The ionosphere at VHF

The following discussion about the influence of the ionosphere on the EME signals took place in the Moon-Net reflector. I found the subject so interesting that I decided to put together all the messages in this page. (Click here for related books)

On 29-Dec-1997 SM5BSZ wrote:
While cleaning up in the book shelf to get some space for the bunches of more recent material, I came across a conference proceedings (RVK 1978, Stockholm) from which I have saved the following information before getting one new centimeter of shelf space:
  *  *  *  *  *  *  *  *  *  *  *   *  *  *  *  *  *  *  *  *  *  *   *
Influence of the ionosphere at VHF:

                                100MHz      200MHz
Change of phase path length     -400m       -100m
Change of group path length     +400m       +100m
Refraction                       0.02deg     0.005deg
Phase change                   -48000deg     -24000deg
Frequency shift                    6Hz        3Hz
Time shift                      1.3microsec   0.3microsec
Polarisation rotation           380deg       95deg
Absorbtion                       0.05dB       0.012dB

The data is valid for a signal travelling one time through the ionosphere at 90 degrees elevation. The integrated electron count
along the signal path is assumed to be 10^17 (100 000 000 000 000 000) electrons per square meter, which corresponds to about 5.5 MHz as the critical frequency for the F layer at 90 degrees.

Due to normal variations (sun activity, time of day,latitude...) the numbers vary by a factor of 10 up or down. At an elevation of 10 degrees all numbers are about 4 times larger.
  *  *  *  *  *  *  *  *  *  *  *   *  *  *  *  *  *  *  *  *  *  *   *

The above is 20 years old, but maybe it is interesting to some members of this group. We immediately see that Faraday rotation should often be many turns at 144MHz, so we can expect it to vary rapidly. BUT we should practically never see any attenuation due to the ionosphere.
The EME conditions do not change due to ionospheric attenuation.

The data comes from Harald Derblom at Uppsala Jonosf?observatorium.

On 30-Dec-1997 W7EME wrote:
This is probably fundamental info to some, but is kind of exciting to me. Leif, your postings are always of such fascinating stuff.

Along the line of Faraday rotation, I am curious about the following:

My limited understanding of the phenomena is that it is a relation or maybe function of my signal path with our magnetic field (our Earth) and the thickness of the ionosphere at the particular moment of ops.? It also appears to me that if my signal was at a 90 degree angle to our magnetic field that maybe no Faraday would occur at all???

For instance if I were an TN8 in the Congo insted of W7, would I not experience any Faraday on my own signal, and no experience with another equatorial station? Or would that be of course only true as stated in your posting Leif, if I were radiating straight up?

I would like to understand all that I know about this?

On 31-Dec-1997 SM5BSZ wrote:

The posting of "normal" influence of the ionosphere was the info as I found it. Presumably the author made some assumption (Location is Europe ?) that was not explicitly stated.

As I understand it, the Faraday rotation would vanish if the signal path is at 90 degrees towards the magnetic field. If it is a cosine function (as I would like to think) the angle has to be very close to 90 degrees before any noticeable effect will show up.

It would be very interesting, if someone makes an expedition near the equator, to hear the result of a Faraday rotation experiment:

If A is the angle between the signal path and the magnetic field and EL is the elevation angle, F = cos(A)/sin(EL)  should be the factor that gives the amount of Faraday rotation. Near the equator F goes through zero regularly.

The experiment would be to check that the own echoes are ALWAYS heard at full amplitude every time F is below 0.001 or so. Assuming 200+200 degrees normal Faraday rotation at 144MHz, the normal polarisation misalignment would then be 0.4 degrees, or less and even during extreme ionisation the polarisation misalignment would be very small, with no loss of signal due to Faraday rotation

On 2-Jan-1998 KK7KA wrote:
Since both path loss changes and sky noise variation caused by the moon's orbit are well known, one might think that the signal and noise received over a given EME path could be predicted to within a fraction of a dB.

