Jump to content

British Os Versus Wsg84 Lat. & Long.


Team Ullium

Recommended Posts

After reading the post on 999 calls and the discussion about the two formats of co-ords it brought to mind the following question which has been nagging at me!

 

For quite some time now I have been downloading cache waypoints from the geocaching.com site in Lat. & Long format .... however just recently I had occasion to seek caches using OS co-ords and found that not only was my signal steadier but my accuracy was slightly better!!

 

I always thought it wouldn't matter which type of co-ords I set my GPSr to ... but after my recent experience (which might just be a one off of course) I'm not so certain now!

 

Most modern GPSr automatically set the datum for OS or Lat. & Long so perhaps the only the other variable might be the North referece...i.e. map or magnetic!

 

Has anyone else experienced this? Or indeed have an opinion on which set of co-ords is best for geocaching in the UK ?

 

Ullium.

Link to comment
.... however just recently I had occasion to seek caches using OS co-ords and found that not only was my signal steadier but my accuracy was slightly better!!

 

From my (admittedly meagre) understanding of how handheld GPS units work, it hard to credit that these things could be connected.

 

Mind you... it's amazing what some people believe: I met a local rambler, the other day, who has just changed the settings on his GPS, in order to improve its "accuracy"...

 

"Well, it WAS on metric distances, right? But now I wentan changed it to feet, 'cos feets is shorter than metres, right. So it's more accurate, now - stands to reason, dunnit?"

 

And so saying, he rambled off...

Link to comment
From my (admittedly meagre) understanding of how handheld GPS units work, it hard to credit that these things could be connected.

 

Perhaps because the different co-ords will be interpreted by different computer algorythms and depending on the efficiency factor of both ... it could well be that one is slightly more efficient than the other ???

Also I understand that the internal conversion algorythms can decrease the accuracy of a waypoint when using the GPSr to convert from British OS to Lat & Long then perhaps this factor is involved??

 

Well anyway it seems to make sense to me Wildlifewriter .... but as I said perhaps I'm mistaken and it was just a one off ;)

 

Ullium.

Link to comment

100% observational effect. GPS accuracy varies from one day to the next, and if it's a good day when you are using a new datum, you might well ascribe one to the other. All the numbers which you see are rounded off from internal values and converted to the datum and units which you want, at the very last minute.

 

Another example: mark ten waypoints at the same spot. On my Geko 201, 3 or 4 of them will have the same coordinates (to 3 decimals). But if you pick one of those and ask the distance to another one, you will get "0.4m" or whatever, which means that internally the value was stored to a different degree of precision.

 

However, this illustrates the same principle which leads some people to think that homeopathic medecines, etc, "work" for them. Since you can't predict in advance how long any given cold, etc, will take to go away, you can't tell whether the treatment is having an effect in any given case. This is true of all medicines for diseases which basically go away by themselves; it's just that for "real" medicines there are large, statistically-valid surveys which demonstrate that there's a real (overall) effect, and for the "fake" ones there are surveys which demonstrate that they don't - all that's left is anecdotes ("I took X and my headache went away"), which is the standard reasoning fallacy known as "Post hoc ergo propter hoc" ("B happened after A, therefore B was caused by A"). See also the MMR controversy.

 

Sorry for going OT but this is my #1 hobbyhorse ;)

Link to comment
Also I understand that the internal conversion algorythms can decrease the accuracy of a waypoint when using the GPSr to convert from British OS to Lat & Long then perhaps this factor is involved??

I understood that GPSrs, and indeed the satellites themselves, use WGS84 internally as their master datum, so I'd expect that using OSGB would decrease, rather than increase accuracy. (Though I'm still an advocate of ridding UK geocaching of the nasty WGS84 and converting to good old OSGB, you realise! ;) )

 

Either way, it'll only apply when manually keying in coordinates - all garmin downloads that I'm aware of use WGS84 and swapping the display settings on the GPSr afterwards shouldn't change anything. And if your GPSr uses the same conversion algorithm as GC.com, then it shouldn't make any difference anyway (assuming the conversion algorithm is symmetric).

Link to comment

I think you are right there Teasel about the satellites and the GPSr's using WGS84 as their master datum as I'm certain I have read that somewhere as well !!

And of course given that one would think the WGS84 datum would be slightly more accurate.

 

Also as far as I am aware not all the conversion algorithms in GPSr's are equal and even the better ones such as 'Grid InQuest' will not reverse convert accurately, by that I mean convert one way then use the result to convert back!

 

I appreciate what you are on about sTeamTraen and I did give that this might be case...what struck me at the time though was the fact that the '100% observational effect' as you put it, was just so pronounced at that given time and as I swapped back and forth from one type of co-ord to the other so did my accuracy and steadiness of signal improve.

