We planned a day to test some of the systems that failed at our last launch attemp. Our hypothesis was that the ATV system with the power amplifier on was swamping the GPS, the 802.11b telemetry, and the recovery node's 2m receiver. So we met at Andrew's cabin and packed everything up and headed to the front lawn of Reed College in the southeast to do a short range (500 ft?) "ground station" to "launch tower/rocket" test.
Of course, not everything went as planned (!).
We met at 9:30am and it took us about 1 hour to load the ground station up in Keith's van and everything else up in Dave's truck. Once we got to Reed, it took us another hour to get the system up and running. During setup, Dave and Jay left for Brian's house to pick up the launch tower and load it into Dave's truck. Apparently the process almost killed them: I think we can officially say the launch tower takes 4 people to move. But somehow they managed to get it in the truck.
Keith, along with Jamey, had the groundstation in his van, with the PSAS field server, the trackmaster 2000 antennas, and a couple of 12V batteries courtesy Glenn to power everything. Keith parked halfway between Elliot Hall's front entrance and Elliot Circle on the sidewalk/roadway that runs in front of Elliot.
Andrew, Jeromy and Jay hung out at the rocket/launch tower side of things by the SE corner of MacNaughton Hall, a few dozen feet from the driveway. This gave us a nice line of site to the ground station.
We wasted a bunch of time trying to get the rocket turned on, which was due to two problems:
- Keith turned the FC on with the launch tower. The 802.11b power amp was off, so the ground station couldn't see that the FC had come up. And Andrew couldn't see the small red LED in the bright sunlight since it's on the OTHER SIDE of the FC power connector (duh) so we ended up pulling the avionics system out to see what was going on. And indeed, it was on, just like we expected. So we decided to do the serial console thing.
- Andrew's brilliant plan was dashed by a stupid - if not blindinly obvious - cabling error: from Keith's laptop, we ssh'd into the LTC from via field server and then ran
apt-get installon the LTC, of course) on the LTC in order to have the ground station be able to have a serial console to the rocket. But it turns out that the cables we have for the MOPS/520 boards - the FC and the LTC - are both null modems and so of course two null modems together make a null null modem which... doesn't work. So we used the CANtalope laptop to serial in and run
run_threadswith the power amps on, and lo and behold, the ground station could now connect directly with the rocket.
At this point Dave and Jay arrived with the launch tower... but to our dismay we found out that the launch rail wasn't on the launch tower. Apparently it was taken off after the launch attempt and not placed near the LT so Dave and Jay never saw it. So we cancelled the test of the rocket on the launch rail (we'll have to do it over again, this time with the rail). Poor, poor Dave and Jay. Thank you guys for all that effort. At least the launch tower is now at Andrew's cabin where it's closer to everything else, including people to help load it ;)
We turned on the power amps and ran run_threads. Without the launch rail, we got very clear ATV and 802.11b connectivity... as we expected because of previous tests, this worked just fine. But with both power amps on, the recovery node could barely hear 2m commands. With the HT only inches away, only about 10% of the DTMF tones were registered by the recovery node. Turning the ATV power amp off cleared up the problem.
The last thing we did was try and get the GPS to lock. We never actually did get it to lock: even with all the poweramps off, the GPS just wouldn't lock despite having a nice constellation according to Keith's handheld GPS. We got the GPS to lock the Wednesday before without any power amps on, but we just couldn't get it this time. There's clearly something marginal about the GPS cylindrical patch antenna and low noise amplifiers... Jeromy offered to have us take the rocket in to his company, Maxim Semiconductor where he has access to RF test equipment like spectrum analyzers.
Lots of people stopped by to talk. Thanks in particular to Bob Ormand, Electronics Guy of Reed College and Andrew's Mentor from Days Long Past, who swung by and gabbed with us and shot the only photos of the day!
Here's what we learned:
- Something is wrong with the the GPS system; it never locked, even with all the power amps off. It always locks with the external flat patch antenna, so this has to be an antenna/LNA problem.
- The 2m receiver on the recovery node is definitely being swamped by the ATV system when the ATV PA is on. Why? Can we filter it out?
- The ATV didn't seem to affect the 802.11b telemetry, but we also didn't have the launch tower so we couldn't really test how the launch rail couples it.
Possible things to do:
- Use a spectrum analyzer to find out who's clobbering who.
- Box the ATV power amp
- Put a filter on the output of the ATV power amp
- Put filters on the 2m receiver
- Put filters on the GPS