I have been working on this chassis a few weeks now and I think I have finally gotten it to where I am ready to show you guys. I tried to base this chassis completely off of 1114’s 2013 chassis. All my work was done by research of pictures and videos . I have not actually seen the chassis in person so keep that in mind. Well the only thing I haven’t really figured out how to implement into this design is the chain tensioners. They basically are a bolt with a nylon tube around it inside a slot pushing down on the chain. Everything is kept simple just like on 1114’s actual chassis. Well, without further adieu enjoy!
P.S. I also I have a sorta remake of their climbing arm in the works. I am trying to model it but it is not as easy to do as the chassis.
Just remember, if you need more pictures or information for reference, never hesitate to PM Karthik or another team member for help. On several occasions they’ve provided CAD files and screenshots of CAD as well.
From what I can see, that is the least rigid sheet metal chassis I have ever seen.
The front is open without any stiffener across, and the back had a nice flanged back plate but you cut the flange out.
Something has to bridge the two sides and provide stiffness, or the chassis dynamic performance will be quite bad.
I’m sorry? Did I forget to mention it is unfinished?
You posted an image of a chassis design on Chief Delphi. The Chief Delphi community has seen many chassis designs, and helps many designers such as yourself improve their designs and methods.
It’s generally frowned upon to copy a design without understanding the engineering behind it. You can understand the engineering behind it by designing, building, and testing a chassis and learning what is right and wrong, and improving it for the next design iteration. Frequently, by doing this, you will learn why they did something a specific way by not doing it on your own design and wishing you had, or finding that the solution works fine in testing and accepting it as good enough. It’s perfectly fine (and highly encouraged) to use another teams design as a starting point when designing something new, but to gain anything you have to make it your own. Short of actually building and testing a physical chassis, you can ask experts and teammates for advice during the design process to iterate to a final design, before possibly building and testing the final product. Chief Delphi includes many long-term FRC members who have designed many drivetrains, and some engineers who have learned the common design pitfalls and are willing to help you.
As someone who has built and tested many sheet metal and other types of drivetrains, your design needs more torsional stiffness. This is most easily provided by adding cross bracing.
You did forget, and if you post a design or CAD on Chief Delphi that is “ready to show”, expect critique whether it’s finished or not.
Thanks for all the support guys! (no pun intended) I appreciate the help, I updated it so it should withstand anything now. Built like a tank win #sostrong #support #yolo chassis aluminum ##
is that a solid metal block. If so you may run into some weight problems.
Let me preface this with the fact that I think that it is coming along great… But(thus letting you know that everything else will be critical of your work)
Have you put any thought into how you would build this chassis if that was going to be an offseason project or something? To me, it seems pointless to design a robot that is impossible for YOUR team to build. Not just a hall of fame team, or a world champion team, but your team. What you have right now looks like it is un-manufacturable to me, but maybe it is just the way that I am looking at it.
If I were you, I would try and figure out what it is that you are trying to learn from this experience, is it just to learn the concept of CAD? Or is it to get practice for the build season? Or Is it to learn how the world champion and hall of fame teams build robots and why they do certain things to their robot?
Once you establish what it is that you might try to learn from the experience, it will guide you as to how you would proceed from here. I did a similar thing last fall. I was eventually just told by our lead mentor to stop doing that and make a robot that we could build, (given only a spec, and a general construction technique) and because I gained more of the experience of how to make a robot that my team could build it helped A LOT more in the build season.
Oh no this is something that I am doing for practice and also for a chassis for a team potential standard go to chassis, (Minus the cut in the front) I built this all using the sheet metal tool so idk how it would be not build-able? care to explain how it wouldn’t?
Just because you can make it in CAD doesn’t mean it can be manufactured (or that it would be sensible/economical to do. (i.e. Would building this chassis require an insane amount of wasted metal?))
You would need some good sheet metal breaks to be able produce all the bends on the side plates with them being so close together.
Exactly. I have seen parts in CAD that would be impossible to build without one particular piece of equipment.
Here is what I’m seeing that is probably a waste of material, and what I’d do about it:
–Those two solid (endcapped?) cross-beams. You don’t need the ends on there; go ahead and remove them (they’ll be a REAL pain, even with sheet metal). Better yet, replace them with a bellypan and some much lighter angled upper crossmembers. If you do need only upper crossmembers, then make them as small and light as you can get away with–and probably not a full tube, unless that were to make some other aspect of the design easier.
–It looks like you have a solid sheet area below the rear cross-beam; that could probably be removed. Ditto for part of the front that looks quite similar.
–No light-weighting. Yes, this is sheet metal–but you can still remove quite a bit of material on the sides of the robot. Triangles are your friend.
