Go Back  PPRuNe Forums > Flight Deck Forums > Tech Log
Reload this Page >

GPS - Can't be wrong

Tech Log The very best in practical technical discussion on the web

GPS - Can't be wrong

Old 6th Jul 2001, 11:40
  #1 (permalink)  
Posts: n/a
Post GPS - Can't be wrong

On a number of ocassions, on a number of threads, a number of people have reported errors with GPS. As yet nobody has explained how this could happen. Can you?

This is my position. There are errors in the GPS system but they are not systemmatic. There is no error that can affect more than one satellite in the same way.

Now move to your receiver. It needs 4 satellites to create a 3d fix. Three for the geometry and one to correct the time error in the receiver.

Clearly you have a problem if you are below this number but my 100 GPS shows this failure.

At this number maybe you are sensitive to an error in any one satellite. However typically you are running more. Up to 12. Unless the reciever is so stupid that it tracks 8 satellites and then simply discards the surplus ones, you must have some averaging process going on. How in this scenario can it possibly be fooled into reporting an incorrect position?

Now I have omitted to mention one problem situation - jamming/interference. I assume for this that no normal interference/jamming will fool the receiver. It will simply lose the signal and thus report insufficent satellites. Not an error - just a can't work situation.

So there we are. Anybody going to take me on on this one?

Old 6th Jul 2001, 12:14
  #2 (permalink)  
Posts: n/a

Satellite errors themselves are highly unlikely, in that the signal contains the mechanism for its own integrity to be checked. In this way a failed vehicle can be spotted by its uncongruence with the other signals and dropped.

Most alleged GPS errors that I have come across relate to two simply-explained problems:

1) The Map data supplied with most GPS units is not nearly so accurate as the aeronautical data. The precise positions of roads railways and rivers is not guaranteed. For this reason, if you follow a road around the edge of a CTR using the GPS, rather than the physical object, you are in danger finding yourself infringing.

2) The other common fault is with the aeronautical data itself. Much as it is carefully checked, a number of high-profile errors do creep in and catch the unwarey.

Does this cover your errors, or are there others you are thinking about?


[edited for typos]

[This message has been edited by 2Donkeys (edited 06 July 2001).]
Old 6th Jul 2001, 15:46
  #3 (permalink)  
Posts: n/a

You see, I think you are right. Whenever I see a discussion about GPS reliability it is actually about the accuracy of the maps, not the GPS absolute position information.

However we have seen remarks like 'there was a problem over ORTAC' which its is not clear whether it is a mapping or navigational fault.

GPS does have errors eg signal bending which could not be detected at the minumum number of satellites. Also you could have 5 satellites all from the same unfavourable direction and thus having multiple atmospheric distortion errors not correctable by averaging.

I can't think of anything else ie if you are doing a VOR-style overlay approach (waypoint coordinates validated) with 7 satellites the GPS will either show nothing or the correct track and distance.

Anybody experienced anything different?
Old 6th Jul 2001, 16:05
  #4 (permalink)  
Posts: n/a


GPS has a number of errors, civil aviation units are known to have the following errors
  • Ephemeris error 4m ( GPS satellite not where it should be)
  • Clock errors 2m
  • Receiver noise 1m
  • Ionosphere 8m
  • Troposphere 3m
  • Selected availability 32m

The RMS average of the above is about 33m, along with the PDOP will give you a position of close to 100 m only 95% of the time.

The atmospheric effects are predicted and corrected to the GPS position, as the atmosphere is anywhere between 130-190 km the amount of slow down will vary. Military GPS gets rid of this error by using two frequencies from each satellite.

You may have more than 12 sats in view, but the GPS receiver will drop out any sats that are less than 7.5 degrees above the horizon (a TSO C129 standard).

The selected availability which has been turned off now used to make the sats transmit a false position so the military of non US alliance members could not use the GPS against the US. The still change the clock frequency which alters the time output in the GPS signal which give a false position.

PDOP is position dilution of precision, ie if you have sats low on the horizon PDOP is high, if the sats are high in the sky PDOP is low.

BTW with baro aiding only 3 sats are required to get a 3d fix, the baro aiding gives the gps unit a psudo sat and solves one of the variables in the four x/y/z/time equations.

With RAIM the GPS unit compares the positions of multiple sat sets and gets the best position and discards the problem ones.

Jamming is unlikely, the US military thought this through very well, the GPS signals are transmitted at a very low power, and actually use the earths background noise for amplification. This makes it near impossible to jam. High power signals are easy to jam i.e radar, low power very hard.

