Skywalker 1900 , gridmission skipping patch and not following waypoints correct

Hi all,

After alway's flying manual(20 years)  i wanted to help a farmer to do some arial mapping. I have build a Skywaker 1900 flying nice and trimmed manual and also easy to fly in assisted flight.

In auto (grid) mission it does not follow each path correct and also skipping waypoints, i did already a auto tune on the plane. after that i created a automission with a box and tune the L1 parameter back from 20 to 16 (15 did not get any more accuracy). Changed also the max banking angle from the std 45 to 50, it can do nice corners but for some reason it does not follow the flight path correctly and even worse skipping waypoints

I use a PixHawk 2.1 with the Here+ gps. using Adruplane 3.8.4 and mission planner 1.3.56. All digital servo's running on 4s. GPS signal is excellent, i am out of ideas what to check next. 

I have set overshoot to 80 meter and leadin to 20 meters to help the plane to get on track on time but it does terrible job at it. Way-point radius is set to 10 meters. See attached screenshot

I am not speaking natural English, if something is unclear please ask.

Views: 602

Attachments:

Reply to This

Replies to This Discussion

I’ve seen this before, especially with the Skywalker. I suspect that your plane is flying in a slip due to the spiraling slipstream from the big pusher prop hitting the tail and creating a yaw.  You can see the plane flying in a slip (crooked) while it’s climbing out just after launch. Try adding lots of rudder trim and see if that helps the tracking accuracy.  

Additionally I think your L1 is still too high. Keep decreasing it until the plane starts oscillating across the path, then increase. I use 60 deg bank limit to keep the turns tight. The 1900 won’t tip stall easily. The 1880 is terrible that way. 

Hi,

Thanks for your suggestions 

At cruse speed ( in manual mode)  i don't have to add any rudder for compensating only at full power, in auto mode i fly at 12 m/s not even half power. 

I will decrease L1 even more and increase the bank limit to 60. Not sure if i have time this week ( work & weather) but i will try those and report my findings here.

How do you know that you don’t need rudder? The plane will fly in a straight line of the ailerons are compensating for the rudder. Check the lateral acceleration in the logs to be sure.  

Bert said:

Hi,

Thanks for your suggestions 

At cruse speed ( in manual mode)  i don't have to add any rudder for compensating only at full power, in auto mode i fly at 12 m/s not even half power. 

I will decrease L1 even more and increase the bank limit to 60. Not sure if i have time this week ( work & weather) but i will try those and report my findings here.

I maiden the plane with just a Receiver the traditional way after build and did trim it using the linkages to get it fly correct on normal cruse speed , so without any kind of controller. And also with Pixhawk in manual mode. But i will check the logs to be sure.I hope I overlooked something 



iskess said:

How do you know that you don’t need rudder? The plane will fly in a straight line of the ailerons are compensating for the rudder. Check the lateral acceleration in the logs to be sure.  

I don't have the correct log file only the logs from the pixhawk controller (telemetry log is on a different laptop).

But i had something on my own laptop as well from a random test flight. Are we looking for the ay ? parameter, i am not that familiare yet with the logs 

Attachments:

Ay is what I use when graphing in real time using the Tuning window in MP. 

I’m not near a computer so I can’t tell you what to look for in the logs. Both logs should have accelerometer data. You are looking for lateral (side to side) acceleration. 

I have got also the tlog files but ,I have no idea where to look form google does also not help much.

If you are near a computer in the oncoming day's and would take a look that will be great.

Hi Bert,

I use the Skywalker for mapping. I looked at your images and here is my 2c worth:

1. The lines followed by the bird were quite straight and running parallel to the mission line. They seem to run right of the mission line on all occasions(Not sure if I'm seeing this right - seems the same side all the time). I would speculate that this due to torque from the prop and not have slip control from a rudder.

2. I use a 150m overshoot and a 70m lead in. I fly at 15m/s airspeed and use an airspeed sensor. If you do not have an airspeed sensor and you make your turning circles very tight, you might invite a stall. I see your ground speed is 20m/s with an wind speed of 5m/s. As mentioned above 12m/s is very close to stall speed for a heavy loaded Skywalker. At 20m/s you move the complete length of your lead in, in only 1 second. There is very little time to maneuver. Give the bird some space.

3. This might just be speculation - the fact that you have every exposure point defined as a way point might have an impact on how tight the line is followed. Maybe one of the plane developers can comment on this. The bird seem to miss a waypoint but the AP moves onto the next waypoint if it is unable to maneuver in time. With my missions, when the bird turns at the end of the 150 overshoot and comes back for the 70 lead it, it also misses the point by 20m but by the end of the lead in line, it is on line and tracking well. I dont use waypoints for exposures. Just line start, end, overshoot and lead in. I would do it like this:

I will download my bird's settings when I'm back at the office and forward the file early next week. Maybe that will help.

Good luck

If you are using the Skywalker 1880 then 15m/s is appropriate, however Bert is using the 1900 which has an undercambered wing. I use 13 m/s as a trim speed on my 1900’s. 

Antonie Kruger said:

Hi Bert,

I use the Skywalker for mapping. I looked at your images and here is my 2c worth:

1. The lines followed by the bird were quite straight and running parallel to the mission line. They seem to run right of the mission line on all occasions(Not sure if I'm seeing this right - seems the same side all the time). I would speculate that this due to torque from the prop and not have slip control from a rudder.

2. I use a 150m overshoot and a 70m lead in. I fly at 15m/s airspeed and use an airspeed sensor. If you do not have an airspeed sensor and you make your turning circles very tight, you might invite a stall. I see your ground speed is 20m/s with an wind speed of 5m/s. As mentioned above 12m/s is very close to stall speed for a heavy loaded Skywalker. At 20m/s you move the complete length of your lead in, in only 1 second. There is very little time to maneuver. Give the bird some space.

3. This might just be speculation - the fact that you have every exposure point defined as a way point might have an impact on how tight the line is followed. Maybe one of the plane developers can comment on this. The bird seem to miss a waypoint but the AP moves onto the next waypoint if it is unable to maneuver in time. With my missions, when the bird turns at the end of the 150 overshoot and comes back for the 70 lead it, it also misses the point by 20m but by the end of the lead in line, it is on line and tracking well. I dont use waypoints for exposures. Just line start, end, overshoot and lead in. I would do it like this:

I will download my bird's settings when I'm back at the office and forward the file early next week. Maybe that will help.

Good luck

Hi ,

Thanks very interested in the settings , i have search every where for this  lateral (side to side) acceleration in the logging it is somewhere i am sure but i have no idea where.

I wil try to find some more info abount sideslip settings etc.

My overshoots and lead in where not that big, so i can increase these also to give the plane more time to get on track.

Here is an old thread on the topic:

https://discuss.ardupilot.org/t/flight-path-offset/3881

Unfortunately the others have long been deleted from the DIYDRONES. 

Hi Bert,

I'll send a PM with a link to the file.

Reply to Discussion

RSS

Groups

Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2018   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service