Essentially, what I’m seeing is that you could use some studying in the direction of how to effectively use sheet metal. I don’t have a lot of experience myself, but the one thing that I keep hearing is “flanges” are what give sheet metal its strength–you’ve got flanges, all right, but you also have a lot of material that probably isn’t contributing a whole lot to the structural integrity of the drivebase.
Is there a purpose to front gap other than copying it from 1114’s chassis? If you took out the front gap and made it like you did in the back minus the cutout in the flange you’d have a pretty solid frame.
Side note: because of your cim cutouts in the bellypan and the flanges on your frame it will be impossible to either install or remove the gearboxes once the frame is assembled. I suggest elongating the slot for the cim to allow space for the gearbox to slide out from under the flange. Details like this are often missed in cad, make sure you put some thought in how this thing will be built and assembled.
edit: EricH is right, triangles are your friend, quoted for truth.
Let’s analyze this from the beginning:
-
Material selection - Alloy and thickness
-I will assume aluminum. Most sheet metal designs in FRC use either a 5052 alloy or 6061 alloy. 5052 is softer than 6061, which makes 6061 harder to bend. This is both good and bad - Many sheet metal shops will only work with 5052 because it is easier to bend. If you use a thin 6061 alloy, it will bend just fine. We use 6061 T6, and bend it on a manual finger brake.
-Common thicknesses are 0.050", 0.063", 0.090", and 0.125". Some designs rely on the materials for strength, but good sheet metal designs rely on the profile of the sections (e.g. making C-channels or triangle beams out of formed sheet pieces) rather than the absolute material strength. We use 0.050" for our chassis and 0.090" for our gearboxes, with a mix of 0.050, 0.063, 0.090, and 0.125 for mechanisms. In 6061, 0.090" is hard to bend, but we do bend the gearbox plates. I’ve seen most 5052 teams use 0.090 for chassis components, but I have no experience with 5052 in that application. -
Important forces/loads:
-Most chassis designs I’ve seen in FRC can handle the standard loads from gravity (weight of robot) and straight line acceleration, although point loads from mechanism attachment can bend thin sheet - it’s better to spread out mechanism loads over a larger surface, using rows of rivets. It’s also possible to use mechanisms for stiffness, which I’ll get to later.
-Good dynamic performance of drop-center drivetrains requires good torsional stiffness. This is the rotation of one driveline relative to the other, and turning forces can pick raise and lower wheels such that the exact wheels you DON’T want to be touching the ground during a turn will be touching the ground, which is bad for dynamic performance.
-Any component which is stressed, or carrying loads which you are considering for chassis performance, should be modeled, anything which is not stressed should be modeled only for packaging at this stage. In some designs, the gearbox is stressed (carrying chassis loads). We (33 2013) use our gearboxes to span the chassis outer and chassis inner, as well as bridge the chassis inner rear and front panels (they are not continuous panels, the outers are). A bellypan as used by west coast drives is often thick plate (0.125" or 0.250") and used for material strength rather than as a section as sheet metal designs primarily use. It will help with racking of the two sides, but not will do nothing in torsion at the thicknesses used for sheet metal.
-Mechanisms can be used to increase stiffness of the chassis, especially if they span the width of the chassis. -
Sections
-The strength of sheet metal designs primarily comes from the sections. A flange can turn a flat sheet into a L or C channel, greatly increasing it’s strength. A box is hard, as it is almost impossible to bend unless it’s more than one piece.
-The two chassis rails are almost always C channels. Cross members can be single C channels, but a full box or triangle section is very strong. Our 2012 chassis had two triangle sections crossing the chassis, one of which was also a ball ramp, and was very stiff. -
Folding sheet metal
-Sheet metal is usually folded on a brake. With a manual finger brake, it’s possible to remove fingers to bend tight corners (e.g. your chassis outer panel). Flanges which are bent several times (e.g. a flange off a flange) are usually hard to bend, there are a few exceptions such as the chassis rear in your design, which has two ~45degree bends and should be easy to bend.
-For some pieces which are small, you can bend them by hand with furnace pliers. We made quite a few ~4" square pieces with 4 flanges (one per side), two flanges had rivets and the other two for strength, to bridge chassis inner and chassis outer panels. These were easy to bend on the finger brake. We also once made a piece which connected to itself, this was tricky but we could bend it mostly by hand with furnace pliers. In general, designs which connect back to themselves should be split into two or more panels. -
Joints
-A lot of stiffness can be lost in joints. Most sheet metal is joined with rows of rivets. There are many options for rivet size (1/8", 5/32", 3/16", 1/4"), spacing (1/2", 3/4", 1", 1.5"…) and material (steel or aluminum). We have used both 3/16" steel and 1/4" aluminum rivets in chassis designs and currently use 1/4" aluminum, but most hand riveters only go up to 3/16". I recommend pneumatic riveters, they’re lots of fun.
