by Phil Brady
Why have these notes been written?
The Ordnance Survey of Britain
The OS Cartographic System
Map References
Which North?
How do I set my GPS receiver?
Where can I get a list of Waypoints in Britain?
Conversion Code (Excel)
Other Implementations
Disclaimer
Why have these notes been written?
There are frequently articles in Usenet news group asking about the Ordnance
Survey national grid in
Questions like :
These notes and the accompanying software are an attempt to
give background information which will either answer those questions directly
or show an alternative way of working which is more helpful in
They have been compiled as a special thanks to all those newsgroup contributors out there who have been so helpful to me in the past.
The Ordnance Survey of
The scales published include 4cm/25Km for route planning, 2cm/Km for motoring or walking and 4cm/Km for walking. More detailed maps for surveying and land ownership matters are also available. The OS web site is http://www.ordsvy.gov.uk and has lots of interesting information. The email address for customer service is custinf@ordsurv.gov.uk.
Road atlases published by the AA (Automobile Association) are also based on the Ordnance Survey grid, are widely available and can often be found at discount prices so shop around. At one time they included the 100Km square identifying letters both in the index and on the maps but this was discontinued in the mid 1990s when the letters were totally omitted. I am pleased to note that the year 2000 edition (on sale from Sept 1999) has reintroduced the letters on the pages but not in the index. Most motoring atlases are based on the OS grid, but they vary in the degree of referencing information they give.
You may also find Bartholomew maps for walkers. These have essentially the same co-ordinate system as OS but a modified reference system. They have references which omit the two letter square of the OS system but give eastings and northings in metres from the OS false origin. See http://www.bartholomewmaps.com
The visitor to
The Ordnance Survey of Great Britain is a Transverse Mercator
projection. This is a particularly ideal choice for a tall thin area like
The OS projection is based on a true origin at 49 degrees North, 2 degrees
West which is near the
From this origin, a square grid is imposed on top of the
‘converging’ lines of longitude and the ‘curved’ lines
of latitude. Points are located by measuring distances in metres east and
north from this true origin. Since this process will introduce distortion
as you move away from the central longitude of 2 degrees west, a scale factor
of 0.9996012717 is introduced which reduces this distortion - two lines of
longitude with zero distortion are thus created. Incidentally, this is
very slightly different to the factor of 0.9996 used for UTM.
Furthermore, to eliminate the inconvenience of negative values of easting, the
concept of a false origin is introduced. A value of 400000 (400 Km) is
added to eastings and 100000 (100 Km) subtracted from northings so placing this
false origin near the Isles of Scilly off Lands End in the south
west. The model uses the Airy spheroid, OSGB of 1936. Readers
wishing to delve very deeply into the topic of map projections are referred to
a publication by a former colleague at the University in
Coordinate Systems and Map Projections, D H Maling, Pergamon Press, ISBN 0-08-037234-1.
See also the OS web site http://www.ordsvy.gov.uk
A full reference which is unique across
The first letter gives the 500Km by 500Km square. 25 different
letters (omitting I) could be given here though only five actually map to
the covered land areas: S for south west, T for south east, N for the north and
ABCDE
FGHJK
LMNOP
QRSTU
VWXYZ
The digits should be split into two groups - in the above example of SS 599 939 the point is 59.9 Km east and 93.9 Km north of the south west corner of square SS.
The measurement of a reference is made easier by the regular grid of squares representing 1Km (or 10Km for motoring maps) on the ground and printed on every map. The references of those grid lines are shown on the margins of the map and, on some, at intervals across the map. To determine a reference on a map with a 1Km grid:
The resulting reference SS 599939 resolves to 100 metres and the procedure is described on the margin of every OS map. If you wish you can use a centimetre ruler (there is one on my compass) to help you achieve more accuracy or you could invest a pound and buy a plastic scale called a romer. Be warned though that a romer is transparent and easily lost - mine vanished a week after I purchased it, re-surfaced briefly after a year and has not been seen since! Note that dividers or special scales are not needed as I understand they are with some US maps.
I can recommend an inexpensive index to the maps which is FREE in good
bookshops and which describes the grid and the coverage of each map. Note
there was a misprint on the back of the 1997 edition -
The margins of the maps also have 'ticks' and the body of the map has blue
crosses to show latitude and longitude. Note that these are relative to
the OS 1936 datum and not WGS84 which is becoming increasingly standard across
the world. Also note that the lines of latitude and longitude are both
curved and are not parallel to the printed OS grid.
You will find references to three different norths in the margin of OS maps. True north is just that - a direct line to the North Pole and the Earth's spin axis. Grid north is as defined by the grid lines on the map which cannot match true north because the map is a flat representation of a curved surface. Magnetic North is the direction indicated by a magnetic compass; it does not coincide with true or grid north and wanders slowly from year to year.
