Errno 3.14

From GAS Wiki
Jump to navigation Jump to search

Errno 3.14

27 May 2017

Launch Location Blue Creek, Utah
Duration of Launch 2:30:00hr
Landing Location Mt. Logan, Utah
Time of Recovery 28 May 2017
Launch Team
Raul Ramirez
Hailee Maxwell
Jacob Yates
Carlos Denoso
Sam Dalrymple
Kaden Ledbetter
Ian Jepson
Scott Glaittli
Hailey Hampton
Nathan Hampton
Radio Team
Ammon Hepworth
Kelly Birch
Recovery Team
Ammon Hepworth
Kyle Hovey
Jack Conrad Kiefer II
Whittney Despain

The goal of this flight was to achieve deployment of the Aeroboom that the GAS team has been developing for some time now. It reviews the data that we have received, the achievements of the flight, and the faults of the flight.

The Goal of HAPCAD, now known as ERRNO[edit | edit source]

High Altitude Payload for CubeSat Aeroboom Deployment (HAPCAD), was originally created by James Gardner, had attempted to deploy an Aeroboom which was engineered and constructed by the Get Away Special (GAS) Team. Upon reviewing their flight report the GAS team identified where they went wrong and was able to make changes for future flights. The first flaw was reliance on time for deployment. James' team did not take into account that the balloon would not have an identical assent rate across flights. Therefore, it deployed much too soon and at too low of an altitude for proper inflation. The second flaw was the lack of direct communication between the payload and the team on the ground.

The GAS Team set out to fix these flaws and in doing so lead to the creation of the Elevated Real-time Research Near-earth Orbiter (ERRNO). The GAS Team's software team was able to fix this first flaw easily by adding a Barometer, called the MS5607, to measure the payload's altitude. The Barometer is accurate up to 120,000 feet. ERRNO, was designed to focus on the Barometer as a primary instrument to tell the payload to deploy at a specific altitude. From the GAS Team's mechanical team, they determined that a pressure reading of 44 millibar (75,000 ft), as described by James in his original report, would be the correct altitude and pressure to deploy the Aeroboom. ERRNO was then programmed to deploy the boom in the pressure range of 30-44 Millibars.

The second flaw was fixed by having direct 2-way radio communication with ERRNO. With the RFD900, the software and the GAS Team's electrical team were able to allow ERRNO to have radio communication. ERRNO has successfully sent and received information at over 100,000. ERRNO's backup deployment system is a manual override received directly from a user on the ground. "b", for boom, is pressed on the computer in communication with ERRNO to enable the back up deployment. The system then confirms deployment by prompting the user to press "y", for Yes. Confirmation is required only to deploy deployed manually.

Mechanical Team[edit | edit source]

Observed Problems[edit | edit source]

In the summer of 2016 we had a few different high altitude balloon flight to test the Aeroboom. During this time we noticed some problems to the boom. The boom was not holding air the way that it was intended. The inner boom is intended to hold all of the necessary air for deployment while the layer between the inner boom and the outer boom is evacuated of air. At this time the air was escaping from the inner boom into the layer between the inner boom and the outer boom, making it difficult to for boom to remain effective for even a short period of a few days. Even shortly after the boom was glued into the box and deployed in the vacuum chamber the boom popped out between 6-10 inches and then continued slowly expanding from there. We attributed the poor ability to retain air first to poor seals and sealing techniques, and then later to permeability of the lay-flat plastic tubing that we were using. To seal the boom we needed to insert the end into the heat sealer for a period of about 5-7 seconds. Afterwards the plastic would stick to the hot metal of the heat sealer requiring steady and delicate movements to remove without effecting the viability of the seal. This by itself led to wildly inconsistent seals, with failed seals leading to wasted material.

