Jump to content

how far out


Recommended Posts

A minute of latitude is approximately 1853m, so:

 

an error of xx° yy.100' in latitude equates to 185.3m

an error of xx° yy.010' in latitude equates to 18.53m

an error of xx° yy.001' in latitude equates to 1.853m

 

For longitude, the length vares approximately as the cosine of the latitude. So, for example a minute of longitude at 51° N is about 1166m, so:

 

an error of xxx° yy.100' in longitude equates to 116.6m

an error of xxx° yy.010' in longitude equates to 11.66m

an error of xxx° yy.001' in longitude equates to 1.166m

 

So an error in the first digit after the decimal point is significant; after the second digit it is almost reasonable and with the third digit it is just about negligible. Of course if the error is greater than one...

 

Hope that helps.

 

Edited to correct speeling mistakes.

Edited by Master Mariner
Link to comment

Master Mariner, what a star.

 

I've been looking for this for a while on and off. Don't know if you saw my thread about having 10 identical GPS units all showing slightly different co-ords when placed side by side.

 

So I was trying to work out the maximum distance the co-ords could describe, if you see what I mean, and now I can finish the job.

 

The geek in me is fully satisfied

Link to comment

Keith, I had seen your other posting about the 10 GPSr's and thought I had commented as I had done a calculation regarding them. Just checking I must have forgotten to post a reply but I still have my scribblings.

 

If you take the centre point of the two latitudes and two longitudes a circle with 7 metre radius will cover all the points. That assumes that one position is actually at a corner point. If none of the positions fall at a corner point then the 7m is reduced. See this thread.

Link to comment

Keith, I had seen your other posting about the 10 GPSr's and thought I had commented as I had done a calculation regarding them. Just checking I must have forgotten to post a reply but I still have my scribblings.

 

If you take the centre point of the two latitudes and two longitudes a circle with 7 metre radius will cover all the points. That assumes that one position is actually at a corner point. If none of the positions fall at a corner point then the 7m is reduced. See this thread.

 

Just added my calculations to the other thread.

 

Came to the same conclusion as yours, 7m circle just about covers it, but one unit does happen to be at a corner and is therefore 7.67 metres from the averaged centre.

 

Not much on TV tonight?

 

Damaged knee stops me going anywhere.

Link to comment

I find that very interesting stuff and answers a question I've had for a while.

 

Does anyone know similar figures when a OS Grid Reference is involved (using OSGB36)?

 

At work, we often take our fixes using the above datum, and end up with a figure such as TL 12345 67890...I was just wondering how much a single error in 4, 5, 9 and 0 would really be on the ground?

Link to comment

If you use a 6 fig grid ref like TL 123 678, then that puts you at the bottom left corner of a 100m x 100m square so your location can be anywhere within that square.

 

So, TL 1234 6789 is a 10m x 10m square inside that one and TL 12345 67890 is a 1m x 1m square inside that.

 

So in your example, if you are out by one digit with the 5 or 0, then you are 1m off. If you are out by 1 digit with the 3 or the 8, then you would be 100m off.

 

Hope that makes sense, but I doubt it. :)

Link to comment

Does anyone know similar figures when a OS Grid Reference is involved (using OSGB36)?

That question gets complicated by the fact that there are three different conversion methods in use.

 

The cheapo GPSrs, such as the Geko/eTrex series, use an overly simplistic 3-parameter datum shift which simply adds a fixed number to the X;Y;Z cartesian co-ords.

 

Intermediate level machines use a 7-parameter shift, which uses the X;Y;Z figures and adds four scaling and rotation parameters to make a better fit over the British Isles.

 

Good systems, such as MemoryMap, and competently written websites such as Nearby.org.uk use the full polynomial conversion method which produces metre-perfect (or better) conversions. The government kindly produces a free software module, which any competent scriptkiddie or codemonkey can very easily splice into any webpage generating routine to show proper grid co-ord conversions.

 

I cannot explain why GC.com continues to produce such garbage OS grid co-ords on cachepages, but the cheapo GPSrs do at least have the legitimate excuse that their basic firmware was written in the bad old days, long long before geocaching happened, when Selective Availability (SA) meant that consumer-level GPSrs had a fix error of something like 35 to 50 metres without some intelligent tweaking by knowledgeable users. The 7 to 9 metre error of the junk grid conversion was subsumed under the massively greater error of the indescribably stupid SA. Those days are long gone, thank Goodness, but the junk which GC.com continues to publish in lieu of decently calculated grid co-ords on cachepages continues to defy commonsense or rational explanation.

 

You can produce metre-accurate grid co-ord conversions quite simply on your own GPSr, even the cheapest ones, by inputting your locally accurate XYZ datum shift parameters in the "user defined" section of your GPSr's menu. Calculating those proper XYZ numbers for your own local area is a bit of a bother, but it's just arithmetic and it ain't rocket science.

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...