[time-nuts] Measuring sidereal/solar time? Re: A Leap Second is coming

[jimlux]

On 12/30/16 9:53 AM, Ilia Platone wrote:
> Brooke,
>
> The problem in radio ground observation can be resolution accuracy, but
> there's also a good transmission into far infrared wavelengths, which
> could require smaller dishes to get stellar images. The problem of far
> IR is the cost of right filters/sensor, which are a bit difficult to find.
>
> Radio objects, on the other hand, can be solved using an interferometer:
> LOFAR interferometers work at frequencies higher than 10MHz, frequencies
> totally transparent to the atmosphere and easily computable even by
> consumer PCs. There is some work done with common PCs using two RTL-SDR
> dongles and two satellite dishes.
>


the earth's ionosphere is hardly perfectly transparent at frequencies 
below, say, 10 GHz.   The effect is small at GPS L-band frequencies 
around 1.5 GHz, but still large enough that you need to either make 
measurements at two frequencies (so you can calculate the effect) or use 
other data, if you want accurate "sub-meter precision" data.

At HF, the effect is huge: during daytime, you might not even be able to 
see the signals you're looking for, either from D-layer absorbption or 
F-layer reflection/refraction.

The real challenge at HF (e.g. LOFAR) is that it's not just a time of 
flight thing, because the propagation is not in a straight line: the 
anisotropic ionosphere bends the rays: and even better, the bend depends 
on the polarization.  For GPS, the signal is CP, and the effect is 
small, so they typically look at it as an overall propagation speed 
effect. At HF, the effect is so large that's not really valid.

The ionosphere is also only stable on a time scale of <1 second: that 
is, on a HF skywave path (and by inference, on a HF "through ionosphere" 
path), signals are pretty much decorrelated at time scales greater than 
3 seconds as clumps of ionization move around.  This is the fundamental 
accuracy limit on things like the ARRL Frequency Measuring Test (FMT).

It's true that you can do the interferometry easily on a PC, but taking 
out the ionosphere effect is tough, unless you carefully choose 
observing time and avoid high solar activity events, etc.

I suppose one can do some sort of inversion process on measured data 
from known sources at multiple frequencies to infer the ionospheric 
structure, but this is a *hard* problem.


If you want to use RF interferometry, I'd go higher:  maybe Ku-band- 
cheap electronics and dishes available.  There's some water vapor 
attenuation, and I'm sure that changes the propagation speed a bit too, 
but it's measureable with radiometry, you can easily tell whether there 
are clouds in the path with a pretty simple Ku-band radiometer.  You'd 
want to throw out days when there's rain.

I don't know if there's any useful celestial sources at Ku-band.  DSN 
uses bright quasars as pointing & timing reference when doing Delta 
Doppler One way Ranging (Delta DOR) but on the other hand, they're also 
using cryogenic receivers with 34 meter apertures- something not 
available to the casual (or even dedicated) amateur.