My ArduPilot Trials Day 2
I started off the day very excited to see how a new day of testing would go. I'd re-read portions of the manual, searched for ArduPilot tuning posts and contemplated the replies I got from my previous post.
The conditions were great and I was ready to toss that sucker up and get some satisfaction.
Instead, I learned a lesson. You see the picture above? That's what you get when you fly your plane using the wrong configuration on your radio. Notice how the tall grass beacons my plane to follow it's shadow... Well it did! Fortunately for me the grass was tall and I was quickly able to see the problem and cut power so no damage to speak of.
After I double-checked the electronics and hardware I reset the model in the radio to the appropriate configuration I got a GPS lock and tossed it up. I waited to engage the ArduPilot, which was in Stabilize mode until I trimmed out the model. Wow, what a great flight. The Stabilize mode checked out wonderfully. My fellow flyers were impressed with how well it did, especially when I used the stabilizer to bring it in for a landing. It was truly great. I lined it up on the approach I wanted, into a slight breeze and cut the gas and it glided in perfectly.
My ArduPilot Trials Day 3
Like yesterday I started of today very excited to try the next test. After flying yesterday, I took home the plane so I could change it's mode from Stabilize to Fly-By-Wire-A.
I was all set and ready to figure out what the heck "tell it where you want to go and it'll figure it out" was all about (that's a slight synopsis of the description of the FBW modes). I got my lock, put the wing on and gave it a toss.
Again, I learned a lesson. Not a new lesson, but one that should have been very fresh in my mind. You see the pictures below? That's what you get when you fly your plane using the wrong configuration on your radio. Notice how the tall grass beacons my plane to follow it's shadow... This time the grass was shorter and right next to a service road! Fortunately for me it was the grass, which was tall enough so again damage to speak of, just a bruised ego as I had watchers and needed a rubber band to hold on the GPS which had popped off.
So just what was that lesson? No it's not "dont fly your stryker" or even "double-check your model". It's Use Checklists to avoid these type of problems. I could have been sent back to ground zero with either of these crashes. I'm now building some easy grease pencil checklists that I'm going to attach to the plane and complete right on the model prior to flight. That should save me from myself and from having to remember to bring the list!
Todays second attempt went very well. Again once I trimmed out and got the model into the AP mode the FBW worked very well. It is different than the Co-Pilot for sure, but I use the rudder to produce a flat turn. The flatter the turn the better it's keeping itself stable. I did not get the chance to do the tests that are recommended for FBW, that will be next flight. For now I'm very happy with the ArduPilot. I again used the AP to land the plane, which was a little off from yesterdays great landing but I blame that more on me and how nervous I was after drilling the plane into the ground two days in a row. It did remain very stable during the approach and landing the line up was a little off so I had to walk a few yards to retrieve it instead of a few feet like the previous day. No big deal.
Next Test, FlyByWire-B, although I'm not sure what to expect in this mode with the motor? I guess it just keeps it constant? If anyone can explain that I'd appreciate it.
Wish me luck!Notice how it's not quite square to the frame of the first picture? That creates visual interest!
In the second shot the shadow almost forces your eye to whats outside of the frame.
The third shot conveys movement. Notice how the grass in the background is in focus yet the foreground grass is blurred as the plane smashes into it...
Last but not least a nice shot after the model was corrected.
Comments
Good info on dialing in APL
http://www.diydrones.com/forum/topics/getting-a-little-confused-on?...
The next day's testing turned out very interesting. First I totally forgot my cameras so I have no video of the flight testing. This could be a good thing as there was one scary moment... Ok, actually not that scary but it did produce a bit of a pucker!
My platform, the Sky Fly Max can get down to a stupid slow speed and just sit there as long as there is a little wind, but you have to watch out how much rudder you give it or it'll turn nose-down a bit, wing over and fall out of the sky...
I've had this happen using the FMA Co-Pilot as well as the APL. I have flown this plane a LOT using the FMA Co-Pilot so when this happened I just let off the controls to let it recover and then realized, oyi RTL mode does not do that!
Fortunately for me, I was quick on the dis-engage and easily pulled out of the dive.
I'm not sure what was going on with that when the stall occurred, it did not seem to be trying to get back to the RTL point at all. After the stall I let it try a few other times and none seemed like the AP was heading to my launch point. So I landed the plane.
I didn't unplug the battery, I decided that I'd try the reboot sequence (toggle off/on x5). At first it didn't work so I tried again, the second reboot actually rebooted the APL. I don't have a view of the APL module so I can't see when the GPS lock occurs. I time it during my initial setups when the wing is off and I have a view. I usually power up and get a lock two times before I put the wing on so I can get an idea of how long it's taking for the lock. I waited about a minute and tested the APL with the plane in my hand.
(actually the plane started up automatically which was something I didn't expect!)
