TBS Discovery frame not stable in Loiter and Position Hold

Log Analysis -- deprecated Copter Logs -- deprecated
1

Hello Guys,

I have build a quad based on the TBS discovery frame.

APM 2.6 + external mag+gps (ublox m8) I get an Hdop of 1.4

After solving several issues on vibration and interference I can now fly quite stable in Stablized and Alt-hold modes. Position hold and loiter are however very unstable, they do accept my inputs but no where close to keeping the quad in place, the quad tends to move all around.

I have been through the log review wiki but I still do not know what to look for, specially in Position Hold mode. All things are within limits (vibration and compass interference).

Would someone please review the log files and guide me through this, I am really out of clues. The only Thing that I find a bit abnormal is the slight offset between AccX and AccY. Could this cause the problems?

I am posting 2 log files the last one also has a part of Position hold in which everything went wrong and the quad ended in the trees! I am also posting a tlog file. The tlog covers the duration of both flights performed on 2015-03-11

(during the first flight I do a save trim, land, disarmed and started over again )

2

I am not an expert but i think you need 4s battery for TBS.

I used to have the same problems using the DJI Naza CTR and after change to 4s from 3s battery really made the deference in stability for me.

I notice your voltage is very low at your logs.

3

I used to run my Discovery weighing 2kg on 3s and she flew fine just slowly.
4s just made her a lot punchier.

Have you run an autotune on the quad?