I am looking some critiques of this design, or possible improvements. This is something we are considering using a modified version of for build season, depending on next year’s game.
If anyone would like a more detailed view, or a different perspective, I would be happy to upload some more pictures. Here are some basic specs on the design.
Specs:
3 CIM Vexpro Ballshifters
Two CIM’s on each side, one BB-775 geared into a CIM-ile gearbox
3.68x Shifter ratio, 17 fps top speed
Six Colson wheels, all run off of one chain
Center wheel directly driven, one #35 chain run on each side (the blue stuff)
Flanged aluminum and tube chassis, around 38 pounds currently
Mostly riveted structure
Drop down pneumatic “bellypan”
Capability to add a dog shifter for an upper mechanism powered by drive wheels
Consider the momentum of a fast-moving FRC robot. Cylinders aren’t really designed to take side loads. Are you concerned about bending those cylinders if hit?
What do you do if you lose pressure? Will the bellypan drop and prevent movement?
Rs775s have been used on drive trains by teams before, but be careful. Running those motors near stall for a long time will fry them- they are fan cooled, as opposed to CIMs, which are passively cooled and rely on a large thermal mass to keep from overheating.
If I remember right, we fried a 775 when we tried using it in conjunction with the CIM-ile on our gearbox this year. Also, am I understanding it right that there is one chain guiding all three wheels? If so, are you worried that the center wheel may not be perfectly aligned with the chain?
Belly pans, especially metal ones, are generally used to connect members together for more torsional rigidity. If you articulate the belly pan you lose more or less all of the benefits of a metal belly pan but you keep all the weight.
Not to mention, you want to resist pushing matches by exposing all of your electronics to anything at ground level? This can’t end well.
I’ve seen dozens of teams and robots with pocketed bellypans and no plastic sheet underneath that did not have any issues with this sort of thing at all. I don’t think pocketing a belly pan is a bad idea - debris is rarely coming from straight up underneath your robot.
But in this case they drop the bellypan, so stuff can touch the floor.
You’re right, it probably won’t affect anything as the force is mainly downwards. Still, my team’s mentors would throw a fit if we didn’t isolate the electronics.
As Chris has stated, dropping your belly pan as a brake will cause your chassis to lose most of it’s torsional rigidity. One good hit and it will become a parallelogram. You are better off pushing individual pads against the carpet. Several years ago, one of the local teams did this. The pads were about 3 " x 3". They glued the wheel tread material to them. Two adults could not push their robot an inch once they were down.
Omitting the pockets in the gusset plates at the corners will allow them to resist hard hits better.
Do you have a separate picture of the “end caps”? They seem to be integral with the gusset on top and have a flanges that attach to the two plates running from to back on each side. It is difficult to see what they really look like but if they are the way I think they are, it will be very difficult to make the bends.
The plates that the gearboxes are mounted to have the bottom flange removed. There is a lot of mass (motor and gearbox) mounted to the side plate. In a collision, there will be a bending moment that can bend the side plate.
Place real components on your belly pan to check the size of the pocketing. The square holes look like they may be too big, forcing you to space the components out too much or making it such that the electrical components are attached to only one of the webs. More surface area will be needed if Velcro will be used to attach the electronics.
It may be advantageous to omit the holes in the top of the front and rear tubes to allow material so that the upper structure can be attached securely.
The tube across the middle will add more rigidity if it is attached to a plate at each end and those plates attached to the gearboxes in two locations.
Where is the battery mounted and how is it retained?
Is the gearbox easily removable for servicing/repair? Are the bolts that attach the gearboxes easily accessible?
Everyone seems to be hitting the major issues but my other gripe is the lack of radius on your pockets. Adding a radius will make it stronger, make it look nicer, and make it easier to manufacture (Lasers, Waterjets, and CNC Mills can not do non radiused inside corners)
Edit: Other than that is a beautiful drive base! Seriously very good job! I feel like some, including myself, forget to say that.
As for the side loads, it is a concern, but each cylinder only has a stroke of 1" at 5/16" bore. They could bend, but it would take a considerable amount of force. These are relatively easy to replace in this design, though I would rather not have to keep extras lying around, since they’re not too cheap.
Currently, there is no redundancy in terms of pressure loss. For our build season design (assuming the game is conducive to this system) we will probably have four smaller “feet” drop down instead of the entire pan. This gives the advantage of having rigidity from the pan, and less room for catastrophic failure. On that note, does anyone have a suggestion on what to do in case of pressure loss? I was thinking maybe surgical tubing to pull against the “feet” or bellypan with just enough force to lift them off the ground.
I have heard many concerns about stalling RS775’s for too long and frying them. If they are too much of an issue, we could replace them with Mini-CIM’s or just take them out altogether.
This is a picture of the first iteration of the drivetrain which is described in more detail in this thread.
We have discussed doing 4 plates instead of the whole belly pan dropping. Part of the reason why we tried the belly pan idea was that we wanted to see if more surface area of the tread on the belly pan created more friction. The laws of physics says that SA shouldn’t matter but that applies to 2 flat surfaces rubbing against each other, not carpet and bumpy wheel tread.
Even so, the failure mode of this design is a robot propped up that can no longer move. I don’t know the likelyhood of bending those pistons, but I would recommend doing heavy destructive testing on it before you consider using the design in a competition. It depends on “when” in a match you would use this though. If it is part of some end game then I would guess that you would be fine. But if they intent is to lock down an area or hold position for long periods of time . . .
On that note, does anyone have a suggestion on what to do in case of pressure loss? I was thinking maybe surgical tubing to pull against the “feet” or bellypan with just enough force to lift them off the ground.
That is a good idea. Surgical tubing, springs, or the like are definitely good there.
If I were doing any sort of actuated belly-pan, I’d do two things:
1: Don’t actuate the actual belly-pan. As has been mentioned, it’s a structurally important piece that you will likely want to mount a lot of important electronic/pneumatic components on. Putting that in close contact to the ground is probably not such a great idea. Instead, have a separate pan on the bottom that is actuated - it doesn’t have to be very heavy (plywood should work fine).
2: Don’t link it directly with the cylinder(s) you use to actuate it. You’re much less-likely to bend a cylinder if you design it such that your cylinder pushes a cam that lowers the plate, rather than just linking the cylinder to the plate.
I agree, this seems like it would be a good use for single action cylinders, defaulting to the retracted position for cautionary purposes so you never get stuck. You would save air as well.
If you do decide to actuate feet rather than the belly pan, a simple hinged element would remove the possibility of side load on your cylinders. Think of a see-saw where the cylinder pushes on a side that isn’t in contact with the ground at all.
Also, I agree on this being a perfect place for a single acting cylinder. If you don’t have any, a simple spring placed over the rod and one port vented to atmosphere can make a double into a single. You get the unique option of having the opposite direction of your typical single acting cylinder as well.
Pneumatic cylinders directly acting on a plate with a short throw will be fine, no reason to over complicate it with a linkage.
Our brake last year was 3x 1" pancake cylinders all attached to a 20x14 plate covered in rough top. Worked great without ever bending the rod on a cylinder.
I will say that you should consider pancake cylinders instead of regular ones from a packaging standpoint they are much nicer for this kind of application.