But this is clearly not the case - there are numerous reports (posted here, in newslettters, etc.) of 144 MHz conditions being much worse or much better than expected.   Many of these come from stations with mechanical or electrical polarity control, so Faraday rotation is not the only cause.  These experienced operators also presumably know that equipment malfunction or local QRM is not to blame.  And I assume that any extreme weather conditions would have been noted.  Rapid changes
in signal, e.g. caused by libration fading, are not at issue here, nor are ground effects at very low elevation; these reports typically cover an entire moon pass.

I had always thought that the ionosphere was somehow responsible, absorbing signal, or concentrating it by varying refraction.  This theory was consistent with a scarcity of "conditions" reports at microwave frequencies.

But SM5BSZ's post shows that it is not true.  If the typical absorption on two meters is 0.025 dB, one might see 1 dB of
attenuation only under worst-case conditions, when combined with low elevation.  And at 70 cm, absorption is presumably negligible, yet conditions reports abound.

Can someone please explain these unpredictable signal variations? Please pardon my ignorance if I have missed something obvious.  I am an EME newbie, just starting to put together a station.

On 2-Jan-1998 W7EME wrote:
I would like to see this string continue. The equatorial ideas are interesting. The possibilities are unique and so is the idea of a
home in the Congo, unfortunately this is not in my lifes goals...hi.

I have spent much time pondering "atmospherics" in regards to the EME path. Have heard on a couple of occasions some strange anomolies arise on my tx. One being the double echo and echos with longer than normal delays? I have spent time chating with some retired service people who have worked high power meteor burst stations for the government and they have supplied some rough ideas on thier theories.

One thot that seems to be shared amongst them is that the solar winds at times produce high ionization of huge size. These clouds may refract radio in such a way that the circuit path is of very great distance. Interestingly these types of echos can be sometimes heard with a lesser length on Top Band. So then I wonder if this is so, then perhaps this "refractioning" of my signal thru this lense is focused in an direction other than the Lunar target? Also an ionized area may not be quite consistant or homogenous (spelling?...hi) and also act as a lense? I have also heard of the scintillation (flutter) on EME but have not experienced it myself.

Back to Faraday, I also do not believe this is all that contributes to my failed skeds. How much is liberation fading to contribute to 144 mcs ops? It must be possible that at times the surface providing ones reflections is not as appropriate as it was just seconds befor??

I also think low angle transmissions are not always providing ground gains at all, and maybe flamed for this statement. It seems to me, and some material I've read supports this, that reflections from the antennas forground could certainly change your path loss for worse? The reflected waves phase and amplitude could just as probably combine with the direct wave in a completely canceling way? Also have noticed an immediate drop in rx signals on several moonsets when the stone is dropping between 10 and 4 degrees at my qth. Is it possible this is just the much greater path thru the atmosphere?

I suppose all-in-all if my body was launched at the moon, at the speed of light, bounced off and returned I would not be very
intellegable afterwords either...hi.

