OSTN02 or OSTN15 Grid Shift - UK Ordnance Survey

Any questions regarding the practice version are asked and answered here.
Post Reply
Luke Benjamin Cummins
Posts: 2
Joined: Tue Oct 18, 2016 3:21 am

OSTN02 or OSTN15 Grid Shift - UK Ordnance Survey

Post by Luke Benjamin Cummins »

In order to go between British National Grid and London Survey Grid a OSTN02 or OSTN15 grid shift file is used.

Documentation here:
https://www.ordnancesurvey.co.uk/busine ... opers.html

Has anyone looked at doing this?

Cheers
Luke
Matthew Monk
Posts: 3175
Joined: Fri Oct 21, 2005 2:46 pm

Post by Matthew Monk »

When I was 12d Support for the UK, I had previously used the OSTN02 grid shift file with 12d's general transformation option (Survey->Conversions->General transformation). This panel can perform transformations using 7 parameters or a grid shift file (*.gsb) in the NTv2 format.

For OSTN02, there was a NTv2 GSB made available- https://www.ordnancesurvey.co.uk/busine ... ormat.html

If an NTv2 GSB file was available, then 12d Model could use that (the NTv2 is more compatible and useful for most software). I can't speak to the ease of integrating the available library into 12d Model.
Luke Benjamin Cummins
Posts: 2
Joined: Tue Oct 18, 2016 3:21 am

Post by Luke Benjamin Cummins »

Thanks Matthew.

I have managed to get this far with a Grid Shift from British National Grid (BNG) to ETRS89 using the grid shift file (NTv2 available from https://www.ordnancesurvey.co.uk/busine ... opers.html )as you suggested). It appears that points line up but this has to rely on a conversion from Northings and Eastings to Long/Lat (BNG to ETRS89) and precision appears to be limited.

I then need to undertake a 7 parameter transformation from ETRS89 to get to London Survey Grid (LSG). I have to do this in Lat/Long. It splits the data set in an odd manner (Greenwich 0 Long and sign convention) but it appears inconsequential.

Conversion Name ETRS89 > LSG
dX 19.019m
dY 115.122m
dZ -97.287m
rX (arc seconds) -3.577824
rY (arc seconds) 3.484437
rZ (arc seconds) 2.767646
Scale 18.60847540

Projection name: London Survey Grid
Projection type: Transverse Mercator
rigin latitude (dms): N051-10-00
rigin longitude (dms): W000-09-30
False easting: 78,250.000m
False northing: -2,800.000m
Scale factor: 0.9999999
Ellipsoid: WGS84

I then need to convert these again to get to Northings and Eastings. This places the points between 300-40m out.

I have tried every combination I can think of with parameter sign convention, lat/long - easting northing and projection changes earlier or later in the process but to no avail. I am unsure if the switch to Long/Lat is removing the precision required or I am doing something else wrong?
Roger Walker
Posts: 33
Joined: Thu Nov 03, 2016 5:43 pm

Post by Roger Walker »

So you want to convert from OS grid to London Grid? And you don't want to use the approx (400mm accuracy method) in the linked document? (section 7.3) http://www.engineeringsurveyor.com/soft ... apping.pdf

So you are trying to convert OS grid coordinates to lat/long and then back to London Grid?

Don't know enough about 12D to say why your attempt isn;t working but you could use ordnance survey "grid inquest" to check the conversion to lat/long. Then check the projection to LSG using a spreadsheet like this one
https://www.ordnancesurvey.co.uk/busine ... sheet.html

300m sounds like the shift grid file is not being used. Just a guess.
Post Reply