Trying new props on my copter today I expected that MOT_THST_HOVER might need adjusting. Indeed - in LOITER, the copter would sink a bit when transitioning horizontally.
Checking my setting for MOT_HOVER_LEARN I found it to be set to “2” for “learn and save.” But there didn’t seem to be much learning and saving going on during my test flight.
Looking up MOT_HOVER_LEARN I was reminding there is an alternate setting “1” for “learn.”
Would someone please explain the difference between “learn” and “learn and save?”
Also - any particulars on what sort of flying is required for this learning to take place?
If it’s set to learn the learned value will be available in the log as CTUN>ThH but it will not update the parameter value MOT_THST_HOVER. You can review the log and manually set this if you like. If it’s set to Learn and Save if will be logged and it will update the parameter value.
Hovering in a non-manual mode will learn the hover value. You can just leave set to 2 all the time if you like.
My opinion is you would use Learn (value 1) if you were constantly changing payloads or take off weight. The learnt hover will be used during the flight but not saved. Each flight will start with the old saved hover value, and learn as you fly.
Normally you would use Learn and Save (value 2) since there’s a good chance you would always leave the same camera (or payload) in place.
EDIT
The hover learning works quite well in AltHold and Loiter and it’s fairly fast too.
So the rise or fall you see during transitions or direction changes could be caused by vibrations or prop wash (for example). Any HNOTCH settings will have to be redone if changing prop sizes, depending on which HNOTCH mode is in use.
Dave - this is all really good information - and makes perfect sense.
I’m testing new 3-blade props. 9"x4.5 instead of the 2-blade 10"x4.5. (on a 450 class quad)
My last value of MOT_THST_HOVER with the 2-blade props was 0.2127801
After today’s tests with the 3-blade props, it’s 0.1924673
This is interesting in part because the published range is 0.2 to 0.8.
I’m using FFT dynamic notch filtering - so I’m guessing unless there’s a new harmonic outside the range, the filtering is still pretty good. The copter was very stable and flew in good tune - even using the auto-tune PID’s from the 2-blade props.
But in a horizontal transition in LOITER, the copter would sink a bit.
I graphed out the CTUN>ThH as you suggested to see what changed through my LOITER test flights - and it indeed stair-stepped down in a couple of steps from the original setting.
I suppose the 3-blade props are more efficient when standing still - but perhaps give up considerable vertical component of lift when pitching and rolling for horizontal flight. Might be a good reason to stay with the 2-blade props if I planned to fly around much in LOITER. But maybe there’s a PID for vertical position that’s too low.
I was curious to see if any significant change in flight time per Ah demonstrated better efficiency - there was some - but maybe within the margin of error. (23 minutes versus 21 minutes)
What I’d really like to find is some nice 2-blade carbon fiber props - but there doesn’t seem to be any in the 10"x4.5 size.
Once again - I really appreciate your analysis and insight - and welcome any other thoughts that may occur to you. Thank you!
Right, nothing to adjust with FFT reference notch.
Yes, motor shafts for those annoying self tightening props are a pain. I have 1 old quad with those. I did find some regular props that fit with nuts but difficult to find.
I have a set of the Master Airscrew MR props with the built in nuts. The quad performs well with them, but the production mold must be getting old because the trailing edges of the props aren’t as crisp as I would hope so the quad now has a very noticeable whistle.
I have had an incredibly good experience with the 12 inch 2 blade MasterAirscrew MR props. They are really nice in flight and perform way better than advertised or expected.