We've had some reports of poor GPS accuracy in 2.9.1 as compared to 2.8.1. This discussion is to debate, test, conclude and if necessary fix issues related to GPS accuracy in 2.9.1.
You need to be a member of diydrones to add comments!
Ran three packs today, GPS in MUCH better shape, now I just have to fix my declination and yaw issues. The config file you sent Stephen seems to work well, no jumping around. I'll test it on my trad Heli as well and confirm. I did also add a ground plane beneath the GPS unit (alu plate).
Still using the pm module.
On machine 1 (trad heli) I noticed the jumping behavior on a newly installed APM 2.5/Ublox and sort of resolved it with a ground plane plate below the GPS unit.
On machine 2 (hexa) I recently upgraded from an APM 1 to an APM 2.5 (APM 1 was running firmware 2.9.1) using the SAME GPS UBLOX module that was on the machine in the first place and the jumping behavior started AFTER I upgraded to the 2.5 unit with 2.9.1. I did NOT notice this behavior using APM 1 and the ublox module. I've not added a ground plane to this machine yet nor have I reflashed the configuration on either of the 2 ublox modules.I am using the PM on the 2.5 setup where I wasn't on the APM 1 setup. I am however using the ESC BECs to power the output rail and the PM to only power the input rail.
I'm going to reupload the ublox config on the hexa, add a ground plane plate and retest to see if that changes the behavior. If that doesn't fix it, I'll probably remove the PM and retest.
I was thinkin ability to install multiple GPS receivers and pick two signals out of 3 or 4 to get redundant GPS signal to APM. In this case if one GPS jumps to a different location suddenly the two remaining would have somewhat the same location and average of that location would be sent to APM. Would this kind of "system" be possible?
Guys, I just spent a weekend with testing problems with GPS accuracy. I found interesting thing, which could help also to others. So:
After uncontrolled flyaway during loiter testing flight last week (with emergency landing due lost of orientation - two broken props) I have recognized, there is a big problem with my GPS accuracy. I have APM 2.5 powered by APM power module, running fw 2.9.1 and uBlox GPS. Location of my quad on map in MP was drifting in unacceptable area (very often out out 30 meter circle from home location), even if quad was sitting at one place under open sky. I tried almost everything. To move telemetry (EU version) and RC antenas far away from GPS sensor, I swiched of RC receiver as well as telemetry unit, I moved GPS sensor to many positions. Still the same result - when I connected USB cable from PC, 3D fix was made almost immediately an stable. When I powered my quad from battery, problems with accuracy were still the same. I read alot of posts describing this "phenomena" regarding "USB powering". And then I found it. I removed completely APM module and connected one of ESCs BEC to power input rails of APM (and I also closed JP1 jumper). And guess what ! GPS 3D fix (inside of my house!) was made almost immeditately after connecting battery. My quad is sitting now already more that one hour in front of my house and his location on MP map, sent via 3DR telemetry is like frozen (but live). Track of quad "moves" is like one thick point - no bigger than lets say 1 meter in diameter. Incredible improvement! Maybe this my research can help also to others. In my case problem was in APM power module. As I removed power module from my quad, my "GPS unacceptable accuracy" issue is over.
I bought a ublox neo (the $27 version) and connected it up to the APM2.5
at first everything seemed ok (connected to USB ) data was coming in from the GPS and the APM said ublox ok.
But then I disconnected the APM from usb and hooked it up to the battery.
GPS lost (NO GPS!) ...back to USB and there was the GPS again.
solution:
(kind of) ....my USB voltage is 4.5V the output data of the GPS has an amplitude of 2.5V . When on USB, the APM finds 2.5V worthy of a "1". When on 5 Volts (via lm7805 from battery) the APM DOESN'T find 2.5V a logical "1" voltage, so the GPS connection is lost......I am building a level converter for TX data out from GPS.
Today I made my first flight with the ublox gps. I had been using the MTK and was not satisfied with the accuracy etc. It would wander around sometimes in a 30 to 40 ft radius and I could never run a battery thru without correcting the position. I did a 7 minute flight this am in loiter mode and never had to touch the radio. The most deviation may have been 1 m but it was very stable. This was in calm environment and will try in some wind, but for now I am very happy. The ublox seems much more stable and it links up very quickly. Well worth the investment IMO.
