[ppnvedado]

starting new post with conversations from another:

me:
too much in programing, so back to 0.
Reprogram board and do all. All seems to work fine, altitude vario and hold position with some variations in altitude. Board covered and foam to reduce light effect.
how important is thermal calibration? never done it.

TheOtherCliff: Thermal calibration mainly helps vertical drift.  In my experience, without it, AV, AH, and GPS modes can drift vertically about a meter per minute.

me: then, it will be done!!!!!

me: Thermal calibration done!!!
I have oscillations in the first flight minutes, not only up and down, horizontal too. Then steady, almost no movement at all, impressive.
I'll do some testing covering the FC and post result. Any other suggestion?

TheOtherCliff: Start with the flight mode you are using and whether the oscillation is faster than 1 per second or slower.  A video would help a lot.

[ppnvedado]

The variations also happens in altitude vario, maybe the barometer, it has foam. I will cover all the FC today and see if there are changes.

[TheOtherCliff]

Baro issues usually cause a slow vertical drift, not an oscillation.

I recall that this is about GPS flight modes.  Some DJI/Naza setups oscillate and there is a firmware that helps if that is the case.  What brand of GPS is this?

Does the oscillation happen in a non-GPS non-baro flight mode?  You mentioned Altitude Vario.  Try the same flight mode without AV.

And a video would help too. 

[ppnvedado]

I´m using a DJI Naza clone (very fast and steady in taking satellites, 15 or 16 all the time). I will try to take a video but I have no action camera yet, and attach my phone to my forehead with tape is a bit complicated  I never let the control away if I see anything wrong (to many crashes)
what is really strange is that only happens when I start flying just a few minutes, less than 5, then everything turns very precise, no movement at all in position hold and regular descend in auto landing after coming home. those were the problems I saw, movements every axis in hold and irregular descending speed in autolanding.
I read about making a home gps antenna to increase accuracy due to serious drift but when becomes steady you can stand by the drone and even pull it (don´t do that at home) (I used a piece of wood I found) to take it away from position and see how returns.
I will do as you recommend and try

[TheOtherCliff]

If there is a horizontal oscillation with a DJI Naza GPS you may find that it only oscillates when pointed in a particular direction.

If you are running 16.09 try this firmware.
https://forum.librepilot.org/index.php?topic=3012.msg21154#msg21154

Read the rest of the thread.  All you have to do is flash the firmware.  No setting changes.  No setup wizard needed.

Please report if that firmware fixes your oscillation.

[ppnvedado]

Hi there:
I haven´t do any changes to post vídeo for you guys to see. The issue I think it´s orientation relates as you said. starts in an specific orientation and if I change it, disappears.
there is a limit in size of video, I can't post cause is too large

My quad it's a little large and heavy. 1655 props and almost 2kg due to an excess of material to assimilate crashes in testing.

[TheOtherCliff]

Standard procedure for videos is to post them to youtube.com or vimeo.com and post a link to that here at LibrePilot.

You should try that firmware I linked to and let us know if it helped.

[ppnvedado]

sorry for the delay, firmware tested last friday and problem solved.
I don't know if all I'm about to said is related but I saw this improvements.

No more oscillations in any direction. 360 degree test and obtain a very steady position hold.
I used to have problems calibrating magnetometers internal and external, no matter what always values over 5%, lots of yellow warnings. After fix, both mags under 1% on the first calibration, same method.

I have the videos but need time to work on them and post in youtube. I´ll post the link.

I'm back to the other problem in the other post. crashing due to control resonance

[TheOtherCliff]

Glad to hear oscillation is fixed.  Video no longer needed. 

Ease in calibrating magnetometer is not related to using new firmware.  I guess you found a good place and a good procedure for calibration.  One thing you must keep in mind is that once started, the calibration is always gathering data, not just when you press the button 6 times.  This means that you can't "set it on you car", etc. in between button pressings.  Also don't set it on ground.  Ideally it should be done in an open field, away from cars and other metal things.  It should be done via telemetry.

Generally a novice should not trust the "onboard" mag and should use the mag on a GPS/mag unit that is mounted up on a pole away (say 150mm min) from mag sources including thick wires and carefully aligned to be aligned with the FC.  Twist all high current wire sets from battery to motor.  Of these, "aux GPS/mag alignment is the only one that should affect the alarms when motors are stopped.

[ppnvedado]

Ok, nothing to do with new firmware. Although I post this info for clarifying.
I have aux only and power zone (PDB and ESC) shielded under metal and on top of it are the FC. My frame is aluminium and motors mounted on aluminium tubes (motor cables inside) The only source of magnetic interference can be the motors 300mm away from the body. I have the aux on a mast that has itself 250mm.
The calibration was done on the same spot same method; open field, more than 6m from any metal except my laptop, and via oplink, may be more or less turns to achieve positions and save, who knows.
Thanks for the info, already post where I buy the GPS to look for the new firmware if they use a revo.

[mr_w]

Metal such as aluminium cannot shield from magnetic interference. Even special materials like Permalloy and Mu-Metal are not so effective in shielding from those types of interference (low frequency em). Only active shielding will do, and that is in fact what twisting the wires will do - by creating counter field that hopefully cancel each itself.

https://en.wikipedia.org/wiki/Electromagnetic_shielding#Magnetic_shielding

[ppnvedado]

Thanks for the info. never is to late to learn something, I also twist and keep mag as far as I can.
I really thank you for the info.