On 3-Jan-1998 OZ1RH wrote:
SM5BSZ wrote:
>should practically never see any attenuation due to the ionosphere.
>The EME conditions do not change due to ionospheric attenuation.
I am still not quite convinced. The data presented is for one pass of the ionosphere at 90 deg elevation: 0.05 dB @ 100 MHz. For two pass eme of your own signal at 10 deg elevation this is 2*4*0.05 or almost 0.5 dB. Still not significant, but the data is probably for an average ionosphere (what ever that is). Suppose the D and E layer is strongly ionized from a solar outburst. If this could increase absorption by 10 times we get 5 dB or quite a significant attenuation. Since I have not read the proceedings mentioned I can't say if this could happen.
I however recall what I have read in some older proceedings (I.R.E. ect.) from the mid 50ties on meteorscatter and ionoscatter. It states that solar outbursts may give more ionization in the ionosphere resulting in more absorption. The point was, that for ionoscatter more attenuation due to absorption is compensated by more scattering _and by attenuation of cosmic noise_.
The average signal strength of meteorscatter signals changes about 6 dB depending on cosmic noise, either noisy sky in front of antenna or absorption of the cosmic noise.
Now this talk on meteorscatter and ionoscatter is mostly for the frequency range of 40-60 MHz, though the texts says conditions gets relatively better for the higher frequencies (60-70 MHz) when there is strong ionization due to solar outbursts.
There is a long way from 60 to 144 MHz, but I still think absorption should be considered in special situations where the ionosphere is strongly ionizised. Someone doing Radio Astronomy or ionosphere research should know a lot about this.
The focusing effect of the ionosphere and the troposphere mentioned could also have influence on eme. You might have noticed that the rising or setting moon sometimes appears visually as larger (or smaller) than usual. This is focusing in the troposphere and applies to our frequency range also. Focusing in the ionosphere has been mentioned earlier in this thread. Short time fading (<1sec) is called scintilation and is caused by focusing/defocusing due to variation of the ionization along the signal pass. It was researched in the mid 60'ties on 136 MHz satellites, so it applies to 144 MHz eme.

I wonder if the thing we call liberation fading (caused by 'rocking' of the moon) is in reality scintilation.

On 5-Jan-1998 SM5BSZ wrote:
The mail below arrived directly from Palle, but I have not seen it from MOON-NET, so I include a complete copy here.

Surely, according to the paper, the maximum attenuation will be 5dB for a single pass through the ionosphere "under extreme conditions" at 100MHz corresponding to 2.5 dB at 144MHz (one over frequency squared).

BUT, and that is the important thing, such extreme conditions are very infrequent and they can in no way be the reason for the great variation of "EME conditions" reported from so many operators.

Stewart Nelson, KK7KA wrote:
>- there are numerous reports (posted here, in newsletters, etc.) of 144 MHz
>conditions being much worse or much better than expected.  Many of these come
>from stations with mechanical or electrical polarity control, so Faraday
>rotation is not the only cause.  These experienced operators also presumably
>know that equipment malfunction or local QRM is not to blame.  And I assume
>that any extreme weather conditions would have been noted.  Rapid changes
>in signal, e.g. caused by libration fading, are not at issue here, nor are
>ground effects at very low elevation; these reports typically cover an entire
>moon pass.

I am not convinced at all. I have been using switchable polarisation for several years now, and my normal procedure for checking my equipment is to send "O O O" to the moon for several minutes while recording the strongest
echo. This level is normally 23dB - degr (degr is calculated the usual way).
If the value differs by more than 1.5dB I know something is wrong. I repeat this test, first transmitting horizontal, then vertical, and check the polarisation angle and signal strength of the echo. Of course this test is made without ground reflections. With ground reflections the signal as measured by this test is quite unpredictable (at my QTH). I have seen values up to about 35dB on a few occasions when ground gain was involved.

I have seen losses of 10dB and more when the elements are covered by thick ice (several centimeters), but I see no change at all due to rain. Once I had about 3dB loss for a while - it turned out to be water inside the tx cable, and
it did not show up as poor SWR!!. Of course "normal" calibration errors in the positioning system are detected the same way.

Only one single time I have observed ionospheric attenuation. That was during the ARRL contest 1994. The signal was then passing straight through an extremely strong aurora. At this time SM4IVE worked OK stations on 432MHz
which is exceptional!!

If we rule out Faraday and absorbtion as the causes for variable conditions, it is difficult to find another explanation. The remaining phenomena do not cause loss of signal, they just modulate the signal, or as we say, cause QSB. In my experience, the amplitude of the QSB does not vary, but the frequency certainly does.

