[time-nuts] Time Nuts at PTTI this past week
Tom Clark, K3IO
K3IO at verizon.net
Tue Dec 12 01:00:02 EST 2006
SAIDJACK at aol.com wrote:
how many M12M iLotus samples did you test to find the one with 30ns
offset? Were these pre-production samples or mass production units?
How did the other ones perform?
Seems iLotus may not test/calibrate/inspect as thoroughly as Motorola
I thought I had noted this in my earlier comments, so I'll try again.
So far, we have seen only one iLotus M12M sample (and one earlier
Motorola M12M prototype). The firmware for the Motorola unit reads
v1.0x6 (like maybe the 6th experimental test version of v.1 ???),
while the iLotus reads v1.0. If/when we see additional iLotus units,
we will run them thru the same procedure and see if a pattern emerges.
Art Sepin (Synergy -- the Motorola & iLotus rep -- see
http://www.synergy-gps.com/) was at PTTI and we discussed the
measurements with him. I have reason to believe that he has already
reported our results to the factory. Rick conveyed the message that he
(CNS Systems) will be happy to perform the calibrations to get rid of
the bias if they want.
The following are my personal comments/observations and should not be
taken as definitive or authoritarian: I have heard rumors that after
Motorola "sold" the M12M line to SiRF, most of the Motorola GPS
technical staff moved to SiRF. SiRF (as I understand it) then licensed
iLotus to produce the hardware (I guess as a build-to-print task). I
am unclear on who of the three has retained the rights/details of the
firmware, so I don't know from whence any future modifications will
come. The preliminary data sheet that iLotus has at
http://www.ilotus.com.sg (and that Art has on his web site) says
"Fully calibrated to UTC at USNO" but I am certain that the M12M has
never been at USNO; I presume this claim refers to Rick's earlier M12+
Since our entire sample is a universe of one, it is impossible to say
why the ~30 nsec offset is there. We do believe that the ~5 nsec
scatter in the M12+ observed in the four-unit 2002 tests came from
slight differences in the group delay performance of the SAW filter
that determines the IF response. Since I have never seen a detailed
block diagram of the M12M innards, I am unable to do any reverse
engineering to aid in understanding.
Of course, since the 1PPS epoch can be set in software, any
manufacturing biases can be easily removed by the user (if 30 nsec is
1. mailto:SAIDJACK at aol.com
More information about the time-nuts