 

I am just wondering if the how the placer has recorded his co-ords and by what method they used to convert from one datum to the other for the sake of putting up his cache page?? Meaning, either one or the other might be in fact a more accurate rendering of the actual global position??

 

Ullium.

Edited by Ullium
Link to comment

<<my signal steadier >>

 

No way could a change of datum effect the signal - that simply depends on the path between you and the satellite.

 

I don't know why UK cachers use WGS84. We have our own very excellent grid system and I believe we should use it. For one thing it is much easier to use an OS map with OS Grid than WGS84!

Link to comment

There is one very good reason why UK geocachers use the same system as everyone else.... it means that UK caches can reference locations outside of the British Grid.

 

The OS grid references only apply to a small square of land and sea in our corner of Europe. Go to France (for example) and there is no OS grid reference for, say the Eiffel Tower. This means that while for most UK caches it would not matter, there are going to be some which are affected. The beauty of the LatLong system is that you can generate locations anywhere on the planet and everyone can reference them and track to them.

 

Having said that, I do prefer to use OSGB while doing UK caches since I find it easier to enter the data. It is grouped into tidy little packets, and not sprawled out with decimal points and degree signs like a LatLong is. Its just a personal thing.

Link to comment

There are 4 (!) things some of you mix up:

 

OS Grid

Lat./Lon.

WGS84

and OSGB36

 

OS Grid and Lat./Lon. are type of grids. NN 35538 71020 (a place somewhere in Scotland) is the very same spot like N56° 48.104 W004° 41.618 IF you use the same map datum for both grids: OSGB36 which is an old british one.

 

The Waypoint 56° 48.097 W004° 41.697 with the datum WGS84 is the very same spot as well. If you're completely confused now you should google for map datum and map grid ...

Link to comment

I remember doing some work on the maths of this about 20 years ago, so very distant memories.

 

Using the deg min or even better deg.decdeg makes it easier to calculate distances between points. There is a direct relationship between metres on the ground and degs lat/lon covered.

 

I have no idea what the maths behind OS grid is. Does one unit specify a particular distance. If given two positions on the grid, is there a formula that you could use to calculate straight line (or great circle) distance between them.

 

Not that I am advocating for the lat/lon solution on the basis that I can use the maths behind it when in the field, just that I know there is maths behind it and have no idea about the logic behind OS

Link to comment
The Waypoint 56° 48.097 W004° 41.697 with the datum WGS84 is the very same spot as well.

Not exactly, I fear. I looks like you took these figures using your GPS unit's co-ordinate converter - which isn't quite accurate enough...

 

In fact, the two "very same spots" are just under ten metres apart.

 

And THAT (of course) is one of the reasons why Geocachers need to use the same co-ordinate system - whether it's convenient or not.

:laughing:

(edited for accuracy)

Edited by wildlifewriter
Link to comment
I have no idea what the maths behind OS grid is. Does one unit specify a particular distance. If given two positions on the grid, is there a formula that you could use to calculate straight line (or great circle) distance between them.

Yep... the blue grid lines on the OS maps are 1 kilometre apart so calculating the 'straight line' (as the cod swims) distance between two points is straight forward schoolboy trigonometry. It's the basic premise of this cache.

Link to comment

Having ignored this thread for several days, I have finally read all the entries. Jumping back to the original question regarding accuracy and the stability of the reading, I suspect that the two algorithms are different in the "damping" they apply to the signal.

I do not know how often the GPSr updates its readings, but say it is once a second. If the readout shows this, the numbers will also change every second. If there is a damping programme in the algorithm, it may actually show the average figure of say the last three or five readings. The results should be no different in terms of accuracy, but they will appear more stable and the averaging may also make the readout appear more accurate. I have seen this effect with readouts of displays of flows in industry. We "stabilised" the flow by putting a damping effect in the algorithm.

As I said, I do not know how the GPSr algorithms are written, but I suspect that this is what you have seen.

Link to comment
I have no idea what the maths behind OS grid is. Does one unit specify a particular distance. If given two positions on the grid, is there a formula that you could use to calculate straight line (or great circle) distance between them.

You can actually. With the OS Grid each 10,000th unit, ie the last digit in eac of the groups of 5, represents 1 meter either in East-West, or North South. And because this grid is equal in both directions (ie no variations such as 1 minute of longitude being a different distance as 1 minute of latitude at any given latitude) then you can pretty much envision it as one big sheet of graph paper.

 

So over short distances (where the curve of the earth doesnt add up to more than a few meters varience) you can calculate distances between waypoints using very simple Pythagorus maths.