Replies
Hello all,
Hope you will find some useful information from my investigation on GPS accuracy:
http://diydrones.com/xn/detail/705844:Comment:1248134
ive had the exact same problem happen to me yesterday with 2.9.1 - is there a solution around this ?
Perhaps this isthe problem im having here ?
http://diydrones.com/forum/topics/problem-with-gps-or
can anybody shime in please ?
Still using the pm module.
hey guys
Built a new top cover for my 2012 steadidrone quad....
will test this tomorow and let you know hte results.
photo 1.JPG
photo 2.JPG
Guys don't know if this helps:
On machine 1 (trad heli) I noticed the jumping behavior on a newly installed APM 2.5/Ublox and sort of resolved it with a ground plane plate below the GPS unit.
On machine 2 (hexa) I recently upgraded from an APM 1 to an APM 2.5 (APM 1 was running firmware 2.9.1) using the SAME GPS UBLOX module that was on the machine in the first place and the jumping behavior started AFTER I upgraded to the 2.5 unit with 2.9.1. I did NOT notice this behavior using APM 1 and the ublox module. I've not added a ground plane to this machine yet nor have I reflashed the configuration on either of the 2 ublox modules.I am using the PM on the 2.5 setup where I wasn't on the APM 1 setup. I am however using the ESC BECs to power the output rail and the PM to only power the input rail.
I'm going to reupload the ublox config on the hexa, add a ground plane plate and retest to see if that changes the behavior. If that doesn't fix it, I'll probably remove the PM and retest.
Hi.
I was thinkin ability to install multiple GPS receivers and pick two signals out of 3 or 4 to get redundant GPS signal to APM. In this case if one GPS jumps to a different location suddenly the two remaining would have somewhat the same location and average of that location would be sent to APM. Would this kind of "system" be possible?
Petri
Guys, I just spent a weekend with testing problems with GPS accuracy. I found interesting thing, which could help also to others. So:
After uncontrolled flyaway during loiter testing flight last week (with emergency landing due lost of orientation - two broken props) I have recognized, there is a big problem with my GPS accuracy. I have APM 2.5 powered by APM power module, running fw 2.9.1 and uBlox GPS. Location of my quad on map in MP was drifting in unacceptable area (very often out out 30 meter circle from home location), even if quad was sitting at one place under open sky. I tried almost everything. To move telemetry (EU version) and RC antenas far away from GPS sensor, I swiched of RC receiver as well as telemetry unit, I moved GPS sensor to many positions. Still the same result - when I connected USB cable from PC, 3D fix was made almost immediately an stable. When I powered my quad from battery, problems with accuracy were still the same. I read alot of posts describing this "phenomena" regarding "USB powering". And then I found it. I removed completely APM module and connected one of ESCs BEC to power input rails of APM (and I also closed JP1 jumper). And guess what ! GPS 3D fix (inside of my house!) was made almost immeditately after connecting battery. My quad is sitting now already more that one hour in front of my house and his location on MP map, sent via 3DR telemetry is like frozen (but live). Track of quad "moves" is like one thick point - no bigger than lets say 1 meter in diameter. Incredible improvement! Maybe this my research can help also to others. In my case problem was in APM power module. As I removed power module from my quad, my "GPS unacceptable accuracy" issue is over.
Hi guys,
Has anyone of you encountered the following?
I bought a ublox neo (the $27 version) and connected it up to the APM2.5
at first everything seemed ok (connected to USB ) data was coming in from the GPS and the APM said ublox ok.
But then I disconnected the APM from usb and hooked it up to the battery.
GPS lost (NO GPS!) ...back to USB and there was the GPS again.
solution:
(kind of) ....my USB voltage is 4.5V the output data of the GPS has an amplitude of 2.5V . When on USB, the APM finds 2.5V worthy of a "1". When on 5 Volts (via lm7805 from battery) the APM DOESN'T find 2.5V a logical "1" voltage, so the GPS connection is lost......I am building a level converter for TX data out from GPS.
Today I made my first flight with the ublox gps. I had been using the MTK and was not satisfied with the accuracy etc. It would wander around sometimes in a 30 to 40 ft radius and I could never run a battery thru without correcting the position. I did a 7 minute flight this am in loiter mode and never had to touch the radio. The most deviation may have been 1 m but it was very stable. This was in calm environment and will try in some wind, but for now I am very happy. The ublox seems much more stable and it links up very quickly. Well worth the investment IMO.
-
1
-
2
-
3
-
4
-
5
of 5 Next