[time-nuts] Re: HP Z3801A project update

[ed breya]

Last night I told it to do a survey, and let it slog through it 
overnight. This morning it seemed to be working fine, and apparently in 
position hold mode.

Then I shut it down for about 45 min to let things cool down to nearly a 
fresh cold-start condition. On power up, it went through the usual power 
routine, and almost immediately began tracking satellites after the 
inner oven warmup, then within a few minutes, the GPS lock indicator 
came on. Also, before tracking even started, the position coordinates 
reported exactly the same as before power down. So, now this RX unit 
appears to be behaving "properly," just like the other.

This RX unit was probably OK all along as it was, but I just didn't give 
it enough of a chance to catch up with everything. The question is why 
these two units, virtually identical, and neither having any long term 
memory, and operated in the same setup and location, would act so 
differently at first. There's no battery backup of the big SRAM, and I 
don't see any NVRAM evident. There are two big Flash RAMs, but I believe 
they only hold the firmware, so are not written to in normal operation. 
Maybe they are?

I've been wondering where the actual, necessary data must be kept in 
order to get things working quickly like this. Since after a power down 
and up, the exact same location was recovered before any tracking was 
possible, it has to be stored somewhere in the RX or the Z3801A. I 
suppose the Z3801A could have some NVRAM - I haven't looked closely 
enough yet to see, but I will next time it's opened up.

Now that this spare RX seems to be verified OK, my next experiment will 
be to swap them again and see how the original works first time up, with 
the Z3801A used to working with the spare. If it acts very slowly, then 
I'd surmise that the Z3801A treats it differently, because its ID number 
is different, and has to figure everything out again. If it still runs 
easy and quickly, then I'd have to think the data is in the RX, or that 
the Z3801A doesn't care, and it is the keeper of the data. If the latter 
though, then it's a mystery again - the Z3801A should have told the 
spare RX right away the location, quite accurately, and whatever else 
was necessary, so it should have been able to get up and go quickly, 
right from the first time.

Ed