[time-nuts] Re: HP Z3801A project update

[ed breya]

I just did another experiment with the new deal. I shut it down for 
about an hour, to get fairly close to a fully cold start. Here's 
approximately the sequence of events and timing:

T=0, power on
  SatStat establishes com
  GPS time 04:00:00 and counting
  Date 01 Jan 2006
  "OCXO warmup"

T+2 min
  "GPS acquisition"
  PRNs listed, blank data

T+3 min
  PRN data starts to show
  GPS time correct
  Date correct 25 Feb 2022
  satellites begin tracking
  "fine frequency adjustment"
  "leap second determination"
  "stabilizing frequency"

T+5 min
  "1PPS synchronized to GPS time"
  Lock indicator on
  "stabilizing frequency"

T+35 min
  "stabilizing frequency"

So, it took about 5 minutes to get started and achieve lock, and the 
date is right. I can't figure out why the 01 Jan 2006 initial date 
allows it to apparently work. I've looked at rollover info online, and 
ballparked some numbers - nothing makes sense to me to explain it 
numerically. I suppose there could be some integer math errors/overflows 
or corrupted data (or even a bug) in the RX, due to forcing different 
numbers, that just happened to make it go.

As of right now, it still seems to work right and get the right answers, 
after a number of power up cycles. It will be interesting to see what it 
says tomorrow. Maybe it will start counting backwards or something.

Anyway, it's possibly fixed, but I don't know, and maybe never will.

I won't be able to do anything on it this weekend - I have to go up to 
my sister's place and try to fix her microwave oven and other projects. 
Sunday night, if there's any point to it, I can somewhat explain what I 
did. I took some notes, but they don't make much sense either.

Ed