. Feedback is more reliable, and motors work more consistantly.

My difficulty is in the transition from work related machines, (where I am capable of spending >$300,000 on a 4 axis lens system) , to home brew projects.

One technique I have found, is to use a laser cutting service. I use polycarbonate mostly, but I am thinking of starting to use acrylic and possibly even wood, because of their dimensional stability.

My curiousity, is where people buy the precision, where they make it themselves, and where they do not build what is necessary, and how they suffer as a result. What services have you found that meet your budget, and what sources for components are inexpensive and reliable.

T.I.A.

Mike Keesling

Reply to
Blueeyedpop
Loading thread data ...

Most folks in amateur robotics don't need much precision. Just note all the robots where the servo motors have been strapped on with double-sided tape. Lunt suggested in his old Nuts & Volts columns using Coroplast (Gatorboard, take your pick) for cheap robot bodies. Now there's low-tech. But hey, if it works...

When precision is needed, I prefer to buy the part, but not from a classic factory automation source, as that would be cost prohibitive. Take the gearbox from a Tamiya 4WD vehicle. You couldn't machine one like it without considerable skill and tools. The kit only costs $15 retail, including the balloon tires. Buy two, mount wheels on one end of the axles only, and screw the motors to a piece of wood or plastic. Voila!...you have a 4WD robot with indepedent right and left motors. Very cheap, very rugged.

Tangent to this discussion is your note about dimensional stability and acrylic. The two concepts don't know each other. "Dimensional stability" is a subjective thing, but few thermoplastics offer it -- at least the ones mere mortals can afford. Many of the kits in polycarbonate are made from 0.118" stock, which arguably is pretty flimsy when under load. Polycarbonate has impact resistance, but is still a flexible plastic.

Best to double-up on the material, or combine with a laminate-type substrate. A 0.032" aluminum laminate over plastic or other material might work. You can buy this ready made if you prefer -- tradename is Alumalite (not Alumilite, which is a casting resin). Alumalite has the same weight as 0.040" aluminum, but is 50 times stronger.

Even a dense expanded polystyrene will help considerably here, while not adding a lot of weight. You can make the lamination yourself. Be resourceful in your choice of substrates. Just as plywood gets its strength from multiple plys in cross-grain fashion, you can often combine two flexible materials to create a fairly stiff one.

Never use acrylic on any flexing member of a robot. Bad. Ptuui! It develops stress crazing over time, and will eventually break. Acrylic is for looks -- always has been -- not strength.

-- Gordon Robots for Less at Budget Robotics:

formatting link
Robot Builder's Sourcebook & Robot Builder's Bonanza

Reply to
Gordon McComb

Could I suggest the niche opportunity of a "home manufacturing workstation"? A desktop 5 axis milling machine with decent precision capable of doing small jobs with soft raw materials. Link it to the net, and as people develop useful gears or mechanisms the designs or cnc files become open source. Get the cost down to where PC's were when people started getting interested in those, and the concept would start to take off. It would save time and shipping costs for lots of stuff, and the people who would make money would be the workstation suppliers and the Raw material suppliers. Just ask HP/Epson/Canon/Kodak how lucrative raw materials supplies are!!

Reply to
andrew

There are a few exceptions, such as the gear mesh between a large motor and the transmission or push rods controlling the swashplate on a model helicopter. But for the most part, I agree that ABEC 5 bearings are overkill for most hobby applications, but I will use them in certain applications (main shaft support, tail rotor drive, etc).

Some folks even build entire airplanes out of Coroplast. Check out the Simple Plastic Airplane Designs for some very interesting designs:

formatting link
One of my favorite quotes from a SPAD builder is that they love election season -- after the polls close, they have enough signs to build hundreds of airplanes.