When the QSB is very slow the signal is completely absent for long periods. The QSB maximum then lasts long enough to give both calls with good margin. Under such conditions it is easy to loose the signal by incorrect tuning. But if the
QSB characteristics of the particular time is known (I guess it correlates with the doppler width at 10GHz) the QSB rate can be compensated by an appropriate QSO procedure.

Summing up, I do not think it is justified to rule out Faraday rotation as the cause for "varying conditions" on 144MHz. When Faraday is unfavourable for horizontal stations, the activity drops, and then the band seems empty also to those who can change the polarisation. My feeling is that this is one of the main explanations for the "variable conditions".

On 5-Jan-1998 G3SEK wrote:
I'm not sure that mis-matched polarization is the whole explanation.
My experience is limited to 432MHz (with mechanically rotatable plane polarization) but there are definitely times when reliable "beacon" stations show little or no variation of signal strength with polarization angle. In other words, the transmitted plane polarization has become uniformly dispersed or "smeared" over all possible angles when it returns. This implies a 3dB loss compared with matched plane polarization.

Under good conditions with very little polarization dispersion, cross-polarized EME signals will null out to about -20dB, but it is far more common for the null to be filled to -6dB or higher.

I don't have any accurate statistics, but have a strong impression that polarization spreading correlates with higher ionospheric activity (eg it is more common in summer).

On 6-Jan-1998 VE7BQH wrote:
My findings correlate very closely if not completely with Ian, G3SEK. My findings are from the experience of 16 years with polarity rotation on 2 Meters.

It is my belief there is some form of polarity distortion present most of the time. A relatively small amount of times (thank goodness!) the polarity distortion is such that you can rotate on a signal with no distinct polarity null other than about 6 or so dB.

I find it VERY common to rotate 10 - 15 degrees from a signal that is barely detectable and have the signal go to contact copy. Obviously,this type of signal change does not equate against a theoretical polarity loss curve.

This relatively small movement in polarity with such a significant change causes me to get "lazy" quite often with my polarity search. In other words, I do not search far enough!

On 6-Jan-1998 W7EME wrote:
Ok so...

It's not really correct to call this a "rotation" during such events??

More like a scattering, signal is no longer polarized in any recognizable form? But just comming in as a handful of toothpicks thrown into the air and viewed as they are falling again??? If such is so, it is unlikely that the returning footprint is as strong per meter^3 as is a "good reflection"....right?

What mechanism does this fall in? And is it a lower atmosphere phenomena?

On 6-Jan-1998 G8MBI wrote:
So if I have sporadic e' down here, or there is aurora in the north, then somehow magically the same percentage of energy is radiated towards the moon then..??...(or even tropo)

Which also implies that the "ionospheric" scattering QSO's that are very common, especially in northern europe also have zero rf content in their link budget as well..

thats a good trick

think scattering and refracting, because thats whats happening....

On 6-Jan-1998 W7EME wrote:
Pardon my persistance, however this seems to be it to me also. Of the handful of posts I am not sure if anyone is or has said whether or not Faraday is the force that is causing this. Does it during these "disturbances" just distroy polarization in the normal sense or is this scattering happening inside or outside the Faraday "zone"? Maybe this is really the question?

Can the coherence of 144 mcs. be simulated using light wavelengths?

Boy I dislike this English language...hi!

On 6-Jan-1998 K6QXY wrote:
I have been reading the discussions on FARADAY ROTATION and the effects of the ionosphere on EME signals. In 1991 we constructed a very large polarity switchable array for 6 meter EME. 4x11el vertical and 4x11el horizontal. We could steer the array in linear polarity horiz. vert. 45 deg. and 135 deg.

