Just posted this version which has an increased performance DCM implementation. This allows the main loop to run 50Hz faster without increasing the CPU load. I've flown this for 30 minutes and it's an improvement in every respect.
Circle mode now seems to be working just fine. This should be the last beta! enjoy testing!
Here is the Link to a download:
No, that just means that you haven't updated your Mission Planner recently. See this post.
Just installed MissionPlaner: 1.0.68
Finally got the time and the weather to load up v42 and have to say its a big improvement!!
And I left everything default. Surpised be this because previous version Ialways had to twinker with stable P or D
Stable is stable like previous versions. Very good.
AH great at low and high alt.
Loiter seems great though needs some finetuning when there is some wind.
Tested RTL several times and the quad just comes back very easy. But altitude goes up and down. As AH pids are great, which should I change to adress the oscillating alt in RTL???
And the finally a mission. My first. Started of with 3 WP. WP 1 was just a waypoint, WP 2 loiter for 10 sec and WP 3 LAND. I took off switched to auto and of it went :-) What a great feeling.
But .... It seemed that the quad did everything I wanted but only at WP1??? So it went to WP1, stayed there for about 15sec and started to descend very slowly until it touched the ground.
I then reloaded the mission and tried again. But nogo. Samething and I had the impression that the quad was very unstable.
I'm doing something wrong??
What I also noticed on the planner is that the current altitude isen't correct. It also jumps up and down. I for example start at 0m and land after a flight with -8m?
A great leap forward.Thanks guys!!!!!!!!
Vincent, the altitude displayed in the mission planner seems to be only the gps altitude. GPS altitude is fairly unaccurate. I also fly fpv with an OSD with gps altitude and often when I land, the gps altitude will display -5 to -15 values.
It might be a good idea to replace gps alt with baro alt in the mission planner ?
Anyway, congrats on the mission ! Hope you find out why it stayed at wp1.
Ok then, it would be nice if alt was more correct.
Here is the log of all the tests I did this evening;
What if they took sonar = ground 0, most sonars rated for 6M. Use sonar to keep the the most acurate alt, 0 above ground when init. At this point you could then perhaps have the baro calibrated because you know where ground 0 is, well which is 20cm off the ground because of min distance on sonar. Then once in the air, use sonar for as long as possible before moving on to not just baro but gps too. So you could perhaps use the average of the GPS and Baro to keep tabs on each other and everytime you get accurate alt data from the sonar (5secs solid data), use it to re-calibrate the baro and average out again with the GPS. Why does APM use the GPS alt, when we rely on the baro for general flight, doesnt make sense.
I am sure one of you guys can articulate this better.
I wish the sonar were good enough to reset the Baro. I used to do that in fact, but many things can cause the Sonar to misread. This can lead to the copter rising indefinitely. Which it used to do.
But I've found the Baro to be the most accurate sensor. Even better than the GPS. So adding in the GPS data would only worsen the reading.
The best way so far seems to be re-init the baro at arming. But I'm open to different idea. The best way is to clone the Git branch and try a few ideas out. Then share the results.
What about APM using GPS instead of baro data?
For what it is worth, my GPS altitude, even with a 9 satellites lock, varies by up to 16m while sitting still.
I think that's more legacy. Doug was the one who got me converted to Baro only.
So APM Planner no longer using GPS data?
Sorry, I thought you meant Arduplane. I'm not sure what the planner is using. The Altitude is reported back via mavlink in Absolute (with GPS) and Local (baro offset from home). You'd have to ask Michael Oborne what he uses.