[time-nuts] Fwd: [hpsdr] SDR experiment for the solar eclipse

[jimlux]

On 8/4/17 7:22 AM, Tom Van Baak wrote:
>> We were originally going to put a 5071A-locked beacon on three ham
>> bands, but decided WWV and CHU would be better sources, and logistics
>> were turning into a problem: I'm going to be doing my wideband recording
>> from a cottage in northern Michigan.  But I'm still a time-nut, so the
>> receiver will be GPSDO-controlled, and there will be a stratum 1 NTP
>> server in the cottage to provide timestamps. :-)
>>
>> John
>
> Hi John,
>
> My favorite write-up about atomic clocks and eclipses (a null result) is at:
>
> http://www2.mpq.mpg.de/~haensch/oldStuff/eclipse/eclipse.html
>
> There you will find a good summary, thorough methodology, and many plots for the 1999 eclipse. Plus they posted all the raw data (H-maser, cesium, rubidium), a time-nuts dream. There is no model for why an eclipse should affect time at the atomic (quantum) level so a null result is fine. If nothing else, it sets an upper bound on measurement precision or a lower bound on clock anomalies, if they exist.
>
> Much more dramatic is what an eclipse might do to the ionosphere, as this may affect both GPS and HF radio. So I'm very please to see the ham community milking this rare opportunity for all it's worth.
>

It is of great interest - the fact that we have an ionosphere (which 
helps keep us alive) also is one of the bigger factors in accurate time 
distribution from space since the medium is refractive - not only does 
it change the speed at which a wave propagates through, but the 
propagation path is not a straight line.

The ionization is almost entirely due to the sun. There is some small 
effect from terrestrial upper atmosphere phenomena propagating upwards, 
and then there's also human caused changes (HAARP and other heaters, 
rocket launched clouds of easily ionizable material, and the occasional 
high energy nuclear reaction in the upper atmosphere/space).

There's a lot that is unknown about the time constants of the ionization 
and deionization, since the usual situation is that you gradually reduce 
the solar input (sunset) and increase it some 12 hours later. And 
there's not, often, the chance to have a "step function" in the solar 
illumination vs horizontal distance.

The real challenge is that it's a short event - with conventional 
ionosondes, the sweep is minutes long (which is fine given the usual 
slow diurnal variation), which is significantly longer than the duration 
of totality. That long slow sweep makes synchronization less critical - 
if you start off by a few milliseconds, as long as your receiver 
bandwidth is wide enough, you still capture the sweep.

It's not going to happen this time, but something where "all frequencies 
get measured every few seconds" would be wonderful. There are some 
ionosondes using a broadband PN waveform.  And there's some interesting 
propagation experiments that have been done over the Cascade mountain 
range using TV and Radio stations as a form of pseudo random transmit 
signal.

But you could do a real interesting science experiment with several 
dozen (or hundred) RTL-SDR receivers, a good oscillator for timing, and 
a data logger spread over several thousand km. In conjunction with the 
existing ionosondes, you could collect a lot of oblique measurements of 
the ionosphere during the eclipse.

Then would come (years of) post processing -