We played with this system for 4 years. Here are my observations:
1--Forget about link budget calculators and models at 50 mhz. they don't work!
2--The ionosphere has huge effects on 6 meter EME. It can make signals disappear completely or it can greatly enhance them far beyond any link budget would calculate.
3--Signals are in a constant state of polarity rotation! We could see movements in polarity say from horiz. to 45 deg. in 6 sec. or less.There is absolutely NO conformity to the movements. They can go clockwise or counter clockwise or not move at all!
4--Without out polarity steering you only see the signal peaks not what is going on between call this FARADAY rotation if you like.
5--We have found many times when signals were completely dispersed ie equally good or bad in any polarity!!
6--We have a version of W9IP(REALTRAK) with 50 mhz. sky data and a link budget calculator based on 50 mhz. data.It is almost never right!! This is no disrespect to MIKE'S efforts just the real world @50 mhz.

On 6-Jan-1998 G3LTF wrote:

   I completely agree with Ian and Lionels comments,Ive operated rotatable polarisation on 432 since we started on that band regularly about 30 years ago as I've always used a dish antenna with a rotatable feed. at times the polarisation angle is really sharp and at others its spread all round the dial.

  I have a fairly quick motor on the feed so I can turn the feed while listening to my echoesso for a start I can see how much its being rotated and I can soon tell which way to orient the feed to give me the best chance of the other guy hearing me( sometimes I will rotate it while calling so at least the guy hears one call sign!!)

Mostly the worst "spread polarisation condx occur during daylight especially summer, in my experience.

I'm less convinced that we can forget ionospheric absorption at 432, I seem to recall some work done by INMARSAT that gave about 2-3 dB worst case one way loss at regions above the eqator at 1.6 GHz. Certainly the long term records of GPS signal strength at the distributed monitoring stations would give the answer if ayone has access to them. We shouldnt forget that the big increase in 432 activity has in general taken place in the quiet sun period so in the next few years we ought to be able to get a better  handle on it.

On 7-Jan-1998 SM5BSZ wrote:
This subject has revealed some quite new (to me) information. I think it is highly relevant to 432MHz EME-ers using parabolic reflectors.

Particularly the sentence:" but it is far more common for the null to be filled to -6dB or higher" is really alarming. We know for sure that any ionospheric influence will scale at least as one over frequency. Most probably one over frequency squared.

If the explanation is some ionospheric phenomenon, Faraday lockout should then practically never happen on 144MHz because the null would then practically never be filled by less than -6dB. We all know this is not the case.

I have only a few years of experience, so I have only experienced a sunspot minimum, but I can tell for sure that the minimum is NOT filled by anything like -6dB on 144 MHz. I have electronic rotation by turning a potentiometer, and I frequently check the polarisation by locating the null which is practically always as deep as the accuracy of the system can detect.

The conclusion MUST be that the absence of the null has some other cause. I propose the following: THE TRANSMITTED SIGNAL IS OFTEN ELLIPTICALLY POLARISED ON 432MHz. At first this conclusion seemed impossible to me, but after having talked to SM4IVE my opinion is that this is a real problem, and that the 432 and above community has
something to watch out for here. I learned from Lasse that a very common feed arrangement is the K3BBP design with two crossed NBS double dipole antennas. This antenna can be described as four dipoles constituting the sides of a square in which the corners are cut out. The distance from the tip of a horizontal dipole is only a few centimeters away from the tip of the nearest vertical dipole. This kind of arrangement is extremely sensitive to symmetry. The large capacitive couplings at the element ends must be very accurately matched to avoid coupling between the horizontal and vertical structures.

The same problem is of course present also for crossed yagis. On this site:

there is some discussion about the problem. There are also some NEC2 calculations for a cross yagi that is not at 90 degrees. In short:

Angle = 88.5 degrees   (misalignment = 1.5 degrees)

Isolation = -10dB (feeding power into H, terminate V in 50 ohms and measure power in this 50 ohm termination)

Horisontal radiation = 73.8% of the total power
Vertical radiation = 26.2% of the total power.

The phase angle between the horizontal and vertical radiation depends on the cable length of the vertical feed cable, and if it is left open or shorted. In case the unused vertical feed is terminated in 50 ohms, less power is radiated because 10% will heat the termination.