The north-south grid lines point west of true north for areas west of the
true origin at 2 degrees west, and east for those to the east but this
difference is only about 0.75 degrees per degree of longitude. For
practical purposes though true north is not very relevant and we are only
interested in grid and magnetic north, the difference between which varies
uniformly across the country with only small anomalies. However, magnetic storms, often coinciding
with the 11 year solar sunspot cycle, can produce large short lived variations
(eg an 8˚ variation in less
than an hour at Lerwick was recorded in 1989).
The variation is small compared with that in many parts of the world but
not quite small enough to be ignored - it was 4 degrees west of grid in Wales
in 1996, decreasing by a degree every 8 years. I hope to be still walking
when it's zero! Whitaker's Almanack documents this variation in
This difference between magnetic and grid bearings is called magnetic
variation in
To add to the confusion, declination is used to describe the latitude of
stars universally (forgive the pun!) and, in
Magnetic Variation - Compass
adjustment made easy.
Every description I have seen describing how to use a compass encourages the
user to adjust his/her compass by adding or subtracting the magnetic variation
after setting it from a map or after taking a bearing. They come with helpful ditties to remember
whether to add or subtract, but which may be invalid in other parts of the
world. It's fine in a nice warm room and
a steady table to rest on, but have you ever tried this in a howling
blizzard? You cannot do it with gloves
on, but you are afraid to remove them because you know won't get them back on
over wet hands. You cannot see the fine
gradations on the scale because of the hail in your eyes (or you need your
spectacles), you've forgotten whether to add or subtract, and you are shivering
too much for careful adjustments anyway.
Oh the joys of hill walking in the
The method I came up with is very simple and utterly reliable. Note from your map where magnetic north is relative to grid north. Now note where that is on the bezel around your compass needle capsule. It might be 3 degrees west and that corresponds to the bottom left of the N. You might be in far flung corners of the globe where it is 23 degrees east and that might correspond with a nick on the bezel to the right of north where you last fell on it. If there isn't an obvious mark then make one temporarily with a pencil or sticky paper until you trust the method enough to put a more permanent mark with nail varnish or whatever. There are then just use two VERY SIMPLE rules:
1. When you have your compass placed on the map, set it according to the parallel north/south lines in the capsule, as usual, but do not adjust for variation.
2. When you are turning round or turning the compass to get the red end of the needle to point to the north mark, get it to point to your mark instead.
The method works whether you are setting your compass from the map in order to follow a bearing, or are taking a back bearing on a distant church to find where you are on the map. It also avoids any ambiguity between rough bearings ("which road?") and more precise ones ("is that faint track the one I want?") in that both are the same. It's a method that's been argued over and I've been told that it can't be right for a variety of reasons – the least compelling of which was that "the military don't use it", but it is effective. Give it a try and you'll never look back. From now on, only mugs will remember ditties like "Magnetic unto grid, subtract - MUGS" and continue applying minute adjustments to the dial in foul weather.
If you think about this principle of never adding/subtracting from the dial you'll also get new perspectives on activities like taking a bearing on the point you've recently left (point white end of needle at the mark instead of deducting/adding 180) or 'boxing' round an obstacle (use E and W marks instead of adding/subtracting 90).
How do I set my GPS receiver?
If contemplating use of a global positioning system (GPS) receiver in
Garmin receivers
In the case of the Garmin 38, 45, 12, 12XL , eTrex family you should set 'British Grid' for format on the navigation setup page. Datum should be set to 'Ord Srvy GB' to avoid errors of up to hundreds of metres. Note that the Garmin holds references in some unpublished internal format but applies appropriate conversions to keyboard input and screen output to reflect currently set format and datum. I choose to set my GPS for metric measurements rather than statute for walking since the maps have a 1Km grid, but tend to use statute for motoring to match the odometer in the car and the road signs, both of which use miles! 'Auto magnetic' for north gives good results and complements my trusty Silva compass, and that was how I set my GPS when I first used it. More recently though I have found it much more convenient to set the GPS to 'Grid' and I then have consistency between GPS, map and the protractor scale on the compass using the method above.
The OS reference is given by a Garmin with a resolution of one metre (but not to that accuracy!) - eg.
SJ 12345
BNG 67890
This example shows we are in the 100Km square SJ some 12345 metres east of the square origin and 67890 metres north. The BNG refers (presumably) to British National Grid. When inputting the grid reference of a waypoint, note that the receiver requires the two letters and five digits of eastings on one line, and five digits of northings on a separate line.
