Big

Stinkin

Marble

Sorter

Mark Lehr, John Ouimet, Leo Johnson

The Adopted Design Process for PLTW® Coursesand table of contents

Define a problem Brainstorm Research and generate ideas Identify criteria and specify constraints Explore possibilities Select an approach Develop a design proposal Make a model or prototype Test and evaluate the design using specifications Refine the design and Modifications Create or make solution Programs Reflections & Communicate processes and results

– ITEA Standards for Technological Literacy

Define a problem

Create a FisherTechnique autonomous machine to sort and collect five different types of marbles, two of each, within

1.5 minutes: steel, clear glass, colored glass, wooden, aluminum coated plastic.

BrainstormingLogic sequence, method detected, method removed

Metal: magnetic, pulled-knockedAluminum: current, pushedWood: weight, fan blowClear marble: light sensitivity, pushedColored: last one drops off

Brainstorming

Brainstorming

• Preliminary testing confirmed that all hypothetical methods of identification and separation should succeed.

Research and generate ideas

Identify criteria and specify constraints

Explore possibilities

Testing magnetic selection & collection

Select an approach Develop a design proposal

•Construct and test for each marble one at a time.•Modify as needed.•Implement programming as necessary.

Designing the hopper system to introduce the marbles one at a time.

Hopper flapper prevents marbles from bouncing

Hopper tested good with propeller as first gate to allow one marble at a time; & works good with magnet selection of steel marble. Propeller blows wood ball off.

Make a model or prototype

One difficult barrier was that there are only four controllable motor and light power ports. We had to test creative methods to power all of the needed active systems. The final solution was to use the nine volt output to continuously power up the fan for the wood ball and the final sensor light for the detection sensor for all three final balls; aluminum, clear glass, and colored glass.

Test and evaluate the design using specifications

Explain the Problem and Proposed SolutionOriginal design used an electro magnet to capture the steel marble, found not enough pull to lift marble off of track Changed to permanent magnet, more predictable and consistent.

Design Modification Sketch

Refine the design and Modifications

Explain the Problem and Proposed SolutionOriginal design was to sense the aluminum marble using the conductivity of the material. The marble would pass between to contacts closing the circuit sending a 1 to the controller. The aluminum marble check would follow the steel so that the aluminum would be the only conductive material left. The marble would not contact the sensors with enough force to complete the circuit. The team tried stripped wire, paper clip ends and aluminum foil taped to the track and the gate face, all materials failing to complete the circuit. One plan was to use the track for a common leg but the track material is anodized thus non-conductive. 

  

Design Modification Sketch

Continued: Explain the Problem and Proposed Solution The decision was made to use reflectivity as the sensing criteria. The team was able to use the same sensor set as was being used for the clear/opaque marbles. A distinct level of reflectivity/opacity was noted for each of the three marble allowing the use of this sensor with a compatible logic loop. This change resulted in the need to use a fan for the wood marble, changing the sequence of separation  

Design Modification Sketches

 Explain the Problem and Proposed SolutionThe original design was to make a check for the wood marble at same sensor pack as the clear/opaque marbles. The need to use this sensor for the aluminum made it necessary to change the wood marble sensing process.  It was decided to use a fan to blow the wood marble off of the track as it passed to the last sensor pack. No marble at that station would result in an empty shuttle dumping to the clear marble bin, not perceived as a problem, minor impact on cycle time. The aluminum marble has a weight close to that of the wood. Some adjustment was needed to prevent the aluminum marble from leaving the track at the fan station.

Design Modification Sketch

Explain the Problem and Proposed SolutionInconsistencies with the feed hopper has caused multiple marble feeds, stuck feeds and marbles leaving the track prior to reaching the appropriate sensor pack. Original design utilized a cylindrical feed into a V-channel, at the end of the channel vertical gate opens and closes to release 1 marble. Timing of the gate has proven difficult but has been managed. The hopper/channel interface seems to be the problem, marbles either bunch up or fail to feed consistently. Original design is shown below.

Design Modification Sketch

Explain the Problem and Proposed SolutionHopper/feed was changed to eliminate the cylindrical hopper and extend the V-channel to accommodate the 10 marble limit. This design proved to be more consistent, but some fallout was still evident. Newer design shown below

Design Modification Sketches

Explain the Problem and Proposed SolutionA cover was added to funnel the marbles into the channel in a more orderly manner, again with limited improvement

Create or make solution

Identifies and moves colored ball to the left

Identifies and moves clear ball to the right

Counts ten balls and stops program

Picks up steel ball magnetically and drops it

Identifies aluminum ball and drops it off the back

Programs

Main Program with timer and counter

When we began this project my biggest worry was I had little experience with the Fischertech equipment, not knowing all of the components and how to best optimize our designs.

