I am having an issue with autotune not responding to RC inputs to reposition, even on full rates, so I have to resort to turning off auto tune, repositioning, and switching it on again to resume. I am using the latest AC 3.2.1, but was also an issue on 3.2. AC is running on the HK pilot micro (hobbyking APM). I have the “nearly all” log below, showing the RC inputs not being accepted whilst autotuning, I checked it with the APM log file analyser.
Log file:
https://drive.google.com/file/d/0B6QFq1-xwWDSOXBGa1hlMWxCdU0/view
Analyser output:
https://drive.google.com/file/d/0B6QFq1-xwWDSTzloOWplYmpfZFU/view
https://cloud.githubusercontent.com/assets/6254189/6318970/2f3f9788-baa3-11e4-83b3-081aa4f50d69.png
OK, this took some digging to find all the information, I think this shows the effect. I have the Taranis logs on the left, and the Arducopter log on the right, the scaling is similar. Notice the small excursion to 1460 on the left of the arducopter and the -60 on the left Taranis log at 78 seconds. If you notice on the taranis logs, approximately halfway between the positive and negative excursions at both ends (78 and 104 seconds) there are multiple larger negative excursions at around 89 seconds, at which time the autotune is still on according to SC (switch c, the autotune switch), yet the arducopter logs miss this somehow, and doesn’t see a significant input, If you look very closely on the arducopter logs there is an apparent 5us square wave before the alt hold kicks in, which follows the inputs according to the taranis. I’m sorry this effect isn’t more pronounced, but I can do another autotune with lots of inputs to highlight this further.
Turn out it’s my own fault, bad mix settings on the Taranis.
