Pocketing Gussets

[cadandcookies]

I was recently looking through some pictures of robots I took for research purposes, and I noticed a rather large variety of ways that teams will pocket their gusset plates for superstructure or chassis-specific components. I ran a quick search and the closest I could find to discussing this was the discussion of 100's offseason robot this year.

As a team that tends to just use rounded rectangles when we pocket our parts (gussets and frame members), I'm curious as to what reasoning goes behind what shapes teams use to pocket their materials, as well as whether anybody has noticed performance (weight/stability) gains from different lightening patterns (specifically regarding gussets, but I suppose it also applies to frame members and the like as well).

So, what pattern(s) does your team use for gusset plates, what materials/thicknesses do you use where, and what has your experience been with these methods?

[AdamHeard]

Decent rule of thumb is that stress wants to flow in smooth, uninterrupted, straight lines.

Think about how your part is loaded (including all twisting and bending) and see visualize if stress is forced to unnecessarily flow around pockets and corners.

A lot of the pocketing I've seen in FRC is actually really, really inefficient. In all fairness, we do a lot of visually inefficient pocketing as well when parts are much stronger than necessary.

[DampRobot]

I wasn't actually the one who designed those gussets, but we have a couple of methods for lightening things on 100.

First, the one you saw in the photo. Basically, we draw out all our critical features, and cut out everything but some "spokes" that run between critical holes and features. It is meant to take load that goes from one feature to another in a straight line like Adam described. Then, we add safety circles around all of the critical features and put in fillets. Perhaps not the most efficient method, but it does look great. A number of west coast teams pocket in ways like this. It also takes a while, so beware.

Second, just plain old holes. Sometimes, when we've got a big long member that we want lightened, we take it over to the drill press and just cut out a bunch of huge holes with a hole saw. If you do them regularly, they don't look too bad, and more importantly, they're a lot faster to do then milling. This is a pretty inefficient lightening method on the other hand. It may not be the strongest either (although I've never seen a break because of the lightening holes on a part like that). This method is good for low resource teams too, as you don't need a mill. We did this on our first iteration climber arms:


Third, don't forget about plain old pockets. For large plates that need to be thick for some reason, or tubing we want lightened, we can set it up in the CNC, and basically take all the material not around the cutout or critical features to about .060". This saves a lot of weight, is faster to design and machine then the first option, and looks real slick (milled alu finish for the win!). Check out this for an example:


I have to remind you, the number one way to reduce weight is to reduce material thickness. If you've got a frame made out of 1/8" 2x1 and you change it to 1/16," you're basically cutting its weight in half. We aren't too gentle to our frames and we've never had a problem with 1/16" thick tubing. Also, very few plates on a robot need to be out of 1/4" plate, or even 1/8." Our bellypan was all 1/16" alu, and is perfectly strong. We probably could have even gone with 1/16" thick gussets instead of 1/8" too.

Most material isn't doing much. Lightening removes that material. Structure you're lightening patterns as such.

[Seth Mallory]

I looked at team 100's side plates today and they look great powder coated. I would recomend that teams with a CNC mill give it a try. And if you are good you can do it on a Bridgeport. Very sharp looking.

[DampRobot]

Originally Posted by Seth Mallory

I looked at team 100's side plates today and they look great powder coated. I would recomend that teams with a CNC mill give it a try. And if you are good you can do it on a Bridgeport. Very sharp looking.

Thanks Seth, it was a pleasure having you over.

[Mr V]

The strongest shape is the triangle so rather than cutting out large rectangles some smaller triangles would typically be stronger.

Take some Popsicle sticks and attach them together at each end with a single pin. In the shape of a triangle you can not cause the shape to rack w/o causing the fasteners or the sticks to fail while a square/rectangle will rack and collapse without much force.

[cadandcookies]

Originally Posted by Mr V

The strongest shape is the triangle so rather than cutting out large rectangles some smaller triangles would typically be stronger.

See, this is the part that I'm a little bit fuzzy on-- I'm well aware that triangles are the (logical) strongest shape to use, but what I'm iffy on is whether there's a significant increase in strength in FRC-specific applications. MY team has always approached it with the view of "we have limited machine time and expertise, and our rectangles haven't broken yet."

We may have to experiment with this during the fall...

[MichaelBick]

Originally Posted by cadandcookies

See, this is the part that I'm a little bit fuzzy on-- I'm well aware that triangles are the (logical) strongest shape to use, but what I'm iffy on is whether there's a significant increase in strength in FRC-specific applications. MY team has always approached it with the view of "we have limited machine time and expertise, and our rectangles haven't broken yet."

We may have to experiment with this during the fall...

The best method we have found is to use thinner material/wall. On thicker parts we resort to lightening on a waterjet(thinner material still costs less and takes more time, so that is default regardless of the fact that we waterjet all our gussets).