Tube Versus Channel

Technical Technical Discussion
1

As a relatively new member to the FRC community (this is my second year in the competition) I have been very closely looking at other teams’ robots, as seeing how different teams design is rather interesting to me. While doing this, both online and at regional competitions, I have noticed several things about frames, which have gotten me thinking about robot design in the future, as my team is likely to design some sort of showbot over the summer for outreach, as well as participating in future FRC seasons.

The main thing that I have been wondering about is structural materials. Namely, I’ve noticed several distinct types:

a) Channel

b) Tube

c) 80/20 or other extrusions

Noticing these differences lead me to wonder:

Why do teams choose these different materials?

I have several theories on the benefits of each.

For channel, it may be for convenience, as, if I remember correctly, 1" channel is the typical kitbot construction. Additionally, this isn’t an all together structurally unsound material. This also may be perceived as cutting weight. It also only requires one hole to be drilled through the material to bolt something to it.

For tube, some possible reasons may be that it is more structurally robust than channel. Also, there is more area to mount onto, which may be a benefit.

For 80/20, possible reasons may be the ease of prototyping with this material, as well as the excellent structural integrity it affords.

I am interested as to why different teams choose these building materials, as well as what benefits and disadvantages are seen in them. If any other team members could offer some advice to this humble sophomore, I would be much appreciative.

If anyone has already posted a thread to this effect (I didn’t see one when I searched, but I may have missed it), I would also be perfectly grateful with a link to that thread.

Also, this is specifically for metal robots. I am aware that some teams make wooden or plastic robots, but as these are rather uncommon, I would appreciate it if any discussion focused on materials for metal robots.

2

I think your hands down favorite/best material to construct a robot out of is waterjetted and bent aluminum sheet metal, since you can design it to the exact requirements you need, given you an optimum strength to weight ratio.

Lacking the fabrication support necessary for the above, 57 has settled on the kit frame with a chassis of 1"x1"x1/16" wall square aluminum tube, mostly held together with more 1/16" sheet plates riveted with 3/16" aluminum structural rivets. It’s pretty lightweight, plenty strong if you’re not dropping your robot off a cliff, and we can get 20’ sticks of the stuff for $6$20 from a local supplier, even on a Saturday. Which is harder to say about AM channel or 80/20 or another T-slot extruded aluminum.

Thick wall AM style channel is heavier and tougher than we need for any framing purposes. Thin wall channel would be flimsier than tube, if lighter. Either of the u-channel options means you have an entire side that can be a pain to attach things to.

80/20 style extruded aluminum is good for prototyping and quickly adjusting parts, etc. Only problem is that it’s heavier than the 1/16" wall tubing that we use, for no more benefit than the quick propotyping.

3

Our frame is all aluminum tubing welded together. 1x1 1/16 wall for most of the upper structure, 1x2 1/16 wall for the base. As well as 1x2 1/8 wall for the base side beams that have the DTs mounted.

We do all of the welding in house and find that a welded tube frame is light and very stiff if done correctly.

It also gives us flexibility of the frame geometry and how we mount other components.

4

Interesting… We use 1/8" wall square tube on a kitbot frame, and every year it seems like we bend or break at least one piece of it tubing – making us leery of going any thinner. (Also, there’s no such thing as a “local supplier” in Naples, NY… Unless you need a cow, STAT.)

5

We used to go with all 1/8" wall as well. We eventually tired of being constantly overweight with frames that took up a distressingly large percentage of our weight allowance. Yes, they would probably fare well in a head-on collision with a small motorcycle, but that’s obviously a bit overdesigned for most FRC games. Our biggest problem was that we usually ended up using the 1/8" in various locations for motor mounting bars, etc. Places that didn’t need anything near as strong as 1/8".

So, we build of the kit frame, and figure on that taking any big impact loads. Anywhere we’re not expecting to get hit gets the 1/16", and we look at using 1/8" if we’re likely to be hit hard and often somewhere.

