As I am relatively new (started in 2017) to FRC, I have never designed or tested a drivetrain to cross bumpy and difficult terrain. My team has some struggles in 2016 with a modified KOP chassis, as we could cross the obstacles but could not turn very well. I’m looking for any tips or processes about drivetrains that can deal with this kind of terrain, more specifically for tank drive. Any examples would also be very much appreciated 
2016 was the most recent year with difficult terrain and arguably the year with the worst terrain. I highly suggest you look at videos and robots that did well that year. The Blue Alliance would probably be a great place to start.
Team 254’s technical binder from 2016 has some nice models of their drive train. It’s very simple, just tubing and axels with a belly pan. There’s no need for the custom gearboxes where you’re at now, but slap something like evo shifters on it and get going.
Just make sure there’s very little gap between the wheels and don’t try to make a perfectly designed out drive train with gaps that “shouldn’t ever get stuck in theory.” like we did and then, of course get stuck on the moat.
So I know in 2016 254 had an 8 8" wheel design which worked really well, and 1538 had a 10 6" wheel design which worked as well. Is there any benefit to going with more 6" wheels as opposed to less larger wheels or vice versa? I have also seen some teams using a steel bellypan, is that necessary or is 1/8" aluminum fine?
2016 we didn’t add wood rails until after week one seeing another team use them but it really smoothed out the crossing. Our large pneumatics wheels, 10" iirc, really made crossing most of them not to bad. I think unique designs i.e. rocker bogey style might be wasted effort. I will say the 8" 4 wheel designs could probably cross more gracefully which was needed somewhat for autonomous.
Larger single steps can be overcome several ways, see 2004. The multiple step problem is something that is a little more vexing.
edit: I can’t recall too much issues with bellypans, but maybe others remember. I think we were high enough it didn’t matter.
edit2: Here you can see our robot that year (and the rails underneath). https://www.youtube.com/watch?v=OpEfWgsq5HU
One of my favorite reveals in 111 2004 robot and how it when smoothly on and off the platform: https://www.youtube.com/watch?v=NHw0j0lFBnwSo, here’s some tips that I’ve picked up over the past few years…
-
Use pneumatic wheels - Make it easier to cross terrain and makes the crossing easier on your drivetrain.
-
Add as many wheels as your length will allow - Whether it’s an 8WD with 8" wheels, 10WD with 6" wheels, etc. you want to prevent your chassis from bottoming out over the terrain.
-
Use a larger drop for your center wheels - With pneumatic wheels you typically need a larger drop. I would say 1/4" at minimum.
-
Make sure your wheels have a leading edge - If a piece of terrain is particular tall, you can risk your frame running into the terrain first. So give your pneumatic wheels the leading edge and let them pull you over the terrain.
In 2016, it seems like 8WD with 8" wheels had a pretty smooth time crossing defenses, but many successful teams had 6WD with 8" wheels as well. As long as your frame has enough clearance, you can use whatever setup best suits your design.
In 2016 we used the kitbase with the 8" pneumatic wheel kit for crossing the defenses. As long as you have clearance and the wheels filled up you should be good.
After out competition we added more leading edge to the front wheels and added wedges as well. As you can see here
The only crossing defense we had issues with was the rockwall, and it was mostly because of the space between the wheels.
Take a look around The Blue Alliance for more teams and what they did that year. I know quite a few teams used more smaller pneumatic wheels or a combination of pneumatic and colson wheels and those solutions worked well as well.
With AndyMark’s release of 8" pneumatic wheels with 1/2" hex bore, I am interested in how WCDs would handle a terrain-based game. Considering that the product page calls it a “perfect choice for building a WCD style drivetrain”, I’m wondering if any teams here did a WCD in 2016 with pneumatic wheels. Are there any special considerations that you implemented to modify the WCD? I’m mostly concerned about the cantilevered axles going over terrain over the course of a season, but I’m not sure if that’s something I actually need to be worried about.
Pneumatic wheels can handle just about anything as Team 3875 proved in this match:
We ran eight 6" pneumatic wheels in 2016 with no issues due the cantilever. We did modify the drive base, putting shock absorbing springs with about 1" of travel on both ends. In hindsight the shocks were an unnecessary bit of extra work, it would’ve been fine without them.
We did some careful planning with our wheel spacing in 2016 to ensure that there was no position in which we could high-center on any of the obstacles. It resulted in an odd frame length, but the effect was great: We had no fear about hitting any of the obstacles at speed.
Our wheels were cheap: 8" harbor freight pneumatics, with press-fit hex hubs that we spun on our lathe. They were extremely heavy, but they allowed us to bounce over obstacles with no issues whatsoever.
