[time-nuts] The Demise of LORAN (was Re: Reference oscillator accuracy)

Magnus Danielson magnus at rubidium.dyndns.org
Mon Nov 16 01:49:10 UTC 2009


Mike,

Mike Monett wrote:
> I said nothing about the type of modulation. The equipment I listed is
> designed specifically to disrupt GPS. Presumably they use whatever
> modulation method that gives the best results.
> 
> However, GPS is spread-spectrum, so it  inherently rejects noise that is
> not correlated with the satellite signal. This means effective jamming
> requires a lot more power than is available from a 9V transistor radio
> battery, and even then, the range is only a few meters.

The power of de-correlation is related to the length of the 
pseudo-random sequence. This relates back to the codes auto-correlation 
shape.

The C/A code has a length of 1023 chips, giving a correlation gain of 
about 30 dB.

The P code has a length of 7*86400*10,23E6 chips (a GPS week) but only 
fraction of that can be used for de-correlation.

Looking at the detailed codes, the vunerability of C/A codes is worse, 
for PRN 1 the suppression is only -22,71 dB at 42 kHz sideband.

But these coding gains comes to no use if the input is saturated for 
most of the times. Simple AGC strategies would allow the AGC to be 
captured by the CW signal for instance. Lack of filtering would allow 
out-of-bandwidth signals to infect the input etc. etc.

Pre-digitalization code de-correlation is less sensitive to jamming, but 
comes at the price of 3 times number of channels analog decorrelators 
and integrators. Then again, it only solves part of the problem.

The tracking threshold of the particular receiver in its environment has 
a huge effect. Support systems such as clock, movement sensors etc. can 
allow for tighter loop filters and thus allow the channels having a 
lower threshold and that would aid in the jamming resistance too.

Regardless, the code-lengths isn't a particular good measure on the 
resistance to jamming. The P(Y) code is only about 10 dB better than C/A 
code in that respect, and M code gives another dB or so.

Cheers,
Magnus




More information about the Time-nuts_lists.febo.com mailing list