As 2Donkeys said problems occur when the data you put in is crap, you get crap out. The crap in this case is the waypoint positions. GPS receivers normally use the WGS84 datum, if your waypoints are in another datum, the GPS will take you to the wrong spot, ie the position you entered but on the WGS85 datum.

This is still a problem today, I understand that Jepp have just issued a notam in the current database cycle that they have removed a number of airports, and waypoints in Mexico as the Mexican aviation authority are in the process of changing datum’s but is not consistent throughout an approach or departure.

GPS tracks are normally more accurate than that displayed on charts or conventional aids. Ground based aids need to be surveyed all the time to get the magnetic variation to apply to the true tracks. Most countries do this only every 2-5 years which can lead to errors particularly with enroute vors. The GPS calculates the earths magnetic field for the latitude, longitude, and time and comes up with a variation which is statistically within 200nT of the real value.

Because of all these errors, the GPS units will not become a sole means navigational equipment because it cannot meet the accuracy, integrity, availability and continuity of service without some form of WAAS.

[This message has been edited by Zeke (edited 06 July 2001).]
Old 6th Jul 2001, 16:18
  #5 (permalink)  
Ham Phisted
Posts: n/a


Interesting answer and I follow your reasoning for most of your discussion; however, I am somewhat confused by your statement that GPS

<font face="Verdana, Arial, Helvetica" size="2">actually use the earths background noise for amplification</font>
Don't you mean that the use of the pseudo-random code by GPS leads to a spread spectrum thus bringing the signal down below the background noise level? If this is the case, jamming becomes difficult because the jamming signal would either have to correlate with the long, pseudo-random code used or use extremely high power (or be very close).
Old 6th Jul 2001, 17:28
  #6 (permalink)  
Posts: n/a

Has anyone else found that the Vor at TFS isn't quite where it is supposed to be on GPS equipped aircraft? i.e. its lat. and long is slightly in error.
Old 6th Jul 2001, 18:59
  #7 (permalink)  
Posts: n/a


Just checked a few live navigational databases, TFS should be on your charts as N28.0024337768554, W16.6879320144653 what does your chart say it is ?

Ham Phisted,

I was not trying to go into too much detail....

The GPS signal for civil use are sent on the L1 carrier at 1575.42 Mz using spread spectrum techniques.

It transmits pseudorandom noise on modulated L band waves with bi-phase shift keying of the carrier.

The signal level is well below earths natural radio noise level, the receiver detects it, and then it uses the predicted P and C/A code to normalise the signal back to the original carrier frequency above the natural noise level.

The receiver only needs to have between 160-163 dbW signal power level to work.

The FAA is presently investigating worm holes in the GPS system, these are thought to be as a result of blocking and/or jamming of GPS signals.

Possible thought on how the worm holes are being created ...
[list=1][*]direct inband interference around 1.57-158 Ghz[*]harmonic interference (UHF TV)[*]intermodulation of two or more signals[*]suppressing or blocking using strong RF to desensitize the receiver[*]sidelobe interference[*]intentional jamming (read military)[/list=a]

Old 6th Jul 2001, 23:03
  #8 (permalink)  
Posts: n/a

You may find the

Global Positioning System Standard Positioning Service Performance Analysis Report (PDF, 284 K)



interesting reference material on availability and accuracy.

Neither seems to specifically address the issue of receiver failure or software issues, which I would guess play a significant part in the overall availability and accuracy of GPS.

[This message has been edited by bookworm (edited 06 July 2001).]
Old 7th Jul 2001, 01:35
  #9 (permalink)  
Posts: n/a

GPS has worked for me for over five years in airline and private aircraft without a hitch except......RAIM not available for about 20 minutes. So, into the hold until it was available.
Works really swell on the western side of the great divide. WAAS will be much better, LAAS better still.
Old 7th Jul 2001, 09:00
  #10 (permalink)  
Posts: n/a


I agree GPS is the best thing since sliced bread, but like anything we should know its limitations.

I have never experienced any problems, however some friends of mine that have flown commercially in the UAE area have had significant problems to say the least, sometime lasting for days.

Old 7th Jul 2001, 23:41
  #11 (permalink)  

Keeping Danny in Sandwiches
Join Date: May 1999
Location: UK
Age: 76
Posts: 1,294
Likes: 0
Received 0 Likes on 0 Posts

Don't know just now I will have a look next time I go to work; it certainly isn't to that accuracy!
sky9 is offline  
Old 25th Dec 2005, 01:01
  #12 (permalink)  
Join Date: Sep 2005
Location: Aus
Posts: 8
Likes: 0
Received 0 Likes on 0 Posts
I have a few GPS questions if anyone can help me. (Zeke- if you're still out there you sound like you know what you're talking about!) Anyone else is welcome, of course!

