RCMF Donations

Enjoy using RCMF? How about a wee donation to help us keep you in the style to which you've become accustomed?

Welcome to RCMF. Please login or sign up.

December 03, 2020, 17:30:20 pm

Login with username, password and session length

GPS readings

Started by JimG, February 07, 2018, 19:15:44 pm

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

JimG

Do any of you have experience with the GPS telemetry module. I am using the Horus with a Tundra for testing and have the GPS module fitted. I have one of the switches set to give me an audio readout of the GPS speed and the GPS altitude. While flying last Sunday I found the accuracy of the altitude highly variable, one flight gave an altitude of 400 feet when it was flying between 50 to 100 feet.
I have the Tx set to record a log for the flight while the timer is switched on so could check up what the readings were during the flight. The recorded readings were what I would have expected for the flight althought the speed  did have a number of readings that were much too low. The main fault I found was the GPS coordinates recorded. I was able to convert these to a suitable format and imported them into Google Earth. According to them the plane was flying around 20 miles to the south east from the flying site and was around 5 to 10 miles from land over the North Sea. Surely the GPS cannot be that inaccurate or is the Tx reading the telemetry wrongly?

Jim
Jim Gill
Dundee Model Aircraft Club

Brian Cooper

LOL... it sounds like it has something in common with my old TomTom sat nav.  That thing would take you to places you didn't want to go to by some very strange routes indeed. Mind you, you saw lots of the UK which you wouldn't have seen otherwise..  :D

B.C.

Pup Cam

It's a consequence of how GPS works.    Altitude readings are never very accurate and are particularly susceptible to varying errors - it's all down to geometry if I recall.   If you want to know your altitude, use an altimeter .....

Alan
Still distracted by a 1953 AJS 16MS and now a 1939 BSA 250 too!

Phil_G

February 07, 2018, 21:21:01 pm #3 Last Edit: February 07, 2018, 21:38:26 pm by Phil_G
Just a thought, once the GPS module is moving quickly its harder for it to get a fix, so allow it enough time on the ground to get a solid fix before launching.  Some modules dont have a 'fix' output as such, its derived from a minimum satellite count, which could be as little as 3 sats.  Launch a fast model with only 3 sats in view and it may not pick up any more.  As Alan says, GPS isnt accurate on altitude, its basically 2D with 3 sats, and more sats are needed to get altitude.  The shape of the earth isnt perfectly uniform and the GPS has an approximate geoid model of that shape, but its not perfect so altitude can never be perfectly accurate.
GPS is not at all accurate for instantaneous speed either, it can measure speed over distance very accurately but point to point does not give anything like true speed.  If you look at any GPS trace you'll see it cutting corners between samples, and often a sample is missed. Sometimes on a track day we've seen as few as 5 or 6 samples round a lap of Donington - join the dots and you get a very much shorter course than a true lap so the GPS displays very low speeds.  Some GPS modules run 10 samples per second, but just one sample per second is typical.  In one second you could have done entry, pylon-turn & exit, and your GPS module never even saw it  ;)
Cheers
Phil

JimG

I understand the delays in measuring speed however the flying was done with the Hobbyking Tundra which is not a fast model. My biggest concern is the positioning error, 20 miles out is a large error and was consistant over several flights. Unfortunately the telemetry doesn't tell you how many satellites the receiver sees. I did check on one flight that the GPS module was picking up the signals before the flight.

Jim
Jim Gill
Dundee Model Aircraft Club

dickw

I have also sometimes seen these strange "miles away" positioning errors with GPS - but my systems are not FrSky.
As Phil suggested, I have found that a GPS module can sometimes need to be left switched on for up to 5 minutes before first launch of the day to get a fix. It is usually quicker after that. It does seem to depend on where you are and the conditions that day.

Can the GPS module "see the sky" OK or are there wires/batterys/other electronics in the way?

Dick
Grow old disgracefully

dart16

Quote from: JimG on February 07, 2018, 19:15:44 pm
. The main fault I found was the GPS coordinates recorded. I was able to convert these to a suitable format and imported them into Google Earth. According to them the plane was flying around 20 miles to the south east from the flying site and was around 5 to 10 miles from land over the North Sea. Surely the GPS cannot be that inaccurate or is the Tx reading the telemetry wrongly?