-If possible, jointing two pieces on two planes is preferred. For example, joining the chassis outer to front panel via the top and front faces is much better than top only, or side face. This is especially important for mid-chassis cross members, as it’s easiest to only attach them via the top or side, but both are required for stiffness.
-There are many cases where an axle needs to be supported by sheet metal. For a light-load application like a conveyor roller, you can just make a hole in the sheet metal and insert a bearing or bushing. For drivetrain axles, it’s recommended to stiffen the area around the hole with another plate riveted to the chassis. This can also help align the axle and constrain it, depending on design. If you make it slightly beefier to hold the entire axle load, you can also slot the holes and use bolts instead of rivets, allowing tensioning. -
Generic drivetrain concerns
-Wheel size
-Wheel tread/type
-Gear ratio
-Power input
-Using required drivetrain axles/components to stiffen the chassis
We often use the drive axles to provide stiffness between the chassis inner and chassis outer. Chain tensioners are also a good candidate for this. If designed right, a gearbox in the chassis can also provide a lot of stiffness between the two rails. -
Manufacturing
Panels are 2d parts and can be cut on a 2d machine, such as but not limited to:
-CNC plasma cutter (not preferred for aluminum, but it does work)
-Waterjet cutter (good tolerance, moderate edge finish)
-Panel laser (excellent, very very expensive tool)
-Turret punch - Limited to specific contours which it has tools for, unless it’s integrated with a laser or other cutter. If you use one, it is best to optimize the lightening patterns to what tools it has - e.g. if it has a 2" round punch, it’s much faster to use 2" round lightening holes than triangles.
-By hand on bandsaw and drill press, this is sometimes impractical for large parts
Panels are then bend into 3d shapes using a brake. This is either automated or done by hand.
-There are many shapes which can be made in CAD which cannot be formed on a brake. Think about the bend order and how you would bend it for complex parts.
There’s no substitute for design experience. No chassis design is perfect, there are many tradeoffs. You really need to build and break it to understand what is wrong with it, but you can catch a lot of possible issues in CAD.
Could you provide a picture of this?
Here are some pictures of the prototype. I don’t have any good pictures of the in-season chassis, plus it’s painted black and hard to see detail on.
The four pictures are:
-Gearbox assembled, upside down. Secured to chassis inner on front face and bottom flange with bolts, significant structure to chassis inner. Secured to chassis outer with four bolts/spacers on front face, bolts on bottom flange to chassis outer flange. I’m not sure we actually bolted the gearbox outer to chassis outer on the bottom flange, we might have thought the four bolts were enough.
-Gearbox assembled, another view. It’s just an AM shifter with 22t output sprockets and new plates. Prototype was servo shifted, production was pneumatically shifted.
-Prototype chassis assembled, with electronics tray and servo shifters. This later saw the bandsaw and became the practice robot chassis (we literally removed the rear chains and bandsawed it to 27" long). Talons also tested, performed admirably without fans.
-Chassis inner assembly without chassis outers or ends. Gearbox inner plates temporarily put in for fit-up, not yet folded. The top hat above the gearbox does very little, almost all of the loads in that area go through the gearbox. In fact, it’s the only thing that bridges the inner and outer chassis in the middle of the span.
There were a few changes between this chassis and the production one, mostly in ride height (added 1/4" ride height) and chain tensioning (slotted all 6 wheels instead of only inner wheels). we also moved the battery for cg and placed the arm gearbox in the battery space.
Just to clarify several things. The chassis Andrew showed pictures of is misleading. He talked a lot about stiffness and then posted a picture of one with very little. That’s because that chassis is specifically designed to reduce weight and increase packaging space by intentionally being flimzy. It makes use of all the manditory things in a drivetrain like the gearboxes and axles for structural support but uses its bumpers (which don’t count towards the robot weight) as the “meat” of the frame providing almost all of the stiffness.
I attached a .pdf to better explain it. I just didn’t want anyone looking at it to get the wrong idea.
Regards, Bryan
8wd WASP.pdf (703 KB)
8wd WASP.pdf (703 KB)
For ten years now, I would often look at 33’s robots and wonder “How the heck do they make weight?!”
Now that I know how that pans out, I’ll have to ask “How the heck do their bumpers make weight?!”
(Mind you, I find their approach fantastic and truly next-level work within the rules. 
)
Does the box tube bumper frame force you to give the robot a smaller perimeter?