It seem'd to respond so I gave it another Chuck and tried again. This time around the plane did fly right over the launch point on 4 different passes. It was a bit windy and without the 3 way switch to go into another mode on the APL I decided to land and call the testing a success.
Observed: I think the APL may be expecting some coordinated turns with the Aileron and Rudder channels. Right now, I've only got the ailerons in the APL so I need to check to ensure that the rudder (CH4) is not configured. I think it is, but my is not even soldered up so I can't go there at this point. I'm still hoping that others respond and tell me if they are using both the Aileron and Rudder via the APL.
Observed: Turn seemed a little slow, but it was pushing against the wind so it may be ok. I'll test again with less wind.
Next Test: Changed the Rudder channel in APL code to 0. Configured simple Tri-Angle way-point test with 3 passes (9 legs). To see how well it turns.
Wish me luck.
Today I'm satisfied that it's now testable again. I had to adjust the rudder and put a fair amount of "left" to get it to fly straight. It does now and I'm happy with the flight.
Because today did not have any wind to speak of I also got a good feel for the APL settings I've got configured. The two pictures show the full left and full right aileron turn bank angles.
The left and right bank turns are not quite the same, but better. I'm on the fence on adjusting the max_head as the turn is much tighter than before and I think I'll setup a test with the ArduPilot traveling between two points. I know the flying area very well and in pretty much any flyable conditions, so observing the travel between the two points will give me a good feel for trying to tighten the turns.
That leaves me to my current quandary... Do I set the APL to RTL or Set the Way Points and just go for it... I still don't have the rudder channel configured so I'll really need to do this all again, but I just want to give it a try.
Think I'll do the RTL and then if I have time at lunch I'll try to switch the modes. My Spektrum DX6 does not have a 3way toggle so I need to reprogram the AP mode. (I really need to buy a reciever or two for the RDS8000).
**QUESTION** - Anyone know if the mission planning stays with the APL after you've re-compiled and loaded updates to the APL program?
Anyway, wish me luck!
@ Brian [Observed: When testing full stick deflection, to the right I get a very nice tight turn but to the left it's much wider. I remember a discussion that spoke of a similar problem so I'm going to go hunt that down to determine where to fix the issue]
It could be torque induced if it's favoring turning opposite your propeller rotation. I've had that issue on over powered gas models in the past... You can verify it by testing while reducing your throttle.
Ed
Well, as I predicted I was able to test but not get back online to report. My flight to day went very well. Conditions were good with no wind, temp was good and plenty of difference for the thermopile sensors to read between ground and sky.
My first test was to confirm the proper operation of the counter roll/pitch movements as I had just previously changed the orientation of the XY sensor to meet what the ArduPilot configuration lists. So it's now in the default setting. I had remembered that the roll was reversed for sure so I only made adjustment to the REVERSE_ROLL setting. Ended up that the Pitch was reversed as well. Settings now show:
#define REVERSE_ROLL -1 // To reverse roll, PUT -1 to reverse it
#define REVERSE_PITCH -1 // To reverse pitch, PUT -1 to reverse it
#define REVERSE_RUDDER 1 // To reverse rudder for 4 channel control setups
*****I am curious, what does the rudder channel not say to put a -1 to reverse the channel?
Is this different? My guess is that it's not any different than the others, but if somone knows differently, pleas post a reply!
I had also adjusted the SERVO_ROLL_P setting, it now reads:
#define SERVO_ROLL_P .007
Which takes me to the test flight. Like I said the conditions were very good. I had no wind so it was easy to see what the plane was doing versus what the wind was doing to the plane.
Observed: The plane seems to climb once the AP is engaged. The plane files well in regular RC mode and it's level. I think I will need to adjust the "trim" for the sensor to get out of the climb.
Observed: When testing full stick deflection, to the right I get a very nice tight turn but to the left it's much wider. I remember a discussion that spoke of a similar problem so I'm going to go hunt that down to determine where to fix the issue. If anyone knows which blog that was please post a link. I also noticed that slight drop in elevation during the long slow left turn. The right turn was so quick that I did not notice the drop in elevation.
Overall I am very happy. I just need to get the left and right turns symmetric, adjust the pitch as needed and add the Rudder to the mix. Right now I am not using the Rudder channel on the ArduPilot. I look forward to having some nice controlled flat turns.
Hopefully I'll be headed out in the am to test again. So wish me luck!
http://www.diydrones.com/forum/topics/pid-nav-troubles?id=705844%3A...
Wish me luck!
Thomas,
That would be a feature on the newer DSM2 Spektrum transmitters... I have a lowly DX6 DSM transmitter. It does some things you should know about, like when two models have reversed channels from each other, it carries over the setting to the new model, you need to reverse the channel to match the new plane. I have an AirTronics RDS8000 and am waiting for a new rx. I'll use that when I get it because it has the 3 position toggle, mine only two position...