Archive for the ‘ Robotics ’ Category

Senior Skip Day?

RoboShock

Sigh.. I procrastinate so much that I even put off blog posts that would take me no more than half an hour if I’d actually tried. Now that I’ve burned a few hundred kilocalories trying to get my head around flight orientation, here’s another summary of an event days past.

May 20th, the second-to-last Friday, was CV’s senior skip day. But instead of going to the beach, I:

  • Attended NCIIA’s Invention to Venture workshop on technology entrepreneurship
  • Attended the OSU Engineering Expo
  • Set up field for Roboshock

Invention to Venture

Given my limited business experience, much of the entrepreneurship discussion was rather theoretical for me, although I nevertheless picked up some concrete ideas I hope I can put in practice in the very near future. As the talks were several hours long in total and covered many different topics, I will summarize what I learned in a separate post (link forthcoming).

I do remember coming away from the talk with the definite impression that entrepreneurship is going to be one hell of a sleep-depriving activity.

OSU Engineering Expo

I don’t know about the others, but the computer science and electrical engineering presentations I listened to did not impress me.

The best of them, in my opinion, were the two seniors who had developed an automated AI class scheduling system for college courses. The system could use a database of courses from any university to organize a schedule for the four years of college around any specific course(s) the user flagged as a must-take, automatically taking care of all pre- and co-requisites. The vivacious senior to whom I talked presented me with a straightforward, working implementation and clearly stated goals. Not a novel idea, but well executed. It’s too bad that the system will not yet be available for use at OSU next year.

The others… had no sales plan (why should I care about your project?), had no obvious applications (you essentially made an RC toy; what are you going to do with it?), had no goals, no direction, lacked expertise (you developed this program; how do you not know what language it’s written in?), and/or was unoriginal (your project does exactly what Amazon does already). Or they just did a bad job of communicating their ideas to me.

I am not saying that OSU Engineering itself is bad. I just know that when I do my Senior Thesis Project, I’m going to do a lot better than that.

Roboshock field setup

The OSU Robotics Club hosts a full-scale FIRST Robotics scrimmage dubbed RoboShock every year. Our team is more or less in charge of the field, so some of our members helped set up the field at Gill Coliseum in the evening.

The actual competition was on the next day (a Saturday) from 10 am to 6 pm. Unfortunately, although our team was scheduled for 9 matches, we ended up doing only four or five due to WiFi interference issues (nearly every team started lagging).

At official competitions, there is a single router to which all teams connect their robots. OSU didn’t bother with that and had each team use their own router. Since the ClassMates are only capable of the 2.4 GHz band and there really are only three channels available for use (1, 6, and 11) for six FRC teams and who knows how many FTC teams (who had their own game going next to the FRC field), interference was inevitable. Hopefully that will change next year.

Advertisements

Raiderbot X, 23 Days to Robot Ship: Lego Minibot!

Minibot proof of concept out of old Lego parts. The purple arms push against the pole, which angles the wheels into the pole to create a horizontal vector just strong enough to keep the bot on the pole. Took about two hours to make.

The bot is climbing up a 40 inch metal pole that is embedded in the concrete floor of our garage (for reasons beyond my imagination). Climb speed is about 6 in/s. Power is provided by a 13-year-old 9V Lego motor that is the wimpiest of the wimpiest Lego motors out there. It is being run on 12V.

The TETRIX motor we will be using in our actual minibot has 21.7 kg-cm of torque. This motor exerts no more than 200 g-cm of torque.

This is the first time in six or seven years I’ve put anything together with Legos. I think I’ve lost one of my tubs—I made this robot out of what few parts I could find. I need more gears!

Credits to my mother for taking the videos!

Raiderbot X, 26 Days to Robot Ship

This is an email I sent to the Raiderbot mailing list. I think it sums up everything that is going on.

Hello Raiderbot,

The bad news:

We are still a week behind largely due to delayed shipping of supplies and bad planning on my part. We have no driving robot, nor do we have a fully functional arm. If we add up all the hours, we have little over 60 hours (2.5 days) of work time left in the season. Also significant is the fact that too few people have been showing up.

The good news:

We have the aluminum tubing required for the chassis and have enough already cut for one chassis, ready to be welded using the jig constructed last week. HyTek has donated more plastic to us than we’ll ever need this year. We received the materials for the arm today and should be ready to start building tomorrow.

The roller claw beta prototype (demonstrated last week) looks very promising. Great job to everyone who helped design and fabricate it!

