Tips on Improving Robot Design

This is an off season project my team has nearly finished. It has served at a great learning tool for new students/ a first time test for pneumatics and Solidworks for the team. Any suggestions for improvements? - Besides that the pneumatic piston is still ductaped to the arm and the manipulator wheels are not yet powered? Any and all critiques are welcome!!

Some specs: We used 1/16" thick, 1" square aluminum tubes, rivets, and various brackets for build. The lift for the arm is powered by a banebots motor and FP gear box. This system is currently attached to a rod via pin to the rod (rod subject to bending…). The manipulator wheels will be powered with a Vex motor.

Just my opinion, but GET THE MECANUM WHEELS OFF!

I know the omnidirectional drive is cool, and looks promising, Any standard tank drive system will easily push you all match long. I think mecanums are cool for any purpose other than competition. Do you know how many mecanums usually make it to Einstein? As far as I know, none. Any team with mecanums are not only easily pushed, but are generally easily defended against, which, while game pending, is never a fun thing.

I think mecanums are just a waste of weight, and drivetrain, so that is my advice to you. :]

EDIT: Also, it looks a little back heavy. How’s the weight balance working out for you?

For this game, how much pushing around was there, really? We had mecanums and didn’t regret it. In fact, they came in handy when in the safe zone, positioning for the score.
In a game like Aim High or Rack N Roll, I’d totally agree with you. But here, mecanums were a viable option. Plus, it’s an offseason project. What better time to explore drivetrain options to see if they’re a fit for your team or not?

I’ll come out and say it, I like it. Yeah, it isn’t pretty and by the sounds of it has some flaws but I like that you took the time to build an entire robot in the off season. Now the $5 million question, what did you learn?

Like all things in engineering there are tradeoffs between this drive system and others. If not being pushed is important to you but moving sideways is not then you must take that into account. It depends on what your strategy is.

Great job, use it, abuse it, and see what breaks. If you want to give the programming team a challenge, add a gyro and have them test field-oriented driving.

Don’t be so down on mecanums, especially for learning. Using that robot as a test bed, you can easily swap the four wheels and experiment with traction wheels, omni wheels, mix of the two, etc.

Mecanums are also a good choice for a demonstration robot for fundraisers and sponsor presentations. Easy on the carpeting, and maneuverable in small spaces.

The more teams experiment with mecanums, the more you will see ingenious solutions for pushing. :slight_smile:

It’s the offseason. Leave them on.

If you experiment with them in the offseason, then when the season rolls around and you build your competition robot, if (and this is a big if) the mecanum drivetrain turns out to be advantageous, you already know how to use it. If it doesn’t, it’s not that hard to turn a mecanum system into a 4WD system and practice with it.

It looks really nice. Integrating the actuation for the wrist into the 4-bar is very slick, however be wary of bending. I think when 148 did it this year they used a piston to push/pull sheet metal sliders. Basically the 4 bar was still very strong because the piston was not actually acting as a structral part of the arm. I would be concerned with the piston getting even slightly bent in a hard colision and no longer functioning properly.

Also, on your bell crank pivot down towards the bottom of your robot there is quite a lot of open shaft on either side of the arm. The further away that shaft is supported from the load the greater chance that it will bend. There is also the chance that it will begin to bend over time and continued use so it is probably worth supporting it closer in on either side so that can’t happen. In any case, it is a good practice to support shafts close to the load as possible.

Overall it looks great! Looks like you are very ready for build season.
Regards, Bryan

The picture is a bit small, so it’s hard to make out many details. The first thing that I saw that I’d change is how the pneumatic cylinder on the arm is mounted – and I don’t mean the duct tape.

Right now, when the cylinder rod extends, it is ‘exposed’ and any side loading on the end of the arm will get transferred back against the cylinder rod. A rough impact may bend your cylinder rod and then you’re in a bad place since replacing the cylinder is expensive – both in time and money.

It’d require a rework of the upper link of the arm, but consider sliding the cylinder back toward the upper pivot so that, when its rod is fully extended, it’s still supported and protected by the fixed part of that arm. I suspect your design is inspired heavily by 148. Look more closely at that robot – its CAD model is available here – and consider carefully why they designed their extending upper link the way the did.

Forgive the derailment, but I see this getting parroted a lot these days and I’m not convinced it’s a valid argument. Folks make it seem as if the only difference between a mecanum wheeled robot and 254 et al is the wheelset. That’s absurd on its face.

So, let’s have it – in your opinion, which teams were capable of winning the Championship but, by deciding to use mecanum wheels, didn’t make it?

We used this logic in 2009 as a reason not to build a widebody robot, because so few had made it from 2001-2008. (IIRC Wildstang in 2001 & 2003 were the only examples we came up with)

Fast forward to Einstein that year 11 widebodies and 1 narrowbody (us). Our logic was flawed because this is really a game/strategy/field driven decision.

