Hello,
I'm not sure if this is the right category to put this under, but I was on RCGroups and someone suggested I repost my questions here. Here is a link to the original post:
http://www.rcgroups.com/forums/showpost.php?p=32171268&postcount=21785
I am preparing for an event in the model rocketry nationals, an R&D project involving a rocket-boosted quadcopter.
I need some help with a major problem and crash today, involving a unique rocket boosted payload section that uses quadcopter recovery.
The controller is a Mini-APM 3.1 using ArduCopter and Mission Planner.
After building a 260 Racer Quad and getting that to work with a Mini-APM and GPS, I made a custom quadcopter to act like a rocket payload section using quadcopter RTL capability. For a future project, it could be launched thousands of feet into the air on a very big powerful rocket.
I worked out the bugs of getting it to fly, and have folding landing outriggers to keep it from falling over when it lands on its base.
Wednesday I finally flew it on top of a rocket. I did not want the props to spin during rocket boost, so the rocket has wire rods that prevent the propellers from spinning until after it separates from the rocket (it is set not to spin the motors when armed). I know that once armed, it has to throttle up in 15 seconds or will automatically disarm. It was launched in Loiter mode, later would be put into RTL mode after ejecting from the rocket and stabilizing itself.
I armed it, heard the beep and saw a solid red LED and it was launched 2 seconds after arming. After the rocket ejected it, I throttled up, but nothing happened. It fell to the ground, with no response. It happened well within the 15 second arming period, the crash to the ground was about 14 seconds after arming.
Someone suggested that possibly the programming might not like a sudden acceleration upwards when the throttle is down, and disarm the motors automatically, Could that be it, or something else?
What can I do to solve this? If it is the above theory, then if I had the motors on low throttle when the rocket took off, would that likely work as far as the Mini-APM is concerned?
Based on replies that I have received, I will look at the logs. In the meantime, if it went into crash mode or otherwise aborted for some reason, what else could I try to solve this? For example, if I throttle up at 10%, then launch the rocket, is that likely to solve the issue?
Images:
YouTube:
PayloadCopter test , RTL landing Side view
https://www.youtube.com/watch?v=7dQG7ZXbrdE
Payload copter takeoff &RTL landing , camera looking skyward
Rocket launch and crash, camera looking skyward
Replies
Update:
Thanks for your reply to my message about the rocket-boosted quadcopter in this link;
http://www.rcgroups.com/forums/showp...ostcount=21785
I have checked the log file. The log file can be downloaded here:
http://www.filedropper.com/rocketcop...-07-1519-44-59
Here is a graph showing errors.
It was in loiter mode for rocket boost. I switched to Stabilize mode after ejection and throttled up, so you can see when I did that. It shows errors GPS-2 and BARO-2. The later error for EKF Check 2, I do not know what that is and it was long after trying to throttle up after ejection from the rocket. The accelerometer sensor info does not show the launch acceleration, I do not think the APM was set to log IMU data so that might be it.
Here is what the Auto Analysis says:
Update: GPS glitch was enabled and baro glitch was enabled. So if those were the cause, would disabling them solve the issue?
I am running short on time, and weather, I may not be able to do a series of tests such as to rocket launch it without props, put a parachute on it, and see if it will throttle up.
I am considering repeating the flight by throttling up to 10% then launching the rocket. Would that seem likely to solve the problem?
After looking at the logs, it looks like the GPS isn't the only sensor that gets confused at launch. The barometer shows a pressure increase, meaning it thinks it is going downward. CTUN_ALT (internal navigation's estimate of altitude) shows a drop to as low as -40 meters halfway through the flight. CTUN_CRt (estimate of climb rate) goes to -2300. I'm pretty sure that's in meters per second... downward. There is just no way the APM can figure out what's going on.
I think the only chance for success is to launch in stabilize mode with throttle at mid level. After ejection, the quad should stabilize and be pretty close to hover. You should then be able to change to loiter or RTH.
Keep in mind that I'm no expert, and I'm just offering this as opinion and theory. :)
I would hate to give advice that might crash your quad, but in the interest of your short schedule I would say yes. Get the APM to ignore the fact that it's going faster than any quad ever has and it should be quite happy.
Great work so far! This is really a cool project. Please report back with more vids when done!
Oh, and it would be interesting to see the acceleration data of the launch.
One possibility is GPS failsafe. If the APM sees that it just moved 800 feet in two seconds, it may get angry.