Testing and Changes[edit | edit source]

Towards the end of summer and into the fall of 2016 we started conducting tests and making small changes to the boom. As poor sealing was the biggest problem, we looked for ways to improve the consistency and quality of our sealing method. We eventually came to find that kapton was a good non- stick material. We decided to try wrapping kapton tape around the end of the plastic tubing before sealing and we increased the sealing time from 5-7 seconds to 35-40 seconds to ensure that Teflon seals through the kapton tape. This method proved both successful and consistent. Under burst testing the Teflon tubing fails before the seal. We conducted more tests to determine a rough idea of the lifetime of the booms with the new seals. We made booms of just the inner boom Teflon as well as booms of just the outer boom Teflon. These test booms we placed in the vacuum chamber for extended periods of time. When the test booms were laid flat in the vacuum chamber the lifetime was about a week. The test booms that were z-folded and clamped lasted at most three days. The new seals helped the booms to hold air more than in the past. However, air was still escaping, albeit at a slower pace. This time we decided to look at the material itself. Every material has a rate of permeability as determined by the pressure difference, the thickness of the material, and also by the nature of the material itself. In our research we found that there were some materials that had a better permeability coefficient than the Teflon we were using. Seran has the greatest permeability coefficient, however it has very low strength. Mylar, or biaxially-oriented polyethylene, has a great permeability coefficient while also having good strength. We spent a good deal of time trying to find lay-flat Mylar tubing. Unfortunately, Mylar is only produced in sheets. We next decided to look into creating the inner boom with multiple layers to create a more gradual pressure gradient across layers rather that a sharp pressure drop. We found that we had difficulty with the seals for this. However, a couple of weeks ago one of our team members had an idea to wrap the entirety of the inner boom in kapton. Kapton has a decent permeability coefficient and it is a material we are familiar with that we use regularly for our seals. We made a couple of booms with a kapton lined inner boom for testing. This type of boom lasted in the vacuum chamber for about a week and a half. Additionally, before the spring flight this year we tested the deployment of this type of boom from a box on the payload inside of the vacuum chamber. During this test deployment the boom immediately extended to the full intended length, which was a large improvement from deployment tests of earlier booms which only extended about 8 inches before slowly expanding the rest of the way.

Other Changes[edit | edit source]

The initial design for the boom called for having some putty at each end of the inner boom to act as a support seal. However, during our vacuum chamber testing we found that the putty would begin to pull away from the sides, thus ruining its purpose as a support seal. We have since modified our construction process as we have deemed the putty be unnecessary and it is taking up valuable space. We are also planning to change how we fill the boom with air from using a metal rod to instead using a syringe to fill the boom with the exact amount of air we wish to use. In addition to this we are looking to make a box that will be reusable from boom to boom. Both of these changes will allow for our testing to be more exact as well as making our booms more consistent.

In conclusion, we have worked to fix the seals of the boom and reduce its permeability by wrapping the boom in kapton. We have removed the putty from the design of the boom and working towards being able to inject an exact amount of air into the boom. All of these changes have proven to be good improvements, and we are hopeful for successful high altitude balloon launch and boom deployment.

The Flight[edit | edit source]

The second balloon launch of the summer, which took place on the 20th of May, was an overall success. There were issues with the boom itself, but it deployed at the correct altitude, which was better than the past launch. Next time changes will be made so a successfully deployed boom will come back down.

What Went Wrong[edit | edit source]

Figure 1 - New boom (close) and old boom (far) in vacuum chamber. Both Kapton booms are fully inflated. The older one is more uniform and cylindrical.

Two weeks before the launch, some members of the mechanical team noticed that the fiberglass supply was running low. Due to some miscommunication, more fiberglass did not get ordered until the week of the launch. The team would have been okay with the amount of fiberglass left if a few troubles had not been run into in the boom making process.

On schedule, a few members met to make the boom for the launch. Everything went smoothly until it was time to take the flat metal rod out. In this portion of the process, the inner boom is inside a sleeve of fiberglass, taped in such a manner that a rod can be used to push both materials into the outer boom. In a race against the clock, heated epoxy is painted onto the fiberglass as it is slid into the outer boom. At the end, the rod is flush with the seal of the outer boom, and a team member must try to grab some of the inner material so the rod can be pulled out without disturbing the rest of the boom. This proves difficult, as trying to pull the rod out frequently results in the inner boom and fiberglass moving with it.

