Standard vtol lost control during alt_hold mode

Hi, I got a standard vtol plane, yesterday I almost suffered a crash. Itā€™s after the transition, the plane flies in manual mode in fixed wing state for a while, then I switch it to alt hold, and right after I switch it to alt_hold mode, it lost connection with QGC, at the same time, I found that I canā€™t control with my RC!
Then I just watched my plane flies with height decrease slowly as I canā€™t do anything, when it almost hit the ground, I found that I can control it with RCā€¦then I saved my plane and got landed.
The whole crash flow is like: mc(normal) -> transition to FW(normal) -> manual mode(normal) -> alt_hold(lost signal!!!) -> free fly for a while???(donā€™t know what is that state) -> RC control regain -> land
After landing, I tired to read the px4log to find out the reason, but, the log stopped at the moment the plane lost connection with QGC.
here is the log:
http://logs.uaventure.com/view/F8oujwUjtMzCXbbjVKkRoa

Could anyone explain this? I donā€™t want to suffer this again, itā€™s hopeless watching your plane slowing hitting the ground. :frowning:

p.s. I set RTL when both QGC and RC signal lost, but the plane didnā€™t act like RTL, so maybe the RC signal had not lost, but why I couldnā€™t control with RC?

the msg info:
2016_08_22_05_28_38: [lpe] baro init 63 m std 32 cm
2016_08_22_05_28_38: [lpe] z resume
2016_08_22_05_39_48: [lpe] xy resume
2016_08_22_05_45_07: data link #0 lost
2016_08_22_05_45_17: data link #0 regained
2016_08_22_05_46_41: Already landed, not executing RTL
2016_08_22_05_46_41: RTL: completed, landed
2016_08_22_05_48_04: ARMED by RC
2016_08_22_05_48_04: [blackbox] /fs/microsd/log/2016-08-22
2016_08_22_05_48_04: [blackbox] recording: 05_48_04.px4log
2016_08_22_05_48_13: Takeoff detected
2016_08_22_05_48_50: Transition to FW!
2016_08_22_05_49_09: FW mode!

after landing, I canā€™t disarm the plane with RC, so I had to hold it down by hand and cut off the battery.

Weā€™re looking into this. Certainly looks wrong but I donā€™t have any conclusion to offer yet.

Based on the log and additional information on gitter it is possible the board lost power.

Additional information provided by @Yuontheway indicates that the RX was powered by the pixhawk, the RX in turn powered camera gimbal servoā€™s. This is not supported as the current draw is much too high.