Trammell

Reply to
Tramm Hudson

Laminates and cutting:

Yeah, I have a lot of good succes with those. My problem comes down to processing. My laser cutting guy gives me a break, since everything I give is properly prepped, CAD file wise, and I spend less than $2.00 a minute for cutting. He can only handle acrylic and wood easily. He does polycarbonate, albeit somewhat grudgingly. I was even considering plywood. Vinyl based products emit nasty fumes, so I avoid them, and he doesn't have the wattage to touch aluminum.

I had a TON of honeycomb Aluminum, carbon fiber, and fiberglass, but again, it comes down to workability of the material.

Do you have any good leads on "affordable" CNC routing? That would open up a world of possibilities.

Materials:

I was aware of the crazing issues with acrylic, but I thought that was more of an issue of repeated flexing into a "zone" where I would never go with acrylic, and U/V exposure. I has always looked at this as a failure mode of acrylic, much the same as steel or aluminum, but taking into account the strength and elasticity of acrylic. Is it more of case that it just won't accept any sort of repeated flex, no matter how little? I had considered making a poly carb skin on acrylic, but I would rather not deal with that much solvent.

My intention was to apply "welding design" techniques, by building "I, H, and T" beams, to minimize flex to a minimum. Is this still doomed?

Thanks

Mike

Reply to
Blueeyedpop

Gordon McComb wrote in news: snipped-for-privacy@gmccomb.com:

For CNC home style look at

formatting link
>> The Workshop. There a 80+ page thread there with several home built CNC's

Reply to
Andrew

Any help on where on the site this thread is? Quite a number of forums...

-- Gordon Robots for Less at Budget Robotics:

formatting link
Robot Builder's Sourcebook & Robot Builder's Bonanza

Reply to
Gordon McComb

Look out for line wrap

formatting link
HTH TJ

Reply to
TJ

Yes!

If I could get > Could I suggest the niche opportunity of a "home manufacturing workstation"?

D. Jay Newman

Reply to
D. Jay Newman

Reply to
Mitch Berkson

Go composite, it's cheap and easy, albeit a little smelly.

Reply to
Shawn Brown

I just sold off all of my honeycomb composites too. 8^(

Reply to
Blueeyedpop

I assume you mean carbon composite (there are a lot of "composites"). The smell I can live with (I machine acetal resin which is real stinky, too), but without a good dust collection system, this stuff isn't nice on the nose, lungs, and and other air passages. I dislike working with fiberglass for the same reasons.

When you route at 30,000 rpm, there's a very agressive blast of air from the tool that spreads fine dust everywhere. Heavier plastics "chunk up" and the sharf is too large to get into your nose, or into your skin. Milling carbon composite, fiberglass, and similar materials causes a thin film of dust that spreads over the entire shop. A real pain.

-- Gordon Robots for Less at Budget Robotics:

formatting link
Robot Builder's Sourcebook & Robot Builder's Bonanza

Reply to
Gordon McComb

Multi-axis CNC mills may be hugely expensive, but a manual metal lathe would go a long way into making the parts you need. The classic 7x10 lathe sold by Harbor Freight and others is a pretty small investment (under $400, including tax), though you might want to go with a Sherline as attachments are easier to get. (You'll also be helping my local economy; Sherline is a neighbor.)

I've made shafts with 1/4-20 all-thread ... it's not very good material, as it is fairly soft. Personally, I'd get some steel rod from the local machine shop outfitters (McMaster-Carr if you're so inclined), and use your new lathe to create the shafts. As you know, there are two basic chores here: cutting the shaft to length, and cutting a groove for retaining clips. You'll become an expert at these in no time, if you don't already have experience in this already. Camfering the ends of the shafts and other nicities can come later.

Depending on the size and weight of your robot, you might look at polymer bearings. Igus is a major manufacturer. You can get the bearings already mounted; a stamped flange-mounted bearing might be only $12; $50-60 for something that can handle the weights and loads of a larger combat robot. They're made to accept a misalignment of the shaft, so you don't have to be so precise in your machining.

