pic: Off Season Sheetmetal 6WD



A sheet metal based 6 wheel drive I designed to get some practice with sheet in case my team decides to use it next year.
4 in plactions with 1/8 in center drop. 25 roller chain is used to drive the outer wheels. Guts from a default Supershift used for the gearbox, resulting in an estimated 3 and 8 fps @ normal load.
All rails are .09 in 5052 except the gearbox flange, which is .125. Total weight is estimated at 34lbs.
And now for my questions: Should I add some sort of lateral support in the middle? It can’t hurt, but is it necessary?
Is a bellypan for this type of design suggested? I have one modeled, but removed it because it was actually quite heavy (1/16 in thick).
Is it necessary to hem exposed edges? My coach says its a necesity, but looking at others’ designs, I have never seen anyone hem them. For those who do use sheet metal, does your manufacturer just do it automatically?
A question for those of you who are familiar with CAD simulations: if I want to test the strength of the frame (not the gearboxes or wheels), what would I use as my fixed geometry? For example, the bottom of the frame, or the faces of the holes the axles contact. Also, what kind of forces can you expect in an FRC match? My calculations suggest that a collision produces about 150 lbf. Is this realistic?
Thank you for taking the time to look at the model and provide feedback.
Finally, a big thanks to Teams 228 and 148 for posting their designs and CAD files from previous years.

  1. Looks really really nice.

  2. We have used .050 sheet metal with no lightening holes in the past with no issues (many of those robots lacked side bumpers, and are still fine). Have you looked at the weight difference of using a thinner sheet without lightening it?

  3. Seems slow. Stock SS’s are geared for 6" or 8" wheels, so with 4" wheels it will be a bit slow. You could re-gear the SS (modifying the last stage would probably be the easiest way to do it), or use 6" wheels.

  4. Instead of the bolts holding the two sides together, you could use a sheet metal plate on the top and bottom. It might be lighter.

  5. If the inside width of the two rails is about 2.5 inches, you can take an AM shifter and use the side rails as the side plates of the transmission.

Tips I’ve learned from James Tonthat (1477) and Sean Cantrell (2415):

If you are fabricating the sheet with a punch, you should use the minimal amount of dyes possible for lightening patterns (228’s frame is a great example of this).

Your flanges look pretty large (might be the picture), I learned it is good to standardize the flanges to .5" and .75" depending on the application.

To make the frame more rigid, a 1/8" cut bellypan such as the ones used by 254/968/973/1323 would be greatly useful.

I’m still learning a ton about sheetmetal so I’m just passing on tips that I’ve gotten. Looks good.

It looks like they did this - all of the lightening slots could be cut by the same punch.

Nope, I was talking about the truss pattern, looks like the ends could fit the same slot pattern as on the rest of the rails.

You shouldn’t need cross supports. My team made a very similar chassis with 1/16" wall 5052 aluminum on the sides and .08" on the front and back. It was very rigid before we added a belly pan. The belly pan is great for mounting electronics, and also serves to add structural rigidity.

I love this design. It looks great!

As a response as to whether cross supports are needed:

Probably won’t need dedicated cross supports for the purpose of cross supports. This may seem confusing; but once you start building off of this frame (i.e superstructure, arm, etc), these new components will act in the same way.

I would say a relatively light (in mass) and strong (against bending) piece across the bottom becomes useful when designing mounting locations for the electrical/pneumatic components.

I think the frame looks well designed, but as always room for improvement. Design is iterative and you can always find ways to lower weight and time for machining. Now that you have something to start from, tweak it to make it better.

Also, you may want to do some research into Finite Element Analysis (FEA). It allows you to determine the stress throughout the design. Precaution: FEA is a difficult skill set to learn. You may want to ask a mentor that is familiar with it to help you. This will really show where the design is poorly designed and overly designed from a structural standpoint.

Keep up the work

i have been working on a (somewhat) similar process with tube and rivets. i identified 4 different drive setups, and designed a frame that could accommodate all of them, with minimal modification. beyond that i saw that there are 2 primary types of manipulators, those that require a large footprint inside the base, and those that don’t, and i set up the frame to accommodate any amount of manipulators within the frame. so my question for you is, how would you adapt this frame style to a variety of challenges?

Flanges - It looks to me like like the flanges are around an inch long, our rookie year, we used 1" flanges and thought it was overkill, this year we’re using 0.75" flanges. Like what Stogi said, we use two basic flange sizes, 0.75 and 0.500. This just makes the brake operator’s life a lot easier, and it looks like you’ve done that.

Gearboxes - I’ll say that using flanges to make the sides of the gearboxes is a lot harder. I have no doubt that our shop and your shop can make those flanges the right size, it just seems to me using bolts and spacer/standoffs is a lot easier.