The information from Ian very strongly suggests that the isolation usually used between the H and V ports on 432MHz feed systems is in the order of 10dB which is certainly inadequate. The isolation should be in the order of 20dB. I do not know if it is common practice to measure the isolation, and tweak the elements for minimum, but it is a simple thing to do.

Ian's observation means that the polarisation practically in use on 432 is elliptic to the extent that usually 1 dB or more is lost.

Any comments ?

On 7-Jan-1998 G3SEK wrote:
Sorry - I really didn't meant to write "-6dB". However, it is definitely quite common to find nulls filling to -10dB on 432MHz, and this is much more common than nulls to -20dB.

I had not considered the possibility of stations transmitting elliptical polarization. That may be possible for some dish stations, and it may affect the observed statistics; but it is not likely for stations using ordinary single-polarized yagis.

Obviously we need some systematically recorded statistics - I will start next weekend.

On 7-Jan-1998 G3LTF wrote:
Further to my comments on polarisation "smearing" and Leifs comments about 70 cm feeds.

My 432 dish feed has always been the same, a pair of  folded dipoles half wavelength apart  fed by open line from a single balun, Pawsey stub type, and about 0.2 wavelength above a 1 wavelegth diameter sheet relector. Thats essentially the same as the NBS EIA  gain standard antenna.

I rotate the pawsey stub and thus the feed rotates above the reflector.  I believe the X polar response of that in the dish is much better than 10 dB, Id guess nearer to 16 or 20dB and it certainly gives a good null on local tropo signals, but I'll check it next time I have the feed in the dish.

This phenomena of polarisation smearing was noted years and years ago on 432 as long time readers of Al's 432 and above NL will know, what we dont have is a good explanation

On 8-Jan-1998 G8MBI wrote:
Ian G3SEK wrote:-

>Obviously we need some systematically recorded statistics - I will start
>next weekend.

and to do that we now need a sensible system to exchange  polarity information.....

which, when you suggested it a few months ago and I excitedly and hurriedly 'seconded it' ...with some constructive input because I could see the issue arriving on 144 as more and more go polarity capable and we start to see all the same effects that the 432 and up folks have known about for years, it then got 'shouted down'...

lets talk about it with all at the paris conference..??

A paper from you Ian..?

How about a letter system .....??..using 0 to 180 only (as thats all thats necessary)..?? using A to T (20 steps) will conveniently be 9 deg steps, that gives accurate representation of 45 deg and 135 degrees for
those using 'X' or '+' polarity and/or using simple polarity synthesis..?

and a few simple guidelines like, to totally avoid confusion, polarity info may only be sent (and obviously optional) with final rogers...or to be even more sure of no confusion in a final, final, 73's period....73 73 TXA 73 73 TXA...would be vertical.....'outlaw' sending of polarity info in contests so that those with no real interest cannot complain about time being
wasted..?...or maybe make it a "sched only" activity...?? again folks with no interest do not get it forced upon them on random..??....although I think we can assume that many with polarity capability would like to hear
it, those without polarity capability need not be subjected to it ...those with need only apply common sense so as not to become a nuisance to those that have not...and those that have not, need not stop those that have from extending the discussion...?


Or to make it even simpler, just send the polarity 'as is' most polarity capable stations will be louder with each other than conventional stations so exchanging the angle in plain language like TX 135 would also be ok, even with the single yagi on 144 I cannot think of a polarity capable station that I could not copy well enough >80% of the time to get that ok...(unless we have very disturbed condx :-)....common sense can dictate when sigs are good enough to do it..

RX info as well for the bigger stations..??...

BTW  as soon as we can get folks doing this with both rx and tx data being exchanged then my prediction is that we will discover that faraday is non reciprocal at times as well?, could be that the polarity distortions have much greater effect than faraday and so confuse the faraday observations though.?....and that stations from very similar geographic locations with same spatial offsets, QRV at the same time, are peak received on totally different polarities..again maybe polarity or ionospheric effects 'over ride' the farday influence..?...that will sink a few more old chestnuts as well.