Example *taps blackboard with chalk :bad: *

Waypoint A is at SN 75392 12345 and

Waypoint B is at SN 75678 12987

this means that B is 00286 Meters further 'east' and 00642 Meters further north. You can use these as the lengths of sides of the right angled triangle to calculate the hypotenuse. 286² + 642² = hyp²

I work this example out to be approx 702.8 Meters. I entered these same arbitary waypoints into my GPS and used its measure distance function on them and it reported that they were 702.3 Meters appart. So what's half a meter between friends? :laughing::D

 

By the way I made these points up from the top of my head, if they are actually someone's back garden I do apologise :bad:

Link to comment

I was out today geocaching and used the OS co-ords and would you believe I found the same increase in accuracy and steadiness of signal....so just as an exercise I switched to WSG84 and again I was all over the place and my accuracy decreased....not by much I admit ... only about six or seven meters .... but that is a lot in geocaching terms!!

 

Does that answer your question Wildlifewriter ??

 

I admit this may all be 100% just chance as one poster suggested...but hey!....I will be using OS co-ords from now on until someone convinces me otherwise :laughing:

 

Ullium

Link to comment
I don't know why UK cachers use WGS84. We have our own very excellent grid system and I believe we should use it. For one thing it is much easier to use an OS map with OS Grid than WGS84!

 

There are a couple of hundred different spheroids in use around the world and even more geodetic datums.

 

If every geocachers of every country adopted their own local spheroid and datum, geocaching.com would become hopelessly complicated and practically unuseable. In some parts of the world the difference is several hundred metres. In many parts of the UK, the difference is something of the order of a hundred metres.

 

I, for one, am glad that the 'Company' decided to adopt the global standard of WGS84. It keeps things nice and simple.

 

Even the OS itself is adopting a version of WGS84, slightly adapted to take into account tectonic plate movements within the UK, and is moving away from the old OSGB36 datum and Airey spheroid.

 

I do, however, agree that a good conversion of a geocaching standard format geographical position into OS grid co-ords is very useful when using the 1:50,000 or 1:25,000 (or larger scale) sheets.

 

Cheers, The Forester

Link to comment
If given two positions on the grid, is there a formula that you could use to calculate straight line (or great circle) distance between them.

One of the many reasons why Surveyors take the time and trouble to convert geographicals to a grid is to simplify such calcs.

 

The spheroidal maths of calculating a rhumb line or great circle from geographicals may crease your brow, but doing a grid distance calc is simplicity itslef.

 

It is quite simply Pythogoras.

 

The square root of the distance is equal to the sum of the squares of the difference in Eastings and Northings. That gives you the grid distance, in metres, between the two points. There is a difference between grid distance and true distance, but it is not significant in practical terms for geocaching type activites.

 

Calculating the grid bearing from one grid point to another is equally simple. It is quite simply the Tangent rule (remember Opposite over Adjacent from school maths?).

 

One thing to be wary of, however, is that Grid North, True North and Magnetic North are all quite different and must be taken into account when navigating accurtely by methods such as Dead Reckoning or Traverse.

 

Cheers, The Forester

Link to comment
I was out today geocaching and used the OS co-ords and would you believe I found the same increase in accuracy and steadiness of signal....so just as an exercise I switched to WSG84 and again I was all over the place and my accuracy decreased....not by much I admit ... only about six or seven meters .... but that is a lot in geocaching terms!!

 

Using OSGB36 instead of WGS84 in a retail GPSr will indeed reduce your accuracy.

 

These units cannot replicate the very complex polynomial conversion algorithm which is needed to rigorously convert between the two datums. In fact, I know that at least one of the major brands only uses a three parameter shift and not even the more rigoruous (but still quite woolly) 7 parameter shift.

 

There are only two ways to improve the quality of a pseudorange fix. One is to upgrade to a GPSr which uses WAAS data such as EGNOS. With good signals, this will give you an accuracy of two or three metres in good conditions.

 

The other method for a static fix is to average out lots and lots of readings:- the more, the better. This helps to cancel out the random errors which are induced by such grollies as clock error; ionospheric and tropspheric errors and somtimes some of the multipth errors. By patient averaging, with a WAAS enabled unit, you can quite easily obtain a fix which is more accurate than the 0.001" Lat/Long resolution of the GC.com standard format for geographical co-ordinates.

 

Cheers, The Forester

Link to comment

I wondered how long it would be before you posted on this subject Forester :(

 

And as usual your comments are well worth reading :(

 

My Garmin GPS 76 marine can access over a 100 datums and about 14 location/grid options ... so when I opt to use the OS British Grid (LORAN TD) how do I know if I am using OSGB36 (which I thought was obsolete now and is not apparently one of my options!!) or the updated OS British Grid ?? Which btw was as far as I'm led to believe was one of the reasons why our OS maps were updated and the Pathfinder Map series dropped in the process!?

