Sheet Metal vs Tube Drivetrain

Our team has recently obtained a waterjetting sponsor and would like to build a new chassis over the summer. We are currently debating between building a full sheet metal chassis or a chassis made out of aluminum tubing (probably 2 * 1). We have noticed that a lot of the best teams, like 254/118 have used aluminum tubing in the past.
What are the pros/cons to using each on your drivetrain?

Also, we were wondering why many teams use chain on their drivetrains, as we have always used belts.
What are the pros/cons to using belts vs chains on your drivetrain?

Do you plan on bending the sheet metal or just using flat plate? Use the resources you have to build the best robot you can. Resources are not just access to a water jet it’s also having people with the knowledge of how to take advantage of the water jet.

We have done both. We found the tube chassis to be a lot faster to assemble. Unless you have the ability to do precision bending, it’s almost not worth it to put holes in your sheet metal first, otherwise your bearing holes or axle holes probably aren’t lining up. Our sheet metal fabrication abilities consist of a shear, finger brake and a press brake that we don’t really use. None of which allows us to drill holes prior to bending and have the holes line up. Therefore, we had to bend first, then mill/drill out the holes.

The tube/WCD drivetrain we did was all milling out the pockets for the bearing blocks.

Sheet metal vs tube is a resource decision. If you don’t have the sheet metal bending capability, you are limited to COTS and tubing/channel/angle to get enough stiffness. If you do, it boils down to the usual COTS-vs-homemade decision - are we willing to take the time to design and build a custom sheet metal frame out of our build season to get exactly what we want, or do we take some COTS (extrusion) shortcuts to get something that’s “close enough”?

Chain vs belt is a bit different. Among the pros of each are:


  • You can make any length you want (multiple of chain pitch, or better, 2x chain pitch), including much longer lengths than belts are generally available.
  • It is narrower for a given amount of load; uses up less space.


  • It does not stretch over time like chain; works better with exact center-center axle spacing and no tensioners.
  • It is lighter than chain.
  • It stretches under shock load and is likely to stay in sheaves even when the belt jumps teeth. This allows it to be used as a torque limiter to prevent broken teeth and smoking motors, at a cost in belt life.

Our team has never had good luck with chain for the drive train. Starting with our 2012 (2011 was our rookie year), we found that belts were lighter, easier to work with, more reliable, less service and maintenance, and seemed to accelerate better. Yes, you do have to have your centers right, but Gates and many others have great calculators to do that.

In 2014, we put our belt drive train inside the tubing. WHAT a difference! The drive train was absolutely fantastic. The mistake we mad was using 9mm wide belts and they did stretch little, but in aggressive negotiations in a defensive manner, they would strip teeth.

In 2015, we use large tubing so that we could use 15mm wide belts and never looked back. Zero issues. Durable. Reliable.

In 2016, due to going over the defenses, we used two sheet metal plates separated by standoffs so that we could put in an eight wheel drive kinda like the Army’s Striker (we feel the Army copied our robot). This chassis was unbelievably durable and reliable, as we dropped tested the robot from 3 feet and it kept on ticking. Our chassis was king this year!

In future years, we plan on going back to the belts in the tubing unless the game dictates otherwise, such as swerve or rugged terrain.

A waterjet should be able to cut through the 1" x 2" tubing. However, using 1" wide tubing means 9mm belts; I would recommend 1-1/2 x 2 or 1-1/2 x 2-1/2 so that you can put 15mm wide belts inside and if you are worried about weight, put in relief pockets in the tubing to lighten it up. Don’t put holes in it where debris can get in, just thin sections from the 1/8" down to 1/32" thick. It looks cool and protects the parts inside.

One last thing. You can buy sprocket material in logs, cut them to length and bore out what ever size hole you want. A typical log is about $80 and you can get 6-8 sprockets out of it. Well worth it.

Any questions, contact us at

I think it depends highly on your bending equipment.
Personally I’m partial to sheet metal because it goes together very fast (put everything in place and pop in rivets), but if your bending tolerances are +/- 0.01", it wouldn’t be suitable for bearings and shafts between bent walls.
Tubes are very easy to work with due to the integrated right angles and geometry and my team has used a lot of 2x1 and 1x1 (both 1/16" and 1/8" wall) everywhere. I don’t know how much of an effect your waterjet’s accuracy and taper will have, but if you want to be more careful you can purchase a 1.125" reamer and use that to finish all holes to a final size.
1.124" would let you do a light press fit, but those have to be bought custom to my knowledge.

It’s not about what’s best, it’s about what your team can produce the best. There are world class teams with good designs using both methods.

If you have a water jet sponsor and no milling or precision bending, you’d probably want to build a plate drive.

pipe bending has to be done carefully and in the right order. If you have not done it before the summer is a good time to work the bugs out. However, the sheet metal frame can also be bent in a contact game year.

Need to decide what is best for the game. Also, just because you have a sponsor for something does not mean you need to use them, it might end up being a disadvantage.

I don’t think if their team were to use tube it would need to be bent. With a water jet sponsor they could rivet and gusset 2x1 and 1x1 to build a very competitive robot.

Additionally, my team used sheet metal this year and had zero issues of it bending and used .090" 5052 sheet everywhere on the robot. When using sheet metal I don’t think it’s a matter of “choosing” this building style based on what is best for the game but designing the robot to meet the needs of the game with bends in the right places/using thicker material. Many teams were successful with sheet metal this year as well as tube.

Generally speaking, teams that use tubes don’t bend them. The bending discussion is for sheet metal (and is needed to get flanges for strength, if nothing else).

There are exceptions–1024 is the one that popped into my head–but they tend not to bend metal in the drivetrain anyways.

