paper: GUS Team 228's 2011 Robot CAD Assembly

[artdutra04]

Thread created automatically to discuss a document in CD-Media.

GUS Team 228's 2011 Robot CAD Assembly by artdutra04

[Akash Rastogi]

Thank you for posting your CAD once again, Art. It continues to serve as a learning tool for myself and many others. Beautiful robot.

[Ankit S.]

That is the most beautiful thing I have ever laid my eyes on.

What CAD software do you use?

[Akash Rastogi]

Originally Posted by BeltSanderRocks

That is the most beautiful thing I have ever laid my eyes on.

What CAD software do you use?

Solidworks

[Ankit S.]

Originally Posted by Akash Rastogi

Solidworks

Thanks.

[artdutra04]

Also, here's a photo of new ramp deployment mechanism (inspired by 190's) we had in STL:



We never modeled this version of the deployment or Minibot in SolidWorks. While we had much earlier designs for both modeled, the constant minor revisions and iterations of these mechanism meant they were literally in perpetual prototyping stage. By the time we were satisfied with the performance of both, we just took them apart, painted them, and brought them to Championships. (Too bad a surface mount capacitor partially broke off inside our DLink radio which caused our weird control issues in the second half of our matches there )

We could have redesigned the deployment to be much lighter, but we didn't really need the weight savings as we were almost 7 lbs under at WPI Regional.

[topgun]

Art,
What was the design intent or advantage with the bent in corners on the hostbot?

Also, can you give us some of the specifics on the sheet metal drive train this year?

Thank you for posting the model again Our CAD team uses GUS models to study design and CAD techniques.

[JesseK]

Thanks for posting this Arthur.

In highsight, were there any mechanisms or dimensions you should have proofed or prototyped more before starting the CAD? (e.g. the roller claw width to maintain grip or roller claw height to ensure it 'pinched' but not too much)

[artdutra04]

Originally Posted by topgun

Art,
What was the design intent or advantage with the bent in corners on the hostbot?

Also, can you give us some of the specifics on the sheet metal drive train this year?

Thank you for posting the model again Our CAD team uses GUS models to study design and CAD techniques.

Go back and watch video of 177 or 1503, both of which picked up tubes from the feeder slot almost exclusively. It's not uncommon for them to catch the corner of their robot on the tower when they enter or exit the lane and get hung up for about a second.

We predicted that tube starvation would be used a lot*, so that would mean we'd have to go into the lane a lot to pick up tubes. Using 2D geometry simulations in a SolidWorks sketch, and then testing in real life with the Kitbot chassis and a tower base, we confirmed that shrinking the width of the robot to 25" and making the corners angled at 60 degrees eliminated nearly all of the hangups going into/out of the lane. The shorter wheelbase from the knocked off corners wasn't a stability issue for us since we went with an elevator/short arm design.

Some details on the drive train:

• Base chassis 0.125" 5052 aluminum sheet metal
• 6WD with 4" modified Colson wheels
• Direct-drive, 0.125" dropped center shaft
• 25p roller chain to end wheels, all live shafts
• Custom 2-speed gearboxes (each weigh less than a stock ToughBox) made from a la carte AndyMark gears, geared for 6/16 ft/sec

Here'a a photo of our competition robot drive train:



While we did design to use an exact number of 25p roller chain links, chain stretch from breaking it in required us to throw in some floating idler sprockets (like this one). This was predicted ahead of time, which is why our chassis standoffs were stepped and had a small Delrin sleeve over the inside end. While we had previously used floating idlers on various mechanisms, this was our first time using them on a "real" drive train (I'm not counting 2009 due to the very low loading that year). They worked flawlessly, never once popped out, and at just a few hundredths of a pound each are the lightest chain tensioner we've found.

Overall, we were very pleased with our drive train this year (once we broke in the gearboxes, the drive train literally required absolutely zero maintenance, repairs, or attention). Even without a belly pan, this chassis was rock solid and had literally no warping or flexibility.

Originally Posted by JesseK

In highsight, were there any mechanisms or dimensions you should have proofed or prototyped more before starting the CAD? (e.g. the roller claw width to maintain grip or roller claw height to ensure it 'pinched' but not too much)

We pretty extensively prototyped the Minibot and deployment, and went through about a half dozen iterations of roller claw designs. We tried wide ones, we tried narrow ones, we tested how much we should pinch the tube. The only thing I would change had we another change was to test the length of the roller claw, as making it about two inches shorter would have been perfect. But we ran out of time to do this, and we pretty much only have one shot per season at sending parts out to our machine shop sponsors (they help us, but it's not like IFI-level involvement where we can keep going back to them again and again with new revisions).

Nothing for the elevator or drive train was prototyped, for those we just did the math, designed and checked the geometry in SolidWorks, and sent it to manufacturing. All in all, we were pleased with the outcome (by STL), as until that weird radio issue popped up about halfway through Friday in STL, we were putting up an ubertube, two logos, and 1st or 2nd minibot nearly every match, and had three qual matches over 120 points.


* Frankly, we were completely surprised at the amount of tubes thrown onto the field after watching Week 0 and 1 events. But after seeing how widespread it was, we also made sure we had good human players and ended up picking up 90% of our scored tubes from the floor near our home zone.

