Altimeter funny business...

Rocket geeks like to know how high their birds fly; over the decades, we have gone from "just guesstimating" to visual tracking with theodolites and trig tables to the tiny barometric altimeters made possible by modern electronics. Most of us - with the exception of a few contest rocketeers - would agree that the trend has been towards increased accuracy and greater precision. Altimeters are now legal in contest events and the qualification scores of TARC teams are dependent on the beeps or flashes emitted by the devices. However, I am often curious as to the accuracy of the altimeters we use; I would guess about a few feet if everything goes normally, but that is a just a "guesstimate". So, being a nerd, I decided to fly 2 altimeters (Perfectflite PNUT, Altus Metrum Micropeak) in the payload bay of Probe-18 at Saturday's HARA launch. Lofting small payloads was what the rocket was designed to do, and I figured the comparison would be useful, fully expecting the altimeters to agree within several feet.

I was wrong...

First, a bit about the flight. Probe-18 was launched in light winds on a B6-4 motor. The field was a fairly flat sod farm, with no significant rises or dips in the terrain. The rocket flew straight, with only a slight of amount of turning into the wind, and looked to have achieved an altitude of a few hundred feet. The parachute fully deployed and the rocket descended slowly; the ejection charge fired on time, though it sounded loud, indicating perhaps a bit too much powder (i.e., a "shotgun" ejection charge). After recovery, the body tube near one fin was observed to have failed, and the fin was chipped at the leading edge. Further inspection revealed a deep gash in the balsa coupler at the bottom of the payload section, suggesting that it had snapped back into the rocket at ejection, hitting the top of the fin; the force of the impact caused the body tube to fail near the fin root.

Probe-18 damage (Click to enlarge).

Probe-18 payload section layout (Click to enlarge).

When the rocket was recovered, the PNUT was beeping out a peak altitude of 257 feet, which seemed a bit low, but, hey, my eyes aren't calibrated. I did not bother to check the MicroPeak until I returned home, at which point I downloaded the altitude profiles from both altimeters. Boy, was I surprised! The MicroPeak recorded a peak altitude of 322 feet, a whopping 65 feet higher than the PNUT value - what was going on?

Probe-18 altitude profile recorded by the Perfectflite PNUT (Click to enlarge).

Probe-18 altitude profile recorded by the Altus Metrum MicroPeak (Click to enlarge).

A "normal" PNUT altitude profile - that of Nemesis at the same launch (Click to enlarge).

Both profiles were similar in that they showed a huge altitude dip at ejection, characteristic of a pressure increase associated with the Estes shotgun ejection charge. That much I could explain. Looking closely at the PNUT profile, I saw that it did not start out at 0 feet at 0 seconds, which was unusual. I then checked the profile from a second PNUT that had flown aboard Nemesis at the same launch; that profile looked normal, with the rocket lifting off and landing at 0 feet. So no insight there. I copied the Probe-18 PNUT numbers into Excel, where I discovered that landing was considered to have occurred at -65 feet altitude. Hmm... 257 + 65 = 322, the number reported by the MicroPeak. Since the MicroPeak produced an expected profile, I was inclined to take it as the "truth"; I found I had to add 65 feet and +0.3 seconds to the Perfectflite PNUT data to get a match to that of the MicroPeak.

Altimeter profile comparison (Click to enlarge).

So what's the deal here? Both altimeters were in the same payload section, with adequate ventilation to the outside air. Both showed the dip in altitude at ejection, though the PNUT had it occurring 0.3 seconds earlier and the dip more severe, going right though 0 to -112 feet. In the MicroPeak data, the dip reached 25 feet, not quite to 0. Could this somehow have screwed up the Perfectflite smarts, causing it to think the ground was at -65 feet? Or did the shock of the impact with the fin give the PNUT a "concussion"? Were my holes to the outside too big? Or was there a gremlin blowing on the PNUT, trying to mess with my mind?  I then remembered a similar occurrence involving the MicroPeak back in January; it too registered a ground level well below zero feet. Back then, I attributed the anomaly to not giving the altimeter enough time to sense ground level before launch, but this time the altimeter was beeping steadily (indicating ready to launch) and the Probe-18 sat on the pad for several minutes before lifting off - Couldn't be the case here.  In the January flight, it was obvious that the altimeter moved in the payload section during flight. Perhaps the PNUT was a bit loose and shifted around in the Probe-18's payload section?

Whether it was a mistake on my part or not, this experience gives me a bit less confidence in electronic altimeters. By looking at the PNUT data, I was able to see something was fishy, and even derive the "correct" peak altitude by adjusting the zero point. But in evaluating a TARC flight, we are allowed to only use the altitude beeped or flashed by the altimeter when it is returned to the observer; no analysis of downloaded data can be used to adjust the score. How often do the altimeters beep out a wrong value?

It makes one want to fly a second altimeter as a check. But that adds weight.