I think SystemS was specifically talking about circular tubing-- in which case what they are saying is extremely true. It doesn’t really apply too much to the discussion of drive trains, but it is definitely applicable in that context. I haven’t had a ton of experience with using circular tubing on FRC robots, but there have been some cool robots who have used it in the past (I think 118’s elevator in 2015 used circular tubing).

We have built entirely waterjeted sheet-metal robots.

A few requirements:

  1. Full CAD of the entire robot,
  2. a decent sized metal break
    3 preferably a pneumatic riveter

Your CAD package should support sheet-metal parts. Make sure to set it up with your breaks bend radius, and decide what your standard flange size will be.

Temporally hold each assembly piece square before installing fasteners.

Make sure you have access provisions for any items that are inside the frame. Gearboxes, pneumatic components, etc.

Large metal breaks are not light or small. Our 48" brake weighs over 1200 lbs. You need a break large enough to bend an entire side of your 120 inch frame perimeter.

I would definitely try some practice structure before build season.

If you don’t plan to go WCD / cantilevered axle, sheet metal seems to be the way to go. That said, if you do want to go WCD / cantilevered axle, I imagine it would be pretty difficult to do so with sheet metal.

Bottom line is, go with a system you are familiar with during the build season. Experiment in the off-season to grow your team’s knowledge-base.

I think 148 would disagree with you.

I’ve worked with both of these frame types before though their are different types of water jetted frame styles ( and If you use a folded frame design, you will need to get it bent by a powerful/precise machine (might be able to be done manually depending on material,thickness and length of part).
Water jetting is nice because you can have a frame cut out and ready to be assembled in about 30min-1.5h depending on the complexity of the design (if you have direct access to the machine that is…)It’s also easier to jet spare parts, and is lighter then tube (If designed correctly). Some of the cons however, is that it is fickle when bending it (you might get a perfect 90 deg bend or not). It needs a lot of bracing to make the frame stable and rigid.
(for my team) when we got our frame bent, it bowed the metal in a arc making a 1/8 gap in the middle of the part when laid on a flat surface.( This was most likely the bending machine’s fault as well as the settings on the machine).

What are the pros/cons to using belts vs chains on your drivetrain?

Belt is better (IMO) it has vary little, too no slack at all. Making it precise. it doesn’t stretch nor does it brake (Vex’s Belts other brands will though)
Chain normally (in my experience) is a flipping nightmare, chains stretch, masterlinks brake, sprockets strip, etc,etc.
Here is a document found on vex’s website “Belt vs. Chain Drive Evaluation” by FIRST Robotics Competition Team 234](“Belt vs. Chain Drive Evaluation” by FIRST Robotics Competition Team 234)(credits to 234!)

Hope this has help in some way:]

Back when I was on 2220, we designed our drive trains primarily as closed tube/channels (tube on each side of the axle). They had some disadvantages, but man they were absolutely bulletproof, and were incredibly fast to design and build (I was able to design one in the style we did for 2014 in about 30 minutes with all the COTS parts downloaded).

It’s been said before in this and other threads: go where your expertise is.

If you don’t have existing expertise, pick a direction that complements your existing skills and go that way.

I really don’t understand why so many people are chiming in about chain being such a problem. Chain is a very, very powerful tool if used correctly. It is even more powerful if you DO NOT use master links. After becoming familiar with the dark soul chain tool, I have to say, it is the best thing ever made for chain. Pushing the pins back in, and using the correct number of links using a chain center to center calculator is about the best thing you can use when designing mechanisms/drivetrains with chain. If you use those two things alone correctly you will never have to worry about tensioning/stretching so long as you use large tooth count sprockets and the right chain pitch for the job. My team used #35 chain for our final reduction coming off our Drivetrain gearbox with fixed direct C-C and #25 chain going from wheel to wheel direct C-C with no tensioners on either and had ZERO issues. We love chain and will continue to use it because it is easier and does not require custom belt lengths for each year or each robot.

The bolded part. Many people don’t get to the part where they actually know how to use chain correctly. Same thing applies to belts. This applies to many things in competitive robotics, but is particularly pronounced in drive trains due to their overall “mission-critical” importance. A single bad experience for a given method or technique can scare someone off of that for a long time, even if it was an easily correctable user error.

I didn’t want to get sucked into another chain vs. belt debate, but:
-Belts are lighter than chains. But, Vexpro pulleys are not lighter than sprockets. A 15mm belt drivetrain weighs as much as a #25 chain drivetrain, generally speaking.
-Belts are quieter than chains, and more efficient. However, 234’s paper only covers #25 chain and not #25, and used an 8020 chassis to do it, so the exact amount imo is unknown.
-When put together with a Dark Soul tool (for #25), chain is extremely strong. Plus, masterlinks are at least $1 per chain run.
-Chain can be assembled on the sprockets, while belt needs pulley removal to go on if you’re not tensioning.
-Belts will slip before breaking gears (on arms and stuff, anyway) depending on tension, whereas chain will kill everything before it dies in a spectacular fashion (ask me how I know).

That’s all I think I know right now. :stuck_out_tongue: I like belts everywhere but the drivetrain, personally, but they require forethought.

While i can not attest to reconnecting chain, I can say with utmost confidence. that belt is by far the easier of the two. Chain stretches making it inconsistent in slop. Belt does not stretch and is more accurate, lighter, cheaper, stronger(depends what kind you get), cleaner , etc. that’s one of the reasons it’s used in engines (and the fact that it can run at much higher speed without oil) Don’t get me wrong. chain does have it’s purposes, but drive wheels without any shielding to the chain is not one. (IMO) It’s wonderful when making a forklift!