[time-nuts] Oooold GPS receiver discussions Austron 2201 as an xample

Magnus Danielson magnus at rubidium.dyndns.org
Tue Nov 10 21:06:22 EST 2015


On 11/10/2015 08:58 PM, paul swed wrote:
> There is a good discussion about old GPS receivers that have been running.
> Its the NAVSTAR proteus thread.
> Very good details in that thread about some issues. Such as the 1024 week
> rollover and that the receiver should still keep working. Though the date
> and time would be wrong.

I was surprised that it resurfaced, as we have been beating that issue 
to quite some depth multiple times here.

> Given that theory I fired up the 2201. This is the unit that had no down
> converter and I home brewed one using another GPS receiver. Put simply it
> all worked.
> But its always been a bit magical. Wait for a full moon, jump 3 times
> counter clockwise...
> Never made sense as to why it worked or really controlling it consistently.
> (Why even care for this pain in the ....? Because it has a really good
> frequency offset measuring system.)
> Over time its become more difficult to get it to work even though the down
> converter is having no problem acquiring and locking to satellites.
> I think and its a theory its because its looking for SVNs but with the
> wrong PRN. (these would be retired SVN/PRN sequences)The trigger was
> looking at the satellite it wanted and realizing the PRN was not accurate
> and the satellite wasn't even up.
> I need to do some more digging but perhaps over time there are fewer and
> fewer correct PRNs that align to the SVN.
> Does this make any sense at all?
> There seems to be no way to tell the system that SVN X = PRN Y

Why should there be? In the GPS signal structure, the SVN number has 
_no_ bearing on achieving lock. It's all PRN codes, and as SVN numbers 
increase with launches, the PRN codes is reassigned.

What is relevant is the almanac indicating the rough orbit of all 32 
PRNs. As you wake up from a long and deep slumber, any remains of 
almanac will be mostly outdated. Once you have locked onto a single 
bird, it takes you 12.5 min to recover the full almanac. If you have or 
don't have a hint of where you are can lead you off track for some time 
if you have old data. Depending on how good the recovery algorithms are, 
the time to lock can vary a lot.

It might be that it just needs to sit there until it locks a bird, and 
then acquire channel after channel.

Some receivers needs a proper reset to get going with good speed.

> The only other fake out approach as a test will be to find a PRN thats up
> and see if that aligns to a SVN.
> Then manually enter that SVN as a search.

No need.

73 de Magnus SA0MAD

More information about the time-nuts mailing list