[quote=“tridge”][quote=“Steven G”][quote=“Steven G”]What do you think?
[/quote][/quote]
did you send me a log file? Maybe I missed it?[/quote]
Sorry, I forgot to…here it is.
Steven
Thank you Tridge.
Best
-S
Thanks! I’ve looked into it and the problem is that the EKF doesn’t estimate wind speed if you don’t have an airspeed sensor. The DCM code can estimate wind speed without an airspeed sensor, so I’ve just pushed a change so that the DCM wind estimate is used when flying with EKF enabled and no airspeed sensor.
Cheers, Tridge
Great!
Thank you.
I’ll have an airspeed sensor soon.
I will try out the FW update via telem. Are there any special settings or procedures I need to be aware of?
Tridge,
The EKF does have the capability to estimate windspeed without an airspeed sensor through use of a zero sideslip assumption, however this calculation is only active for flight without a magnetometer, or if there are no GPS updates and the filter is dead-reckoning.
github.com/diydrones/ardupilot/ … F.cpp#L766
This could easily be extended to operate for the no-airspeed sensor case ( I have successfully flown with no-airspeed sensor and no magnetometer and it was able to form a wind estimate after a few turns).
Regards,
Paul
I have all sensors except airspeed. Sounds like I should just stick with 3.0.0 for the moment, right?
ok, for the moment I’ve just patched AP_AHRS_NavEKF.cpp to use the DCM wind estimate for reporting purposes if we don’t have an airspeed sensor. Once we change AP_NavEKF.cpp to estimate wind with no airspeed sensor in normal flight we can take that change out.
Cheers, Tridge
I’ve just released APM:Plane 3.0.1, a bug fix release for the 3.0.0 release.
This release fixes two bugs:
[ul]throttle failsafe for aircraft using PWM level for failsafe detection
wind reporting with EKF enabled and no airspeed sensor
[/ul]
The throttle failsafe fix is a critical bugfix, which is why I am doing a new release so soon. The bug was found by Sam Tabor, and he posted the bug report before the 3.0.0 release, but I didn’t notice it in the release preparations.
The bug only affects systems using PWM value as the sole method of detecting loss of RC control. I hadn’t noticed it myself as my planes all use receivers which stop sending value PWM frames when the RC link is lost. In that case failsafe worked correctly. Receivers that keep sending PWM frames but with low throttle levels are common though, so this is a very important fix.
Many thanks to Sam for reporting the bug, and my apologies for not noticing it in time for the 3.0.0 release.
Cheers, Tridge
PS: The new firmwares will appear on firmware.diydrones.com and in the MissionPlanner firmware screen in about an hours time after the autobuild is complete
[quote=“tridge”]
I assume you are referring to fetching the parameters on initial connect? To fix that you need to ensure your radio has the flow control pins properly connected. If flow control is working then you should find that parameter fetch is very fast.
If you can’t connect the flow control pins then try setting ECC=0 in the radios.
Cheers, Tridge[/quote]
Yes, the fetching of parameters upon initial connect is very slow. As I recall, I have ECC=0 set. I’m using MinimOSD, would that make any difference? I do not have the MOSD Tx connected. I’ll take a look at the flow control pins but I think I have all set up as in the Wiki.
Thanks for your help,
Steven
[quote=“tridge”][quote=“priseborough”]
This could easily be extended to operate for the no-airspeed sensor case ( I have successfully flown with no-airspeed sensor and no magnetometer and it was able to form a wind estimate after a few turns).
[/quote]
ok, for the moment I’ve just patched AP_AHRS_NavEKF.cpp to use the DCM wind estimate for reporting purposes if we don’t have an airspeed sensor. Once we change AP_NavEKF.cpp to estimate wind with no airspeed sensor in normal flight we can take that change out.
Cheers, Tridge[/quote]
Wind speed is working fine with 3.0.1.
Tridge, et al,
In the attached tlog, please note between WP 5 and WP6, the plane is very subtly porpoising. It seems to almost stall (see AS) then the autopilot lowers the nose to gain airspeed. WP5 to WP6 is essentially a power off glide. I am using Pixhawk with 3.0.1. I’ve never seen this behavior in the past so I thought I would bring it to your attention.
Thanks,
Steven
Steven,
Do you have the flashlog file? It will be a lot more useful to us than a tlog. I see you had the default value for LOG_BITMASK, so everything should be there.
Paul
I heard Steven got banned from the forum so he won’t be able to reply. I hope to see him back here because I found him to be a very active participant and overall contributor to this forum.
I’ve asked the moderators to unban Steven.
Hopefully he will be able to upload his SD card log so we can take a look.
Cheers, Tridge
I just updated my arduplane firmware from 3.00 to 3.01, now when i roll clockwise with my apm, mission planner role ccw in flight data, (i have updated mision planner also),
[quote=“Steven G”]Tridge, et al,
In the attached tlog, please note between WP 5 and WP6, the plane is very subtly porpoising. It seems to almost stall (see AS) then the autopilot lowers the nose to gain airspeed. WP5 to WP6 is essentially a power off glide. I am using Pixhawk with 3.0.1. I’ve never seen this behavior in the past so I thought I would bring it to your attention.
Thanks,
Steven[/quote]
Hi Steven !
Great to see your flight log, your plane seems very very efficient ! You just consumed 1400mAh for almost 16KM of flight.
Which airplane did you use for your flight on 10th April with Arduplane 3.0.1 on Pixhawk ?
Or was it a simulated flight ? 
Regards
Rana
Hmm, beyond visual range flight over populated areas and above 400’ altitude - this is not the sort of thing that we should be promoting.
Besides the safety aspect (I’m not familiar with the legality situation over in the US), I reckon based on the number of large homes with swimming pools and ocean views in the area, that their owners could have a lawyers number on speed dial.
I try the new code, 3.01 with my Skywalker 2014/1800. APM 2.6+. I notice the DO_Jump still not working. Still a dummy way point is needed
That is very strange - it shouldn’t need it. What was the mission file you tried?
Cheers, Tridge
Hi Tridge
Attach herewith the tlog and the mission file.
I’m testing the new 3.01 code in my new Skywalker 2014/1800. APM 2.6+ external magnetometer, 3DR power module, RDF 900, T motor 2814-710kv with 4s 6000 mah battery.
I am quite happy with my current setting. It hold altitude within 8 meter and track the grid line well. I’ll try to fine tune the PID for better performance.
In the means time, if you happen to see my tlog, I will be grateful if you can give me some advise on the tuning of my plane.
Thanks for your great job and look forward for the development of the pixhawk controller. I have one on hand and will put it to my trusty Skywalker air frame for testing and get familiar with it. Thank you once again.