We have located a machine shop (Viper) that has the capabilities and is willing to manufacture our swerve modules (for the drive base). Spencer Hedrick is currently waiting to hear back from them regarding the cost and time of fabrication. We will know by Friday what the decision is. If the price is affordable, we hope to have the modules (and one robot) finished by the end of next week.

The scary stuff:

We need our first robot finished by the end of next week and our second robot by Wednesday of the following week. Please realize that although next week is only Week 4, it is already Week 4. It was our original plan to have the swerve modules finished first, but because the machine shops will not finish fabricating the swerve modules by the end of next week, we need to have everything else finished and nearly perfected by that time. This includes the:

  • Roller claw,
  • Four-bar linkage lifting mechanism (with milled aluminum tower),
  • Electronics board, and
  • The attachment of every necessary motor, sprocket, wire, belt, chain, pneumatic, and pulley we need on each of the two robots.

If all 46 members of our team meet and work consistently, this is entirely possible to accomplish in less than a week. At the current rate of attendance, however, this will not happen. Please also consider the fact that I haven’t even begun to describe what we need to do regarding the business plan, T-shirt design, battery cart, robot cart, scouting, pit build, pit panels design, minibot design, etc. There are too many jobs and too few people who come to meetings!

I am fairly happy about the attendance of the longer (Monday/Friday/Saturday) meetings; we got a lot done last weekend. However, attendance of the shorter meetings has been disappointing at best and unproductive at worst. Because we end two hours earlier, we need more people and dedication to maximize the use of our time in the shop.

It is our (me and Lipi) job as program managers to keep track of what resources we have to work with and keep the team on track for a successful season. Here, a “successful season” means we have completed the following (among other things) by Monday, 2/21, which is 26 days away:

  • 2 Robots
  • 2 Minibots
  • New website
  • Pit
  • Battery cart
  • Robot cart

The request:

We hope we are doing a decent job, but in being effective leaders, we depend on each of you. Regardless of your individual jobs, regardless of how small it is, regardless of how much it may or may not seem related to robots at all, everything you put into the team brings us closer to success.

We would like to go to every meeting knowing everyone will be there, but this has not been happening. We completely understand if you have other commitments (we do, too), but please let us know (by email, word of mouth, phone call) if you will not be there—then at least we can plan things accordingly so we aren’t caught short at the end.

Currently, the programming subteam is the only subteam that is (far) ahead of schedule. (Mechies, are you seriously going to let the proggies outdo you?) Congratulations. :)

Again, the roller claw is looking great. Please be there tomorrow to help finish building the arm and test it on the kitbot.

Cheers,

Lipi Gupta/Soo-Hyun Yoo

Raiderbot X Meeting 7, 35 Days to Robot Ship

Meeting time: 15:00 to 19:00
Hours logged: 47.86

We attached our 2005 game arm to the kitbot. It works! It’s a pretty big arm even though it doesn’t quite reach 10 ft. I can’t wait to see our four-bar linkage arm finished.

Ordering parts… We need another cRIO, but I can’t find cRIO-FRC. NI has something that looks similar, but I’m not entirely sure it’s the right one.

ES and DP are building a battery cart. I’m surprised at how complex they are planning for it to be, at least in the electronics and software aspects. It sounds like a great project. That said, I will need to keep an eye on them to make sure they don’t go for too much. I think they can finish, though.

Raiderbot X Meeting 6, 36 Days to Robot Ship

Meeting time: 10:00 to 18:00
Hours logged: 44.05

Nothing markedly eventful happened today. Logging here for the sake of logging.

Spent forever soldering the wattmeter leads to Anderson plugs (iron isn’t powerful enough for 6 gauge wire), but the wattmeter works. Battery test bot pulls 30 A with wheels freespinning on toat, 120 A at stall. Circuit breakers are breaking!

Filled out purchase orders. Searched but could not find AS5040 encoder breakout boards anywhere other than on their official website. We need to order photoswitches as well.

Subdomain randomagically works! I needed to change “NameVirtualHosts cvhsrobotics.net” back to “NameVirtualHosts *”. Emailed Mr. R and got sandbox subdomain registered.

Talked with Will about goals for next Wednesday.

Raiderbot X Meeting 5, 38 Days to Robot Ship

Meeting time: 10:00 to 18:00
Hours logged: 36.74

Programming

The idea this year is to overload the bots with sensors of all kinds. Current sensors on every wheel, arm joint, and main circuit breaker to graph power consumption. Proximity sensors to avoid collision with other bots and facilitate scoring (automate entire process!). Encoders, gyro, and accelerometer for inertial measurement. And of course, line trackers for line following. I foresee lots of debugging, though it should help that the programmers have switched to Java+Netbeans that allows for serial communication with the cRIO.

