Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

How does standalone GPS receivers ensure an acucrate clock?

Status
Not open for further replies.

kel8157

Full Member level 2
Joined
Nov 14, 2007
Messages
132
Helped
5
Reputation
10
Reaction score
3
Trophy points
1,298
Activity points
2,017
I understand that for a cellular phone based GPS, one can calibrate the clock against the synch channel timing from the GPS/WCDMA to under 0.1ppm..
However for standalone GPS receivers, there is no cellular clock to calibrate against. How do the designer get arround that?
 

Mityan

Full Member level 5
Joined
Jul 11, 2012
Messages
281
Helped
48
Reputation
96
Reaction score
47
Trophy points
1,308
Activity points
2,835
For GPS receiver there no cellular clock, but maybe there is a space segment - the network of satellites to calibrate against, how you think? It just receives their signals
 

kel8157

Full Member level 2
Joined
Nov 14, 2007
Messages
132
Helped
5
Reputation
10
Reaction score
3
Trophy points
1,298
Activity points
2,017
For GPS receiver there no cellular clock, but maybe there is a space segment - the network of satellites to calibrate against, how you think? It just receives their signals
my understanding is an expensive TCXO is used to calibrate against cellular clock, then used to receive the GPS signal.. so before the GPS signal is received and decoded, how do you ensure the clock is good? GPS signal itself suffers doppler, atmospheric distortions and many more..
 

Mityan

Full Member level 5
Joined
Jul 11, 2012
Messages
281
Helped
48
Reputation
96
Reaction score
47
Trophy points
1,308
Activity points
2,835
Your understanding is not right.
The GPS receivers are very cheap, so they dont have a good TCXO. They extract time info from navigation data, but before this they syncronize with the system by receiving the code (C/A for civil, P(Y) for military). 1.023 MHz of C/A code provide 1 us precision but in general much better is achieved.
If you know what time is it now and where are you then you know doppler shift for every satellite because they accurately move on the definite track.
If you dont know your position and time you should find the right satellite signal by looking all over the possible doppler and time shifts (ambiguity function 2D).
All other distorsions are overcome by pseudo random code accumulation.
 

RCinFLA

Advanced Member level 2
Joined
Aug 31, 2010
Messages
590
Helped
187
Reputation
378
Reaction score
186
Trophy points
1,323
Activity points
5,698
GPS receivers have a pretty good TCXO. Usually they are +/- 0.5 ppm spec which are better then ones in most cell phones.

The TCXO does not run continuously and when the GPS receiver is off, the time keeping is a standard 32.768 KHz crystal just like in a PC or digital clock. These are about 100 ppm accuracy.

GPS has many levels time keeping that can accelerate lock times. The first level is being able to predict what satellites are overhead therefore being the best bets to run the correlators against looking for their PRN code. If location is changed significant, like flying on a jet plane to another continent, then this predictiion can actually degrade lock time because many of the satellites may not be in view causing the correlator resources to be wasted.

Part of the satellite location prediction involves setting expected frequency Doppler shift. There can be as much as +/- 3 KHz Doppler shift for satellite coming or going over the horizon. For 1575 MHz, that is equivalent to +/- 1.9 ppm offset in frequency. The digital correlators have a limited frequency lock range of between 300 to 500 Hz so the receiver must 'tune' frequency bins by moving the synthesizer across an assumed frequency error range based on estimating satellite location. If not finding lock after doing all the frequency bins, the software may assume the satellite is not there and may move on, allocating the correlators to another satellite. The worse the time accuracy, the wider the frequency search range must be because of the inaccuracy in predicting the satellite location and its frequency offset.

A GPS receiver keeps track of the last time it was active and computes how wide the time inaccuracy may be based on the 32 KHz clock error. It will widen out its frequency search range the longer it has be turned off to account for the possible time error. This can lengthen time to lock.

For a CDMA cellular network with full GPS assist, satellites visible, precise time to <1 msec, and frequency AFC is used to achieve almost instantaneous lock on satellites. If absolute time is known to better then 1 msec along with satellites health then the GPS receiver does not have to demodulate the 30 baud data which can take several minutes to derive all necessary information to lock on to other satellites. The correlator bandwidth can also be reduced to improve sensitivity.

Newer GPS receivers have more channels and correlators so they can search for more satellites simultaneously and spread more correlator 'traps' over time and frequency to speed the lock process. Tightening down the correlator bandwidth improve sensitivity but increases the number of frequency search bins. Having more correlators offsets the downside of larger quantity of freq search bins, speeding the searching process while accomplishing improved sensitivity.
 
Last edited:

kel8157

Full Member level 2
Joined
Nov 14, 2007
Messages
132
Helped
5
Reputation
10
Reaction score
3
Trophy points
1,298
Activity points
2,017
Your understanding is not right.
The GPS receivers are very cheap, so they dont have a good TCXO. They extract time info from navigation data, but before this they syncronize with the system by receiving the code (C/A for civil, P(Y) for military). 1.023 MHz of C/A code provide 1 us precision but in general much better is achieved.
If you know what time is it now and where are you then you know doppler shift for every satellite because they accurately move on the definite track.
If you dont know your position and time you should find the right satellite signal by looking all over the possible doppler and time shifts (ambiguity function 2D).
All other distorsions are overcome by pseudo random code accumulation.
Yes, i understand the correlators used to find the satellite etc, the bigger the TCXO deviation, the longer time one needs to search for the visible satellites..

And thanks RC for the explanation. :smile:
 
Last edited:

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top