Gluonpilot

I've been testing the setup on the bench for my new airframe and have noticed the odd glitch in the barometric altitude. Everything is running well but every once in a while (maybe 3 times in a 15 minute test) the altitude reading is very high (in the order of 6000m). This is a little annoying in the GCS as the graph scales to suit. It does however result in a brief full down deflection on the elevator.

I have a v1o board with the BMP085 sensor. I'm running a seperate power system for the autopilot (Castle Creations BEC). I have a BMP085 that I use with an Arduino and don't see the same glitches.

Dave

Hi Dave,

This was indeed a problem with the old firmware. It should be fixed now in the 0.8 (beta) firmware. Could you check?

I'll download it this evening, I'm currently running the last 0.7 release. I normally avoid Beta versions of software when I'm trying to sort out other stuff.

Dave

I downloaded the latest versions yesterday and had a couple of 15 minute tests. Didn't see any glitches in the elevator servo. Problem has gone away

Dave

Great!

Hi,

I confirm the same problem with the v10 board, even with the latest firmware (0.8BETA3, but was the same with 0.7.2). Never have this problem with the old version of the board , but there is another barometric sensor.

In garage I can click on the graph and use the mouse scroll to modify the scale of the graph but when flying this is difficult.

Are you using the battery voltage input? How many volts?
I never had this issue with v0.8, but another customer reported a problem which only occured when using a 4s lipo... Very strange

Yes, I am using the voltage input, and on that board I use a 4s lipo but also tried to use the voltage from the balancing connector (and monitor only for 3s for example) but the problem appeared again. Worth mentioning it was a very hot Sunday, 30 Celsius.

I looked at the logs and appeared to be connected with the temperature, when measured temperature passes over 39 C, it logs -2 C and the height raises with approx 6500m.

So, I tried today to simulate the problem on the bench, heated slowly the board with an hot air gun from approximately half meter and waited, 37C, 38C, 39C ...

I've put an attachment. The log from the control station is interesting only after 9:30.

Now that is an interesting find! Thanks! The other guy having the same problem was located in India, where it is 39°C even without a hot air gun

I'm going on vacation but I'll definately look into this problem when I get back.

The only strange thing that puzzles my is that the guy from India only had very short spikes of high altitude... in your sitaution it seems to last.

Forgive me for uttering another solution but Ardupilot just had a temperature related barometer bug, could they be conneted in some way? It also jumped to 6000m

http://diydrones.com/forum/topics/unrel ... c-altitude