Team 955 will now be using Git for version control. Git repository created here on Github.

Rockwell Automation Photoswitches

There is a wiring diagram on the back of the sensor box. L.O. and D.O. are light output and dark output, respectively. Power supply should be between 10.8 V and 30 V. See this Chief Delphi thread for more information. In the end, hooking it up to the cRIO is as simple as soldering brown to positive, blue to ground, and either white or black to signal of a PWM cable (which is true when either is light/dark?) and connecting the combination to Digital I/O. It is NOT necessary to hook up bleeder resistors from the signal lines to VCC as it says in the datasheet.

The datasheet is very useful:

  • Green LED on: sensor powered.
  • Green LED off: sensor not powered, output active, SCP* active.
  • Yellow LED on: output on.
  • Yellow LED off: output off.
  • Yellow LED flashing: output SCP active
  • Orange LED on: margin** > 2.5.
  • Orange LED off: margin < 2.5

* Short-Circuit Protection
** The orange LED shows that the signal strength is at least 2.5 times that needed to trigger an output. Signal strength is greater with higher voltages (obviously). This can be useful to overcome dust on the lens. The sensor’s sensitivity can be adjusted also by turning the knob on the front panel of the sensor. With an oscilloscope, I found that at maximum sensitivity, the photoswitch is triggered by almost anything (e.g., the table, paper, my red 955 sweater, my jeans) even at 10.8 V. We will need to test and calibrate the photoswitches so they trigger only over the reflective tape.

The datasheet recommends that the distance from the sensor to target to be at least 6 mm.

Standardized ordering process

Blanket POs from Will and Mr. A. Co-PMs and division leaders are the only ones authorized to order. Use order form on team website.

Minibot

TETRIX parts list:
2 – Motor Mount, W739089
1 – Motor Shaft Hub, W739079
2 – DC Motor, W739083
1 – Gear Hub Spacers, W739090
1 – Tetrix Resource Kit*, W731900
1 – Battery Charger, W739059
1 – 12V NiMH Rechargeable Battery, W739057
1 – Power Switch, W739129

* The Tetrix Resource Kit contains the parts shown here. Unfortunately, I haven’t found a detailed list containing only the Resource Kit parts, so we have to use the detailed parts list we found earlier here to identify the kit parts.

The default kit seems to have enough parts in it for one minibot. Once the minibot team has a reasonably complete idea of how they want to build their two minibots, they’ll let me know of any additional parts they’ll need. We’ll see exactly what parts we have when the kit arrives (hopefully) later this week.

Sent an email to PHRED about the possibility of exchanging FTC parts for FRC parts. They apparently no longer plan to build a minibot.

Raiderbot X Meeting 4, 39 Days to Robot Ship

Meeting time: 15:00 to 21:00

We are two days behind schedule. I started a WOT (Weighted Objectives Table) discussion of the various lift concepts at 3:30. We started with seven concepts:

  • Single-jointed arm
  • Multi-jointed arm
  • Telescope (elevator)
  • Four-bar linkage
  • Eight-bar linkage
  • Scissor
  • 60/40 scissor

We narrowed it down to the four-bar linkage and the elevator and chose the four-bar linkage after a qualitative analysis using a pros/cons list for each. But I still wonder if this was necessary, because it took another 30 minutes of our time and I ran 45 minutes past my proposed deadline of 4:30. Why couldn’t we have reconsidered the two options using the three factors we had rated most highly? Will says some people just need more stake in things? Anyway, my bad for not calling them on it. Areas of improvement:

  • Do not let the loudest voices steer the discussion off-course.
  • Constantly think back to previous discussions and decisions. People were putting the WOT factors under the pros/cons lists, which is nonsense.
  • As always, be consistent, even when calling out for hands when rating the WOT factors. This is no time for jokes.

Will then separated us into four design groups for the lift, claw, base, and minibot.

Four-bar linkage: Mr. Groom found that making the shoulder pivot points’ distance shorter than that of the elbow pivot points allowed the claw to point down when the arm is down and point up when the arm is up. This should allow us to reach both the ground and the top peg as well as everything in between. This makes sense if we think about the four-bar linkage as a trapezoid and shifting the two bases—the linear displacement for the two lever arms is equal but the arm lengths are different, so angular displacements differ.

Minibot: make it as close as possible to a battery with motors glued on.

Preliminary design review planned for Saturday 17:00.