Servers by jDrones

Hexcopter Flips on First Flight, and second

I agree, this weekend, or earlier if possible. I have my Octo almost ready to fly. I will reload and re-calibrate everything from scratch to see what happens.

So as a postmortem here with trim values for roll, pitch and yaw at 1000 with centered sticks and no trim on the radio you are asking for full authority one way on those axis. It just happened to roll crash because it’s less stable on the ground in that axis with your landing gear. After you do a radio calibration check the trim levels for these channels to confirm they are somewhere around mid range (~1500).

10-4 mid-range, not zero. I will learn more about what those values mean.

I just started over and reloaded the firmware, did accel, level, compass, ESC, and radio calibration. All of the RC trim values are in the 800 to 900 range and are blow the RC min, no idea why but I will gain have to manually set them all at 1500

I did a dry run, i.e. a few minutes with motors running and no props on, the log is dated 1-12 and is posted to the link above.

There is something wrong. No way should the trim values be set at those values with a proper radio calibration. Think about it, the yaw, pitch and roll sticks are centered at the end of the radio calibration (as directed) so they should be ~1500. Not sure what you have going on there.

Apparently the problem is with Mission Planner, and it needs to be updated.I have been advised to download a Beta version.

@spartan27244 have you double checked the motor numbering diagrams? Because even though you select the correct frame, if your motor numbering is different to what the Pixhawk expects, the hexacopter will flip. That happened to me the first time i tested my hexacopter.

Yes sir, it is all correct. Dave and Mike had it correct with the RC_TRIM values being set incorrectly. If you will read here Copter-3.6.4 released! you will see there is a bug in mission planner that did not set the TRIM values properly during radio calibration. I used QGround Control instead which set them all at 1500.

For a radio in MODE 2 AETR, does parameter …
RC1 = channel 1 i.e. roll
RC2 = channel 2 i.e. Pitch
RC3 = channel 2 i.e. Throttle
RC4 = channel 4 i.e. Yaw

I see in the documentation that Pitch is reversed in that low values are forward, high values are backward.

With that in mind, if my copter drifts forward when stick is centered I would adjust the RC2 channel trim to a higher value, correct?

Moving slightly to the left adjust rc1 trim value higher?

etc,

Thanks.

No. You do not manually adjust trim for any reason. Not sure how else to emphasize this. When a proper accelerometer calibration is done the craft will stabilize in assisted modes.What flight mode are you referring to when you say “drifts forward”?

Stabilize, Alt Hold, and Auto Tune. As far as trims I see this in the Wiki which refers to 2 methods for setting trim values, one using the radio trim’s.
http://ardupilot.org/copter/docs/autotrim.html

I also see where I could cut the vibration in the Z axis a bit.

You can expect drift in Stabilize and Alt Hold this is normal. No reason to trim for that as local conditions are variable. Personally I think trimming is a wast of time. Simply use the sticks for this. If you want it to hold position use a GPS stabilized mode.

1 Like

Thanks for the input.

Well, it now works like a charm. I used the Beta of Mission Planner to do the Radio Calibration so the trims were set properly. This time there was very little drift, very manageable with sticks. Graduated to Loiter and it held position and performed beautifully. I then went on to auto tune, but I do not believe it completed because my battery was low and I had to land. Thanks for all of the help.

You can select an individual axis for Auto Tune and do them sequentially if you can’t get thru all of them on one battery charge.

After you changed the frame type to Hex, how did your copter fly? Did you start off with the default PID values? I just finished one that looks VERY similar to yours and after watching the video I’m a bit leery about pushing the left stick. I’m getting a little to old to do the build-fly-crash-repeat cycle anymore. Thanks!!

So after the frame change it took off just fine.
I was using the default pids and did need to be tuned. I had no issues other then the normal testing and adjusting after that.

Thanks for the info! If the weather clears today I’ll be able to give it a try. Did you use the Auto Tune for the fine tuning?
Thanks again!

Morning Tom.
Yes I autotuned it,
But I started with the stock pids and ran the tune with an aggression level of .05. Doing all axis at that aggression level. I then did it again at .07 and then .1.
In each subsequent tune it finished faster then the previous. I assume that’s because its close, then closer then spot on.
I also reviewed each log and looked for errors. Nailed down any error that appeared and had the logs peer reviewed. A second set of eyes was very helpful just to make sure i didn’t miss anything.

Watch the vibrations because there is a new feature…well newer anyay called vibration compensation and I had to major crashes due to that.

Apart from that I don’t have any real trouble. I pay real close attention to current and voltage monitoring because of premature RTL’s that I have had. But once thats dialed in it’s not a problem anymore.

Do your best to balance the machine…Out of balance can drive the motors harder on so ends and less on the other. This just adds to a poor tune.

Not sure what else I can think of.

Thanks!!! I really appreciate the detailed info. I’m glad you mentioned that vibration compensation thing. I’ll need to take a closer look at that. I also need to start paying more attention to logs. I’m real lazy with that. This hexcopter is the only drone I have with round tube arms held in place with clamps. Everything else I have, (all quadcopters), has either square tube or flat arms that can’t rotate. I’m a little concerned that the round tubes might rotate in the clamps, (especially at the motor end), and I’m wondering if it would be worthwhile to glue them in place. Is this something that is commonly done or would this just be overkill? Thanks again!

Servers by jDrones