Hi,

I am new to this space and looking at doing some co-ordinate conversions from
British National Grid to (GPS) WGS84 in software.

I am aware of OSTN02 and OSGM02 and the implemention available from the ordnance
survey but I need to do something more compact and not necessarily aas accurate
(these algorithms involve large lookup tables).

The process I have implemented (based on "A guide to co-ordinate systems in the
UK") is:-
BNG Eastings & Northings -> OSGB36 latitude longitude:
Using an inverse Transverse Mercator based on Airy1830 elipsoid.

Then OSGB36 latitude, longitude, height to WGS84 latitude, longitude, height:
Using a Helemert transform based on Airy1830 and WGS84(GRS80) ellipsoids.

The missing piece of the puzzle for me is how to derive the height value for the
OSGB36 long, lat, height co-ordinate (where theOSGB36 height is the height
relative to the Airy1830 ellipsoid) from the BNG Easting&Northing height which
is based to the ODN (Ordnance Datum Newlyn).

In generating the OSGB36 values (via Helmert transform) for the followig
WGS84/BNG ODN co-ordinate pair listed on the ordnance survey site
(http://gps.ordnancesurvey.co.uk/viewpassiverec1.asp?recnumber=C1SO3639) it
looked to me like the Airy1830 surface might closely conincide (within 2m) with
ODN.

Datum
BNG ODN 336210.713 m 239341.107 m 252.145 m
OSGB36 52° 2' 54.5770'' N 2° 55' 48.3459'' W 253.744 m
WGS84 52° 2' 55.6120'' N 2° 55' 53.7540'' W 303.830 m

If there is only ~2m difference between Airy1830 Ellipsoid surface and ODN Geoid
surface and this is pretty consisitent across the extent of the BNG then I feel
I could safely use the ODN height value as the OSGB36 height value feeding into
the Helmert transform without to much loss of accuracy.
If there is a more eccentric relationship then obviously I need another
approach.
Does anyone out there know the nature of the relationship between the Airy1830
Ellipsoid and the ODN Geoid?

Best Regards

Craig