As you probably already knew, this past weekend was the very exciting Mini Raceday Weekend. We posted the official Sweepstakes freeroll and backhill times, but thanks to the efforts of a couple of our more dedicated alumni, we have bonus data! Specifically we have the calculated speeds just before the chute and at the fourth window of Porter Hall.
The columns represent the speed of the buggy entering the chute (Chute), speed at the fourth window of Porter (Window 4), and whether or not the buggy needed to be pushed before the fourth Porter window.
Official BAA Times
| Team | Roll | Buggy | Chute (mph) | Window 4 (mph) | Pushed |
| PhiDelt | 1 | Perun | 25.6 | 11.9 | P |
| PiKA | 1 | Banshee | 35.9 | 17.1 | |
| PiKA | 1 | Raptor | 27.7 | ||
| SigEp | 1 | Kraken | 30.3 | ||
| AEPi | 1 | Kamikaze | 27.1 | 9.7 | P |
| Apex | 1 | Phoenix | 22.7 | 9.7 | P |
| CIA | 1 | Equinox | 33.7 | 16.2 | |
| CIA | 1 | Icarus | 31.6 | 13 | |
| CIA | 1 | Orca | 29.8 | 9.1 | P |
| CIA | 1 | Ascension | 23.6 | 9 | P |
| SDC | 1 | Malice | 35.2 | 15.8 | |
| SDC | 1 | Vice | 31 | ||
| SDC | 1 | Havoc | 30.5 | 9.8 | |
| Fringe | 1 | Bissa | 34.1 | 13.4 | |
| Fringe | 1 | Beacon | 12.5 | ||
| Fringe | 1 | Balius | 10.8 | P | |
| SigNu | 1 | Krait | 30.7 | 11 | P |
| Spirit | 1 | 1 | 33.9 | 12.7 | P |
| Spirit | 1 | 2 | 28.3 | ||
| PhiDelt | 2 | Perun | 24.2 | ||
| PiKA | 2 | Banshee | 35 | 15.7 | |
| PiKA | 2 | Raptor | 28.1 | ||
| SigEp | 2 | Kraken | 30.3 | 10 | P |
| SigEp | 2 | Pandora | 17 | ||
| AEPi | 2 | Kamikaze | 26.5 | 9.2 | P |
| Apex | 2 | Ember | 27.6 | 10.7 | P |
| CIA | 2 | Equinox | 31.1 | 11.6 | |
| CIA | 2 | Icarus | 31.7 | 12.7 | |
| Orca | 14 | ||||
| CIA | 2 | Ascension | 23.8 | 8.1 | |
| CIA | 2 | Impulse | 27.2 | 10.5 | P |
| SDC | 2 | Malice | 35.1 | 16.1 | |
| SDC | 2 | Vice | 31.5 | 11.9 | |
| SDC | 2 | Havoc | 30.4 | 10.9 | |
| Fringe | 2 | Bissa | 33.6 | 13.7 | |
| Fringe | 2 | Beacon | 33.7 | 12 | |
| Fringe | 2 | Baleus | 28.9 | 9.3 | P |
| Spirit | 2 | 1 | 33.2 | 15 | P |
| Spirit | 2 | 2 | 30.9 | 11.1 | P |
| Spirit | 2 | 3 | 30.7 | ||
| PhiDelt | 3 | Perun | 24.5 | ||
| PiKA | 3 | Banshee | 35.2 | 17.1 | |
| SigEp | 3 | Kraken | 27.2 | 7.2 | P |
| AEPi | 3 | Kamikaze | 26.9 | 8.6 | P |
| Apex | 3 | Phoenix | 30.5 | 11 | P |
| CIA | 3 | Equinox | 33.3 | 16.4 | |
| CIA | 3 | Icarus | 30.4 | 11 | |
| CIA | 3 | Orca | 28.4 | 9.6 | P |
| SDC | 3 | Malice | 34.8 | 15 | |
| SDC | 3 | Vice | 32.8 |
The take home points are:
- PiKA is the fastest downhill and back up
- SDC is 2% slower than PiKA on the downhill, but 3rd fastest uphill
- CIA is another 4-5% slower than PiKA on the downhill, but depending on the buggy are between SDC and PiKA going back up
- CIA, Fringe, and Spirit have almost equal downhill speeds, but Fringe and Spirit are scrubbing speed in the chute.
- Spirit’s down hill times would suggest a greater rollout than they are getting
Speeds were calculated by capturing video using a Nikon 1 series camera, all of which have the ability to take 640 x 240 movies at 1200 frames/sec. The cameras can only record 3 seconds at that speed, so the time it takes a buggy to pass between nearby landmarks (e.g. pairs of trees or parking spot lines) can be precisely measured. After some scouting, our intrepid alumni found useful spots on the course where every buggy takes more or less the same path. They placed the camera where every buggy could be filmed with the same perspective. Each frame is a normal 2D image, so objects in the background remain stationary. This allows you to open the movie in your video editing software of choice and note the time at which the nose of the buggy reaches each landmark. If the distance between the landmarks is known, the time can be converted to a speed.
For comparison, Sweepstakes uses the FinishLynx system on Raceday, which is typically set to take a 1000 x 1 image 1000 times/sec. The finish line software then assembles a series of these 1D lines into a distorted 2D image, where the horizontal axis is time. If you know the length of the buggy, you can tell when the nose and tail cross the finish line and convert that into a speed, however you can only observe one line. The system is optimized to calibrate its clock to real life and not vary much. This allows for accurate split times by placing a camera at each split point. The internal clocks used in consumer-quality cameras are not as accurate, so even though we used multiple cameras focused on different physical locations, we cannot correlate events in each well enough to get useful split times.
More work is needed in calibrating against either another measurement method or something of known speed before comparing speeds obtained our way with other sources, or, for that matter, with historic times. Regardless, comparing 2 teams on the same day should be safe – the systematic errors should affect everyone equally. The video files used to generate these times will be made available on cmubuggy.org for the viewing pleasure of all interested.
—
This fabulous data is brought to you through the ideas and efforts of Shafeeq Sinnamohideen and Tom Wood.