Replies to This Discussion

DeveloperPermalink Reply by jasonshort on September 27, 2011 at 1:33pm

Can you post your PID values? That could be Attitude_I ramping up but I'm not sure. Will have to look.

And your props are balanced and low pass filter pads soldered and enabled?

Thanks!
Jason 

• ▶ Reply

Permalink Reply by John C. on September 27, 2011 at 1:43pm

Thanks for the reply Jason.  Props were decently balanced a few crashes ago, I will pull and re-check them.

 

IMU is one of the EARLY ones, Foxtrap I believe, that had the pads already connected.

 

Stock load of 2.0.45 with eeprom reset.

 

 

 

• ▶ Reply

ModeratorPermalink Reply by Mike Pursifull on September 27, 2011 at 5:00pm

Vibration?

• ▶ Reply

Permalink Reply by John C. on September 27, 2011 at 7:54pm

I've got all three screw holes held down fairly snug against a piece of medium density foam about 3mm thick.

 

Is there another way to measure or see if vibration is causing this?

 

Thanks for the input!

 

JC

• ▶ Reply

ModeratorPermalink Reply by Mike Pursifull on September 27, 2011 at 7:58pm

I think a bench test would show it. See http://diydrones.com/forum/topics/apm-planner-gyro-s-read-updside-d... for a video of a quad demonstrating a fairly extreme case of vibration sickness. This was solved with better isolation.

 

I don't know that you have this problem, but it is a nice demonstration of the aliasing effect, and worth watching. 

• ▶ Reply

Permalink Reply by John C. on September 27, 2011 at 8:11pm

yes, definitely worth watching, and we can see that my results are not as extreme.

 

I'm going to re-balance the props, and see if securing it with zip-ties helps.

 

Thanks for the feedback Mike.

 

JC

• ▶ Reply

ModeratorPermalink Reply by Mike Pursifull on September 27, 2011 at 8:22pm

That you only have an issue along one axis suggests to me, if it is, in fact, vibration, that it might be from the side of a screw and not from contact with the frame. Imagine a single contact along the side of a screw introducing vibration for a single axis, back and forth. The effect would be less extreme that what is shown in the video. Well, it is a theory. Many of my theories are wrong. That's ok, I just make more of them, and throw out the bad ones. So I'm looking forward to whatever you learn.

• ▶ Reply

Permalink Reply by John C. on September 27, 2011 at 9:24pm

So, you got me thinking.

 

I went and re-balanced the props.  They were fine.

 

I re-checked the mounting screws, and noticed that the bottom of the front right screw holding the board down was touching the top of the connector that feeds an ESC.  May or may not be something.  Again, screws are not cranked down, and it's on 3mm of decent foam.

 

Then I went ahead an plotted the pitch as well.  I have a spot level on the top platform, and it is extremely close to level with the IMU.

 

Red is roll, green is pitch.  Same red roll error that we've been talking about.

 

My question is, why does the pitch show a constant deviation from 0?  

 

When I arm it, that should set it to 0 again, regardless of the actual plumbness of the IMU.

 

• ▶ Reply

ModeratorPermalink Reply by Mike Pursifull on September 27, 2011 at 9:28pm

I assume that you have this on the floor or a bench. If you rotate it 90 degrees, do you still see roll?

• ▶ Reply

Permalink Reply by John C. on September 27, 2011 at 9:47pm

The data from above is a live flight this afternoon.

 

I just plugged it in on the floor, and it definitely thinks it's tilted back and to the left a bit.  That is regardless of physical rotation.

 

A "level" command in the APM planner squared it all back to zeros now.

 

Looking back through the manual, I thought that the level command was performed every time the quad was armed.  Looks like the 10 sec. arm and 20 sec. disarm is the way to set the accels in the field now.  That explains the pitch.

 

The roll is still concerning though.  We'll get her in the air again tomorrow AM, and get another take on it.

 

• ▶ Reply

ModeratorPermalink Reply by Mike Pursifull on September 27, 2011 at 10:06pm

You know, I'm wondering if there is a balance issue along the roll axis that is causing a roll oscillation? I cannot tell the scale of the data, is there any visible rocking? How well balanced is the frame? If you are using a + frame configuration, just stuffing your fingers in #3 and #4 and seeing if it wants to flip should rule out that theory. 

 

I don't know that I help much, but I am enriched from these dialogs ;)

 

 

ModeratorPermalink Reply by Mike Pursifull on September 27, 2011 at 10:09pm

but i think that is maybe the wrong direction.... if it was heavier on one side, I am not sure it would present in the way you described, but not knowing how the 4.2 P affects your frame/motors, I kinda created a base-less theory that an imbalance along the roll axis might result in a tiny vibration/motion resulting from rapid corrections.... well, it is completely made up...

• ‹ Previous
• 1
• 2
• Next ›
• Page