wiki/ news/ 2005-08-20 - A spectacular flight of LV2, but with an unhappy ending

2005-08-20 - A spectacular flight of LV2, but with an unhappy ending


  1. 2005-08-20 - A spectacular flight of LV2, but with an unhappy ending
    1. Summary:
    2. Friday
    3. Saturday
    4. Launch of LV2
    5. "Recovery" of LV2
    6. Data
    7. Conclusion


We finally launched "Launch Vehicle No. 2" (LV2) with a complete avionics system to over 18,000 ft (above ground level) on August 20th, 2005, outside of Brothers, Oregon. Technologically speaking, the flight was a stunning success in the sense that the avionics worked far better than we expected. Unfortunately, due to a failure in the nose cone ejection system, the rocket was destroyed when it hit the ground at over 500 mph.

This flight proved:

  1. an ARRL 802.11b telemetry system (IEEE 802.11b run in channel 1 with a power amplifier under an amateur radio license) worked over several miles to a vehicle accelerating at over 15g's and over Mach 1
  2. our Linux-based PC-104 586 flight computer and the associated "sequencer" software sequenced the flight correctly
  3. our inexpensive, "open hardware" cylindrical microwave patch antennas on the rocket worked beautifully with our 1.253, 1.575, and 2.412 GHz radios. Since they are less than 1/8 inch thick, the patch antennas conform to the airframe and add no extra aerodynamic losses.
  4. our ground network, including a Linux-based router and java based control (launchcontrol) and viewing (rocketview) software worked without any apparent failures
  5. the 1.253 GHz 3W FM Amateur TV broadcast system worked far better than expected, with little snow or fade throughout the flight.

We experienced three failures:

  1. The nose cone never deployed, which eventually destroyed the rocket,
  2. The digital video camera on the ground accidentally recorded the amateur TV to a digital memory card instead of to digital video tape. The memory card soon filled and stopped recording before the flight, so the amateur TV broadcast was never recorded.
  3. The backup, or "away" ATV system stopped recording for an unknown reason. This meant that we have no real recording of the ATV transmission (again). (We do have the 1st 3 seconds courtesy of a third party who also recorded part of the ATV transmission.)

For the complete list of pictures taken at the launch, please see the PSAS gallery album for this launch


Friday morning around noon the heavy infrastructure - launch tower, generator, tables, chairs, antenna mounts, etc - left for Bend. We were pretty scattered this year, with people leaving at random times in the morning. We cross-referenced everything to bring using the CurrentBringList, and shockingly, only one relatively minor thing was forgotten (because, of course, it wasn't on the checklist! And yes, now it is on the checklist).

We arrived early evening at the Brothers site, and dumped the launch tower at an away cell south of the usual Oregon Rocketry flight line. There was some confusion over whether we could set up launch control as far away from the launch tower as we wanted, so we didn't do much setup work on launch control until early Saturday morning.


Another huge success for PSAS was the fact that we did a setup, launch, and teardown in essentially a single day: because we really didn't get much setup Friday evening, we did almost everything on Saturday. We've never been that light on our feet before, and it's a testament to refactoring our infrastructure to be "prebuilt" before we get out there. For example, the launch control table and the launch tower computer.

We set up two pop-up launch control (LC) tents about a mile away from the launch tower. We were that far away because we wanted the receive antennas on the ground to be more than 10 degrees off of the vertical null in the rocket's cylindrical patch antenna pattern. At a simulated apogee of 18,000 ft AGL, a mile gave us about 17 degrees off vertical assuming the rocket was going straight up.

launch control flight discussion

The launch tower (LT) and launch tower computer (LTC) were set up in parallel with LC, which meant that by about 1:00pm on Saturday things were basically ready for launch. The only setup notes we have are:

  1. We need a bubble level on the launch tower to help us level it.
  2. We had a remarkably hard time with the signal strength between the LTC and LC. We think this was because the wire parabolic dish antennas (we call them "BBQ grills") didn't work as well over the bumpy terrain (they really should have been higher off the ground) and - perhaps more importantly - they've been really abused over the last few years. We need to replace them with high gain helical antennas (Yagis?) with external covers so we can throw 'em in the back of a pickup and not be worried. When the wireless link finally connected reliably, the LTC BBQ grill was about 10 degrees to the right of where it "should" be pointed as sighted along the pickup arm.

Remarkably, since we were ready to proceed with the launch, we bolted the airframe together, including the prepackaged avionics module, and did a few power up tests according to the CurrentCheckList: turned on the avionics module, focused the ATV camera lens, and taped it down. Then we slid the aeroshell on and checked the picture again to make sure the ATV prism was on correctly. Brian and Maggie put the nose cone on, slid the motor in, and we were ready to go!

Photo_2005_0820_115540AA.thumb.jpg Photo_2005_0820_120654AA.thumb.jpg

At the launch tower we had a few problems, mostly related to 1) being able to reliably talk to the launch tower, and 2) the polarization of the uplink radio antenna. It turns out that the circularly polarized 2m turnstile antenna that allows us to send emergency 2m commands to the recovery node is horizontally polarized at the horizon. And since the 2m "long wire" antenna on the airframe of the rocket is vertically polarized, the 2m commands never got through. Once the rocket left the ground, we were sure the turnstile antenna would work correctly, but we never realized it would be unusable on the ground. We used the 2m comm radio (which has a vertically polarized J-pole antenna) to arm the rocket on the tower and everything worked as expected).