@Stogi, while yes, it would be better to use standard punches, their shop may just have waterjet or laser cutting. You can punch those lightening patterns out if you had indexable tools. One of our sponsor’s shops has a punch and we even designed our parts so it can be punched out by them but they ended up throwing it on their laser since it was so much easier for them to nest it and just throw it on and cut, not worrying about what tooling they had in the punch and tool paths or anything.

Overall, it looks like excellent work. It looks like that gearbox shaft isn’t supported by a bearing but I’m sure you’ll fix that :wink:

apalrd:
I have actually not looked at other widths of sheet metal, with or without lightening. As for the speed, I’m not sure where I would get gears to replace the final stage to get comfortable 6:1 and 15:1 ratios (Up from the current 9.4:1 and 24:1). McMaster’s options seem to require more machining than I’d really like, so I’d appreciate suggestions.

Akash:
With the issue of the lightening holes, my team has not yet contacted possible sheet metal shops to inquire about a possible donation/sponsership, so I am unaware of their capabilities. And yes, the flanges are actually 1" wide, so I’ll cut those down a bit in the next iteration. As for the bellypan, won’t a 1/8" thick one be fairly heavy? I think I will add the 1/16" one I had back in though.

tim-tim:
I am trying to learn how to use SolidWorks’ FEA tools, but unfortunately, I do not have a mentor that I can ask about it because none of them have any CAD training - I’m entirely self taught. If anyone here has some suggestions or tips about how to learn, I would happily give those a try.

Hawiian Cadder:
I think that this frame is already quite versatile, as the wide-open space in the bottom can accommodate either a manipulator or an electronics board. In addition, the pre-punched rivet holes provide ample opportunity to mount manipulators above the drivetrain entirely.

James:
The gearbox shaft is actually already supported in two locations, and from what I’ve heard, adding a third bearing is a bad idea. In addition, my team has cantilevered wheels off of toughbox and supershifter shafts every year now and never had any problems with it.

Finally, I’d like to re-ask my question about hemming: do other teams that use sheet metal hem their exposed edges in the interest of safety? Is cutting your hand open once the pieces are on the robot even a concern, especially with sheet metal of this thickness and chamfers on corners?

Thanks for the responses, criticism, and compliments.

Buy AndyMark gears a la carte, and deal with getting the correct ratio via the third stage. Most common two-speed gearboxes in FIRST are geared for about 5/13 or 6/16 ft/sec. We were almost able to completely eliminate the third stage reduction this year, but ended up using a 40:28 reduction mostly to space off the shifter shaft from the output shaft (to avoid alignment issues from having three bearings).

We always leave 1"x1" square holes in the corner of our frame (see photo below), to allow 1"x1" box tubing superstructure to be added to the robot, even if we don’t need it. This year having these mounting points was very useful for our redesigned Minibot deployment mechanism.

Stick with the single flanged gearbox plate if you can. This is the third year we’ve been using this style gearbox, and they work great:

http://www.team228.org/gallery/142/slideshow/img_252c5-4dee2.jpg](http://www.team228.org/media/pictures/view/6843)

If you’re working with 3/32" or thicker, you don’t need it if they deburred the parts. Many sheet metal parts are automatically deburred at the machine shop by running the flat sheets through a brushing machine; this adds a cool brushed look and eliminates most burs. The more important thing through is to eliminate all of the sharp corners on your sheet metal parts. I usually specify an 1/8" chamfer to knock them off.

If you’re working with thinner sheet metal (up to 0.062"), particularly varieties of steel, you should hem the edges if the entire part/assembly is heavy. I did some work a few years ago with large stainless steel outdoor weather covers for industrial electronics and controls, and the edges on those parts had to be hemmed or you’d slice your hand open lifting the covers.

Can you tell me where to find these CAD models? I have found a couple, but not much.

A note on FEA:

Just because it is in your program doesn’t mean you should use it.

Unless you have someone adequately trained advising and explaining how it is done, the information that the program spits out can be unreliable. You have to know how to structure your mesh properly, choose appropriate elements, and know the limitations of what FEA can and can’t reasonably do.

I recommend that you try to figure out other ways of calculating part deformations in your robot if possible, although sheet is often a problem with that. Given the size of robot parts, I suggest you get a sheet sponsor who really believes in the education aspect of the program to make you some test pieces such that you can guage the strength of different profiles before you build your final robot.

Beyond that advice, you can always overdesign and iterate down until its lighter.

Autodesk has an awesome (new?) program out: Algor. I worked with our engineering mentors, and it gave similar results to the analysis done on their program (blanking on the name). Also, I then re-ran the simulation in Inventor and got the same results. This was with 3/32 and 1/16 sheet.

Any mechanical engineering mentor should be able to help you out with Algor should you decide to use it.