About Fading

[T3sl4co1l]

If you're confused by atmospheric fading (mostly in shortwave radio), consider this:

Imagine looking out on a body of water, a river or lake or whatever. You're watching the sun glint off the gentle waves.

Imagine that surface, upside down, and at an altitude of 30 km or thereabouts. You're sitting inside an inverted fishbowl, watching the waves on the surface.

The waves are much wider apart than for water, because the atmospheric layers are much less dense than water. Also, instead of sunlight, we see in radio waves.

And you're watching the whole sky through your pinhole of an antenna, so the intensity is just whatever waves are glinting off the ionosphere as it sloshes about.

Tim

[chris_leyson]

Instead of sunlight in your analogy maybe think in terms of coherent light instead, then you could see the interference patterns. It's amazing to watch a spectrogram of selective fading on HF on say a wideband DRM channel where the power spectral density is supposed to be uniform. The deep notches that wander back and forth across the channel can sometimes be quite narrow and other times you can loose the whole channel. I like the inverted fishbowl analogy 

[T3sl4co1l]

Yeah, even better is the image of a laser speckle pattern -- it's probably not as familiar of an experience, though!

The pattern of light cast by the reflection from a pool is also good to think about (but again, missing interference).

The interference pattern suggests a possible remedy: use an array of several antennas, say odd fractions of a wavelength apart.  Attach a receiver to each, preferably with synchronized LOs.  The combined detector outputs (should be fine to mix with a regular audio mixer) will be in phase* and, as the patchy "light" cast on your array moves about, one area will dim while the other gets brighter, and your antenna array spanning those patches will pick up a nearly constant amplitude.

*Unless multipath from layers (different altitudes) occurs, in which case some echo or notching can occur in the audio band.

This is called diversity, and is the basic reason why it is so wonderful.  Downside, multiple receivers and a huge amount of space required for all the antennas!


I wonder if this analogy is actually physically correct -- I think the time and length scales ought to be typical of gravity waves among atmospheric layers?  But, I know there's a lot of plasma physics going on in the actual ionosphere.  It could just as well be that layers come and go, as radiation showers down erratically.  Or a combination of both.

Tim

[mark03]

A modern perspective is that multipath is a *good* thing.  It enables > 1 independent channel (still sharing the same frequency) between TX and RX, using MIMO techniques.  With this point of view, old-fashioned diversity reception is the spatial analog of repetition coding in time:  not particularly efficient, and you can do much better using proper error-correcting codes in both space and time.  Hence the rash of cool-sounding journal articles one or two decades ago about "space-time coding"

Classic case of taking life's lemons and making lemonade!  Not so practical on HF though.

[tggzzz]

It is simpler than that: if there was no multipath then cellular phone systems would not work!

[rhb]

A better analogy to fading is rotating a tuning fork close to the ear.

[bson]

That's a nice analogy.

But what's truly amazing is it's the equivalent of putting an omnidirectional 100W light on my roof, use it to modulate my voice, and someone 3000 km away can demodulate it and we can carry out a conversation by watching our reflections off the sky.  THAT is freaking incredible, and would be even if the sky were a perfect mirror!

[rhb]

Back when I was licensed I had a 6L6 and a rock with an RF-2200 portable for a receiver.  I was quite amazed that someone far away could tell when I turned the 6L6 on and off.

The sad part is most people don't see the magic.  Long ago I got a used 7" reel to reel.  I had a Bob Dylan tape and a Leonard Cohen tape.  I would sit and listen to the music and reflect on how the tape was made, the head was made, the head AC bias, amplifiers and speakers and was just enthralled.  I made the mistake of mentioning this to someone.  Their response was, "What's the big deal? It's a tape recorder.  You turn it on and sound comes out."

[T3sl4co1l]

Back when I was licensed I had a 6L6 and a rock with an RF-2200 portable for a receiver.  I was quite amazed that someone far away could tell when I turned the 6L6 on and off.

