Replies to This Discussion

Permalink Reply by Curt Olson on July 21, 2009 at 1:07pm

What kind of radio system were you flying with? 2.4Ghz or 72Mhz? It's been my experience that 2.4Ghz is more solid with respect to potential interference and noise introduced by all the added electronics onboard.

I'm trying to imagine why the ardupilot might lock you out even though you had the auto/manual switch set to manual. Two possible scenarios come to mind.

1. You are flying with a 72mhz system and radio interference caused the ardupilot to think that the manual/auto switch got set to auto.

2. With the "remove to fly" system, the code can override the mux signal from the receiver and forcibly drive the servos. That's a little scary, but properly coded, it should never happen in flight ... unless the remove to fly shunt was installed from the start?

3. Some sort of short or loose connection that manifested itself in the middle of the flight.

It would be interesting to take all the parts back home and do some post crash analysis. Did all your batteries have good voltage? Is the ardupilot unit functioning ok now? Does your receiver still work and how does it range check?

You can never say for sure if a problem you discover caused the crash, or was caused by the crash, but it might be worth going through all the systems to see what works or what doesn't. Even double check physical things like control surface hinges, linkages, etc. These things have been known to fail in flight too. If the autopilot does something oddball, it could put forces on the airframe that are way outside of normal and could cause a physical failure in flight.

With one of our early wings, the autopilot put the wing in maximum (allowed) down elevator configuration. I let it go because I had a lot of altitude and was hoping it would catch and fly itself out of the dive, but it kept forcing full down elevator and ended up doing a forward somersault at about 100 mph before I returned to manual. That was really weird and we got really lucky that we didn't shed any parts ... but our wing is designed to be bullet proof in a marine environment so it's a pretty tough bird.

At any rate, my advice would be to do as much post crash checking and analysis and debugging to try to figure out what happened and why ... otherwise it's likely to happen again on your next airframe. And if there is something going on with the basic ardupilot design (which is probably unlikely, but all things are possible) then lots of other people would sure like to know about it.

Regards,

Curt.

• ▶ Reply

ModeratorPermalink Reply by Gary Mortimer on July 21, 2009 at 1:41pm

Bad luck that only looks like a few minutes work with some glue to fix though.

• ▶ Reply

Permalink Reply by Michal on July 21, 2009 at 2:04pm

Daniel - "FMA Without sensors" - is that mean that your EG didnt have FMA sensors during flight?

• ▶ Reply

ModeratorPermalink Reply by Chris Anderson on July 21, 2009 at 2:30pm

You flew ArduPilot 1.0 without a FMA CoPilot? That's a guaranteed crash--you had no stabilization system at all. Sorry to hear it, but I'm confused by why you would attempt that.

• ▶ Reply

Permalink Reply by Daniel Fidalgo on July 21, 2009 at 3:02pm

ok,
I really do not have a logical answer about flying without the co-pilot and sensors ... But I am waiting for the co-pilot and while I do not have it, I have done some tests just to see how it runs the model, always very high and I stabilized the aircraft through the Ailerons (as if doing the work of co-pilot...). When detected an attitude less "favorable", I just turn off the ardupilot.
Okay, I know, there seems to be all stupid but does not seem to have been the cause of the crash.

The only command "delivered" to Ardupilot was the rudder, the throttle was not linked to ardupilot.


Curt Olson:

I fly on 72mhz, and your scenario number one seems very plausible. Once I have some time available I will check all the components .
Either way, thanks you guys for all the quick answers

It seems that now would I even have to wait for FMA sensors not be?

• ▶ Reply

Permalink Reply by Reto on July 21, 2009 at 3:26pm

What you need now for sure is this:
Sorry for your EG, but Gary is right: your fixed EG will be a much better suited testbed for further experiments than a brand new glider! Good luck and keep on trying.

• ▶ Reply

Permalink Reply by Daniel Fidalgo on July 21, 2009 at 4:07pm

Yes I know, with a bit of glue is like new! Actually I should take up now that is broken to go with a mod similar to the Sr. Bryan Cuervo which is excellent. But for now I will continue with the easy patched, and channeling the budget for electronics ;)

• ▶ Reply