The socket head linked may be better known as button head if you are purchasing from McMaster-Carr or Copper State, as noted on that product page.
Related, we’ve used the AndyMark perforated polycarbonate for the last two years on our Robot in 3 Days robots, and it’s good stuff. Very easy to attach to a kit chassis if that’s the direction you’re going.
It can be done. Normally it’s due to a series of impacts over time that start cracks and then one final break when they all link up. Here’s some pictures of Apex’s 2016 belly pan, which cracked through on the 107th match. In this particular case, at some point in time prior to the event the aluminum angle that had been supporting the battery got bent, resulting in the battery resting directly on the polycarb. As we went over defenses it was basically a jackhammer, and eventually it gave way in spectacular fashion. And yes, it was definitely polycarbonate.
high density polyethylene plastic .11 inch thick super strong and cheap. $19.99 a roll 22 inches wide by 10 foot long. non conductive and the plastic slid awesome over the outer works for stronghold. we had rivets bolts and screws break off with no tears breaks or cuts to the plastic.
Why is your image so large? Too big to look at the whole thing without opening in a separate tab.
Sorry. Took it down.
That looks like what I would expect a polycarb failure to look like. As you said, one or more cracks that expanded and linked up until it split. But I wouldn’t call that shattering - it just split into two pieces. I can believe the forces it was under… In 2013 we had a thick aluminum bellypan supported by 1" angle, and when we fell from the top of the pyramid it would bulge and bend like it was paper. Remove the battery and sit there with a mallet for 10 minutes to get it back where it’s supposed to be and we were good… I can only imagine how bad it would have been with polycarb. On the plus side, it acted as a cushion to help protect the battery from those big falls 
The only time I’ve seen polycarb shatter was when additional chemicals were applied. Specifically Locktite. It works on a chemical level to weaken polycarb, and you’ll see spiderweb cracks start to appear radiating out from the hole where it was used. Overtighten or give it a decent impact at that point and it’s done.
Polycarb has poor fatigue strength and is bad at preventing crack elongation. We use polycarb for combat robot arenas and while failure is uncommon when it does happen it often looks like what was posted above. That’s why using two layers of poly is now the standard.
No problem, plenty of us have done it before. Thanks for changing it to a link instead. CD images can’t be effectively scaled to display in posts unless they started as such. This is why I use imgur, it has a built in scaling modifier you can add to your links that I find super convenient.
In 2016, DiscoBots used rivets with large heads to attach the plywood upper structure and the polycarb belly pan pieces to the chassis tubes.
We have been using 6mm baltic birch for past three seasons, doubt we will ever go to something else. It’s extremely easy to add all our mounting holes straight to the board with our CNC router. The baltic birch takes threads very well, even just using the drill/tap combo bits.
We are experimenting with 1/16" Garolite from McMaster 9910T17, which Adam recommended in a thread. We just mounted it yesterday, so don’t have any testing. In fabrication, the partial transparency made it easy to align on the chassis rails and hand drill the rivet holes.
We are debating between Baltic Birch and Garolite. The tradeoffs seem to be:
Weight Garolite
Rigidity Baltic Birch
Frame stiffening: TBD
Ease of cutting, Baltic Birch. Both are easy to cut, but the fiber particulate from the Garolite are concerning.
Clearance: Garolite - nice for low clearance and components that we use zip
ties.
Ease of mounting: Garolite, due to the partial transparency
We used 9mm Baltic Birch in 2015 and 2016. Super easy to work with, especially if you don’t have CNC tools, and inexpensive.
In 2017 we used 1/16" 7075-t6 laser cut sheet. I was worried it would be too flimsy, but it worked fine for SteamWorks.
I’d use either option again. I also want to play with 4mm dibond
On FRC 865’s 2017 robot we used pocket 1/8th (0.125in) 6061 T6 aluminum. It held up flawlessly through 2 district events, district championships & championships. If you have access to machines to do it, I would highly recommend it.
I’ve been expecting them to chime in but 1640 has a pretty cool belly pan with carbon fibre or fibreglass laminated to baltic birch. It ends up being a hybrid between both materials your comparing.
My team has been making bellypans out of 1/4" polycarbonate for many years now. We could probably get away with 1/8" but the 1/4" adds strength. We also double this as our electronics board since we want something non-conductive to minimize chances of shorting something out. By having 1/4", that gives plenty of thickness to tap and have a good amount of threads that you don’t have to worry about hardware pulling out (even 1/8" would probably work for tapping). It does give up ground clearance but for flat field games, it doesn’t really create any issues.
Unpocketed metal belly pans can easily be .063”. This is the belly pan 2791 has run for many years. If you aren’t pocketing, and you’re using metal, you shouldn’t go thicker than this for the belly pan, it’s very inefficient.
Not 1640, be we at 1339 have done carbon fiber on 1/8” Birch several times, including our 2015 bot and this year’s off-season project. The pros are: light, stiff, strong, very thin, looks cooler than anything else, and teaches students the layup process that they otherwise would not experience in high school. Cons are: expensive, messy, time consuming, possible failures in layup that could result in a wasted board, and working with the fibers safely (and without itch).
We will probably just do 5mm Birch again.
What grade (# of ply’s?) of Baltic Birch are people using for their plywood belly pans?
Four answers already given:
I’d asked for grade/# of plys, not thickness. For the same thickness and type of base wood, there are many configurations with multiple variables including # of plys/layers and type of adhesive used. I’ve found the latter can greatly affect whether you can laser cut a sheet in one, clean pass.