Complications

Wednesday, Thursday, and Friday of this week presented some significant challenges. Wednesday was our flight requirements meeting with CSBF, which was generally straightforward–we discussed our altitude and duration targets as well as various interface issues, and CSBF gave us details about our specific balloons.

The meeting reminded us, however, that we didn’t have a good idea of our total gondola weight–when we had filed our application, we didn’t yet know how much the new solar power systems and detector electronics would add. After the meeting, then, CSBF helped us weigh our various components, and we got a nasty surprise–we weigh about 3400 pounds, or 1000 pounds more than our estimate!

Alan attaches the load cell.

Alan attaches the load cell.

2485 pounds for the gondola...

2485 pounds for the gondola...

Weighing a solar panel

Weighing a solar panel

237 lbs. *per panel*--no wonder they're so hard to move!

237 lbs. *per panel*--no wonder they're so hard to move!

Mark does the grim arithmatic.

Mark does the grim arithmetic.

At this weight, we will have to use a larger parachute. Thankfully, we can still use a “light” balloon–if we weighed more, we’d need a balloon with thicker panels, which would cut our altitude further, and hence there would be more atmosphere to absorb the gamma-rays. (I’ll have more information about the balloons themselves in an upcoming post.)

One problem remains: Our rotation assembly, from which our gondola hangs, has not been rated to hold our heavier weight! The major concern is “chute shock”–when the parachute opens, the gondola sustains a sudden deceleration of up to 3 Gs, so a two-ton object weighs six tons for a very brief moment. (CSBF tells us chute shock has been much reduced–back in the 70s and 80s, chute shocks could reach 10 Gs! McBride related the story of one gondola which broke from the chute and free-fell 100,000 feet to Earth. They had to dig the remains out of the ground…)

It’s possible our rotor will work at our weight; we’re awaiting the results of the mechanical analysis to know for sure. Still, we have options for dropping weight if we must. The solar panels are the largest source of new weight–250 lbs. apiece! (Comparable panels designed for balloons are more like 50 lbs.) While we do need their power, we’re mainly testing them for a future, longer flight, so we could cut them down substantially if needed.

Thursday I worked on the flight code. Our disk data storage has a weird problem which I’ve now determined is serious. Basically, we’re losing a significant amount of the data written to the disk in a systematic way. The problem is somewhere in the data buffer code–there are two “bins” for data, one of which gets filled with new data while the other is written to disk. They get swapped when the “new data” buffer gets filled, but the swaps sometimes miss in a strange way. Debugging this is complicated, because several independent processes can write to the buffers. The memory is protected, as it must be, so the processes don’t overwrite each other, but it appears the protection scheme we have in place is not right, yet.

I also feared I had found a major bug in the pointing code–I couldn’t get the gondola’s pointing system to move–but after three hours I discovered I simply hadn’t commented in the part of the code that drives the appropriate motor. This fact was obscured for a variety of reasons, but still: sheesh.

Finally, on Friday we had a large amount of unexpected noise appear on several of the detectors. It seems to be in the electronics, from what I gather from Mark and the engineers, but I don’t have more details at the moment.

The machine shop is finishing up our new heat-sink parts for the card cages today, so we took our weekly day off today and will install them tomorrow. I wandered through Ft. Sumner and took a bunch of pictures; other than that, it’s been a day of some much-needed vegging out.

About these ads

2 responses to “Complications

  1. That sounds like a rough couple of days. Hope the disk storage issue can be resolved. Are all the reads and writes locked by mutexes?

    If all else fails, just do a ‘cat /dev/null’, I’m sure it’s there somewhere! :>

    We’ll be crossing our fingers that the NCT balloon doesn’t pull a Genesis…

    • Right now the protection is with binary semaphores–that’s how the code I started with implemented the protection. I’m going to try switching to mutexes for the read/write cycle and see if that helps…