Wgs84 To Os Co-ords ?

January 31, 2005

greetings,

as a newbie to gps etc i could do with a nudge in the right direction as regard converting wgs 84 co-ords to british grid os co-ords...ie N 51° 45.487 W 000° 25.886 equates to British Grid: TL 08357 07771 how do i convert the north & west co-ords to the british grid co-ords ??????????????.i have the n & w co-ords i want to convert but have no idea how its done.any help would be great

bright blesings...third-degree-witch

January 31, 2005

I use Grid InQuest available here ( requires free registration ) to convert to and from. Excellent app recommended by the OS .

1.4k · January 31, 2005

Your GPSr can do it for you, but not very well.

For complicated reasons which I'll not go into here and now, you will get an error of about 6 metres if you use your GPSr to do the conversion.

A much better way is to let the government do it for you.

Go to The Ordnance Survey's GPS website and use their Co-ordinate Converter. It's on the clickable topline toolbar at that website. I think you may have to register (I've been using that site for so long that I can't remember whether I had to sign in the first time), but it's free and there are no hassles or advertising involved.

You can convert co-ordinates from Lat/Long to Ordnance Survey grid Eastings and Northings or vice-versa there and the accuracy is superb.

There's also a downloadable program which you can run on your PC, but that's probably not worth bothering with for most geocachers.

Cheers, The Forester

January 31, 2005

Go to The Ordnance Survey's GPS website and use their Co-ordinate Converter

I've just been on the above page and can't get it to work for me. Maybe I fick but , I tried to convert N 55 21 537 W 001 53 221 it's for a new cache, so I already know the OS grid Coords, got them off new cache page from GC.Com

I used Degrees and decimal minutes ( That's right isn't it ) and it didn't come out the same as on my new cache page.

All help gratefully received.

Kevin

January 31, 2005

The OS website is the most accurate (sub-cm I believe), but is fiddly to use and will not provide you with grid letters (ie you'll get 470267,122765, not SU 70267 22765).

For a quick one-off conversion, there are many websites which will do this with varying degrees of accuracy. <plug>Mine is

here and falls into the "medium accuracy" band (a couple of m either way - same as your GPSr)</plug>

Alternatively, for a friendlier system for converting batches of coordinates, check out Waypoint Workbench, which has all sorts of features to help cachers. It too falls into the "medium accuracy - good enough for geocaching" category.

January 31, 2005

Many thanks Teasel, tried yours and it seems to be the easest to use.

Kevin

January 31, 2005

HI T-D-W,

Why do you want to do the conversion, knowing the area you are in I can think of a couple of caches that are listed in OSGB but have had a few problems.

Nick

February 1, 2005

greetings,

thanks for all the helpful replies,im trying them all in turn <_<

nick...ive just done a 15 waypointer cache and wanted to plot the waypoints on a map,i prefer the british grid over the lat/long system....saying that the whole cache is done using the lat/long system.i can get my head around the british grid but the lat/long thing goes over my head :rolleyes:

February 1, 2005

Convert OS grid to WSG84

This works perfectly doing it this way but his application to do WSG84 to OS Grid has a bug and fails. No download required - all on-screen so easy to click and use.

His story as you scroll down tells of how he developed his little application because of problems with driving round Milton keynes roundabouts. Well now you all know.

I do like people who sit down and work it out .... :rolleyes:

1.4k · February 2, 2005

Convert OS grid to WSG84

Grid Inquest performs the Helmert Transform to within 20cm.

Woohoo!!

Methinks he has completely missed the point.

The reason why Grid InQuest is capable of <~20cm accuracy throughout the UK mainland is because it does NOT use Hermert's Transform.

It uses a polynomial tranformation algorithm which cannot be replicated over wide areas with a 7 parameter Helmert transform.

Edited: 'cos I meant 7 parameter, not 7-fig

Edited February 2, 2005 by The Forester

1.4k · February 2, 2005

ive just done a 15 waypointer cache and wanted to plot the waypoints on a map,i prefer the british grid over the lat/long system....saying that the whole cache is done using the lat/long system.i can get my head around the british grid but the lat/long thing goes over my head

If you are plotting onto an OS Explorer map, which has a scale of 1:25,000, with a fine pencil such as a 0.3mm tip, a typical error of a low end GPSr's co-ordinate shift would hardly show at all.

0.3mm on a paper map at 1:25,000 scale is only 7.5 metres on the ground, which is slightly more than the probable maximum error shown by your GPSr conversion. Not many people can plot to an accuracy of a third of a millimeter, so I'd just go ahead and plot the grid co-ords which your GPSr's conversion shows you.

Alternatively, do a sample conversion at the www.gps.gov.uk website and then note how much you need to add to or subtract from the GPSr grid co-ords to make them metre perfect for that part of Britain.

Cheers, The Forester

February 2, 2005

It uses a polynomial tranformation algorithm which cannot be replicated over wide areas with a 7-fig Helmert transform

Did anyone else hear a whooshing sound? No? Just me then...