The sad part is most people don't see the magic.  Long ago I got a used 7" reel to reel.  I had a Bob Dylan tape and a Leonard Cohen tape.  I would sit and listen to the music and reflect on how the tape was made, the head was made, the head AC bias, amplifiers and speakers and was just enthralled.  I made the mistake of mentioning this to someone.  Their response was, "What's the big deal? It's a tape recorder.  You turn it on and sound comes out."

There's a story in The Pleasure of Finding Things Out, where Feynman befriends an artist and they're having a discussion about beauty.  The artist finds beauty in the colors and shapes of a flower, and finds a deeper analysis annoying at the very least, if not outright ugly.  The scientist replies, he can see the flower not just for its superficial beauty just as the artist does, but for everything it's made of, the interactions of cells, the molecules they're made of, and the atoms and subatomic particles and fields they are in turn made of.  It's beauty all the way down.

Different things for different folks, I suppose.

Tim

[rhb]

The Romans had many gods and priests to intercede upon behalf of the public.  Now we call the priests "scientists and engineers".

[chris_leyson]

Must admit I like Tims' inverted fishbowl analogy with maybe waves or ripples on the surface. It got me thinking about what the ionosphere might look like if we could model the reflecting surface at any particular frequency. I know it's not going to be an abrupt boundary, maybe more like a gradual change in refractive index for radio waves. One thing is for certain, it's going to take one hell of a lot of measurements before you can even think about generating a model.

Last time I listened out on HF, probably a decade ago, I was keen to see how many DRM stations I could pick up.
One thing I noticed, looking at a "waterfall" display, was that if you zoomed in on a 10kHz wide DRM channel you could literally see selective fading within the channel, it was quite captivating and I was quite surprised by how narrow the fades or notches are. DRM is COFDM modulated so the transmitter has a flat power spectrum across the channel, it makes for a good "illumination" source.

You would intuatively think that the same mechanisms that generate fades would also generate peaks, but I can't say I've ever seen them. Switching to a linear display rather than log woulg help, it's only going to be 3dB at most.

On the subject of COFDM, a long time ago I counted nearly 100 ALE or STANAG transmissions on HF during a NATO exercise in just a few hours, next week the Vostok 2018 war games are on so I'm going to have to listen out.

[rhb]

As a guess, I suspect that in some cases the fading is related to ripples in the ionosphere induced by fluctuations of the geomagnetic field from solar wind.  When doing a geomagnetic survey you need to monitor the field fluctuation during the course of the day so that corrections can be applied.

This is done either with a recording system or by taking repeat readings at some point during the day. By it's very nature, a proton precession magnetometer is a low frequency device, so it won't see the sort of fluctuations that would lead to fading.

It would be very interesting to have a large array of vertical incidence ionospheric sounders collecting data across a square a thousand miles on a side and sending them to a central location via the internet.  A very low power direct sequence spread spectrum signal with a correlation period of 30-120 seconds should produce interesting data.  By applying compressive sensing a few thousand sites should be able to produce a very high resolution picture.  If you could get enough processing gain to do it with 10 mW out or less, you might be able to get FCC to allow it.

If WWV goes off the air, the sounders could all operate on 5, 10 & 15 MHz using individual PRN sequences that had low coherence.  Use GPSDOs to control timing and frequency.

[T3sl4co1l]

Or you can just look at HAARP data*.  They (and related literature) is probably at a high level of understanding these things.

*I haven't, and... have no idea what kind of public research they have out there, actually.  I've got to imagine they understand layers and conductance better than anyone though..

Tim

[rhb]

I was thinking of real time imaging of the ionosphere. Reduce the data and stream the image over the internet.  I'm not aware of any program that has that capability.  It takes a lot of sounding stations and prior to the development of compressive sensing, would not have been practical.

At present there appear to be fewer than 2 dozen HAARP sites.  But they're using enough power to affect the ionosphere.  That's a rather different line of research.  It's also apparently upsetting the tin foil hat crowd.

There is an abundant literature going back  a long way.  There's a humorous anecdote.  In the 50's the US was testing some early spread spectrum gear which was very highly classified.  Of course, they also knew that the Russians could hear it.  One of the researchers attended a meeting and presented a paper on ionospheric sounding using noise autocorrelation to measure the boundaries.  By all reports the Russians believed it.