Several years ago Dan Burns and I started discussing an engineering project for which students build a crash cushion to investigate momentum and impulse. Using only a few sheets of paper and hot glue student groups design crash cushions (similar to water barrels or guard rails on roadways) to lower the force experienced by a cart rolling down an incline that crashes into them.
Since then we have both completed the project with classes albeit differently. We also presented at the Summer 2016 AAPT meeting in Sacramento, all of the materials discussed there are here. I wanted to share the project here again, with the tweaks my partner Jon Brix and I have made to it since.
My colleague Matt Miller continued the project in Conceptual Physics this year although opted not to use the Vernier sensor that I had last year. He opted for the resettable Drop N Tells I bought years ago instead as it is more visual for the younger students. He set up a ramp and used a lightweight impact car that had an additional <200 grams of mass added. Miller adjusted the ramp set-up until the 25, 15, 10 and 5-g sensors were consistently tripping. His students were challenged to design the crash barrier that did not trigger all the sensors. I believe he set the grading up this way:
C = triggering the 15, 10 and 5-g sensors
B = triggering the 10 and 5-g sensors
A = triggering only the 5-g sensor
Extra Credit earned for not triggering any sensors.
Dan uses a PASCO Smart Cart while I use Vernier sensors. The first year I tried this I used a low-g accelerometer because it was what I had. Through a Donors Choose grant I was able to purchase the higher 25-g accelerometer, 3-axis accelerometer and a Wireless Dynamic Sensor System (WDSS). The wired sensors require some coordination to prevent the cord from catching but are workable. I found that the wireless WDSS made for easier set-ups but would disconnect occasionally. Both the WDSS and the 3-axis sensors were almost too accurate and the graphs produced were difficult for students to interpret. I opted this year to use the single-axis 25-g accelerometer because even collecting 500 samples per second the peak accelerations were easier for students to determine.
In the past I've used a wood ramp and a big heavy dynamics cart that then travels along the flat lab bench into a wall. The transition from ramp to flat tabletop caused additional acceleration peaks so we opted to have the cart run directly into the wall from an incline. The heavy cart and steep ramp produced a high acceleration that exceeded the accelerometer's limits. We decreased the ramp angle and still occasionally "missed" the hit because the time of impact of the cart against the wall was so short. This year we opted to use a low-friction (not smart) PASCO cart and track from another colleague. The 120 cm long track was raised above the table by one textbook and pushed against the wall. A box of weights (over 30 lbs) was pushed against the higher end of the track to prevent it from moving. In initial tests the conservation of momentum caused the track to move quite a bit when the cart struck the end of the track.
This year student designs proved very successful. Because of the light cart and small incline students were able to reduce the acceleration of the cart at impact fairly easily. Usually the designs that "failed" did so because the cart passed underneath the crash cushion and still struck the wall. I had only one set-up in the classroom so groups took turns testing their barriers and collecting data. We stored a trial of the cart running into a book at the end of the ramp and then printed out graphs for each group with their trial on top of the control data. Here is an example of the data student's received with the control (green) and their trial (blue):
Using this information students were to write a Claim, Evidence, Reasoning (CER) conclusion to answer the question: "Was your crash cushion effective?" After grading these conclusions we realized a few things:
1. Students did not agree on what made a crash cushion effective. Most students realized that decreasing the force, as shown on their graph as a decrease in acceleration, by increasing the time of impact made for a successful crash cushions. A few more realized that stopping the moving cart without letting it bounce back was also good. Yet many students considered their crash cushions ineffective if there was any acceleration, even if they reduced their force by more than 50% .
2. Students do not know what is fact vs. opinion. This must be going around recently. Students often stated opinions or qualitative observations in place of specific measurable data. "Our crash cushion was good because it stopped the cart slowly."
3. Some students did not understand the graph axis, significance of peaks, etc. Referring to the example above, some students incorrectly described the "time of impact" to be just over 2.5 seconds for the control trial. They did not understand or forgot the fact that the cart had to roll down the ramp before the impact.
4. When in doubt, students are prolific. I expected three, maybe 5, sentences from students yet often received a full page. While grading these conclusions I often crossed out over half of what was written because it was superfluous. They seemed to just keep writing and praying for partial credit.
Before handing back their conclusions I reviewed the CER format with students and showed them a few pictures of correctly written (short) examples of their peers. I showed them a few sample graphs from their trials and reviewed the significance of each peak. In the future this will be done the day after to give students a chance to correct their CER conclusions before turning them in.