February 2, 2005

What bout goint he other way OS to WSG84?

often have cache all plumbed in then look at map and want to go to a grid on the wonder?

Any ideas or will i hear another woosh to?

February 2, 2005

should be going at the top, someone must have swapped the keys about sorry!!!

February 2, 2005

What bout goint he other way OS to WSG84?
often have cache all plumbed in then look at map and want to go to a grid on the wonder?

Any ideas or will i hear another woosh to?

Well... I knew what Forester was talking about, but I'm b______d if I understood any of THAT.

Why do show-offs like LostAgainAndAgain keep trying to bamboozle the rest of us with this gratuitous technical jargon?

It's not big and it's not clever....

1.4k · February 2, 2005

I'll try to simplify it.

There are basically three different ways to convert a WGS84 Lat/Long to an Ordnance Survey Lat/long.

The simplest is the one used in the basic plain vanilla (or custard, in the case of the etrex) GPSrs. It is simply an arithmetical addition to the X,Y and Z Cartesian co-ordinates. The advantage is that it is ultra simple. The disadvantage is that those XYZ shifts are only valid for one place.

A great improvement on that is the 7 parameter Helmert transformation. This has the XYZ shifts, but also has parameters for scaling and rotating the co-ordinates system. This gives it a much wider area of validity, but does not take into account the quirkiness of the Ordnance Survey's survey network.

Most countries have a basic datum point, usually at an astronomical observatory or a central trig pillar. They accurately measure the geographical position of that point by astronomical observations and then adopt that as their datum point. The survey network is then developed across the country from that point and everything can be referred back to that point.

Not so in Britain. Our national triangulation network developed in a higgledy-piggledy manner and was re-adjusted several times as errors were identified and distributed across the network of triangulation pillars. There isn't a single 7-parameter datum shift which works correctly across the country. Instead the OS split the country up into (about 11, I think) zones and adopted a 7-parameter dataset for each of those zones. It was an unsatisfactory arrangement as soon as GPS came along because although co-ordinates made sense within the triangulation network itself, things started to get flakey when you tried to do accurate work with GPS (or its forerunner Transit or the Russian Glonass) co-ordinates and to tie those co-ords into terrestrial co-ords derived from the trig pillar networks.

For several years the OS would not divulge to the commercial survey community what those datum shift parameters were. Instead they charged a lot of money to convert your co-ordinates for you. Some Surveyors back-calculated what the 7-parameters were for their area of interest, but it was an unsatisfactory method because survey measurements made to millimetric precision were being hung on the wobbly clothes-lines of the network's interface between GPS co-ords and terrstrial ones.

A concerted effort was made by several government mapping agencies in the UK to devise a computer program which would do the conversion calculation anywhere in the UK to decimetric accuracy. That program is the Grid InQuest which you have seen some people referring to. At first it was for sale for about a hundred quid, but recently they've made it public domain and anyone can download it for free.

GridInQuest is by far the best of the three methods. It is so accurate because it is not geographically limited to a small part of the country. It even takes into account the errors in the pre-existing survey networks in the offshore islands such as Lewis, Harris, Orkney and Shetland. It works in a completely different way to the Molodensky and Helmert methods of the two inferior methods. It has a huge number of polynomials which it Uses to model the shape of the network and accurately mesh it with the WGS84 co-ordinate system.

The new system also makes our national survey network much more compatible with our international neighbours. Previously, tying in the British triangulation network with those of places such as Norway and France was a real headache. Now we have the European terrestrial framework tied in together and it is closely tied to the global WGS84 system.

1.4k · February 2, 2005

Why do show-offs like LostAgainAndAgain keep trying to bamboozle the rest of us with this gratuitous technical jargon?

Personally abusive name-calling is out of place on this forum.

It's not big and it's not clever.

February 2, 2005

Personally abusive name-calling is out of place on this forum.
It's not big and it's not clever.

It appears that my shot hit its intended target.

... excellent.

February 2, 2005

What does gratuitous technical jargon mean?

February 2, 2005

What does gratuitous technical jargon mean?

Needless use of long words, where short ones would do the same job.

Point taken - I accept the implied rebuke.

February 2, 2005

Will re-phrase (it looked good in my head!!!!)

Been to Teasels' converter, wonderfull tool.

Would there, by any chance, be a similar site that goes the other way, for when i'm plotting a track in my GPS, so I don't have to keep going thru the menus to change between Lat/Long and Grid?

Or does the OS site one do that? Not been there yet?

February 2, 2005

Would there, by any chance, be a similar site that goes the other way, for when i'm plotting a track in my GPS, so I don't have to keep going thru the menus to change between Lat/Long and Grid?

Serious answer: it depends how much of this you intend to be doing. For multiple waypoints, Chris 'n Maria's waypoint workbench as mentioned above, is very good and more than accurate enough for this purpose.