[Jamie Kalb]

I can't wait to check out the CAD (I don't have Solidworks on this computer) but I have a few questions about the bot. First, did you guys do all the painting in house? If so, what exactly did you use to paint it? In the past, when I've used spray-paint, it gets scuffed up pretty quickly (and doesn't look nearly that nice to begin with, either).

Second, could you explain the chain routing on the elevator? My team used a two-stage elevator this year in VEX with one continuous loop, and we're looking for possible new methods to experiment with for next year.

Looks amazing, as usual. Thanks for posting the CAD!

EDIT: Took a look at the CAD. Outstanding work, and like Akash said, a great learning tool! Thanks again for posting it.

[artdutra04]

Originally Posted by Jamie Kalb

I can't wait to check out the CAD (I don't have Solidworks on this computer) but I have a few questions about the bot. First, did you guys do all the painting in house? If so, what exactly did you use to paint it? In the past, when I've used spray-paint, it gest scuffed up pretty quickly (and doesn't look nearly that nice to begin with, either).

Yup, we did all the painting in our shop (which is shared with the school's wood shop, which causes us to be extra careful because of the dust).

1. Remove all foreign contaminants (grease, machining oil, dirt, etc) from the parts, for which acetone does a great job. If you can't use acetone, denatured ethanol (ethyl alcohol) also does a good job.

2. Select the proper paint. We use enamel spray paint, as it forms a much more durable coating that's almost as strong as powder coating when fully cured. We experimented this build season, and used a "hammered finish" black paint and a plain orange paint, and liked both.

3. Carefully paint it. Not much to it, just don't make it run. Since we're in a snowy northern climate, we have to do this indoors during the build season, which means we usually do all painting at the end of meetings (with the dust ventilation system on, and after almost everyone else left) and let it cure overnight.



4. Let it cure. Even though the paint may claim to "cure" in a few hours, wait at least 24 hours before doing anything intense with it.

5. ???

6. PROFIT! Or at least trick people into thinking you have a powder coated robot. ;-)

Originally Posted by Jamie Kalb

Second, could you explain the chain routing on the elevator? My team used a two-stage elevator this year in VEX with one continuous loop, and we're looking for possible new methods to experiment with for next year.

There's two ways to power elevators, one with a single loop of chain (continuous), and one with one loop of chain per stage (cascade). If you want to learn more about the basic design of these elevator styles, Andy Baker has a great presentation about this here. While we utilized the "cascade" type of elevator, we modified the design to not need spools and only use continuous runs of roller chain (with springs to automatically tension it).

This is the view of the back of the robot. The two chains on the bottom are counter-rotating and each go to one side of the two parallel sets of roller chain used to lift the first stage. The vertical roller chain you see here is what's used to lift the second stage carriage.



So how does this work? Let's consider the outer (non moving) elevator member link 0, the first stage elevator link 1, and the tiny second stage link 2.

Link 1 is powered up and down by a loop of roller chain; the sprockets for these are fixed to the top and bottom of link 0. The chain/cord is bolted to the bottom of link 1, such that any rotation of the chain/cord turns into linear motion for the first stage. This loop of chain is powered directly from the gearbox. On our 2011 robot, these roller chains were on the front side of our elevator.

To power the second stage up simultaneously, this loop of chain is required to have two fixed points. One side must be anchored to the top of link 0. This then must run over a sprocket located on the top of link 1, then run down to a fixed point at the center of link 2. To have it powered down (not rely on gravity), you need to continue this with another section of chain starting from the fixed point on link 2, around a sprocket on the bottom of link 1, then back to the fixed point on link 0. When link 1 is moved up and down, link 2 will automagically do the same thing (and vice versa).

We created all our fixed chain locations using 25p roller chain and #4/40 bolts through the gaps in the roller chain. We also mounted springs in the elevator to both automatically tension the roller chain, but also to help absorb the impact (along with rubber pads on the hard stop) if we run into the hard stop at speed.

[Jamie Kalb]

Wow! Thanks for such a detailed response. I'm still amazed that the paint came out so well.

Yeah, I had seen that Andy Baker presentation before, but never fully understood quite how the cascade system worked. Your explanation makes a lot of sense, with very useful pictures, as usual. My team had some interesting situations with not being able to predict the postion of the middle stage midway through the travel, so the cascade routing is something we might experiment with.

One more question: do the springs in the chain also act as counterweights to offset the elevator weight, or are they just there for tensioning/shock absorbing?

[Jeffy]

Originally Posted by Jamie Kalb

One more question: do the springs in the chain also act as counterweights to offset the elevator weight, or are they just there for tensioning/shock absorbing?

A spring inline with the chain routing will provide chain tension. However, it provides a very equal force in each direction. The result being that it doesn't give any help to the elevator going up or down, just keeps the chain tight. (imagine if it were a string and you pulled on both ends with equal force)

[Jamie Kalb]

Originally Posted by Jeffy

A spring inline with the chain routing will provide chain tension. However, it provides a very equal force in each direction. The result being that it doesn't give any help to the elevator going up or down, just keeps the chain tight. (imagine if it were a string and you pulled on both ends with equal force)

Makes sense. Thanks!

[artdutra04]

Originally Posted by Jamie Kalb

One more question: do the springs in the chain also act as counterweights to offset the elevator weight, or are they just there for tensioning/shock absorbing?

We installed 2x ~8 lb constant force springs between the second stage carriage and the top of the fixed stage to prevent back driving.