The frame is made of welded 1 in square tube with 1/16 in wall thickness which is bolted together at various spots. It still need some supports.
The wheels are on dead axles. Between the pillow blocks and the frame I can put spacers to change the wheel position. Currently the center two wheels are dropped about an 1/8 in.
Any suggestions on how to improve it?
It looks like you only have 2 CIMS powering your drivetrain. Did you do this to save weight, or do you have another purpose for them? I’m of the belief that there are enough motors available in the KOP that 4 CIMs (or at least a CIM and a FP) should be used on any drive train (with 2K9 being an anomaly).
Replace all the 1/16th wall aluminum with 0.035" wall 4130 steel square tubing. You’ll improve chassis strength, stiffness, maintain the same weight, and your welder will have an easier time.
With a steel frame you could also use weld-in nuts which could save you weight and make it easier to wrench on the robot 
Are you using Riv-nuts to hold on the axle pillow blocks?
Instead of using washers to lower your middle wheels, try drilling the holes in all 16 pillow blocks off center by 1/16", and flip the middle ones over so the axel ends up being 1/8" lower than the outside axels. Adding washers isn’t as clean/sturdy.
Also, to lower your CG (which is ALWAYS important), you might want to mount the pillow blocks on the top side of your lower frame members, that way the whole frame is lowered a total of 2 inches. And try mounting your electronics (Battery/CRio/etc) on the lower frame level, not the upper one. This will lower your CG and give the manipulator teams more area to work with. Win Win. 
Instead of welding the pieces together, I would recommend using these Brunner Connectors. We’ve been using these for quite a while. Just make some 1/8" thickness aluminum gussets and some self-tapers in order to secure the pieces together.
Self tapping screws?
Just how would this maintain the weight? 4130 Steel is about 2.83 times as dense as 6000 series aluminum. You’re not even halving the wall thickness in your recommendation. This change would add a significant amount of weight. I would not recommend it.
I’m sorry, I should have been more explicit, my brain is not 100% in this caffeine-free day 
3/4"x0.35" square 4130 steel tubing would roughly maintain the weight of a 1"x0.065" aluminum square tube. The benefit is that 4130 steel still has a YS of around 70-80KSI when welded, whereas welded 6061 aluminum will be total garbage, around 8-10KSI. With further optimization a far more efficient chassis is possible if you use welded 4130 vs welded 6061 because the yield-strength:density ratio is roughly five times higher with welded 4130 compared to welded 6061.
Excellent suggestion. Simple and if a jig is used, very repeatable.
From my many years in the Medical Imaging field performing maintenance on very heavy mobile systems, I have to disagree here, to a point.
With the pillow blocks on the bottom of the main rails, the axles put the pillow blocks and their mounting hardware in compression. The weight of the robot is transfered to the wheels without being suspended by mounting bolts.
Now, if the pillow blocks were moved to the top of the rails, then the mounting hardware would carry the load to the wheels, thus creating an additional point of failure, unless designed properly, which would add weight.
Now to take this a step further, why use pillow blocks at all?
The OP said these were dead axles. If that is the case, then just mount them directly through the frame. Granted, you will not lower the CG quite as far, but I believe, based on the picture and description, it would be more than sufficient.
Along the same lines, we’ve used 80/20 quickframe which is nearly identical to the Brunner stuff but the pricing is lower. We normally rivet 1/16" lexan to tie the frame together. We’ve had good luck with this method the last couple years.
I’m not sure if I just missed, but what’s the total weight?
Have you considered using sheet metal/lexan for your outermost part?
What about tensioners for your chain/belt?
It looks like you only have 2 CIMS powering your drivetrain. Did you do this to save weight, or do you have another purpose for them? I’m of the belief that there are enough motors available in the KOP that 4 CIMs (or at least a CIM and a FP) should be used on any drive train (with 2K9 being an anomaly).
My team uses only two CIMs on our drive train in general so that we free up the rest to use on manipulators. depending on what we plan to do with the robot I may or may not add the additional CIMs.
Replace all the 1/16th wall aluminum with 0.035" wall 4130 steel square tubing. You’ll improve chassis strength, stiffness, maintain the same weight, and your welder will have an easier time
.
The reason I used the materials I did is because my team has a very, very, very low off season budget and all the materials are scrap from this year’s robot.
In terms of welding we have a sponsor who will weld our aluminum for us and they really do not have a problem welding it.
Are you using Riv-nuts to hold on the axle pillow blocks?
We are using 1/4-20 bolts and nylock nuts. I just didn’t drill the holes because Inventor crashed.
Instead of using washers to lower your middle wheels, try drilling the holes in all 16 pillow blocks off center by 1/16", and flip the middle ones over so the axel ends up being 1/8" lower than the outside axels. Adding washers isn’t as clean/sturdy.