4901 ran a chain-in-tube setup with the center direct-driven, but chose to support the other end of the 8" pneumatic wheels due to the loads. A couple teething problems with chain at Palmetto, but once dialed in it got through Orlando and SCRIW and demos with little maintenance.
So, the inevitable question, What about tracks?
We ran tracks with great* success in 2016 (once they were sorted out.) The key to using tracks in a robot and it still being able to turn well is 2 fold. First you still need a drop center, though not as large a drop as pneumatic wheels, (AM rhino tracks have this) and you need a very low center of gravity (don’t build a tall robot and put your battery on your belly pan)
The main advantage of the tracks is smoothness of crossing and the lesser need for ground clearance (in a situation where the obstacle can be mounted by both tracks at the same time)
Some disadvantages though, it does require more energy to turn, even if it has a low CG and is stable when turning. (as compared to a standard drop-center 6 or 8 wheel drivetrain) It takes up more space inside the robot, it is heavy, and the tracks may want to jump off if they are not properly tensioned and guided.
That said, I would choose them again for the smoothness and ease they afforded us getting over the defenses.
EDIT: and please just use belts similar to those sold with the Rhino units (we used Brecoflex belts that we had lying around). Those tracks that are sectional and you put together however many links you want do not have much traction and are not smooth at all.
One word - SWANK
254 had a pretty sweet drivetrain in 2016. Well, they do every year, but in 2016 it was just super KISS. 8 inch wheels, simple wheel spacing, and a slight raise to the front wheel. All on a 2x1 WCD. Slap your own gearbox on it, modify the size and be done with it. I like to follow Corsetto’s golden rule number 1.
For a lot of varied terrain, it’s a matter of clearing obstacles in various states. When it’s just one ramp or obstacle, you can usually do this by changing wheel sizes, paying attention to spacing, etc. without having to get too weird about it.
The important thing to remember is to design the drive to go over the obstacle at all angles - NOT just head-on. It is relatively easy to make a drivetrain go up a ramp straight on, but it can be a lot harder to handle driving at an angle or even riding sideways along the slope.
2016 was almost a worst case scenario with all the different terrain options - and I think that year demonstrated that you essentially never have to resort to tracks. Ultimately you can always add more wheels, vary their size / spacing, etc. In 2016 there were a lot of 10WDs with 6" wheels, which filled the driveline out so you couldn’t high center on stuff. The 8WDs with 8" wheels had a similar principle.
We used a Rhino chassis for Stronghold. Had absolutely no issues with terrain, robustness through competition. We didn’t have the technical expertise to hang but ended up in the quarters as a tough support team mate. I had to get my team to understand it took more than a “great” bot to get a team win. That’s a win in my book!
We will be using 8" pneumatic wheels (8 wheel drive) on a West Coast chassis design this year using a custom designed gearbox based off Vex SS.
We’ve just rolled out a new, separate all female team, so standardizing design and parts between veteran and rookie teams is our major hurdle at this point. We are mentoring our rookies so a mutual understanding of how things work is paramount this year. The logistics are killin’ me. The team is 7535, Purple Poison! It will definitely be a bit weird facing our rookie team at the district comp in Dallas. 
Or, for those whose build budgets budgets weep at such thoughts, you can also shrink the robot. Which is not to say a clipped-corner, 88" frame perimeter drivetrain would’ve been great for some alternate universe like “2017, except fuel matters”. But that experience did sell me on the merits of building smaller than the rules permit to ensure things clear.
Clipping the corners also proved valuable in some situations where we got sideways on the defenses; some teams that built to max dimensions couldn’t spin out because a corner would hit the divider.
We used Rhino Tracks in 2016 and weren’t a huge fan. Part of this was our own fault. We chose the tracks since we didnt have the experience of custom tank drives. The team was basically green that year and didnt under the entire bumper backing and perimeter rules, which we ended up having to make a new sub frame last second around the tracks.
We also mounted the tracks to our frame so we had essentially no ground clearance so we weren’t able to get over the ramparts or rough terrain. They did work well at keeping us from getting caught in the moat and rock wall, giving us 0 issues in that aspect.
On the other hand, I know we snapped 2 or 3 belts during season, constant issues with the 5mm belt off the gearbox to the pulley, and we ended up having to modify the tensioner to keep the tracks from loosening throughout the match.
If the game this year is anything like 2016, we are probably going to use pneumatic wheels to save money and since we actually know how to build a custom drivetrain now.
See also: Twerve