[time-nuts] Odetics 365 date/1024 week problem fix
golgarfrincham at yahoo.com
Mon Feb 27 00:22:09 UTC 2012
I have owned an Odetics 365 GPStar Plus Time And Frequency System
for some time and it works fine except for the annoying date problem that
apparently has something to do with the fairly common 1024 weeks limit
where older GPS receivers display 2012 as 1992 and the day of the year
is not correct. I couldn't just change receivers because there appeared to
be some routines in the firmware that checked the receiver model and
interpreted the data for specific models that Odetics could have used in the
late 1990s. This 365 used a Magellan 5000 OEM receiver.
I recently got a good deal on a FEI-Zyfer (Odetics) 565 that looks like the 365
and it was sold as-is because there was just a backlight but no display and the
keypad was dead. What it turned out to be was the display contrast was turned
way down and the 'beep' was turned off. The unit worked properly but as it was
received it did look dead. I bought the 'dead' 565 hoping that maybe I could
compare the two units and find a way to fix the 365 to display the date properly.
Getting a fully operational 565 was a bonus.
What was very interesting to me was that the 365 (~1998) and the 565 (~2005)
had basically identical motherboards. The only difference I could see was one
27c010 eprom and the receiver and its adaptor board. I put the 565 eprom and
receiver into the 365 and it became a 565 displaying the correct time and date
and working exactly like the 565.
The receiver in the 565 was a Motorola M12+ on a Synergy Systems adaptor board
that made the M12+ look like an older and larger Motorola UT+ board to the 565. I
tried a UT+ in place of the M12+ and it worked fine and the unit identified it properly
as a UT+ V2.2. I then duplicated the eprom and made an adaptor board for the
UT+ board because the Synergy Systems adaptor board for the M12+ would be
too difficult to copy. The adaptor board I made just has the UT+, a 7805 3-terminal
5V regulator, a back-up super cap, and a 7404 buffer (1 inverter driving 3 in more
in parallel) for the 1 PPS out. The RX and TX from the UT+ receiver are also
connected so the 365 with the new eprom and receiver sees all the proper signals
and works as a 565 should. I've attached a photo of the M12+ factory made board
and the UT+ board I made so you can see what they look like. I thought that others
who have run into this same problem might find this information helpful.
More information about the time-nuts