| Author |
 Topic  |
|
|
DJackson
United Kingdom
20 Posts |
 Posted - 17/05/2012 : 10:30:05
|
Hello,
I have been using MM for a while, but have recently noticed that position readings of waymarks displayed on my Garmin are plotted at at different locations by MM.
I have a Garmin60CSx. It is set to
position format: British Grid, map datum: Ordnance Survey GM. Distance: metric units.
I have set the default MM projection as follows using info given in a previous post:
In MM go: File -> System set up -> Edit default projection.
Click on the 'Projection system' tab and select 'UK Ordnance Survey National Grid OSG'. From the datum tab select 'ORD SERV OF GREAT BRITAIN 36: Scotland, Shetland Islands'. Then click 'OK'
I also created the same projection using:
Utilities -> Latitude and Longitude utilities -> Create projection. Give the projection a name (e.g. "OS Scotland"), click the "Projection system" tab and select "UK Ordnance Survey National Grid OSG". Then click the "Datum" tab and select "ORD SURV OF GREAT BRITAIN 36: Scotland, Shetland Isles". Then click OK.
When uploading waymarks to MM, using the above projection, there is a discrepancy between the grid references of waymarks displayed on the Garmin and the position of these waymarks when uploaded to MM. I am using a base map of OS map tiles NG61SW.tfw & NG51SE.tfw.
For example: a waymark showing as NG60669 BNG11670 on the Garmin display, when uploaded to MM is positioned at location 160664 m 811681m (or in lat/long 57o8’0.504N 5o57’24.455W), according to MM's bar in the bottom left hand corner showing the geographical coordinates.
There seems to be a consistent discrepancy such that MM plots the waymarks about 4-5 m further west and 9-11 m further north than the reading on the Garmin. Or, put the other way, if I were to manually plot points in MM using the Garmin display, they would be approx. 5 m further east and approx. 10 m further south than where MM puts them.
Is there something wrong with my Garmin, or are my settings incorrect?
If anyone can help with this, I will be so grateful.
Dorothy
|
|
|
Irvine
United Kingdom
51 Posts |
Posted - 18/05/2012 : 09:36:30
|
Dorothy
Have you tried marking a waypoint on a fixed point you can identify on an OS map to check the accuracy of the plot in MapMaker?
I looked at some waypoints I took last week with My Garmin 60 and the position plotted in Mapmaker is not the same as the OS coordinates shown by the Garmin. However as one of the waypoints is on a feature marked on the OS map I can tell it is only 2m away from the true position.
It might be the case that MapMaker is doing some clever adjustment during the download process to adjust for different map projections.
Irvine |
 |
|
|
mapmaker
470 Posts |
Posted - 18/05/2012 : 15:53:04
|
The conversion from lat/long to UK National Grid is not a straightforward formula. It involves a complex set of adjustments which are different for different parts of the country. Regular hand held GPS devices do not apply these corrections, thus setting a GPS device to display National Grid coordinates will always result in an approximation.
Internally the coordinates in the device are always in lat/long and that is how they are sent to Map Maker, which then converts them to National Grid coordinates, applying the corrections.
If you go to the Ordnance Survey coordinate converter at http://gps.ordnancesurvey.co.uk/etrs89geo_natgrid.asp and enter your coordinates 57o8’0.504N 5o57’24.455W and convert them to the National Grid the result is 160663.761, 811681.230, which agrees with Map Maker.
|
|
|
|
DJackson
United Kingdom
20 Posts |
Posted - 19/05/2012 : 16:31:06
|
Thanks very much for the replies. It looks like working in Lat/long would be more accurate generally?
What I want to do is transfer a line that I have drawn on an MM map, at 150m radius from a fixed point, to the actual terrain. Could anyone advise me on the most accurate way of reading off positions on this line on the MM map and then using a Garmin GPU to locate those positions on the ground, where they can be marked with stakes?
Thanks for any suggestions. |
|
|
|
Neil MacKenzie
United Kingdom
5 Posts |
Posted - 20/05/2012 : 11:38:12
|
It would seem to me that you are perhaps compounding the error margin by creating the points in Mapmaker, downloading to Garmin, marking points on ground, then uploading same points back to Mapmaker. This last move may contribute to the error. If you had not done the upload exercise would you have known about the error? Can you check the accuracy of the point location created in Mapmaker on ground using another method – say the old fashioned way of compass bearing and meter tape via triangulation from known locations on OS map?
Neil
|
|
|
|
mapmaker
470 Posts |
Posted - 21/05/2012 : 10:58:29
|
The most recent Garmin units can read GPX files. In Map Maker under "Utilities - Latitude and Longitude utilities - XY file to WGS84 lat/long - vector" you can convert your DRA file to a GPX file and then simply copy it to the Garmin and use this data as waypoints on the ground.
The garmin GPS60CSx does not read GPX files. To upload points to this and other devices, first export them from Map Maker as Shape file (*.shp) in lat/long coordinates. The use the free GPSU utility program (www.gpsu.co.uk) to upload the data. |
|
|
|
DJackson
United Kingdom
20 Posts |
Posted - 30/05/2012 : 20:52:00
|
Thanks for the replies.
Following Irvine's suggestion I waypointed 4 posts surveyed by a topographic survey company to within 1m. The points, when recorded by my Garmin60csx and uploaded to MM, were plotted 6-8 m NW of the reference points.
The GPU accuracy was +/- 3 m.
I waypointed the posts twice.
1st: with GPU set to lat/long (decimal degrees) and converting data with projection WGS84 global/UK Ordnance system national grid OS
2nd: with GPU set to OS grid, and converting with projection ORD SURV GREAT BRITAIN 36: SCOTLAND, SHETLAND IS/UK Ordnance system national grid OS.
The second set is 1-3 m closer to the reference points, but as I took them about 10 minutes later, I imagine this could be due to external factors eg changes in satellite position?
However, would a 6-8 m discrepancy in a consistent direction be attributable to built-in GPU variance?
Doing this process in reverse i.e. transferring coordinates from a line drawn in MM on an OS map to my GPU (by manually programming waypoints in lat/long into the GPU)and finding the points on the ground gives a similar discrepancy of 7-9 m i.e. points on the ground are 7-9 m SE-ish from where they appear to be on the OS map.
Is there a problem with the Garmin readings, the conversion to and from MM, the OS map registration or me?
It is probably the latter, but what am I doing wrong?
|
|
|
| |
Topic |
|
|
|
GPS readings and MM plots differ
Forum Locked
