GPS Errors

[ditchcrawler]

I checked my 8 channel 7 year old h/h gps against my newer 12 channel h/h which I use with my Yeoman Plotter & was pleased to see that the lat/long positions only varied by 1or2 thousandths of a minute ie 6to 12 ft.However the altitude reading varied by nearly 200 ft.The old gps was only recoding 3 decent satelites.Is this the reason or are gps's not so good for altitude

<hr width=100% size=1>

 

[tome]

Sounds to me like the order of error you'll get if 1 GPS is set to OSGB36 datum and the other to WGS84.

You are quite correct that height is the weakest part of the GPS solution due to satellite geometry. vertical errors will be in the range of 1.5 - 2 x horizontal.

Regards
Tom

<hr width=100% size=1>

 

[Guest]

My old Magellan Meridien and the new eTrex literally plod along together on near enough the same Lat / Long .... the old Magellan gave up showing altitude ..... or as it called it then 3D fixes a long time ago. It seems to stay on 2D now except in exceptional circumstances !

But the old Magellan has a habit of losing its lock often - so has sadly had it's day as a serious tool. It also runs flat a set of AA's before it locks on ....... so I can only run it off the 12V.


<hr width=100% size=1>Nigel ...
Bilge Keelers get up further ! I came - cos they said was FREE Guinness !
/forums/images/icons/cool.gif

 

[Birdseye]

Geometry of the satellites must be part of the answer, but like you I have wondered about the remaining inaccuracy. I suspect that it comes from the mathematical models of the shape of the earth used in each of the datums (OSGB<WGS etc). Your GPS height reading will be relative to the surface of this model - not the actual surface which varies quite considerably.

For example, sea level in the Atlantic is different from the US coast to the European one, and the level rises again quite sharply if you go into the Med. The Earth is not round but is an oblate spheroid with lumps.

In any case, the accuracy of the GPS is theoretical in the sense that what use is a GPS accurate to 3m (which it isnt by the way) if the chart is to the nearest 20m (more in many areas) and there are datum inaccuracies, and satellite geometry inaccuracies.



<hr width=100% size=1>

 

[gjgm]

there are different gps for different purposes, and generally the use is for horizontal not vertical location, so thats where your money is being spent. I believe garmin summit has focus on altitude that when correctly set up can be within 7m, and there are dedicated aviation gps- but general gps are only to within about 70m vertical. So, yes, inpart beacuse of the geometry, and partly because demand/cost of greater accuracy in vertical doesnt merit it.


<hr width=100% size=1>

 

[tome]

Gjgm

Sorry to disagree, but there's not a GPS receiver available which can provide anything like 7m vertical accuracy without a source of differential corrections (eg EGNOS when operational next year). In cases where 4 or more satellites are available the receiver will (unless forced into height-aided or 2D mode) solve for the 4 unknowns of lat,lon,height, and time.

Satellite geometry is the sole reason for weakness in the height solution and demand/cost isn't a significant factor.

Regards
Tom

<hr width=100% size=1>

 

[Talbot]

I have an old Magellan Meridien and thought at the beginning of this season that it was throwing out time as towards the end of last season it took forever to lock on, and at the beginning of this season it was only registering 1 satellite. But I persevered, and left it switched on for a weekend, and at the end of that checked it again to discover that it is now working perfectly. Obviously the data has now all been updated. It locks on very quickly and seems to work better that it ever did before. My meridien was built in 1994, and I thought that its days were past. Boy am I glad that I kept on with it!!!

<hr width=100% size=1>

 

[BrianMoffat]

OK Chasps,
In order to test one GPS unit - you require a minimum of three, independant, units.
I can accept that if the H data is MINUS 10 metres - you are sinking, but if the same data is + 10m, then you should contact Air Traffic Control -

The H data is immaterial when applied to marine purposes - and is normally not calculated as Sea Level is a datum for satellites.

brian.moffat@wanadoo.fr

<hr width=100% size=1>

 

[Guest]

But what sea-level ?? As the Med / Red Sea etc. are different levels and the other matter is that the sea surface is not actually uniform height !!

OK - so I'm nit-picking and being smart - but please take it in the fun it's intended.


As to the 8ch vs 12ch ......... I have left it on a lot in UK when boating and it doesn't seem to want to update very well. Maybe I should rig it alone well away from other influences like ebgines starting stopping etc. which sometimes cause it to switch off !

I'll be sad to see the old Meridien go down the swanny - its seen my boat through quite a lot ..... but I also have to say that the little eTrex I bought is a bloody marvel !! Only down side being that you cannot remote the antennae, the higher model eTrex's you can - but not my base model.


<hr width=100% size=1>Nigel ...
Bilge Keelers get up further ! I came - cos they said was FREE Guinness !
/forums/images/icons/cool.gif

 

[AndrewB]

Re: GPS Errors - Altitude

I thought that, but on checking have noticed the 'altitude' given at our locked-in marina berth rarely varies by more than 10ft.

As well as satellite geometry, are the signals not still degraded slightly? I though I remembered reading somewhere that the US military can get accuracies to a couple of centimeters if they are needed.

I've noticed no very obvious benefits in moving from an 8 to a 12 channel GPS. It doesn't seem so long ago that all we had was a single-channel SatNav system. Now that WAS bad!

<hr width=100% size=1>

 

[tome]

Re: GPS Errors - Altitude

Andrew

Now that SA has been switched off the main sources of GPS error are ephemeris (predictions of where the satellites are in space) and ionospheric effects which alter the measured ranges. GPS transmits on 2 frequencies, but our civilian receivers only receive one of these. With high end dual-frequency receivers it is possible to minimise but not remove ionospheric effects.

