http://i.imgur.com/PbWR4xk.png
Practice CAD for an offseason prototype. Team 3266 has always used the kitbot chassis in some form every year and is finally moving toward our own chassis design.
Special thanks to Justin Foss from Team 2168 and Dave Powers from Team 228 for sending me your robot CAD’s. It’s been a HUGE help!
Specs:
Any ideas for improvements?
Looks good! Definitely see the 228 influence in this design. Some suggestions:
Good call on the standoffs between the rails, to provide some intermediate support. Using them to double as chain tensioners would be a logical next step.
Any particular reason the ball shifters are up so high? Lowering them would help CoG immensely, and open up the possibility of doing a direct drive. Live, direct driven center axle with structural dead axles in the corners is a great way to do this style drive. Doing a 228 style “S” shaped inner rail can help facilitate mounting gearboxes directly to the chassis rail, although this would sacrifice the non-unique inner/outer rail pieces, which is a very nice touch to the current iteration.
Related, the ball shifter mounting looks like a potential weakness. If you look at your frame from the front, you’ll notice that the rectangle formed by the chassis, ball shifters, and upper cross member has a sheet metal bend at each corner. This structure will be susceptible to parallelogramming, as is any structure that puts leverage on a bend like this. A threaded cross tube like used on the kitbot would be a better alternative.
The hole pattern on the flanges looks irregular and uneven in parts. One of the greatest things about sheet metal is the ability to standardize things like this, so you can mount parts readily. My team does a 1/2" spaced 5/32" hole pattern, which allows parts stamped with it to be used for #8 through holes, #10 tap holes, or 5/32" rivets, without any additional drilling.
To save a bit of weight, you can go with an uneven flange on the front and back C channels, without sacrificing much structural integrity. Here’s a good example.
What thickness is used? 30lbs sounds a bit heavy for what’s pictured, I’d bet that you could get away with a thinner gague metal, and probably more aggressive lightening as well if your sponsor has laser capability. As was highlighted in the 1114 inspired chassis thread, you can use bumpers to reinforce a chassis.
Just a piece of general advice from a team that did sheet metal for the first time this season, work closely with your sheet metal sponsor to learn how you can design parts to make life easier on them. We ended up having to redo a lot of parts because the bend radii our sponsor uses was different than what we modeled.
Only a few things, overall, it looks really good. I especially like the way you make the side panels (the ones that hold the wheels) non-unique parts, along with the detailed design of those sheet metal parts.
Someone with a bit more experience than me in sheet metal design would probably be better qualified to answer this than me, but could you make the chassis a bit more stiff in a few areas? I’m mainly looking at it being a large rectangle with braces only at the corners. I feel like it might have a tendency to turn from a nice right rectangle to a parallelogram.
Also, are you doing exact c-c with the wheel chains?
I’m digging the design overall. It looks like there’s a lot of smart detailed design in that frame.
Looks like a nice chassis. However, here are some questions/issues to consider:
Are your center wheels dropped? It’s hard to see from the render…
If they are, is it enough to turn (you’re currently back heavy - the robot, as is, will have a tendency to essentially be a 4WD)?
Again, the bracing - it looks like a tap at the corners going sideways will collapse your chassis.
I noticed an unusual pattern in your front bracket - the hole pattern on the top doesn’t seem to be regular. Design feature? Also, if, for some odd reason, you plan to mount a structural element for the rest of the robot right in the middle of the front bracket, what prevents the bracket from collapsing down?
Is there provision for chain tensioning (that’s the question that kills my chassis designs over and over again)?
Is there a particular reason for the non-direct drive design choice? You now have a single point of failure in the chain going from the gearbox to your first wheel. Remember that when you use master links, you need to design for a FOS of at least 6.
Does the chassis need to that thick? Having the wheels in a well is good for aesthetics, but when it comes time to maintain the drivetrain, that’s just a hassle. As you make the chassis thinner, you reduce the amount of chain you need (thus removing a lot of steel from the drivetrain). You also reduce CG, and increase the amount of vertical volume for the rest of the robot.
Great looking drive train. This looks ALOT like our chassis from this year. We used the same wheels and transmissions and the sheet metal looked just like that too. Though we painted it blue, which is clearly better 
.
I agree that the transmissions can and should be moved down. Put ours in between the flanges on the inside plates. They are the same orientation as you have them so just move them down and you should be good!
Don’t direct drive these to your 4" wheels! That is wayyyyy to fast and has no torque. You will need to drive them with sprockets from the transmission. Ours were 12T/15T and 22T, we switched halfway through the season.