Although, this reminds me of the single biggest problem of using 1/16" wall tubing. If you’re mounting something by bolting through the tubing, you need to be cautious about over-tightening your bolt, or you’ll squish and deform your tube distressingly easily. We’re constantly reining in the gorillas on the team that have a bad habit of severely over-tightening bolts and such.

6

That’s another fun problem with using tube, you need to plan your attaching points pretty early on. Welding (or riveting) a mounting tab to the tube is a good way to go.

We built a welded steel frame using mostly thinwall square tube and some angle and strap for this year’s robot…no problems at all with the frame.

7

You can either drill a large hole on one side of the tube so that you’re only bolting through 1 face, you can use large washers, be careful with your tightening, some teams have had success with rivnuts, or you could try to find some other way to mount to the tube entirely (instead of bolting).

8

(Last thread derailing, I swear)
Rivnuts are definitely cool and we’ve used them before. Did I mention we have gorillas on our team? Last year they cross threaded a bumper attachment screw in a rivnut, then spun the rivnut. We missed a match getting that bumper off and replacing that rivnut. So we’re running reversible bumpers now and are a little more cautious of rivnuts.

9

It appears like most teams use welded aluminum extrusion, if not using the kitbot frame. The most common sizes appear to be 2x1x1/8 on the base, and 1x1x1/16 on the superstructure.

My team has took the rather unusual step of using rivets last year to assemble the frame. We used almost exclusively 1/16" aluminium tubing, with 1/8" gusset plates that riveted the whole thing together. A variety of sizes of rivets were used, from 1/8" aluminium to the huge 1/4" steel. The rivets were a weight saving measure after trying to use steel bolts.

I would recommend rivets to a team with an accurate mill and a desire to save weight. With an accurate mill, the gussets and the frame members can be made so the frame is perfectly straight (we have had some problems with crooked welded frames in the past). Also, rivets allowed us to use almost exclusively 1/16" tubing, which saves weight, but is near impossible to weld. I estimated that this saved us 5-8lbs.

If you do decide to rivet (or to bolt and then are forced to go to rivets), keep in mind that you have to size the rivet for how much material it is gripping. Not just any size will do. Also, 3/16" rivets work in #10 holes.

10

Although they aren’t very well known, rivnuts can make for a very neat solution in my experience. When I went around asking for rivnuts at SVR, people looked at me like I was crazy. I’m glad to hear someone has heard of them.

11

We also like aluminum tube. Really, the best thing to use is 1 x 2 aluminum tube, welded. 1/8 in thick for drive base, and basically 1/16 for everything else. If used correctly it is perfect. Minimize the amount of frame material you need. Look at any 254 robot, and you will see how simple the frame actually is. It makes for lighter robots, and you really don’t need anything more. Also, 1/4 in, waterjetted, aluminum plate goes very nicely with this. It makes building 20 times easier, and reduces the amount of tubing you need. Plus, you can make custom gearboxes. Really, it pays off to get a 2D sponsor, as they can turnaround parts amazingly quickly.

On the subject of sheetmetal, while a lot of good teams use it, there are plenty of good robots built out of tube, as long as you streamline the design process(254 again comes to mind). The nice thing is that the robot comes together lighter, more quickly, and more accurately than an aluminum tube robot. On the other hand, you need to have a good team of CADders, and have the sponsor to be able to do it.

12

2 questions.

  1. Is your 1/4" plate .25", or is it like ours and ranges from .22" to .24"? That made our gearboxes this year a little difficult to get fitting correctly (if you’ve seen the gearboxes, you know what I mean)
  2. Besides custom gearboxes, how else do you use it in your frame? We have a lot more frame right now than I’d like, but I can’t see how 1/4" plate would make it significantly better.
13

If your 1/4 plate is .22 to .24, you ordered a gauge thickness rather than 1/4 plate.

1/8x1 tube is virtually the same weight as 1" 80/20.

This year we used a lot of 1/8x1 angle. Half the weight of the equivalent channel & allows bolts to be properly torqued without crushing the section.