For plywood, I'd go with only "aircraft ply," which any good hobby store will have. Or, as long as you're ordering from McMaster, think about that plastic/aluminum laminate I told you about. McMaster also has ABS and non-expanded PVC. A typical 0.250" ABS sheet will have a tensile strength of some 6000 psi. It's thick enough that you could mount bearings directly into it. A little arbor press (also Harbor Freight) could set them.

If you're using Bruce's cutting service, ask the max thickness for materials like ABS. You might need to go with two thinner sheets, laminated together. There are ABS solvent cements that will literally fuse the two sheets together. It can be applied with a small paint roller.

-- Gordon Robots for Less at Budget Robotics:

formatting link
Robot Builder's Sourcebook & Robot Builder's Bonanza

Blueeyedp> Yeah, that is about the state of the discussion I had with someone else on

Reply to
Gordon McComb

Well, The machine shop route is out, but I may have to resort to using the tools at work. This house was my choice, so my wife gets the garage for her stuff. Next house will be her choice, and I get the garage. I gave away all my cutters a few years ago, swearing I would never touch another mill or lathe ever again.

I have a bucket full of IGUS bushings, thankfully enough, as well as a bunch of bearings. Enough at least for this project.

Yeah, .25 allthread is spooky stuff. Trying to ride the fine line between junk, and working assemblies is fun.

AFAIK, Bruce filener doesn't do much besides wood and acrylic. I won't give him any vinyl based products to cut. He has to use high quality plywood, clear through, without voids, because it gives his laser grief.

The conclusion I have come to, is that there is no cheap way to do this. There is expensive, and very expensive. I figure, generally speaking, I am looking at about $40.00 to $60.00 per axis on a respectably sized walker. Where ever the materials are cheaper, they have to be doubled up to strengthen them. As far as working the material, there is nothing cheaper than fileners.

Thanks Mike

Reply to
Blueeyedpop

Appearently, milling carbon fiber creates conductive dust which can foul electronic circuits.

We gave up on our honeycomb fiberglass. The smell can best be characterized as "angry cheese". Bad Ju Ju.

But... If I wanted to do 10 of these 'bots, striking a mold and going with fiberglass layups wouldn't be so bad. I know my way around that stuff pretty well. It comes down to whether I want to suffer through fiberglass...

Mike

Reply to
Blueeyedpop

You seem to get all the neat stuff.

I was thinking that *maybe* heat treating the all-thread would make it a bit stronger. Since you're not using the threads as a leadscrew, any deformation of the threads would have a limited impact -- the nut would just be hard to put pn. At least in the all-thread I've found at Home Depot/Lowe's, it's a pretty soft metal, to aid in thread cutting by the manufacturer. Terrible for shafts that take a load.

You might ask about the ABS. It doesn't release chlorine gas like PVC, and is routinely used in laser cutting -- in fact, it's often considered safer to cut than polycarbonate (Lexan). ABS has very good dimensional stability, and is easier to work than polycarb. It is pretty dense, though, so it will add to the weight of your robot. I'd be surprised if he would be willing to cut thicknesses above 1/8" (i.e. focus). So, for

1/4" or 3/8" thickness, consider the layering approach I mentioned.

-- Gordon Robots for Less at Budget Robotics:

formatting link
Robot Builder's Sourcebook & Robot Builder's Bonanza

Reply to
Gordon McComb

I meant anything that you can do hand lay-up with, make a mold out of foam and off you go. No dust until you sand it.

Reply to
Shawn Brown

I hear you, but when you don't have access to CNC tools, you can't beat it! Plus you can do body designs that look like something other than a box.

Reply to
Shawn Brown

It's just that I spent several weeks in the heat of Los Angeles summer in an aircraft hanger, building full scale P-51 mustangs out of the stuff. After

120 degree heat, resin soaked pants, and exposure to an open vat of acetone, I decided that I did not want to play with the stuff ever again. Up until now, I have only violated that once, and with Kevlar and Epoxy, so it doesn't quite count.
Reply to
Blueeyedpop

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.