I fly redundant flight computers about 95% of the time. Today, I was reviewing the data from two EasyMini computers that took a ride in my DarkStar to somewhere around 10,500'. Both EasyMinis started logging data above 500'... thus skewing the final apogee report. I own 6 of these EasyMinis and I've noticed several of them, on occasion, having this same issue. These two that I use in my DarkStar are now consistently starting above of 500'. Anyone else having this issue with their EasyMini flight computers?
I think this can be explained by how the firmware works, and that the reported height values are AGL.
As you may know, EasyMini uses a Kalman filter to track the state of the flight. This takes the raw barometric data to update the "state" of the flight, which is contained in three values: height above ground, vertical speed and vertical acceleration. One piece of a Kalman filter is a predictive model of how the underlying system works. In our case, we've got Newtonian physics as a pretty darn good approximation of rocket flight; given the previous height/speed/accel, you can estimate the next state by simple arithmetic.
Because we don't know when the motor is going to be lit, or how it will burn, the EasyMini Kalman filter makes a rather rash assumption -- that acceleration is constant. This is a really good model when trying to estimate when apogee occurs, but not so hot when the motor starts burning. The barometric data is quite noisy, especially when under thrust or accelerating through mach, so the Kalman filter takes quite a bit of convincing whenever the acceleration changes.
EasyMini assumes that the rocket is flying once the Kalman filter value for height is greater than 20 meters (about 66 feet). As the Kalman filter takes a while to adjust to the fact that the motor is burning, it is only slowly adjusting the estimate of the height of the airframe, and the reported Kalman values will lag the actual airframe state by a long time.
To avoid filling up the eeprom with ground data, EasyMini doesn't start logging flight data until it figures out that the rocket has left the pad. It does save 160ms of data from before the launch detect, so you'll see that much data, then the 'boost' line, and then the rest of the flight.
EasyMini records the raw barometric pressure value during flight, not the cooked up Kalman values, so what you see in AltosUI is what the pressure sensor actually recorded. This lets you see the Mach transition effects, along with changes in internal pressure caused by ejection charges. EasyMini also records the barometric pressure at the launch pad, which allows the ground software to compute AGL values for all of the logged flight data. That's how you see the flight start 500' above the ground; it's not that your rocket started the flight from that height, it's just that the data recording didn't start until the rocket reached 500'. I've reviewed a couple dozen EasyMini flight logs and see boost detect at anywhere from 30' to 800', as you'd expect, it depends on how much acceleration the airframe is under, and could also depend on the size and shape of the static ports on the ebay.
I don't think there's anything "wrong" here, just that the boost detect is taking a while, which is normal. If you want more flight data recorded, adding an accelerometer to the system makes the Kalman filter much more responsive as it can quickly adapt to acceleration changes. I reviewed another several dozen TeleMetrum flights and all of them start at 0 AGL.
