I wanted to bring to your attention a terrain following bug. My configuration doesn’t use terrain tiles, and is setup to use a tfmini lidar sensor facing the ground. Terrain_enable is disabled.
Yesterday after a previous day of 3 successful flights, I continued to tune my quadcopter on Pixhawk1 firmware. After some ballooning the previous day on landing I lowered the landing speed, and concerned that the barometer which is definitely getting propeller wash and it’s current position, I went searching for additional terrain following land and rtl mode lidar parameters. After switching to enabling 2 or 3 of those, I don’t recall the exact ones I don’t have my drawing near me. I do have my ground station messages which one reads Failsafe: Terrain data missing. Even though terrain_enable is false.
Additionally, the propellers on the ground in althold mode are not acting right. The are not in sync and they are making strange noises, not responding to throttle input until many seconds later then wanting to flip. GS messages Potential thrust loss on 2 and 3. Yesterday I got to the point of taking the motors off trying to remount them because it sounded like they were flopping around which just doesn’t make sense I believe the software switches I triggered made it not work. I wish I had saved the params before I made the switch I just assumed they were harmless settings that would simply make my landings better. I might be able to follow up with param upload later, but it seems like you should be able to figure out which parameters I’m talking about based on my description.
I’m a little annoyed today because I plan to unsolder motors two and three and move them to another location to test if I’m continuing to get errors on two and three which really is a complete waste of time if it’s a software issue but I don’t know that it is. It’s holding me up from finishing my project so I don’t have a choice but I’m hoping if you guys know there’s something in the firmware that is a bug and beta 7 that you’ll stop me
Also, if these terrain errors are in fact not supposed to be occurring because they are tile related somebody really needs to put a unit test in based on the common schema, that searches source code to ensure if a particular group of functionality is disabled in the common schema properties list that other areas in code don’t override those parameter settings elsewhere. Could save your devs an enormous amount of time in the future given the entire ArduPilot codebase is based on a tree based key/value pair static properties file, sitting on top of an OS high level api layer using if else statements. It’s entirely doable in unit tests and could cut down your buglist substantially. Preventing human error.