Also, to lower your CG (which is ALWAYS important), you might want to mount the pillow blocks on the top side of your lower frame members, that way the whole frame is lowered a total of 2 inches. And try mounting your electronics (Battery/CRio/etc) on the lower frame level, not the upper one. This will lower your CG and give the manipulator teams more area to work with. Win Win.
I would not have thought of drilling holes off center. In general we have used washers and other spacers because we really don’t have access to precision machine tools due to the fact we build out of one of the team members garages.
The battery, compressor, and speed controllers are all below the top of the frame, I put the CRIO on top for easier access. There will be a cover over it. I left the front Area pretty open and thought that that was enough space for the manipulators.
whereas welded 6061 aluminum will be total garbage, around 8-10KSI.
I have never had an aluminum weld break on me. We also use a professional welder on our frame.
How strong are the brunner connectors and 80/20 quickframe system. How much do they cost?
I appreciate the suggestions greatly and will try to implement them where possible.
I did not notice this post while I writing my response.
What is shown wieghs about 48 lbs.
What do you mean by using sheet metal and lexan for the outermost part.
I wanted to put tensioners in but my 5 year old DELL laptop could barely run inventor long enough to make this render.
I would like suggestions on how to implement tensioners though.
What gear ratio are you running in the Toughbox, and what’s your chain reduction?
If a designer knows and understands the limits of a welded 6061 frame, this is not a downside. We’ve been running welded 6061 frames for years, and the one broken weld we had was just horrible, awful, horrible design. Please stop saying how awful welded 6061 is every chance you get. There are a lot of advantages to teams only having to stock primarily one type of aluminum (we still stock a lot fo 2024, 7068 and 7075 for shafts however).
48 lbs seems a touch on the heavy side. Is the battery included?
Instead of using tube stock for your outside frame, can you replace them with bolted on sheet metal?
What we do is two pieces of nylon bolted through a spacing tube, and to tighten them we just tighten the bolts down, to loosen we loosen them.
Clearly welded 6061 can be used with great results, many teams (like yours) have. I am merely suggesting that there could be a more optimal design. There are very few deviations from bolted aluminum chassis, welded aluminum chassis, and folded sheet aluminum chassis. I am just trying to encourage students and mentors to re-evaluate how they make their robots. Of course there are other factors like material cost and availability, and those do matter.
I have a suspicion that many teams just “make the chassis like we always do” without ever trying to optimize it or even try something different (if only on paper as a theoretical exercise). A CAD model is great for that, the model could made with different configurations: one with aluminum tubing, one with steel tubing. Pick a few load cases that you know the aluminum one will pass, then optimize the steel one until you have the same factor of safety and see what the weights are. It’d be a reasonably in-depth undertaking, but excellent practice for anyone who wanted to do it. I know SolidWorks has multi-configuration support, and some of the simulation packages even have built-in optimization function that are very easy to use. In the past SolidWorks has donated free copies to FIRST teams.
The 80/20 quickframe is fairly rigid when you tie everything together. out of the 3 years we’ve used quickframe we’ve never had a failed corner connector or had our frame come apart. The biggest selling point for us is being able to build a robot frame extremely quickly. Price wise the quickframe can be found for about $10 for 8’ and corner connectors are like $2.60 a piece. Price can vary depending on your local 80/20 dealer we tend to get our stuff from the 80/20 garage sale ebay store because of availability.
Good luck with your prototype!
There will always be a more optimal design. In fact, most people using aluminum tend to overbuild their chassis’s as well.
I think you would have a hard time selling a change from aluminum to steel to many teams, not to mention quite a few air-frame firms and racecar teams.
Aluminum has a couple benefits that aren’t immediately apparent.
#1 Cutting aluminum by hand, generally speaking, is far easier than steel
#2 Drilling is usually much easier
#3 Aluminum is usually used in wall thinknesses that support tapping directly without the use of additional hardware
#4 Because aluminum is generally large in size, it’s that much easier to cheesehole and to make up for mis-located/designed components because you have extra leeway.
Our team learned to weld aluminum in a week, then taught the students how to do so as well. It really wasn’t that bad.
That said, we built our frames out of 80-20 the first year, thin-wall steel the next two years, and aluminum the last 2. I wouldn’t say any were heads-above the others. They all had their own strengths and weaknesses.
Back to the frame (I hate to make an off-topic post that isn’t at least a bit on-topic), frankly I think you’ve overbuilt it a bit. I think you can remove a lot of your cross members and retain your rigidity and strength. Think about mounting your transmissions by their base (on a flat plate). This will remove the need for upper-long-ways beams internally, and you can move your cross-supports down onto the bottom portion of the frame.
Neat Design!
