Re: Relativity is a pseudoscience II. The Hafele-Keating HOAX,

On Wed, 18 Sep 2024 15:53:44 +0000, rhertz wrote:

======================================================================
The HP 5360A Computime, introduced in 1969, measured time intervals of
0 to 1000 seconds with 1 nanosecond resolution. The internal crystal
oscillator offered a measurement accuracy of 1 part in 10^7. It could
also be connected to an external 1 MHz or 5 MHz reference oscillator
to improve long-term accuracy and stability. It could average multiple
time interval measurements to reduce noise and improve accuracy over
a series of measurements. It had a 9-digit Nixie tube display.
**********************************************************************
>
>
This is a link of a site with huge amount of data about the HP 5360.
It worked with one of the first programmable calculators ever done,
the 5375A Desk Calculator Keyboard.
>
You can see that it had 11 nixie tubes in the display, but it was based
on statistical averages (new techniques), to present data like time
interval.
It was also a frequency meter (0 Hz - 320 Mhz).
>
>
https://www.crowave.com/blog/2022/11/24/computing-counter-hp-5360a/
>
Check this link.
>
https://dopecc.net/calculators/hp/5360a/
>
>
When introduced in 1969, was 100 times more accurate than its
competitor, BUT
it provided time differences between two sources IN A SINGLE SHOT.

You agree, therefore, that measurements of the time differences
between the 1 PPS outputs of the cesium clocks was quite feasible.

So, to ACCUMULATE DIFFERENCES during the time flight of 50 hours, A
LARGE NUMBER
of single shot measurements HAD TO BE TAKEN, and such set of
measurements (> 4.500
in this case), HAD TO BE STATISTICALLY PROCESSED outside the ensemble.

These thousands of inter-clock measurements were required for
correlated rate-change analysis.

And that is one of the things to be questioned heavily, because it was
when the
data cooking took place (in Washington DC, once the experiment was
finished).
Data "cherry picking) at its best.
>
NOT FOR NOTHING, THE SUMMARIZED DATA (AFTER COOKING IT AT WILL),
APPEARED IN A SECOND
PUBLICATION BY 1972, WHILE THE ORIGINAL 197 PUBLICATION ONLY HAD A
CONFUSING GRAPHIC, NO MORE.

There was not enough room in the short article to publish thousands
of data points. Nowadays, one expects the data from such an experiment
to be made available on some public server. For example, I've spent
enjoyable hours going through Kepler data and LLR data, testing out
various ideas of mine. But in 1971, there was no Internet, and data
storage was relatively expensive.

The stability of the HP 5060A Cs atomic clocks was not enough (in the
short and long terms) to
provide such accuracy DIRECTLY in 1971. The published results were more
a product of desktop
calculations and corrections THAN a real, continuous, full time
accumulative measurement for 50 hours
or more.

Actually, the data was analyzed using an existing computer program
used by standards laboratories specifically designed to perform
correlated rate-change analysis. You cannot expect these calculations
to have been performed manually.

Today, 53 years after, the experiment could be done with just one
measurement for the entire flight,

Yes. And modern re-enactments of the H&K experiment clearly
demonstrate relativistic effects.

but not in 1971, when the measurement instrument (without long term
memory) was invented at HP, in
their golden years.
>
So, I sustain that such experiment was an HOAX, in particular for the
interests of the MIC (Navy, etc.)
to support the experiment, and get more funds for time measurement, use
of Cs clocks all over the military, etc, etc.

The published graph looks rather crappy, but even without correlated
rate-change analysis, I can perform a simple one-tailed statistical
test on the graph. I enlarged the image and measured line centers on
the screen. I can exclude the null hypothesis (that there is no
difference between the accumulated times on the westward versus
eastward flights) to better than a (p < 0.05) level of significance.
I don't QUITE reach a (p < 0.01) level of significance. Basically,
the data supports the assertion that the accumulated times on the
westward flight are greater than on the eastward flight.
Correlated rate-change analysis allows CONSIDERABLY more to be said
about the magnitude of the measured effect.