Yes, it overides WP_SPEED and will fly at the new speed set by the DO command until another one is executed as I recall. I looked at an old WP file and I had speed in the 2nd column. Try it in SITL and see.
Just upgraded to 3.6.5 and it seems the dataflash log download is so slow now. Anyone else experiencing this?
(Fix) Connected the battery and the log downloaded like normal.
@rmackay9 should i do any additional testing or data capturing? run something in debug? Trying to see how to help make some progress on this. Create an issue on git?
thanks!
FYI weāve got a fix coming in 3.6.6 that will change this behaviour so that do-change-speed commands will not directly change the WPNAV_SPEED parameter. Instead the new speed will only persist until the vehicle exits Auto mode.
Also by the way, the 1st column affects what speed is affected. So 1=horizontal, 2 = speed upwards, 3 = speed downwards.
I think I will need some help from @tridge to get to the bottom of this. One thing that is making it difficult is that we canāt reproduce it (two of us have tried). Iāve added the issue to our Copter-3.6 issues list so we wonāt forget about itā¦ but I think I need some help from other devs more familiar with the low-level code to get to the bottom of it.
Thanks Randy. Iām still not clear on the configuration for the columns for the command. Placing the horizontal speed in the second column with the 1st column at 0 changes the horizontal speed but Iām not getting expected behavior for vertical speed in the SITL so I must be getting it wrong. Enter a value of 1-3 in column 1 to specify the mode and then the speed in column 2? Or using 3 columns?
@rmackay9 thank you. If you guys have a ādebugā version of the f/w that logs to the card during the init time (or information how to build one) i am happy to upload and send the results. The issue is definitely occurring during the init time as ECSs never get initialized.
BTW, for my own sanity sake i built 365 branch locally as well uploaded apj file and got the same error. trying to see which incremental change may have caused this issue.
And as i had a sick day off today i did the binary search on commits between 3.6.4 and 3.6.5 and found the commit which broke matek compass support.
broken repo state:
due to this commit:
Unfortunately its a huge commit so cant really gradually track it down by partial compiles.
The last good repo state for matek was:
WPNAV_SPEED parameters first column:
What about 0 = all ? (Just a thought)ā¦and 1=horizontal, 2 = speed upwards, 3 = speed downwards
Thank you so much Randy ! This is great news!
Cheers
Henri
It looks like we actually accept both ā0ā and ā1ā as meaning horizontal speedā¦and the number is actually specified in the mavlink spec for the DO_CHANGE_SPEED command. ā0ā = āairspeedā, ā1ā = ground speed, etcā¦
So if one wants same new speed up, down and horizontal, should align three different commands?
Oh wait sorry! Iāve just realised that the ability to set the vertical speed is not included in Copter-3.6. Itās only in 3.7. We may include it in a point release because itās tied in with the RTL_SPEED bug but I canāt make any promises because the merges from master back to 3.6 are becoming quite difficult.
I did some testing of the stable v3.6.5 copter release on a Pixhawk v2.4.8 setup. No issues seen. Love the SD Card reliability improvements! Great work!
@rmackay9 any hopes for a fix fox matek405wing/beitian 880 combo?
again i can offer to test/validate at individual source code change level (recompile) if trial and error is the best option at the moment.
@maciek01, Itās still on our 3.6 to-do list but not much progress has been made. Iāll bring it up again at the weekly meeting and see if we can get someone to look into it more.
I just wanted to let you know that we havenāt forgotten about this problem and Iāve finally been able to reproduce it. Iāll chat with Tridge tomorrow and hopefully we can get to the bottom of it in time for Copter-3.6.7.
Great - thanks, truly appreciate!
Weāve got a fix now as well so I expect we will start the 3.6.7 beta by the end of the week. Txs for the report and details higher up in this thread!
thanks. i see the fix. happy to validate it once it hits the branch