20_LV2_Erected_to_Launch_Position.thumb.jpg 15_Keith_and_Sara_Work_on_Launch_Tower_Communications_Note_Mt_s_in_Background.thumb.jpg

We were ready for launch, so Holly called in to have the Oregon Rocketry folks activate the 1-hour high-altitude FAA waiver.

Photo_2005_0820_141641AA.thumb.jpg Photo_2005_0820_141517AA.thumb.jpg

At this point we had several instances of human error. First, Jamey forgot to uncomment a line in our prepare-for-flight shells script, which made the GPS never give a "sane" value because the FC thought it was supposed to be in Portland, not Brothers. After about 5 minutes waiting for the GPS, we looked into it, fixed the error, and restarted the launch sequence. Not to be outdone, Andrew proudly one-upped Jamey when, with 7 minutes left to go in the hour-long high-altitude window, Andrew tried to type in the "arm rocket" DTMF command over the uplink radio and his fingers instead dialed the "emergency shutdown" command... which, of course, instantly shutdown the whole frickin' rocket. AAARGGGHH! So we missed the first high-altitude window.

We called into to the Oregon Rocketry folks again, looking for another high-altitude window. They called the FAA to see what they could do... and luckily, we got a 2 hour window from 3:45 - 5:45. Amazingly, we had already power cycled the rocket and were up and ready to go... 10 minutes before the waiver!

For the first time in the collective history of the project, we waited around with nothing to do before a launch. We were stunned (and happy).

At 3:45pm the window activated, and we pushed the launch button on the java launchcontrol software.

Launch of LV2

The countdown went exactly as expected and LV2 leapt off the launch tower. Nick tracked the rocket with our "TrackMaster 2000" shoulder-mounted receive antenna array.


Everyone else watched the launch, or huddled around the three laptops running rocketview or the ATV monitors. Dave got some amazing photographs of the launch:

206_lv2_liftoff_edited.thumb.jpg 207_lv2_liftoff_edited.thumb.jpg 208_lv2_launch_edited.thumb.jpg

... as did other folks:

lv2_launch_zoom.thumb.jpg fred_lv2_launch_DSC_5343_cropped.thumb.jpg 26_LV2_Blasts_Off_1.thumb.jpg 28_LV2_In_flight_1.thumb.jpg

Glenn reported that the ATV video was fantastic - almost no snow and was clear the whole flight.

We saw the telemetry stream down on rocketview, which was exhilarating to see in real-time.

We even saw the commands to fire pyrotechnic charges right at apogee.

What happened next is unclear, but for whatever reason, the nose cone did not come off the airframe. This meant that no parachute ever deployed... and thus after apogee, LV2 keeled over and started descending ballistically. We received telemetry all the way to approximately 300 m above the ground, just a fraction of a second before impact, which gave us the GPS coordinates of the impact site. Prior to impact the rocket was still accelerating. The final velocity at impact was more than 560 mph.

lv2_descent_zoom.thumb.jpg 100x75-rv.png

Update September 6, 2005: We think the gun powder recovery charge didn't burn completely, which in turn didn't pressurize the recovery module enough to blow off the nose cone. For more details, please see the postmortem analysis page.

"Recovery" of LV2

Three separate recovery teams trekked out to the last known GPS coordinates... and there was LV2, embedded in the soft dirt. Had it not been for the GPS, we would never have found it. It was embedded below the level of the sage brush and very little of the rocket was actually above ground.

We're trying to figure out what exactly happened. Did the pyrotechnic charges fire? We think they did... but then why didn't the nose come off? Was the gunpowder charge bad? Did the nosecone get jammed on the recovery module? We hope we'll know more after a careful analysis of the remains of LV2, which we found about 0.7 mi from the launch tower based on the last GPS coordinates received from the rocket.

33_LV2_Dug_Into_Ground.thumb.jpg dscf1809.thumb.jpg

We dug up the rocket to discover it was totally, and completely, destroyed. Nothing was left... not a single circuit board was left intact, including the compact flash card we were hoping to recover. We headed back to the launch control tents. It was, frankly, a pretty depressing end to a stunningly successful launch.

dscf1836.thumb.jpg dscf1855.thumb.jpg dscf1851.thumb.jpg 35_LV2_After_Ground_Removal_1.thumb.jpg

We looked over the remains at launch control, and were pleased to see that Baker the Sock Monkey Astronaut had survived -- barely -- the crash. We're sure that if Maggie has her way, he'll be rebuilt stronger, faster, and better for next launch.



We got telemetry data, and put some analysis at data


We tore down camp and headed back to Portland. Despite losing a person-year or two of work, we have to admit we were elated with the launch. Everything worked beautifully, except for the one system we didn't expect to fail.

So on one hand, congratulations to everyone for all the enormous success of the day. We'll be processing the data we received on the ground for many, many months to come. And on the other hand, condolences to everyone for the loss of a lot of hard work.

So... anyone up for LV2.3? Or how about LV3? ;)