[time-nuts] TymServe 2100 and the 1995 GPS issue
gerhard at wittreich.org
Tue Jun 9 17:41:53 UTC 2015
Very good point...It is a new oscillator that has been sitting in a box for
10-15 years. I'll give it a few weeks to settle in. The other very
interesting thing I noticed was the variability in the oscillator control
voltage before and after installing the new Heol Design GPS. The moving
average of the standard deviation of the oscillator control voltage before
and after installing the Heol Design gps was striking. The variability of
the control voltage for the Heal Design is nearly 1/3 of that for the
original Trimble gps. That suggests that the pps signal from the new gps
module has 1/3 the variability vs the original gps module. I did consider
that this could be due to the new oscillator aging but when I do a linear
regression on the standard error of the control voltage before replacing
the gps module the p-value for the slope of the line is about 0.5
suggesting that the null hypothesis, that the slope is actually zero,
cannot be rejected. Therefore, one would need to conclude that there is no
change if control voltage variability vs time prior to the installation of
the new gps module and, consequently, there was no change in the stability
of the new oscillator (since everything else remained unchanged).
Unfortunately, I do not have sufficient run time on the new gps module to
reach any meaningful conclusion of variability vs time. If anything it
still appears to be dropping at this point. I can further update once I
get additional run time.
I also noted a very marked improvement in the satellite signal strength
reported by the TS-2100. While I don't have any statistically sound data
my observation is that the signal strengths are significantly higher and
there are always more viable satellites in view. In the past I would never
see more that 6-8 with 4-6 having sufficient signal strength. Now I
routinely see 10 with 6 - 8 having sufficient signals strength.
[image: Inline image 1]
Note: The straight lines represent the mean value and are not linear
regression although the linear regression and the mean line are very nearly
the same for the original GPS data. I suspect the same will happen once I
get some run time with the new GPS. Both imply that there is little time
dependence, in the short term, on control voltage variability.
Gerhard R Wittreich, P.E.
On Tue, Jun 9, 2015 at 8:59 AM, Esa Heikkinen <tn1ajb at nic.fi> wrote:
> Gerhard Wittreich kirjoitti:
> Once locked I did notice
>> something interesting but maybe not important. The control voltage to the
>> oscillator (timing > util > tfp 0) went from a very consistent 0xba6f to a
>> 0xbb02. That's about 0.2% of scale. I'll watch it over the next few days
>> to see where it settles in.
> Sounds like oscillator aging to me, sometimes they do larger jumps, stay
> there for a while and then again... Also it would not be suprising that
> this would happen just after the power was off and oven was cooled down for
> a moment.
> Also, brand new oscillators seem to age more rapidly at start and then
> settles down when older but the effects of aging will never stop
> completely. But they seem to get better and better when aging.
> I'm also waiting my MTI OCXO to arrive...
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to
> and follow the instructions there.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 26685 bytes
Desc: not available
More information about the Time-nuts_lists.febo.com