We use mostly 1/4-20 bolts. A little heavier than other options, but you don’t have to worry about the inexperienced over torquing them. Carroll Smith has a great book about fasteners “Nuts, Bolts, Fasteners and Plumbing Handbook”

It is all about the details & trade offs.

14

Another semi-endorsement for rivnuts. They are really cool things, in theory. However, beware of using in 1/16th wall stock – very easy to spin them, and then you have a mess. Installation is also an art! Too much force, thinking to make a better seat, tends to mess up the threads, especially in aforementioned 1/16th wall. Still, they’re pretty neat and tidy when done well!

15

1 question that I have that I have not seen mentioned is price. What is the cost of the tube versus the kit bot c-channel, which is included with the KOP. If it is expensive, I don’t see much bonos to it over C-Channel, which works fine for most uses (at least for 3397). We used it on our chassis and tower and electronics mount. It is light and pretty durable. So, unless it has a VERY good benefit, I do not see the point of buying the tube.

16

Do you have any advice or books/white papers/etc to recommend concerning designing for 1/16" tube? We’re pretty much all 1/8" tube or angle for everything on our bots. I’ve heard great things about 1/16", but we’ve warped welded 1/8" chassis occasionally so there’s some hesitancy.

Price:
From what I’ve found it’s really very, very location-based. We used to buy everything from McMaster until we discovered the joy of local surplus stores. If you think you might be missing something, introduce yourself to a local machining company or two (or ten). They’ll know. Different profiles definitely offer a lot more design flexibility. Certainly I’d think 1/8" C-channel on the tower is over-designed. Maybe angle if you want to stay with 1/8".

17

The chassis could use 1/8", but on stuff that doesn’t need to be that thick, like the upper structure, 1/16" is usually fine. But it depends on how you design it! We like making upper stuff from 1/16" tube and riveting 1/16" sheet or angle brackets to hold it together, and attach stuff to it.

It’s all in the design…

18

We use almost all 1"x1" square tube. In some places we use 1"x2", but very little of it.

If you look at our 2012 robot, the lower level of the frame is 1/8" wall, mainly because we have our drive wheel bolts mounted through it. The front and rear pieces where the shooter mounts are 1/8" wall. All of the rest is 1/16".

The robot frame is almost all welded by students. They practice a lot in the fall with welding scraps of the 1/16" wall to get good at it. It is easy to burn through.

19

I was about to write about 2815’s construction methods of recent years, but this really is what we use–the last two robots (with three blue banners between them) have been constructed of 1"x1" square tubing acquired from the Lowe’s near where I work (which is not as cheap as Kevin’s aluminum–I want your source!), attached with gussets of various sizes and shapes (we use some angle as well) with rivets and attached to the AndyMark C-Base kit frame. For all the nail-biter moments these two robots have given us, the structure has been rock solid.

It’s been many years since I worked with 80/20 (and then as part of an elevator rather than a frame); the stuff can definitely get heavy. If I were building a sliding mechanism, or a part that I knew would require a lot of quick adjustments, I would consider it…but I doubt I’d want to build a whole robot out of it.

We build our structures so the 1/16" tubing is inboard–the only exception was our original roller claw in 2011 (which broke when our drivers smacked it into the player station glass at speed). From there, it’s just been judicious application of 1/8" rivets (sometimes by overzealous freshmen) to hold it together. If you get to St. Louis, look us up–there’s no secret sauce to our framing setup, and we’d be happy to show it.

We haven’t tried welding any of our parts–call it a bit of paranoia about breaking parts in the heat of competition. Granted, we could duct tape it these days…but the idea of just throwing rivets into a new piece of metal has its appeal.

20

Where to buy? 80/20 has other shapes. Price is better than local hardware store. Most places have metal supply companies where you will get the best price. They typically have large minimums so you need to buy in bulk. If you explain what are doing they will often waive the minimum charge, but you will still need to buy all you need at once.