 

(A slight aside....I see now that when I change from Lat & Long to OS British Grid that the datum changes automatically....this was a source of error for many users in the past as they sometimes left the Datum set to WSG84 when changing to the OS British Grid option!)

 

Also, any algorithm which triangulates using Lat & Lon co-ordinates is surely less accurate than one using OS British Grid co-ordinates for obvious reasons!!

 

So where I agree about there being a risk of reducing my accuracy if I use the old OS British Grid system (though even then I am not convinced!) .... I am not that certain that I am using the old grid system!? Remember that most modern operation systems on GPSr have the option of downloading updates!! And one would imagine they would be aware of recent events when preparing those updates!?

 

Ullium.

Edited by Ullium
Link to comment
when I opt to use the OS British Grid (LORAN TD) how do I know if I am using OSGB36 (which I thought was obsolete now and is not apparently one of my options!!) or the updated OS British Grid ??

 

One more shot at this, and then I give up... :(

 

You are getting yourself confused, and using terminology outside its correct context:

 

1) What Garmin tend to call "position format" is the way in which the unit displays position information. This may appear in Lat/Long format (in various guises), or you may choose to display it in a format corresponding to the British Grid. (or Finnish Grid or whatever.)

 

2) OSGB36 is a map datum. It is not any sort of a grid, still less a positional format.

 

3) There is no "updated" British Grid. What there IS, is a new transformation model, known as OSTN02. This model is more accurate (for geodetic purposes), and more numerically stable, than the previous one.

 

As Forester has already pointed out, handheld GPS units do not use OSTN02 when converting position formats applicable to Great Britain. They use an approximate method, which is faster and more mathematically economical, but which is also marginally less accurate.

 

For this reason, it may be a mistake to have your GPS unit set to British Grid format, when searching for a cache.

 

It is DEFINITELY a mistake to have it set that way, when placing one...

Link to comment

Thanks Wildlifesongwriter :o

your patience and knowledge does credit to you !!

 

I really was aware of your point 1) and I was never in any doubt of your point 2) but on point 3) I fully admit I hadn't fully understood even though I remembered reading something about it earlier on this year!

 

So let me congratulate you on your obvious clear understanding of this subject and thankyou for sacrificing yet another ten minutes of your very valuable time from your life to enlighten us all with such a clear explanation... and in such a gracious manner too :D

 

Ullium.

Link to comment

I would like to thank all the posters on this thread for their instructive and sometimes humouress replies....I am very much in your debt :o

 

I have now a better understanding of this subject now and you have all convinced me that the WSG84 datum and Lat & Long co-ords stand a far better chance of being slightly more accurate than anything else :lol:

 

You asked back a bit Wildlifewriter what I meant by steadiness of signal ... and I forgot to explain!

I like to use the map page on my GPSr as it seems to be more reflective of what is happening within the unit itself....by that I mean when I get a spike I see it immediately or if I am only locking on to three satelites or less then the indication on my screen flicks about! Plus I like to see my tracks as they show when I have lost signal and when I have got it back!

 

I guess I was just fortunate these two days in that I was getting more than four satelites when on OS and lost a few when I switched over to WSG84 :D

 

Again thanks guys :D

Ullium.

Link to comment
Fangsy wrote:

By the way I made these points up from the top of my head, if they are actually someone's back garden I do apologise

No, you seem to be ok there. The first one is just off a public footpath and the second one is by a track, at Ystradowen.

 

They both seem to be under tree cover, though, so accuracy isn't going to be an issue there. :lol:

That's Lucky! :lol:

 

They're not too far from a cache called 'Twrch Toilet' by the looks of them actualy. Still need to get to that one :lol:

 

Thanks for checking! :huh:

Link to comment
brain hurts trying to make sense of the thread!

 

I know what you mean nobby.nobbs :lol:

 

should have my gps set on wsg84?

 

As Forester said the answer is yes ....but not with OS British Grid Co-ords...only with Lat & Long....well that is what I understand is meant!?

 

That was one of the things I pointed out about the modern GPSr's is that they automatically select the correct datum for whatever co-ords you select!

 

I am now convinced that despite the British OS co-ord system being more accurate as far as our map reading is concerned....the fact that the Satelites use the WSG84 datum means the GPSr does not lose out on accuracy by having to convert to the British system!!

 

I hope that is correct?...no doubt someone will correct me if it isn't :rolleyes:

 

Ullium.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...