wiki/ news/ 2005-07-03 - System Test at Gabriel Park

Goal: Do a final "pre-launch" test of the complete LV2 system

Who and What

Who: Andrew, Jamey, Tim, Bart, Jay, Keith, Ryanb, Frank, Dave, Holly, Richard, Eric

Previous evening prep:

Here's the schedule for the test day:

10:00 - 11:00 Hardware prep at Andrew's workshop: get the launch tower and rail and ground station and rocket all ready to go
11:00 - 12:00 Set up at location
12:00 - 16:00 System test and picnic
16:00 - 17:00 Quick teardown and drop off at Andrew's workshop

Where

We spent the day in Portland, first at ?Andrew's workshop and then at Gabriel Park, in SW Porland near Multnomah. The launch tower was on 37th avenue on the east side of the park and the rocket was in the middle of the partk in a parking lot near the park's maintanence buildings (or whatever they were).

Agenda

How it all went

For some unkown reason, PSAS activities either have to be done on cold dark winter nights, or in the middle of freezing central Oregon deserts, or on blisteringly hot days. We certainly all got our share of the 1 kw/m2 of solar radiation on Sunday, which caused lots of dehydration and a few sunburns. Note to selves: bring lots and lots and lots of water next weekend :)

We met at Andrew's house at 10:00am and took about 1.3 hours to load everything, including the launch tower into Dave's truck, the ground station table and equipment into Jay's SUV, and the launch tower computer table and equipment in Andrew's van. Overall it seemed pretty efficient - we're certainly getting better at this. Note: a "to bring" checklist would have really helped here, one which differentiated between a system test and an actual launch.

We arrived at Gabriel park and scouted for locations. We finally ended up at the northeast corner of the park, where we set up the launch tower and a tent over the LTC and various test equipment, and a parking lot near the park maintenance buildings and tennis courts in the center of the park.

Launch_Control_Tent.thumb.jpg Rocket_Tent.thumb.jpg Andrew_with_TrackMaster.thumb.jpg
The launch control and launch tower tents - note in the last photo of Andrew and trackmaster you can see the launch tower tent in the upper right of the picture. It was about 300 - 400 feet away from the launch control tent.

We took an hour or so to unload and get everything connected, and in the process we realized we forgot the trackmaster at Andrew's cabin. Doh! We sent Bart back to get it, but it put us back at least an hour and a half and so it wasn't until about 2:00pm that we first turned everything on and started testing.

We once again ran into the problem where we power up the flight computer (FC) but the wifi power amplifier isn't on, so we couldn't talk to the rocket after it was powered up from the ground station. Jamey walked his laptop in ad-hoc demo mode up to the LT and turned on the power amps by hand - we were finally up and running.

Things seemed to be going well: the GPS certainly had no problems getting satellites and locking. But it seemed like we were dropping wifi packets left and right, especially when the IMU was in high data rate mode: we were getting about 1,500 packets out of the usual 2,500. We took quite a while to try and figure out what was going on. Weirdly enough, Jamey could get all 2,500 packets on his laptop with a dinky little orinoco card sticking out of it if he went about 30 feet in front of the ground station... that led us to belive that 1) the rocket wifi power amp is really working, and that 2) we clearly had some kind of wifi receive problem. We tried switching the trackmaster and the "bbq grill" antennas... only to find that they were switched around in the first place! We defintely need to label our cable ends, and not rely on cable tracking to make sure we're right. But strangely enough, switching to the right antenna made no difference in the packet loss.

At the same time, we saw two instances of the ATV displaying complete "static". Unplugging the shore power cable made the ATV pop back. Weird. We thought it might be due to the spurious noise from the launch tower's 12Vnom to 19V DC-DC converter, but we switched to an AC-DC converter (albeit manufactured by the same company) and it did the same thing. Then, it stopped. Mysteriously, we never got this to reoccure. This isn't a flight-critical issue; we can solve the problem just by turning shore power off.

Jamey left around 4pm, and it took Bart and Keith about another hour to get set up to continue the tests. Finally, Keith futzed with the N connector that went into the field server's orinioco card, and suddenly we got our 2,500 packets. We knew the N connectors on the thick 9913 coaxial cable weren't quite the right kind, and we were careful to plug them in, but apparently during all the set up the center pin had pulled out of the barrell conector. the trackmaster wasn't even connected! So the entire time we were receiving data, it was through the field server's Orinoco card's antenna, not the trackmaster 2000 with it's helical antennas and power amp. So: wifi mystery solved.

Throughout the whole day we probably had a few dozen people come up and ask what we were doing... and one boy in particular was pretty excited about what we were doing. So he brought back his own rocket - some kind of medium sized Estes model rocket - and launched it for a us a few times. And he never lost it, which was astounding.

Model_Rocketeer_Views_PSAS_Rocket.thumb.jpg
Model Rocketeer

The rocket was working as planned, with a few exceptions: the aforementioned ATV blackouts, and some freezing in run_threads, the software running on the rocket. We think they were due to "operator error" in the sense that we gave it an out-of-sequence command, at which point it locked up. Also, the avionics power system node doesn't report the correct battery current, which is annoying. We know why it's not, but apparently the code to fix the problem hasn't been burned onto the node.

We ran two start-to-launch runs, and they went flawlessly. It was past 6pm, and we were all dehydrated, tired, and sunburned, so decided to call it a day.

It's pretty clear that we "could have" launched on July 9th. Besides some cabling repair, and some software fixes, we really are ready to go. But because we just didn't get the time to do the full suite of system tests we originally outlined, we didn't feel comfortable saying that we are ready -- even if we might actually be! Going out to central Oregon and scrubbing again just won't do it. So we decided to call it a "no go" and make next weekends launch another system test day. We packed up, got back to Andrew's workshop at about 7:30pm, and it only took about 30 minutes to unload everything.

Thank you to everyone who helped - everyone really worked their butts off, and the despite the issues we encountered, we were all pleased with the final results.

Notes for the Next System Test

  1. Use the CurrentCheckList
  2. Use the CurrentBringList
  3. Would have been handy: stepladder, ATV and rocketview display at launch tower, lots more water, lots of business cards