Thanks for your help, I take it that it is down to my pid settings that my quad won't hold altitude.
I have tried a number of settings some seem to improve the flight but not correct it fully.
The problem is as follows, the quad flies very well in normal attitude mode with manual thrust (mode 1 on my controller) then I flick to mode two (identical to mode 1 except thrust is altitude vario) when stable, centre the throttle and the quad will sometimes slowly loose altitude, sometimes slowly gain altitude but once or twice it has gained altitude fast (3 or 4m over a second), the quad does react to increasing or decreasing throttle when gaining or loosing altitude slowly but hardy reacts when gaining altitude fast (causing me to flip back to mode 1 to bring the quad back down to a more suitable height).
I'm at a loss what to try next, my quad is a 250 with 2204 motors and 5x3 props running 3s battery with revo controller (I have GPS but have disabled until I get the altitude sorted, then I'll sort out mag and only when they are both solid will I attempt gps). I have foam over the baro and have looked under a scope and it's the correct baro, please can anyone in the community suggest anything or share there setup if it's a similar quad?
Thanks for any help!
Mark
If you are using the Librepilot GCS, note, while the quad is in mode two, how the altimeter's values change. If, for example, you set it to hold altitude but due to perhaps a barometer problem the barometer readings are changing to reflect increased a climb, then the quad will react to the non existent change by reacting with a descent in an attempt to get back to the altitude it should be holding.
One thing I might try would be to remove the propellers (so it won't change altitude), setting the quad down somewhere with the Openpilot board still wired in but also hooked up to your comp so you can see the GCS readings. Set the quad to any arbitrary RPM and set it to hold altitude. In the scopes, you should theoretically see a constant RPM because the quad should think its maintaining its current altitude with whatever RPM the motors are spinning with. If, however, the scope shows the altitude changing (though its not actually since the props are off and its on the ground), you should expect to see a RPM change in an attempt to go back to the altitude it thinks it should be at but already is at in actuality.
In reality, these barometers are not flawless and perhaps may not have the resolution to stay steady within a meter for example. So I'd say the main problem is how big are the fluctuations are. Secondly, you report that sometimes the quad's altitude suddenly changes while in flight. Well, when just the board is plugged into the computer so you can see the scope for the barometer, are there ever sudden fluctuations in the barometric readings?
I'm not familiar with Librepilot or quadcopters but these are really just my thoughts on the matter.