[HighAltitude]

With a new year, I'd like to thank the Librepilot development and support crew again for all their help. 
As I accumulate more experience with using the GCS, I'm increasingly impressed with its capabilities.  Like any complex and comprehensive software, the learning curve to using its many features is long and steep.  The articles and instructions in the online manual are a big help, but the support of the Forum is indispensable, especially for an old geezer who started aero-modelling when U-Control handles were the prevailing technology for controlling one's aircraft.
With the info and advice from the Forum, one of my quads are now flying semi-autonomously in Velocity Roam and Position Hold and the other in Position Hold and Return To Base flight modes, albeit with some oscillation after returning to hover over the Arming location. 
I look forward to exploring other capabilities, such as getting my Micro minimOSD to display flight data on my ground station screen and goggles via my modified Mobius/docking station/transmitter, so I'm sure I'll have more questions in 2017

[TheOtherCliff]

The quad with the oscillation: Does it have a DJI Naza GPS?  An authentic one or a clone?

I have an oscillation in VelocityRoam on my 500 with authentic DJI GPS, but my 330 with OP GPSv9 has no oscillation.

[HighAltitude]

My two quads are similar - 450 mm frames, Multistar 22 series 900-100 Kv motors, Simon K 20A ESC's, Orange receivers, Revo FC boards, so I can experiment with different configurations while maintaining a backup air worthy vehicle. 
The oscillating quad is set up with 3 flight modes - Manual, Position Hold, and Return to Base.  RTB is configured per the Manual, with optional settings of a 3 m altitude offset (since I'm flying over a 5 acre open field). Take Off Location set to Arming Location, and Return to Base Next Command set as Hold.  All other parameters are default.
My procedure is to take off in Manual mode, switch to Position hold, which is pretty much rock solid, switch back to Manual to reposition the vehicle 50-100 feet from take off location, stabilize in Position Hold, and switch to RTB.  As "advertised" the quad climbs about 3 meters, flies back to above the take off location, but seems to overshoot.  It then tries to correct, but overshoots and starts oscillating. 
I recollect there's a reference somewhere in the Manual about "braking", which may need retuning, or some experimentation with PID settings may be necessary to stabilize in the Hold mode after returning to base.
At least I have a large field with a foot and a half of soft snow to minimize damage to my aircraft if I get it wrong, as long as my fingers stay warm enough.............

[DocHardinger]

Mine is oscilating too, but only if i use rtb with hold. All other gps modes are great also rtb with land works great. I got a neo m8n gps.

[TheOtherCliff]

The Hold of "RTB with Hold" should be the same as simple PositionHold or even as VelocityRoam with sticks centered.

I find that if I am motionless when I start VelocityRoam that maybe it does not oscillate.  If I engage PH from fast flight it will oscillate.  If I start VR with it motionless it may not start to oscillate, but if I move it and stop suddenly it will start oscillating and get worse until it is using the max angle (25 degrees IIRC).

I also notice during the oscillation there are steps that seem about 5 per second and that matches with the GPS frame rate.

I am looking into what is wrong here and how to fix it, assuming that the issue is with the DJI GPS being so filtered (to get smooth PH) that there is a big lag in position/velocity data.

[TheOtherCliff]

Just an update.

I put the DJI GPS on another working LP GPS quad and it works fine.  It looks like the issue was a bad mag calibration.

I found that the oscillation happens more when the quad is pointing in one direction than another.  The mag re-calibration did not fix it.

[HighAltitude]

If it ever quits snowing long enough for me to pack out a takeoff/landing zone in my hay meadow, I'll recalibrate the sensors in quad A and try my other (more or less identical) quad B.  Both have seemingly OEM DJI NAZA GPS/Mag units.

[TheOtherCliff]

I found that the oscillation happens more when the quad is pointing in one direction than another.  I have a firmware fix that is the only thing I know that helps.
https://forum.librepilot.org/index.php?topic=382.msg20502#msg20502