Magellan Blazer 12
Paul Slootweg has kindly sent information on the Blazer 12 GPS system:
"This GPS received includes support for the OSGB grid system. To
configure, simply use the 'menu' button and select the 'Setup' menu.
It will then offer the option to select 'COORD SYS' (coordinate system) and
choose the OSGB option. Then select the 'MAP DATUM' and
choose GRB36 (OSGB)
If you pick OSGB it won't allow any other datum than GRB36, which avoids the
possibility of positional errors. It imposes similar restrictions with
some of the other grid systems to ensure the positional data is correct.
The unit also allows the selection of magnetic or true north.
The display will show OS references
SJ 1234
6789
consisting of 2 letters and 2 groups of 4 digits - i.e. 10m resolution,
which seems more than enough for my use."
Paul believes that the same settings will apply to other Magellan systems.
Where can I get a list of Waypoints in
The motoring maps published by the Ordnance Survey office give the full reference and are slightly dearer than other brands but still good value.
Christian Schmadalla has set up a web site with zip files holding waypoints
of motorway junctions and major cities in the
Another useful site was http://www.ITatWork.freeserve.co.uk/waypoints.htm where John Seymour had a listing of over 5000 waypoints in Britain including places to eat, motorway services and junctions, car parks, petrol stations, munros, etc., and which were suitable for upload to a GPS receiver. Sadly they no longer seem available.
There is excellent information on walks in the
http://www.lakedistrictwalks.com/
Phil Newby has a list of GPS waypoints for
http://www.36haroldstreet.freeserve.co.uk/GPS.htm
Satellite Navigation systems for cars also, of course, have waypoints but do not in general give OS references.
With an OS map and a GPS system which will display OS references there should be no need for conversion of OS reference to latitude and longitude or vice-versa - simply 'think OS' - the references are very easy to read off the map scales. I can think of only two instances when one might need the conversion:
Conversion routines are available as functions in Microsoft Excel (tm) spreadsheet format These perform conversion between longitude/latitude in OS 1936 or WGS84 datum, OS references and eastings and northings as well as giving additional related functions such as computing bearings and distances. Full examples are given, including a worked example of a distance/bearing calculator.
With Excel in the office XP suite you will need to lower security levels to 'medium' before loading this.
This code was based on an original version written in Quickbasic which is essentially obsolete now.
Accuracy of these routines is well within that of a consumer quality GPS, even without selective availability, well within one's ability to read a 1:25000 map, and probably within the limits of stability of paper maps. I have found that they agree, invariably to the exact metre, with the conversions performed by my old Garmin 38 GPS, and Garmin confirm that the accuracy of their conversions is about 1 metre. This meets my recreational needs. They have not been designed, written or validated for professional work, though they are in use by a borough in Kent for plotting their drains, have been used for microlight championships, are recommended by the British Atmospheric Data Centre, and at the time of writing are about to be used by those involved in coordinating rescue services! I have also received an unsolicited email from Ordnance Survey describing it as a fine piece of work.
Implementations of the code for other platforms are referenced below.
The latitude/longitude to OS reference conversion routines use the algorithms published by a former colleague at Swansea University, Derek Maling, though I must claim credit for any errors in implementation.
Other Implementations and Sites
I frequently receive requests for this code written in C or C++. Russ Hore has done some work in converting the code and might be able to help if you email him. Chuck Gantz has also produced some C code which we are hosting for him in a zip file but I have not examined this. Please email any comments directly to Chuck.
Russ has also provided an online web convertor which does latitude and longitude to OS reference at http://www.russ-hore.co.uk/
David Harper has lots of interesting astronomical data at http://www.obliquity.com including a web based OS reference to latitude and longitude converter at http://www.obliquity.com/astro/osgb.html and a distance and azimuth calculator at http://www.obliquity.com/astro/distance.html
Mark Crossley has kindly made an OSGB/Lat,Lon (GRB36) converter for the Psion available here.
http://www.streetmap.co.uk/ does
conversions of
Frederik Ramm has written a Java version at: http://www.remote.org/frederik/projects/software/
If you have an interest in OS references and microlight flying then try the site by Richard Meredith-Hardy http://www.flymicro.com/records
The information and the program code given are believed to be accurate and the opinions are my own, though no warranties or guarantees are given. Liability will not be accepted for any way in which this information or the code is utilised.
Please report any inaccuracies, alternative views or suggested improvements to this document or the code - I may be reached via email but no spam please to me or any other contributors.
phil dot brady at hotmail dot co dot (well, you know the short form for
My thanks go to the many contributors who have helped improve this document and provided useful additional information.
Phil Brady
Updated 14 March 2008.