Comments

• 

  Dmitry Prokhorov November 15, 2015 at 8:53am

  Stefan,

  After "git clone https://github.com/diydrones/ardupilot.git" command you need to perform "git checkout Copter-3.3.1". W/o quotes.
• stefan jakobsson November 15, 2015 at 8:48am

  how do you compile copter 3.3.1(I use ubuntu)?  I have done as described in :http://dev.ardupilot.com/wiki/building-px4-for-linux-with-make/ .

  But that takes the master build.I want to compile a stable version.

  The command "git clone https://github.com/diydrones/ardupilot.git" clone the

  master build.So how do I type so I can make a build of 3.3.1?
• Phil Horner November 14, 2015 at 11:36am

  I'm back. Thanks guys. your posts are great.

  Yes, although I'm in Papamoa NZ, the location with the Pixhawks GPS's are stated to be between 40 - 50 kms out to sea in the Bay Of Plenty.

  Although not inside, the GPS,s were booted up in a confined courtyard approx. 20 ' in diameter and with a limited window to the sky.

  Typically, it is now raining outside so I can't check anything, however you both have given me some invaluable information. 

  My course of action will be as follows.

  1/ Go a friends farm (Wide open spaces) and recheck the discrepancies.

  2/ If problem resolved i.e. caused by multi pathing.

  3/ if unresolved do manual magnetic declination and re calibrate mag.

  4/ Then finally look at SBAS if nothing has improved.

  I'm so grateful for all your replies to date. As soon as this crap weather improves, Ill get out there and test and let you know how I got on. Again, thank you.
• Developer

  Bill Bonney November 14, 2015 at 10:41am

  Here, for reference, is what happens to multi path GPS signal. The copter was sat on my bench in the house the whole time. :) GPS is not reliable unless it has a clear vie of the sky, it's worse as soon as the vehicle starts moving as well indoors, stationary it seems to figure it out to some degree.
• Developer

  Bill Bonney November 14, 2015 at 9:17am

  Where you testing with the GPS units indoors? With multipathing it can think its in quite a different place, you need a clear view of the sky for GPS to work. If all three units are indoors they would be seeing the same signals probably reflected in through the window. In anycase the phone will also be reporting it's location based on the WiFi access points detected in your area. Google Street car builds a location index based on the detection of wireless access points as it drives around, triangulated with real GPS. This is how your phone can get such a quick lock, even indoors. Though it's not very accurate, it better than nothing.
• Phil Horner November 14, 2015 at 3:27am

  I will do as you suggest in the morning. It's half past midnight. I'll leave a comment for you in the morning as to how I got on. Thanks for replying.
• 

  Clifton Hipsher November 14, 2015 at 3:26am

  There is nothing wrong with your aircraft.  The "error" is with the phone.
• Phil Horner November 14, 2015 at 2:19am

  It has just been suggested to me to manually do the magnetic declination. What is the GPS_SBAS_MODE?
• Phil Horner November 14, 2015 at 1:53am

  That's a good point, unfortunately I'm by myself at the moment and it's nightime with kids running round the street letting off fireworks so I can't drop it off down the street and come back to the computer to see. I'll have to do it tomorrow. One thing I did notice although it probably has no bearing on it, is that the image used in mission planner is an older one than that I'm using in Google Earth. I'm wondering if there is a setup I've missed in Mission planner. I've recently updated mission planner, but I don't know if this problem was there in the previous version. Sorry, sounds like I'm rambling.
• Phil Horner November 14, 2015 at 1:19am

  LOL. No, they where all activated separately and none were operating together. I noticed the problem with one, then tried the other two in turn all resulting with the same Long and Lat position. So although they were all totally independent they did give me exactly the same location position.