As of Week 3..

So we’ve hit a bit of a quandary – both blog posters have been away from robotics for the last several days. I’ve been busy with the many things I have to do in college, while Jack is on an exchange trip to France. Hence… not many updates. Here’s where I’m guessing we’re at based on email correspondence and the single meeting I’ve been to. Between the massive snowstorm and college I have only been in the shop once – enough to film this video.

Drivetrain

This is more or less done. Our prototype drivetrain exposed a flaw in our design – after extensive use, the set screws holding on our drive pulleys slipped, sliding the pulleys out of alignment and snapping the flanges off of them. Since the prototype was meant to be quickly disassembled, the set screws were not held with Loctite. The quick fix was to simply Loctite the screws in the final version. No torque is being transferred through the set screws; a hex pattern has been milled (yes, milled – they look really funny but they hold) into the pulleys for this purpose. It’s not ideal, but it should work. Cotter pins are sitting at the ready if those fail. The shafts were found to have too much “play” between the snap ring grooves – they were then remachined I believe. It’s awesome how easy hex bearings make this kind of drive go together. We just have a ton of hex stock and turn down the ends to accept a snap ring. Anyone can do that in 5 minutes on a lathe. Thanks Andymark!

Some small modifications remain – we may need to cut down the end of a CIM shaft to make room for a belt, and the CIMulators have not arrived yet – but the drive is more or less done. I’m sure since I was last there it’s driven along the floor.

The chassis, with arm tower, is significantly lighter than Shaker’s 2010 bot. We shouldn’t have problems making weight.

Gripper

A wooden prototype pinch claw was mocked up. It seemed to work pretty well with surgical tubing for grip. Not the strongest thing in the world, though. With some refinement it would be a great system. Meanwhile, our friends from Team Number Removed have been working diligently on sending us their roller claw. It was designed as a test bed with an adjustable height top roller but we will probably use it as is. If not, our friends from Other Team Number Removed used some of the data we got in prototyping to develop their claw. They were VERY satisfied with the results. Their design is simple enough to construct that we have a backup just in case our sheet metal claw for whatever reason doesn’t work.

As of now I have no idea which claw the team has decided to use. We’re keeping our options open and hopefully designing our claw to be interchangeable. This way, we can put off the decision until as late as after our first event.

Arm

We made a stupid simple mockup of a one-time telescoping arm joint out of some PVC, a wedge-shaped piece of Delrin, and a piece of Polycarbonate. It worked perfectly. The PVC slides down via gravity, pushing the wedge under the polycarbonate. Once the arm is fully extended, the polycarbonate “snaps” behind the edge of the wedge piece, creating a simple one way latch. Vigourous tests with shaking the arm and repeatedly inverting it were unable to reset the system. If it needs more help, we’ve got tons of surgical tubing.

We have been waffling between a Banebots arm gearbox and a custom one I made in CAD. Originally Banebots was the way we were going, but now my arm gearbox appears to be favored. It has the advantages of not needing to be shipped from a supplier in addition to allowing the robot to self right.

Minibot

Waiting on new motors. We got a Vex based minibot to work just fine. If we had extra motors I would love to do transmission modifications, but it seems out of our reach at this point.

Day 7: Commitments and Conundrums

Sorry that this post came out on Day 8…

With our drivetrain out for welding, the Design Committee formally met to commit to an arm design. With a short discussion about complexity, ease of build, and ease of scoring, the committee unanimously decided to go for the single jointed arm design. All subsystems except the gripper were formally decided on – my arm gearbox is currently planned for production soon. Details are being firmed up on the sensitive joints of the arm over the weekend – but we’re all on the same page and concept for sure now.

Then the discussion turned to the gripper. Since day 1, everyone has been in favor of a roller claw. The design is cool, effective, and proven useful. However, we wanted to make sure we were making a smart decision, so the Design Committee drafted a Weighted Objective Table. We started talking about features independent of any design that a claw would need. After brainstorming some specifications, we then went piece by piece and weighted them based on our perceived importance. Here’s what we came up with.

Criteria ranked by priority, 3 is highest.