Jim


Sounds like you have two different datums being used here.  Google earth uses WGS84  what datum does your GPS use and how did you 'convert' the readings?

FrankS

I've been using the aerobtec altis and SM Modelbau GPS logger and have found them quite accurate (they both have bariometric altimeters for height and vario), the Aerobtec does take longer to get a GPS fix, my friend also has the SM Modebau one too, and on downloading his data and converting to google earth I was able to see exactly where he crashed into a fence and the terminal speed  :''

JimG

Quote from: dart16 on February 08, 2018, 12:07:47 pm
Sounds like you have two different datums being used here.  Google earth uses WGS84  what datum does your GPS use and how did you 'convert' the readings?

The log readings are of the format 246.1868W  5629.0413N
Using Excel I first removed the trailing letters to give a straight number then divided them by 100 to give decimal degrees  2.461868   and  56.290431. The longitude was then made negative -2.461868 to fit being west of the zero longitude.
Actual values for our field from Google Earth are -2.770412 and 56.484366.
As far as datum surely it would use the same zero points as standard.

Jim
Jim Gill
Dundee Model Aircraft Club

Kevin Fairgrieve (NSS)

I find the GPS useful for plotting flights for a bit of fun.



Kev

Patmac

February 08, 2018, 23:45:48 pm #10 Last Edit: February 09, 2018, 12:34:26 pm by Patmac
Quote from: JimG on February 08, 2018, 19:47:56 pm
The log readings are of the format 246.1868W  5629.0413N
Using Excel I first removed the trailing letters to give a straight number then divided them by 100 to give decimal degrees  2.461868   and  56.290431. The longitude was then made negative -2.461868 to fit being west of the zero longitude.
Actual values for our field from Google Earth are -2.770412 and 56.484366.
As far as datum surely it would use the same zero points as standard.

Jim

I was thinking degrees & minutes so I moved the decimal place two positions left then divided the 6 digit to the right by 60 giving -2.769780 & 56.484051 which is only about 55 metres from your field's values. Seems strange because although it gives reasonably accurate result it clearly isn't a correct conversion to degrees & minutes.   $%&

Edit - The above was subject to a late night brain fart. Realised this morning the log reading format is basicaly degrees & minutes if the decimal point is ignored. The section that I multiplied by 60 converted it to degrees.
Pax vobiscum

PDR

Not sure about the standard Horus software, but in OpenTx the GPS data format is a setting you can choose in the General Settings section.

PDR
There are no shortcuts on the long, hard road to success. But if your dad's rich there could a limo service...

Steve J

Quote from: JimG on February 08, 2018, 19:47:56 pm
The log readings are of the format 246.1868W  5629.0413N


The format is probably NMEA GGA, so that is 2°46.1868' W  56°29.0413' N which appears to be the side of a track next to a flying club.

Steve

paulinfrance

I have the same problem with my Futaba module, whatever setting I use with it and my Tomtom it is at least 100 metres out,,,  ( Ground / Ground )
Mode 2 THE only way to fly

JimG

Quote from: Steve J on February 09, 2018, 23:17:42 pm
The format is probably NMEA GGA, so that is 2°46.1868' W  56°29.0413' N which appears to be the side of a track next to a flying club.

Steve


Thanks Steve, this gives a much better result. Only around 80 meters off so an acceptable accuracy. Now to see if I can convert the readings to this format in Excel.

Jim
Jim Gill
Dundee Model Aircraft Club

Steve J

Quote from: JimG on February 10, 2018, 11:19:43 am
Only around 80 meters off so an acceptable accuracy.


Acceptable accuracy would be 1-2m horizontally and a bit worse vertically.

Make sure that your GPS is the right way up and it has a decent view of the sky.

I find that GPS altitude is generally OK, but can go AWOL from time to time. The following graph has the altitudes from a Spektrum altimeter and a homebrew GPS that uses a £7 5Hz module in a reasonably quick model.

[attach=1]

Steve

Steve J

Quote from: Steve J on February 09, 2018, 23:17:42 pm
The format is probably NMEA GGA, so that is 2°46.1868' W  56°29.0413' N


2°46'18.68" W  56°29'04.13" N looks better so maybe the format isn't raw NMEA.

Steve