With the need to separate 5 configurations of marbles I was concerned that we would have enough sensing methods to differentiate some of the marbles that were similar in characteristics.

That turned out to be no problem at all we were able to develop sorting methods and logic in no time. Building a structure that work flawlessly was not such a slam-dunk. Our initial build was not too bad, stable and functional. Unfortunately our loading chute and dispensing control, not so good. There was a big problem with repeatability, balls stacked up, got stuck, and fed 2 at a time during one run then perfectly the next. This appears to be our weak link.

Another problem that propped up was with the output values of the photoresister. They seemed to fluctuate over time causing the logic loops to fail on occasion.

After working through this project, and observing other groups I will have one very important thing to take back to my classroom, don’t let the students get too frustrated. This is a tough design challenge, there will be times when some kids may want to throw in the towel. I will have to find ways to plant ideas and leads to help some teams move forward.

I look forward to playing with my new Fischer kits when I get home, see if I can improve on our design.

Reflections & Communicate processes and results

The team tried to stay with the original process determined during brainstorming. But when the aluminum ball wouldn’t easily register carrying an electric current, the successor plan worked very well. The final design helped us to minimize the needed number of motor/light power ports. It was very handy to have the final sensor check for three of the balls. The propeller first gate worked out fairly easily also. The most difficult task was building the feed chute. The light balls would be pushed through two at a time and the heavy balls would erratically not fall quick enough to not be caught by the propeller. We ended up with a consistently successful design. Everybody was impressed with the fan blowing the wooden ball off. The programming turned out to be just four simple subroutines and two minor lines to start the ball flow and count to ten to stop. I fell that the best part of the project was the proactive ease that the three of us respectfully worked together.

Reflections & Communicate processes and results

Brainstorming – It was very important to have tried the various examples of the Fishertecnik components. The design activity was directly related to using the components (ie. Interface, programming IDE, motors, fans, track, structures, sensors – photo resistors, photo transistors, lights, electromagnets, permanent magnets). Without having done these initial projects the activity would not have been possible.

Examining the marbles to be sorted for this project was the first step. Each marble was paired with a sensor which could be used to distinguish it from the rest. Steel-Magnet, Aluminum-Conductive, Glass-Light Transparency and Wood with no Transparency. Each sensor was tested to determine the adequacy of the separation approach. Using these facts and the ability to use Motors/Tracks for movement determined the final solution to the problem.

Sequential Development – First the hopper and gate were configured, then the magnetic separator was conceived, third the aluminum separator was devised and last the photo sensors were paired with the device.

Roadblocks and Redesign – The first problem encountered was with the electromagnet. It was not strong enough for what we intended. This led to a redesign utilizing a permanent magnet with a swiper action to drop the steel marble. Next, the aluminum was tackled. This took the most amount of time since we could not get the aluminum to conduct. Everything from copper brushes to aluminum foil was tried and to no avail. The track would not conduct and unless we pressed down on the aluminum we could not get a closed circuit. At this point we resigned ourselves to use the photo-sensor for the aluminum and utilize a fan for the weight differential of the aluminum and wood ball. We got very lucky that this worked.

Reflections & Communicate processes and results

Final Solution – The hopper and gate became problematic when trying to dump and test all the marbles at one time. Up to this point, only a few marbles were tried at one time so no problems were encountered until overloading the hopper. This required several redesigns and still is not up to par and giving 100% repeatable results. Our program was also unique. We tested one sensor/motor at a time and put these pieces into subprograms. When the final program was built, we simply used subprograms that had already been tested. Required very little time putting it all together.

Lessons Learned and Improvements - A completely redesigned hopper is necessary. We need to deliver 1 marble at a time with out bunching up at the gate. In addition, the gate should be redesigned. The propeller is to susceptible to problems. A small window letting in 1 marble at a time will solve 90% of the total problems. The aluminum needs to be measured directly. Currently the track we are using is aluminum. It would be possible to use a different track made out of conducting metal to sense the aluminum. Another team tried this approach and it worked quite well. Since we would sense the aluminum, then instead of using wind provided by a propeller to knock the wood off, then we can sense the non-transparency of the wood with the glass beads to finalize the design.

Reflections & Communicate processes and results