The same problem occurred in this boom making process, and despite the best efforts of team members, the epoxy hardened too much and the rod could not be pulled out. The whole boom had to be scrapped. With just enough fiberglass left for another boom, other team members attempted to make a boom the next day. To avoid the same issue, a smaller flat rod was used to push the inner boom and fiberglass into the outer boom.

This resulted in the fiberglass significantly stretching, allowing a significant amount of space between the inner and outer layers (see Fig. 1). This made the boom unfit to use, as the fiberglass, which is what gives the boom its structure, was not a good cylinder. One team member then suggested using an old boom, and they were able to scrounge up a two-month-old Kapton boom that was decent to launch. Some air had made it into its outer boom, but the Kapton boom was still in pristine condition compared to the boom that was just made.

One issue that was foreseen with the old Kapton boom before the flight was the bit of extra air in the outer boom. It caused the fit in the boom box to be tight, and the team accepted that the box might be slightly open when tied. When the boom came back cured in a kinked fashion, the team was not surprised. Ultraviolet light got into the box during the flight, as it was not closed all the way.

Another significant issue the team ran into was that the payload ended up at around 9700 feet on Logan Peak. Seeing from this experience, wind patterns and predictions thereof cannot be trusted. Allowing the payload to be completely at the will of the wind and when the balloon pops has already presented itself as an issue when a payload was completely lost. Thankfully, this time the tracker on the payload continued to run much longer than expected, allowing a few team members to execute a rescue mission. Although it worked out this time around, it should not be expected to work out so well in the future.

What Went Right[edit | edit source]

There were quite a few aspects of the launch that were encouraging. The fact that the Kapton boom used had been sitting on a shelf for not only months, but had been left in a vacuum chamber for a week, and was still in good condition, was incredible. The wire cutter issue that occurred last time, which resulted in the boom not deploying, was remedied. Extra precautions were taken, such as putting on a second set of wire cutters, to make sure the wire would cut. The coordination of team members going up to Logan Peak was incredible and showed an immense amount of initiative and devotion to the project. Overall, the launch was successful in testing the wire cutters, as well as giving the team another project, which is figuring out how to control balloon landings.

What Next[edit | edit source]

Learning from this experience, materials will be replenished well before they run out. Booms will be made in more advance to the launch date, since the Kapton has worked so well in keeping air in the inner boom. To remedy the issue of pulling out the inner boom, a notch will be cut in the metal rod to ensure the material can be grabbed, and no more booms will have to be scrapped. To fix the boom curing prematurely, either the box itself will be lengthened, or the booms will not have extra air in them from sitting for a few months. The issue with the balloon ending up wherever the wind takes it will be remedied by creating a sort of cut down system to pop the balloon on command. Making these changes will certainly put the team that much closer to a successful deployed boom.

Electrical Team[edit | edit source]

Figure 2 - full schematic of payload before secondary wire cutters were added
Figure 3 - schematic of the attached secondary wire cutters

During the previous balloon launch, the payload, despite three separate attempts of boom deployment, was recovered with the boom box still tied closed. After testing the wire cutters in temperatures similar to those found in the boom's typical deployment range and unable to recreate the error, it was determined that a boom switch and secondary wire cutter should be implemented. The satellite's designs already included a boom switch for verification of deployment. The boom switch rests just beneath the lid and reads open upon successful deployment. The only changes necessary for this implementation was adding a small notch next to the camera for the boom switch's wires to leave the inside of the box and writing additional code for the software to test whether the switch was reading open or closed. For the secondary wire cutter, a little more work was necessary, which included designing and implementing another PCB (printed circuit board) with the necessary components attached.

Changes Made[edit | edit source]

The purpose of the secondary wire cutter was to work in a loop with the primary, until the boom switch read open or a predetermined amount of time had passed. First the primary wire cutter would run for the designated three seconds, and then switch off. The secondary wire cutter would run for another three seconds, before also switching off. These two wire cutters would then increment their time by half a second and continue the same process until both had ran for five seconds. At that point, if the boom switch still read closed, the code would return a message detailing that the boom had failed to deploy. A manual deployment could then be sent to attempt the process again.