You shouldn’t discount something just because it hasn’t worked yet, if you do that you will never innovate and refine something that is close to working because you instantly call it a failure. There are too many influences besides drivetrain style that effect who makes Einstein, so don’t mistake correlation with causation.

I have no doubt there will be a team that creates mechanum, drives it well as a mechanum and makes Einstein. (I just won’t predict this year without seeing the field.)

Try telling 525 that mecanums are bad.

I agree, but here is not the place to be arguing mecanum vs tank.

It looks like most of your weight is on the back wheels. This may lead to problems because balance is extremely important for a mecanum drivetrain. Otherwise, great job. Its a very clean looking robot.

Make a y bracket and attach to frame to lock in the gripper when it is in the down position picking up a tube. In the event of a hit on the gripper from another robot the bracket will off load the forces from the upper axle down low to the floor preventing your arm from being bent.

Clean up and label your wiring. Invest in a crimper and some Molex connectors.

Keep the CG as low as possible and try to maintain a balance design.

When selecting your gearing try to use paired ratios in your transmissions so you can easily change gear ratios to make your mechanism slower/faster

Wow. I didn’t expect this to derail the thread. We’ve had experience with mecanums, and it wasn’t so good. I’ve seen a lot of unexperienced teams with a lot of potential use mecanums not knowing what to expect defensive-wise, and end up getting pushed around by the simplest of drive trains, disabling their ability to function. Just my opinion, that’s all.

Let’s not derail this thread further.

I agree with Roysturr. Adding a y bracket and attaching it to the frame would help a lot.

These connectors I linked to in an earlier thread is incredibly useful for wiring. Many teams (including ours) have learned to love them. You should definitely consider using them in the future.

Our robot has done a lot of sitting and slow driving around. So we haven’t encountered any problems as of yet-- thanks for pointing that out! Also, thank you for pointing out other mecanum flaws. This will serve as sufficient motivation to work on CADing other drive trains!

-I have learned that rivets are amazing and often better than bolts. (saves: weight, time re-tightening loosened bolts, and awkwardness of trying to hold wrenches at funny angles)
-That the robot manipulator should always have full control over the game piece// and be able to protect it when driving around. (did learn this from 148!)
-To look work on efficiency and look at the big picture. Our actual design from the Logomotion game could not pick tubes up off the ground.
-And of course, that planning a design in Solidworks avoids countless hours of frustration and re-fabrication.

All of your posts have been very helpful! I never thought of adding a y-bracket to protect the arm.

We were lucky enough to be stationed next to 148 at the Dallas Regional last year… we were very inspired by their robot-- and worked on building a simplified version in offseason. I am trying to build a simplified version of their octonum drive in Solidworks right now… Though I dont think we will be able to implement it without machining the parts. I see that you guys built your own base… Is that something that you would advocate? (note: my team is fairly new at Solidworks and have never had anything machined). This was our Solidworks design for the robot I posted earlier

http://i.imgur.com/uJwgQ.jpg

One of the things that is always preached at work is the concept of a minimum viable product. That is, what is the bare minimum to get done what you need to get done. Along these lines the EWCP took it upon themselves to figure out what minimum actually is. We published two blog posts, a general analysis of 2010/2011 and an in depth analysis of 2011. They are available here and here. We hope you find these interesting and helpful.

I really like your off-season persistence to build a robot from scratch taking into account the best practices of powerhouse teams… GREAT JOB!

Looking back, I wish we did the same thing from scratch, but, what we did was keep working-on/improving our rookie robot for the two off-season competitions we competed in…

As you can see below, we’re kindred-spirits in our admiration of the 148 Robowranglers - Raptor model… :wink:

My rookie year was 2011, and our entire team converged on mecanum. I’m personally a big fan of any team who goes for tank. Although I have seen some large improvement over the development of the tank drive. I have fallen in love with 8 wheel tanks although that’s mainly because it looks cool. You can integrate wheels with tank so I agree with SuperNerd.

Keep up the good work guys. It is very inspiring to see young teams putting in the extra effort to learn what it takes to make a great robot. I have watched several other young teams in Michigan blossom over the last coule of years doing exactly what you guys are doing now.

They say that: “Imitation is the highest form of flattery.” Personally if I was a judge, I would give team 148 an Engineering Inspiration award for posting design details and inspiring other teams to attempt such a cool mechanism. I know of a couple other teams that have completed very similar exercises.

My challenge to you guys that have built these, is to share some of the wealth you have learned. Possibly a short white-paper (1-2 pages) on some of the good, the bad, and the uglies you found while replicating the general design intent.
Doing the white-paper will help document the pros and cons for your team as well as help others attempting similar ideas in the future. It is also good practice for documenting your designs in clean concise manner to help present to judges at competition.