GPS mast length?

[cootertwo]

Here's a question I haven't seen asked before. If you have your GPS puck/antenna, mounted on a mast, will the length of the mast effect the way the quad behaves when in an area where GPS signal is iffy? I'm experimenting with a F330 frame, and an APM 2.7. My GPS is on a mast 6 inches above the APM. While holding loiter under tree cover, the quad will tweak around pitch and roll, as the GPS drops and finds sats. I don't notice that on rigs that have the GPS mounted closer to the CG. Maybe it's just some PID's need adjusting? Just wondering, that's all.

 

[Carapau]

I dont know the apm system. In the DJI system you have to give the coordinates of the GPS puck in relation to the CoG- is this the same with the APM?

 

[Old Man]

Nope

 

[fltundra]

If you have the puck to low you run the chance of magnetic interference with the compass. I mounted mine just high enough to not pick up any. Plus never liked the idea of the puck stuck up like a flag pole above spinning props.

 

[cootertwo]

That's what got me wondering. No way to tell APM where it is.

 

[Old Man]

I have GPS installations in use that run from being flat mounted on a frame plate component (APM 2.6) to on masts varying in length from 4" (Pixhawk) to 7" with a JIYI. Had DJI's with long masts as well. The one that seems to function the worst is the surface mounted APM, having very slow lock times. The rest work fine with the Pixhawk installation the best of the bunch. Ya'll know which ones had a tendency to "get lost".

I believe both APM and Pixhawk know where they are using a satellite fix and perhaps use compass calibration to fix themselves relative to the the flight controller, if such referencing actually occurs or even matters. GPS is used for establishing location, height, direction, and speed and likely has little or nothing to do with flight control, with the controller handling everything to do with aircraft stability through gyros and accelerometers. A few centimeters location difference between a GPS and a controller should not have any impact on anything. For "real" UAV's GPS is used only for positional and speed references necessary for navigation, and target referencing.

 

[cootertwo]

Interesting subject. I know how using a GPS mode will sometimes cause wandering, when not in a clean open space, as sats come in, and drop out. And I have some mounted very close to the flight controller, and some on a mast. But this little F330/350? I'm playing with seems like it wants to lean in the direction it wants to go, rather than move in a flat horizontal motion, like some of my other rigs. I was just wondering if the longer the mast, the more it would tend to tilt, instead of a flat drift. Imagine if the GPS was on say a 6 foot mast, and all at once it found different sats, and wanted to be 6 or 7 feet from where it is, would it cause the craft to lay over to get there? or not? Like, is there a "leverage" factor?

 

[JoeBob]

GPS is accurate only to about 1 meter. Your UAV is going to have some wander within that range. (I've always wondered why my RTL landing position is always 1 meter from the takeoff position. Multiple flights from the same location always seem to return to the same landing location one meter left.)
http://www.gps.gov/systems/gps/performance/accuracy/

The mast is used to distance the GPS antenna from electronic interference (ESCs, VTx, motors) and the magnetic field generated from current running through wires. One of the APM developers suggested 10cm as a good distance.

 

[cootertwo]

Thanks Old Man (like me), and JoeBob. It was just something I was wondering about. I think all my test rig needs, is some loving attention from Mission Planner. Found some entries about setting the lean angle in Loiter mode.

 

[cootertwo]

Wow, never noticed this before, just hooked up to Mission Planner, and with only my ceiling fan running, my barometer was reading 95 !!!
Turned fan off, and it went to 2. Guess I need to cover my controller with some foam!

 

[fltundra]

Wow, never noticed this before, just hooked up to Mission Planner, and with only my ceiling fan running, my barometer was reading 95 !!!
Turned fan off, and it went to 2. Guess I need to cover my controller with some foam!

I'm surprised they don't have the barometer better isolated from the factory.

 

[Old Man]

Earlier 3DR APM baro units were not insulated but they referenced this in their Wiki and provided an instruction to add a small piece of foam over it.

 

[JoeBob]

Current production Pixhawks already have a small piece of porous foam on top of the baro inside of the case. I notice a small improvement in AltHold by covering the unused openings in the case. (Anything without a cable plugged in gets a cover of blue painter's tape)
View attachment 26255
Sudden stops from fast forward flight create an artificial drop in pressure (no more wind in your hair) which fools the Pixhawk baro into thinking that it is higher than it is, so it drops. Sonar and Lidar will give you more accurate AltHold, but working them into the code is an on-gong process.

 

[violetwolf]

The baros are also sensitive to light, so dark foam if you have it.

Good idea with the tape

 

[Riche]

I still have a lot to learn about drones, but I have some GPS experience.
And field of view has everything to do with an accurate location as well as the limitation of the hardware and how the signal is processed.
So I would totally expect GPS coordinates to fluctuate with your proximity to obstructions. Not to mention if you do not have a lot of satellites in your field of view there is another issue called "bounce" were you have a signal literally bouncing of and object.
Just to give you and idea of how to establish a decent sub centimeter location, a minimum of a 3 hour satellite observation has to be processed and compared to other know base stations. There is a quick way that we get sub centimeter locations is with a system called RTK it allows for a GPS rover to be compared to a GPS Base in real time. It is still subject to all the above issues.
FWIW, I still have a lot to learn just thought I might be able to contribute a little.
I am going to be really interested in GPS accuracy and LIDAR as my knowledge of multirotors grows.

 

[Old Man]

You are one of the few that understand the importance of real time kinematics. Thanks for the post.