From there, we went through the chart with four manipulator concepts. First was a hinged-jaw roller claw designed to articulate around the contours of tubes. Second was a simpler rigid roller claw similar to the popular 2007 design. Third was a wide pneumatic pincher, also like 2007. Fourth was an “inside tube” grabber that we didn’t really take that seriously. One by one we ranked and agreed upon how each design would perform each specification, only assigning a value with the consensus of everyone. After about 20 minutes of methodical categorization, we summed up all the values and came up with this surprising result:

Items ranked 1-5, 5 highest.

This result was quite surprising. Everyone thought a roller claw would be flat out better, but it looked like a pinching claw had a significant lead on our WOT. We double checked every number and we agreed with each individual assigned value, as we had spent a long time working out all the values. This unexpected result got us seriously thinking about our direction. After twenty minutes of discussion, we reached the following conclusions:

• As long as we commit to a design and refine it, the roller claw or pincher claw choice probably won’t make or break our season.
• The hinged roller is unacceptably worse than the piston pinch in nearly every category.
• Since we’re getting roller claw parts from (team name removed), we can spend the week or so until those parts get here building a pinch mechanism, testing them side by side by the end of next week, and deciding with more information on hand.

Overall, this semi-quantitative look at the engineering design tradeoffs in various manipulators was definitely an eye-opening experience and helped change our direction from here on out. Battle of the Manipulators starts today!

As for the minibot – we smoked our motors. Less than half a second of “stalling” them, kind of sort of, and poof. The Tetrix parts are annoying… I’m thinking we’ll end up being conservative and direct driving the wheels with the motors instead of overgearing. Marginally slower, but we need reliability. Luckily, one of our brightest students is heading up that subgroup so I don’t have much to worry about. I’ll put up pictures later…

-Chris

Another Snow Day…

We had another snow day, so I designed a potential arm gearbox to pass the time. This also meant conditions were too hazardous for me to try for sponsorship.

It uses the guts of two Toughboxes and a CIMple Box. 12 tooth pinion on the planetary, one 12:50 reduction, two 14:50 reductions, one 14:56 reduction. This brings the box down to 780:1. With a 16 tooth hex sprocket on the output and a 54 tooth sprocket bolted to the arm, that’s an easy 2600:1 or so. Perfect for a single joint arm!

For more info see this Chief Delphi thread.

-Chris

Day 4 – Simplify, Simplify

Today started with a refinement of the roller claw. We took the roller claw, brought the rollers closer together, then slid some tubes in and shook them. Violently. The claw had no problems holding on to either the square tube or the white round tube, but failed to grip the tube at the thinnest point of the red triangular tube. The popular solution for this problem is to spring load the roller claw. This way the rollers would always be in contact with the tube. Alternately, looking at Team 100’s 2007 robot, we determined they had flexible pulleys that would react against the tubes, holding them with tension. No reason we couldn’t do that as well. We will probably spend some time making a “full scale” “prototype” out of grippy stuff later.

We reconsidered a conclusion we came to on Day 1 about loading off the floor – we now consider it a lot more important. A robot that can’r pick up off the floor is severely limited, more so than ignoring the top. This has some pretty big consequences design wise…

From there, the team leaders had a serious discussion about arm designs. A lifting design is all but dead with the 84 inch rule. Lifts are complex and difficult and Shaker does not have time for complex or difficult this year. A four bar linkage design looked promising, but we determined that picking up off the floor and reaching the top row were mutually exclusive with a four bar. Concepts to solve this problem included a four bar with extending links, and an elevated four bar with a “mini arm” that passes tubes from the floor to the mechanism. These were both deemed to be too complex. Our biggest fear is trying to do too much and going with a complex and convoluted solution – and that is exactly what we thought these designs were.

While debating whether or not we would give up the floor or the top row, I did some sketching in Solidworks with shapes and found something out. With a roller claw that can rotate tubes to change their orientation, we figured that a four bar linkage’s orientation-holding feature could be simulated with a driver manipulating the game piece. We watched videos on TBA of Team 100 spinning tubes for a good scoring angle. So… why not make a single jointed arm with a gas-spring deployed claw?

This design quickly gained quite a bit of traction among team members and myself. It is simple – one joint and a passive deployment mechanism. It fits in the box with room to spare. It does not require telescoping of any sort. No decisions made yet, but I think this is the direction we’re going to go. I’m willing to bet this is the “ah-ha!” moment of my team’s 2011 season.

On to Friday…