Needs MS-Excel on PC, though.

February 2, 2005

greetings to all here..

well ive attempted to try all the different ideas and have to say that Teasles convertor gets my vote for idiot proof usage..ive now been able to plot my new cache route 'witches grave' (GCMN5R) and very pretty it looks to

a big thankyou to Teasle

1.1k · February 2, 2005

FIFTEEN MICROS!!!!!

I hope that Deego will award 160 points for this one for the COTM tables!

February 2, 2005

the 15 micro waypoints was a challenge but it all works as ive tested it out.situated in the Ashridge woods it should keep you all on your toes and hopefully amused i thought it deserved a cache at the end so i have hidden one in the 'grave'.

February 2, 2005

If the cache is as difficult to find, as its web page is difficult to read...

... this one could take a while.

February 2, 2005

difficult to read ?

February 2, 2005

Another Plug, you can use my Coordinate Convertor

to do single conversions in either direction.

Just paste in a Lat/Long into the box eg your 'N 51° 45.487 W 000° 25.886' is read automatically, I hope it should understand most formats. If you paste in one that doesnt work then I will try to add a understanding for it.

This should provide the Easting/Northing and the Grid ref with letters.

Actully one of the main features of the site is actully to provide to lots of sites with more information about the location mentioned.

Barry (PS for anyone wondering I use the Helmert transformation)

February 2, 2005

Actully one of the main features of the site is actully to provide to lots of sites with more information about the location mentioned.

Barry (PS for anyone wondering I use the Helmert transformation)

Sorry Barry, but your converter isn't even close ...

Example: N52º 26.0000 W1º 39.0000

(Near Meriden)

Correct grid ref: SP 23892.1 81762.0

Garmin conversion: SP 23893 81767

Your web conversion: SP 23942 81813

Approx errors in distance-

Garmin:5m

Yours:71m

... Not really a geocaching tool, then. :unsure:

February 2, 2005

Sorry Barry, but your converter isn't even close ...

... Not really a geocaching tool, then.

Opps, it wasn't that bad in any of the testing I did! :blink:

Guess will have to do a bit more tweaking, hopefully just a misplaced comma or something...

Thanks for letting me know, Barry

February 2, 2005

Example: N52º 26.0000 W1º 39.0000
(Near Meriden)

Correct grid ref: SP 23892.1 81762.0

Garmin conversion: SP 23893 81767

Your web conversion: SP 23942 81813

Now outputs SP 2389381764 which is much closer to the mark!

Again thanks for helping be to spot this, Barry

PS I'm using a Module, which included a rather anoying bug/feature, a fudge factor (of what you seeing here) of 50m in each direction, which it said was to account for the fact that the 'point is within the square based as the E,N coordinate' what ever that means. I thought I had removed it, but not on the online version :unsure:

February 2, 2005

Tried the waypoint workbenc before, cann't get it to run on my box, only have MS works, which doesn't seem to allow .XLS files, looks like the thing i need tho

am still working on it

February 3, 2005

Now outputs SP 2389381764 which is much closer to the mark!

Much better. Do LOTS more testing, though, because...

I'm using a Module, which included a rather anoying bug/feature, a fudge factor (of what you seeing here) of 50m in each direction, which it said was to account for the fact that the 'point is within the square based as the E,N coordinate' what ever that means.

I know what it means. If you like, put the source code in a file on the web, send me a link by PM and I'll take a look at it for you.

February 3, 2005

Simon from "The Wildleys" geocaching team once coded a converter tool (CoordinateTranslator.exe) that referenced the Grid InQuest DLL.

His site has gone down now along with the private link to the tool, but maybe the tool is available for public peer review elsewhere.

February 3, 2005

the module I'm using is

http://search.cpan.org/~pkent/Geography-NationalGrid-1.6/

although I have removed the fudge factor as put in on line 244 of NationalGrid.pm, and fixed a bug on line 73

this module doesnt use the Hermert Transformation, Ive just coded that from the OS spreadsheet

the main problem with this fudge factor was it only applied it on Lat/Long to GridRed, and not the other way

Also by removing it it also matches up pretty well with other convertors, I had run my testing on my local machine which had the patched module, I just forgot to patch the online one.

I've tried batches of about 100 coords around the country and only twice does it venture much over 10m. In fact I've run it against the 6500 odd trigpoints and didnt notice much of an error there, but didnt quantify it, but will do if want to make sure.

Barry

February 3, 2005

tried batches of about 100 coords around the country and only twice does it venture much over 10m

Quite like this tool. Just tried it and the 5th OSGB digit waves around within +/-3 on the ones I've tried which is well within GPS margin or error. My old Garmin Map12 waves around +/- 1 with good sat spread from 4 or more sats - so with a bit of care when it says we're standing on top of it - we are!

Wgs84 To Os Co-ords ?

Recommended Posts

Join the conversation