During the week leading up to this most recent flight, the schematics for the payload were found and studied. From there, the base components that corresponded to the primary wire cutters were copied and integrated into the drawn circuits. With very little time until the flight, a temporary PCB was designed, etched, and then the necessary components soldered in place. Upon testing, both wire cutters worked as prescribed and the PCB was taped into the satellite a night or so before the balloon flight.

Problems[edit | edit source]

With implementing the boom switch, no significant problems arose. The schematic had been designed intending to use the switch, so only the code needed to be written and tested. No major changes were made to the payloads design for this implementation, and everything with the boom switch ran smoothly while in the air.

While attempting to add the secondary wire cutter, due to the limited time and access to materials, a couple complications arose. The foremost of these complications involved determining a place to attach the newly designed PCB. Upon discussion with Mechanical Team, it was determined that the battery was to be ousted from the base of the payload and a place for it was carved into the interior wall of the Styrofoam insulation. This battery was then padded with more Styrofoam in hopes that the extra outside layer would help to keep it warm. The PCB was then taped to the bottom base plate of the payload next to the camera.

During flight, when the boom attempted to deploy at the specified pressure of 44 millibars, the payload instead lapsed into a strange loop where it would restart itself, read that it was within the correct pressure range, and then retry heating up the primary wire cutters before immediately shutting off again. Scrambling to figure out what could be causing this problem, the only thing we noticed out of the ordinary was that the battery voltage was reading very low. Attempting to manually deploy the boom did nothing to stop this loop, and, for the most part, we just had to wait until the balloon lifted the payload high enough that it would exit the pressure range for boom deployment and we could attempt the send another manual confirmation. However, after about ten thousand or so more feet, the voltage on the battery started rising again along with both the eternal and internal temperature. There came a point where the payload broke out of its loop and sent a confirmation that the boom switch was open, indicating the boom successfully deployed. We don't have an accurate reading on exactly where this point was, since the boom deployment information that was to be written on the mini SD card was corrupted.

Possible Solutions[edit | edit source]

Our current hypothesis is that there was not enough voltage given by the battery available to keep everything running and also heat up the wire cutters, hence the restarting loop. Though we are currently uncertain, statistically, if there is a correlation between the temperature and the battery voltage, we did observe that the voltage started rising again once the payload started heating up. It was during this point that the payload stopped restarting and successfully deployed the boom. Until we can further analyze the collected data, we have determined to better insulate the battery and change the pressure range for deployment such that it is between 15-30 millibars instead of the 10-44 millibars it was before. In the future, we also hope to redesign the PCB's such that the secondary wire cutters can be better implemented with all the other components, reducing the amount of empty space on the boards and compacting them down to the necessary size for our final design.

Conclusion[edit | edit source]

ERRNO went really well. Apart from small issues that the flight had, as low voltage, late deployment, and boom not inflating. We did have a larger issue that software will begin looking into, the data that was on the SD card was corrupted. Software will begin looking into why this happened, as the theory right now is that the SD Card was not formatted properly before flight. The battery voltage stopped recording data so it is impossible to tell if temperature has an affect on power with our data. However, our research online suggests that temperature does play a massive effect on how battery voltage comes into affect, but the next flight will be able to tell us that. Software has also multiple image captures for the payload. When the box opens, it will be delayed for about 5 seconds. After the 5 seconds it will take 5 pictures consecutively so the GAS Team will have the chances of capturing a photo.

Software is also working on image retrieval from a radio using Slow Scan TV. The only issue that the GAS Team can see is that SSTV uses 100% power utilization. Which can be a massive problem if the battery voltage is too low.

Data[edit | edit source]

altitude compared with pressure over time
light intensity over time
ultraviolet light intensity over time
temperature in degrees Celsius over time
voltage compared with temperature over time
gyroscope data over time