Too much power and pitch compensation is bad thing... not enough is bad too.



No ground crew today. Down view camera was installed.
Front camera view of todays flight testing of PID settings for throttle cruise 80 & throttle max 120, pitch compensation .30 Test battery a 2250mAh 3s lipo had 1 cell going bad,

I replaced battery and retested, but crashed flat and hard when I panicked and set Mode to FBW instead of Manual..
Systems damage report:
Power train,
APC 7x4 EP prop broken off both sides, replace with APC Gas 7x3 $1.95
Motor mount ripped off, $2 flexable superglue and carbon fiber, New mount for contest $5
2250 mAh test battery has one bad cell now, $60 new 4s 2650mAh 40c for contest

Structural,
starboard side,
Tears in foam on both sides of vertical fin up to spars, $1 Liquid Nails or lo temp hot glue.
New foam body $21 for contest.
Control horn ripped off, but still on wire z bend, $0.50 superglue and add some carbon fiber.
Stripped servo plastic gears, replace set $2.49
All other electronics are fine.

Made several test flights, my observations:
1. Don't flight test with a weak or old battery
2. make smaller PID p changes, over correcting in pitch and over powered throttle in FBW mode
3. Start FWB mode at about 150' when testing throttle or pitch PID changes, 50' higher than normal
4. Train left hand Learn to use Mode switch when panic ensues, place in Manual vs FBW and relax
5. Don't ground test with a weak or old battery
6. Flying in snow is fun!
7. Don't panic!

Views: 132


Developer
Comment by Doug Weibel on March 11, 2010 at 7:45am
Mark,

I looked at the config values you supplied in the other thread and I would suggest a different systematic approach based on first finding the appropriate pitch and roll gains, then nav gains then throttle.

- Route the throttle directly to the rx for manual control. Set the throttle value at what you would eventually like your cruise throttle to be.

- Fly FBW at 200'. Use a flight profile that makes it easy to see what is going on. I will usually switch to FBW, allow it to fly straight and level a bit, give full left stick and hold for a full circle, straight and level for a bit, then a full circle the opposite direction.

- I collect the telemetry data and tune the PID's for a reasonably clean step response when I give full roll stick and a reasonable hold of the commanded value. Tune until your data looks like a classic PID response with noise. I had a nice graph of data I took the other day I'd pass on but I think I deleted it ;(

- While you are tuning the roll response you need to also work with PITCH_COMP. Find a value which pulls enough during turns but does not cause significant climb.

- Repeat this for pitch.

- I would suggest for a Stryker that you will need HEAD_MAX set to at least 4500. Otherwise your high speed turn radius will be very large.

- After you get your pitch and roll response good try AUTO and work on the nav gains.

- Finally after gettting the above sorted out route the throttle through the AP and work on the throttle related values.

Just my 2 cents.

Developer
Comment by Mark Colwell on March 11, 2010 at 11:13am
Thanks Doug,
For guide to follow, especially pitch comp & nav and in the proper order, My testing has been for just a few seconds in FBW where I was getting results that were confusing, due to low battery, aft CG, and operator errors.
Rebuilding Stryker II now, found spruce spine was broken in 3 places, replaced with FG tube, removed about 30g of duct tape. Glue is drying now, replaced servo gears, re-hinged elevon, next I will troubleshoot TM stream, need to get the TM working well as video doesn't give enough info. That may take all day...
Comment by Jack Crossfire on March 11, 2010 at 12:18pm
The plane crashes but the goattee is invincible.

Moderator
Comment by Robert Krogh (hooks) on March 11, 2010 at 1:34pm
Stubborn as a donkey!
You will win.
:-)

Moderator
Comment by Brian on March 11, 2010 at 8:55pm
It's no wonder you were having troubles. Even the mighty Strykers can not endure the power of a hair Twister!


Developer
Comment by Mark Colwell on March 11, 2010 at 11:47pm
It got caught in tail rotor when testing a T-rex 450 once, Ouch...

Moderator
Comment by Brian on March 12, 2010 at 7:20am
Not sure what would hurt more, that or the croud of folks watching when that happened!

Developer
Comment by Mark Colwell on March 12, 2010 at 7:40am
I was in my kitchen, holding heli and testing new ESC when wind got me. Yes scissors were needed. Lucky the mains blades didn"t get me too.
Comment by Pbreed on March 12, 2010 at 9:51am
When tuning things I set up to put the gains on one channel of the RC transmitter so I could adjust constants in flight. I had a simple code utility that let me pick what variables to assign to the gain knob.

Paul

Developer
Comment by Mark Colwell on March 12, 2010 at 10:08am
Paul,
In flight PID tuning would be great, but I don"t know how to set it up, can you post code or algorithm? This would be very useful! Also a self learning mode that determines PIDs based on RC pilot flying a standard pattern. But the AT328 may not be able to handle it.

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

© 2014   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service