Ephemeris errors can only be removed in real-time by static measurements which determine more accurately the ephemeris and ionospheric errors. These are transmitted as range corrections for each satellite, and are known as differential corrections. With differential techniques it is possible to get the horizontal accuracy down below 2 metres, and more expensive receivers can be corrected to less than a metre. This will become possible after April 04 when the satellite based augmentation system (SBAS) becomes operational in Europe - it's known as EGNOS and will transmit wide area differential corrections. If you have an SBAS capable receiver you will see a big improvement.

More advanced techniques such as real-time kinematic rely on two receivers (one fixed) being in constant communication over short distances (<20km) with the static one transmitting sophisticated corrections to the mobile. With RTK techniques it is indeed possible to improve accuracy to the 2cm level. However, tests with military receivers have shown that the accuracy is not much better than our civilian receivers in normal standalone mode. Incidentally, it's a well known fact that during the first Gulf war there weren't enough military GPS receivers to go around so the US bought civilian sets and switched off SA!

The altitude in your marina may not appear to move very much, but if you observed it closely over time you would see a much greater variation, unless it is operating in a fixed altitude mode.

Regards
Tom

<hr width=100% size=1>

 

[DavidTocher]

As I understand it the GPS height is relative to the geoid used e.g. WSG84 The geoid doesn't match 'sea level' exactly. This means that for example one can be sailing 20m above or below the geoid and still be at sea level.

To get a GPS fix one needs to solve a set of equations for x,y,z & t which requires four sats. With more sats there is a volume of uncertainly (a 3-D version of the 'cocked hat' from three bearings). It would seem best to use this extra data to minimise the error in the lat and long dimensions as terrestial navigators cannot do much about controlling their height.

<hr width=100% size=1>

 

[mldpt]

I still have on board and can be used a Raytheon fixed GPS only uses three ch, now that is taking a long time to locate itself It has been suggested that the memory battery has failed I bought it in 1992 so if there is a battery then I am sure it has failed but I would have thought that it would have a rechargeable battery charged when the unit is powered up.
Regards Mike


<hr width=100% size=1>

 

[Oldhand]

Re: GPS Errors - Altitude

Andrew,

I presume your reference to a single channel SatNav system was the Transit system. I can understand yacht installations being "bad" but the system wasn't that bad if used to its full potential. If you had a dual frequency reciever and specialised systems to provide the dead-reckoning information between fixes, such a doppler-sonar and gyro-compass specifically selected for accuracy, then the Transit system could be used navigate to better than 200m accuracy. I don't think that is particularly bad.

The whole crux of the system in obtaining accurate fixes was the accuracy with which the vessel's track was known during the 18 odd minute period it took for a satellite pass. I guess yachts didn't provide accurate track data unless it was used in conjunction with an other system such as Decca and Loran.

<hr width=100% size=1>

 

[AvanLoon]

To learn more about GPS this site could be helpful: http://www.trimble.com/gps/index.html


<hr width=100% size=1>

 

[Birdseye]

Think you have them mixed up, David. GPS measures the height relative to the speroid, which is the mathematical model of the shape of the Earth built in to the particular datum that is being used. So you will get different altitudes from your GPS if you switch from osgb to wgs etc.

However, the spheroid is not the same as the geoid, which is the equipotential surface of the earths gravitational field - for which read notional sea level. In the UK, this sea level is about 50meters above the spheroid, so even at sea level your GPS should read an altitude of 50m or 200ft. Add on to that any geometrical inaccuracies, tides, etc



<hr width=100% size=1>

 

[Guest]

Re: GPS Errors - Altitude

Having 'grown up' with Transit system in my commercail days ....... I can honestly say that thank-god we moved on !!!!

Most of my time with Transit was on Gas Ships plying the China Sea from Borneo / Brunei to Japan and boy was it bad !!! The coverage used to go in fortnightly cycles ... from reasonably well time-spaced fixes ..... to a bunch one after another and then nothing for hours ...... in fact it could run to 10 -12 hours at times !!

The best bit about the Magnavox as we had on those ships - was you could do great-circle and Compass error calcs !!!

Of course because it was early days ..... the Shipping Companies and Masters didn't trust it and we had to do all sorts of trad nav as well ....... anyway it passed the watch time !!! One of the memories was a shoal and a wreck on it ..... we used to watch the E/S trace and when the rise of sea-bed indicating the shoal and wreck started to print - it was port 10 to the new course !!!! For those interested it was a British Ship that went down a long time ago and is about halfway up the Palawan Passage !!!!

Transit ......... UGH !!!!!!


<hr width=100% size=1>Nigel ...
Bilge Keelers get up further ! I came - cos they said was FREE Guinness !
/forums/images/icons/cool.gif

 

[Guest]

Well I never !!! nm

*

<hr width=100% size=1>Nigel ...
Bilge Keelers get up further ! I came - cos they said was FREE Guinness !
/forums/images/icons/cool.gif

 

[DavidTocher]

Yes - you are correct on the geoid - ellipsoid ( a slip of the brain/fingers!) but the heights etc are calc from time differences from a set of sats signals. The solution to the equations is a 3-D set of coordinates x,y&z relative to a world ellipsoid. These can then converted to lat, long and height for a given datum which can be a local ellipsoid or geoid or whatever one chooses. As I suggested; with more sats there will be a more than one xyz solution. It is possible to then minismise the error in the prediction and one could choose to min the lat long error at the expense of a poor height estimate.

<hr width=100% size=1>

 

[philip_stevens]

Re: GPS Errors - Altitude

I hope you weren't on watch at the time she went down ;-)

We, or should I say the company I worked for, had a 2nd mate who "used" Rondo Island as a waypoint.

The CEO got woken up when the red light kept flashing in his forward window - the ship was up on the rocks! :-o

<hr width=100% size=1>regards,
Philip