Regarding strength: what the heck are you all saying? This chassis is going to be EXTREMELY rigid! I know this because ours was the exact same chassis (almost) and it was the strongest chassis I have ever seen or built. The supports in the rails is a good idea also.
The thickness is fine, we used 1/8 sheet on our chassis and it kept it very strong! Think of it this way, if you don’t have a working drive train ob your robot, you’re gonna have a bad time. Make your drive train the bat part of your robot.
You don’t need to change much. It’s actually weird how close it is to ours…
You can also get a little more rigorous with the lightening pattern. Try triangles and maybe even isogrid. (Unless your sheetmetal sponsor cannot do these types of cuts.
You’re going to want a bellypan to make your chassis stronger and less prone to warping.(Or add cross members) I know FRC 2168 makes their bellypans into their front and back members as one piece. If you do this, make sure your sponsor can do the exact bends that you need. You don’t want gaps in your assembly. Another tip on the bellypan, I know 2168 does this as well as many other FRC teams but on the bellypan make a cut on both sides so that both gearboxs can slide out the bottom of the chassis without having to remove the bellypan or superstructure. I hope Justin doesn’t mind me posting this but this is what I was talking about with the bellypan.
http://1.bp.blogspot.com/-cGjobU8Ro1M/UQyBb_UUvoI/AAAAAAAAASQ/mPXwpvXU470/s1600/2013-02-01_21-04-35_265.jpg
Most of us (I think) in talking about the weakness of the chassis are talking about the lack of a bellypan. Without that, the rigidity of the chassis is determined by the play in the fasteners holding the chassis together. Now, you wouldn’t notice this play in competition, because you have a superstructure on top, which probably took out that play. Or, you were lucky, and didn’t stress the fasteners enough to have them slip.
I left the ball shifters high so the inner and outer channels could be non-unique to try and cut down on manufacturing time.
…but I didn’t think about the higher CoG, that’s a great point. I’d rather increase manufacturing time (which isn’t much anyways) than sacrifice performance.
The hole pattern is actually consistent, for some reason the render removed random holes. On the front/back channel has .192" holes spaced out 1" for 3/16 rivets. While the side channel alternates between .192" and .257" spaced out 1". That way we have the option of 3/16 rivets, 10-32 or 1/4-20 bolts.
And the entire chassis is made of .100" aluminum. The reason being that our machining sponsor has that readily available. And our drivers have always played aggressive defense when needed so I don’t mind using thicker metal
Yes the chain is all center-center distance, and I planned to use idler sprockets as tensioners to make up for stretching. But the idea to use the cross bracing as tesioners would be much more efficient. I’ll re-design and add cross-bracing tensioners.
Yes the center wheel is dropped .100"
I chose not to direct-drive one of the wheels because my target top speed was between 16fps and 14fps and directing driving the wheels would have a calculated adjusted speed of 20.5fps! Way to fast for our driver to handle! 
Could you explain more about the thickness of the chassis? I’m not quite sure what you mean. and I’m unfamiliar with the term FOS, what does that mean?
Thanks everyone for your help so far!!
FOS=Factor of Saftey. Basically, chain working strength is less than spec sheets say, due to the master links.
You can do a direct drive and slow the top speed down using another stage of gearing. A good way to set this up with the ball shifters is to split this final reduction between the gearbox and the chassis, with the driven gear cantilevered off the wheel’s shaft, remaining with the chassis when the gearbox is removed.
Also, it’s certainty possible to keep the non-unique inner and outer rails and do direct drive, 228’s S rail is only one way to do this. PM me your email and I can send you our CAD from this past year which did this.
With 0.1" aluminum, you can definitely get away with way more aggressive lightening. The hexagonal pattern on 2168’s chassis is a great way to do this.
By thickness, I refer to the height of the chassis. From what I see in the render, the top of your wheels are under the top of the brackets that make up the chassis. My worry is that when it comes time to fix something with the drivetrain (if it happens), you’ll have to reach down between those brackets, which may be a hassle to do, especially because you’ll have the rest of the robot on top.
That really isn’t an issue. Having the wheels below the height of the flange is actually a nice thing to have, as you can mount things (I like pneumatics tanks because they are long and skinning) right above the wheels on ridiculously thin lexan. If you’re worried about access, either access underneath or use velco for the lexan (which is what my team did).
If you use the three stage ball shifter you can get very similar FPS with either 6" or 8" wheels depending on which output gears you choose. That’s what we are hoping to build with our test robot next fall.
The chassis we built that was very very similar to this one was very strong without a belly pan. Though we did put one in just because we needed to put electronics on top of it, the chassis was very strong without it.
