I had seen there were some issues fixed with this and am running 4.6.3, custom build as I am running Matek AS-DLVR-IC2.
I think the video with a good hud shows it clearly as it indicates “flare” as soon as the rangefinder is acquired. I have been trying to resolve this for a few months now. I included a lap of the QGC generated loiter to show airspeed was in good order, perhaps a 3 m/s headwind on landing.
I’ve got the log as well but the vid shows the problem clearly.
What is the rangefinder range? In my experience you want 10-12m at a minimum for small planes so plane has enough time to cleanly adjust path or perform go around.
There were some issues with flare transition when using rangefinder that are fixed in 4.7.0
6.5° is fairly steep unless you have a draggy airframe or reverse thrust.
.bin dropbox link. There are two flights, the first the land position was out and I went around and landed manually.
The 6.5 is steep. Ranger 2000 heavy with manual flaps and crow. I initiated full flaps when final started, but no crow. I had been trying lower slopes to get to the point where the plane used power on final but that approach wasn’t working, still early flare no energy so I went back to 6 degrees.
Rangefinder kicks in at 10m consistently, I tried increasing the RNGFND1_MAX_CM 1500 4 on a suggestion from AI. but behaviour was the same as when it was set to 1000.
Perhaps my problem is 4.6.3. I saw some fixes to rangefinder flare around 4.6.x so I thought I would be fine with the 4.6.3 custom build I had to do for the Matek AS-DLVR-IC2 to work.
Immediately after the rangefinder engages at 9.81m, LAND_fh (height for flare timing) shows -167m. This may be a problem with the terrain correction, so you might want to try setting TERRAIN_ENABLE to 0.
I think i found the problem. Even with rangefinder_max at 8, there were some 5m readings in the stream as the finder acquired. This blew up the the land.fh. Here’s the stream from the rangefinder, after the last stamp all values were good.
I notice the wiki for the Benewake TFmini suggests a rangefinder_max of 6, this might not be appropriate as I see the 5m anomaly showing up in many logs.
but behaviour was the same as when it was set to 1000.