Sunday, April 27, 2014

Failures on Launch Day: Discussion


In the launch, a number of failures prevented our successful recovery of our payload as well as the failure to recover sensor and position data remotely during the flight. In the next section, I will detail exactly what went wrong, how we failed to catch the problem before launch, and what kinds of tests can be done in the future to prevent a repeat of our payload's failures.
\subsection{Launch Performance}
Our payload suffered from an issue with the FTUs implementation, an issue with lack of range testing the AC4490 radio, a sudden issue with the MicroTrack AIO that was not properly dealt with on launch day, and a mechanical design flaw for our cell tracker payload.

To start things off, the FTUs design fundamentally worked. The FTU was designed to activate on a timer, turn on (burning through the cord it was attached to), then after 4 seconds turn back off. In order to test the FTU, we used a 12 second timer before the rope would burn. Having successfully tested the FTU, we completed the foam shell that would surround it, attached the proper length of cord to it, weighed it, and placed it in the "complete" section of our work space. This was our first mistake. We should not have placed the FTU in the complete section without setting the timer to the appropriate amount of time. The day of launch, I woke up and remembered that the FTU's timer had not been set to 90 minutes. I therefore changed the program of the FTU to set its timer to what I believed was 90 minutes the morning of launch. The second (and critical) mistake with the FTU was that instead of converting correctly from minutes to milliseconds, I miscalculated by a factor of 60 and accidentally set the timer to 90 seconds. When I told the team about the first FTU issue, they trusted that I had made the correct calculation, resulting in our payload severing its tether to the balloon before launch, exactly 90 seconds after I turned it on at the launch site. We decided to proceed with the launch immediately, and bypassed the FTU, tying the parachute to the balloon with one line rather than one line leading to an FTU and the one below it leading to the parachute.

Our radio testing for long range was so sparce as to almost be non-existent. When we tested the AC4490, as mentioned in the testing section, we found that the radio had a range in the parking lot of only a couple hundred feet. Rather than dedicate the time to explore the issue, we assumed that the radio wasn't working because of the non-optimal radio conditions present in a parking lot. We should have instead dedicated more time to long range testing the AC4490 (particularly when we were still protoyping it on a breadboard. More thourough testing would have allowed us to potentially maintain radio contact with the payload, allowing us to receive GPS coordinates throughout the flight and recover the payload, as well as collect a copy of the data from the flight remotely.

Before launch on launch day, we turned on the MicroTrack AIO and the APRS system on the computer and attempted to receive the signal from our MicroTrack. After failing to receive a signal, we called Dr. Pawlowski over to the ground station to help us trouble-shoot our issue. We could not diagnose the issue, and once again assumed that for some reason that would be fixed with better radio conditions, the MicroTrack was not functioning well. We decided to go ahead with our payload's launch immediately anyway. We did not receive any signal from the MicroTrack during our balloon's flight. We had tested the ground station and MicroTrack together in the same configuration the previous afternoon without any issues, so it's unclear what caused our setup to fail the day of launch. One possibility is that the antenna did not have a good electrical connection with the MicroTrack as a result of it hanging freely below the payload. To resolve such an issue, it would have been a better idea to secure the antenna mechanically to the MicroTrack AIO. Additionally, we should not have launched the payload without having received a signal form the MicroTrack.

Finally, our cellphone tracker suffered a mechanical failure that clinched the failure to recover our payload. Our cell tracker was the simplest and among the lightest of our payloads. It was simply a cellphone encased in a foam shell along with a hand warmer (to preserve battery life). The cell tracker was the lowest of the payloads, hanging below all others. To hold both halves of the foam shell together, we wrapped around the equator with duct tape, which also ran through a loop in the cord securing it to the other payloads. We followed the cell tracker's GPS coordinates after launch to a field, where we found the cell tracker sitting by itself with a part of the duct tape sheared through. We had not tested any of the payloads with the exception of the assembled primary payload for ability to resist forces from the load bearing cord. Had we done so, it would have been easy to see that securing the cell phone via one string running through duct tape along an orientation duct tape is designed to tear would be a bad idea. In order to circumvent this type of failure, future teams should test to see how well their payloads are secured to the line, and emphasize mechanical support over light weight.

Written By: David Juenemann

No comments:

Post a Comment