1. The yanks "switched off " SA in 2000, can they "switch it back on"? i.e is it really still a GPS error?

2. If P code is available only to the us military, why is it transmitted on L1 along with the C/A code that we all have access to?

3. What makes L2 different (other than the frequency it's transmitted on) if the P code is transmitted on L1 as well?

4. May I have a laymans definition of "ephemeris"?


Wow- I just looked at the date when I read my reply- how much of a nerd am I- posting on pprune questions about GPS on Christmas Day The things pilots do for that dream job!

Merry Christmas everybody.
dashtrashtoo! is offline  
Old 25th Dec 2005, 08:43
  #13 (permalink)  
Join Date: Apr 2004
Location: Europe
Posts: 467
Likes: 0
Received 0 Likes on 0 Posts
Thumbs up

Very interesting thread, but a bit to scientific for me as a user.
I only know as a pilot, that the navigation system uses the so called hybrid position:
Each IRS computes an hybrid position that is a mix IRS/GPS position called GPIRS. Among these 3 GPIRS hybrid positions received by each FMGC, one is selected according to a figure of merit and a priority. The selection is performed using the following hierarchy :
onside GPIRS position
opposite GPIRS position
If the GPIRS data do not comply with an integrity criteria, the GPS mode is rejected and radio position updating is used, "GPS PRIMARY LOST" message is displayed on ND and MCDU scratchpad.
In other words the GPS position could be the most accurate, but if its outside the integrity circle, the system will reject it and revert to the less accurate rad nav, weird isn't it?

As long as GPS primary is in use, all usual required navigation performance are met, that's what the book says.
Another interesting thing is also that (subject to authority approval) having GPS Primary one can shoot a non precision approach even if the ground facility is U/S.
Does anybody know what sort of prove the operator has to provide in order to be allowed to use the vertical profile in managed mode out of the box?
popay is offline  
Old 25th Dec 2005, 09:49
  #14 (permalink)  
Join Date: Dec 2005
Location: Forest of Caledon
Posts: 212
Likes: 0
Received 0 Likes on 0 Posts
dashtrashtoo! I'll have a go at answering your questions:

Q1. Yes, they can switch SA on again if they want to. It's their ball and they can take it home if they want to. In fact, Dubya has declared that he has the right to withdraw the service from public use if he wants to and that he might do so in time of (unspecified) national emergency. That guy does sometimes behave like an angry 12 year old with a gun so he can't be trusted not to do something stupid to the civil availability of GPS, but it's very unlikely that his minders will let him do so because full accuracy is so useful for so many people and so many companies.

Q2. The reason why the grownup receivers use two frequencies is so that the ionospheric delay can be measured in realtime. The satellites use a fundamental frequency of 10.23MHz, which is tweaked slightly to compensate for the effects of General and Special Relativity. The L1 frequency is 154 x 10.23 Mhz and the L2 frequency is 120 x 10.23MHz. The two signals are transmitted exactly simultaneously, but arrive at different times. The delay caused by the density of the ionosphere is proportional to the square of the frequency. (L1/L2)^2 By measuring the time difference between the twosignals you can identify and eliminate the pseudorange error caused by th ionosphere.

Cheapo receivers, such as the ones fitted to cars and the small handheld jobs that you can buy for a hundred quid or so, only receive L1. To compensate for the ionsopheric delay they use a bog standard mathematical model of the ionosphere which takes into account such things as latitude and time of year. It's a bit like ISA, it dosn't tell you what the actual parameters are on the day of the race, but it does help somewhat. The model typically reduces the error by about 50% but the dual frequency trick eliminates it almost totally.

Q3. Nothing. See A2 above.

Q4. Think of the ephemeris as being a flight plan for each satellite. It lists the orbital parameters for each satellite so your receiver knows which sats to listen out for. Each satellite transmits not only its own ephemeris but also that of all of its mates. With about 30 satellites in the constellation at any one time your 12 channel receiver needs to know which ones not to bother with and it needs to know which ones produce the best geometry for a good fix. As soon as it gets lock on one satellite the receiver reads the ephemeris for the whole fleet and therefore knows when a new sat is going to rise above the horizon.

Apologies for any typos in the above. I've had a few swigs while making the brandy butter.
Low Flier is offline  

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off

Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service

Copyright © 2024 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.