If a sensible system can be adopted I will modify the yagi mount here to have continuous polarity rotation instead of simple switching and join in....been thinking of doing that anyway to circumvent the daft italy contest rules where I have to count my 18 elements twice...

Following is observational, common man, input/conclusions for the discussion:-

It could also be that Leif sees fairly consistant echoe returns because bypolarity search matching he is tuning out the majority of polarity effects coming from the ionsphere and not simply correcting for faraday rotation..??..could even be with variable transmit capability as well, that leif will find the optimum tx/rx position for minimal ionospheric attenuation effects rather than doing what he thinks he is doing which is simple faraday correction..??...

I can rather easily see that even a distorted polarity match could be found from the same geographic location, BUT not be found a short distance away....given the ionosphere is a non uniform entity. I have repeatedly measured a significant difference on leifs signal vertical versus horizontal when he transmitted circular (in the early days he does it less now)...I have also seen it on K1CA but not often (never qrv :-)...and HB9JAW but again not often (not on long with the capability)....supports the distortion theory..but does not eliminate the non perfect tx system....

maybe leif sees this more clearly because he can test echoes rather efficiently with respect to polarity .....simple 'switchers' like sm2cew and myself and mechanical rotators like ve7bqh and g3sek cannot do that..because either we cannot find the exact match or cannot move fast enough to find it.??

1296 folks must have input here..??...

Although unable to measure accurately the precise polarity angle here I can by extrapolation from measurements at 0 and 90 and long monitoring periods confirm that the troughs of 144 stations are highly variable, at times of high ionospheric disturbance the 'smearing' of polarity is obvious even with a simple system on strong stations and a severe dip in signal levels on smaller ones is also these times faraday (but perhaps not 'pure' faraday) has a trend to either lock up completely or move much more other effect as yet un mentioned here is that also at times of high disturbance short periods (often only 20 to 30 minutes) of outrageously strong signals are received as well, with 4 yagi guys and even my own echoes at many, many Db's over 'normal' or those theoretically achieveable (perfect polarity match or not)....anyone else see this.??

The last point leads me to conclude that whilst the majority of signal variations are in fact polarity issues, that the ionosphere has a de-focusing or indeed focusing effect on signals, leading to an artificially widened (and less efficient) or narrowed beamwidth (rarer, but therefore more efficient), this would also seem a fairly logical conclusion given the totally non uniform nature of ionosphere and it's 'layered' structure...could even be a troposphere effect.?

It should also be said that conclusions drawn from only a few stations are dangerous, if you live in OH or SM2 it would be very easy to deny the existance of sporadic E you cannot 'see' it and if analysed on currently known and explainable pure physics it does not exist.

The sporadic e phenomenon is in fact a good comparison because it too remains unsolved as it appears that there are upto 4 effects overlaid with each other making even scientific analysis difficult/ it is with tranmissions through the ionosphere, difficult to differentiate one effect from another..what is clear though is that polarity capability will overcome the majority of them...although I continue to believe not all.

How do I get all my signal to the moon when it rises 'behind' the FAI over northern italy...hmmmm.....and even if the level of reflection is too weak to be detected in 9H1 or EA8 at higher elevations in sporadic e openings, is there truly no scattering or at least refraction of my signal anymore..??....does it not at least heavily imply that my EME path efficiency is severely degraded...??...

On 8-Jan-1998 SM7UFW wrote:
Graham F/G8MBI wrote:

>and to do that we now need a sensible system to exchange polarity

... and to SHARE information, I'd like to add.

I propose a web site where you could enter QSO information after a QSO.
Input could be:

- time/date
- my call
- other stn call
- polarity sent by me
- polarity received at my QTH

Advantages of this system would be

- information would be shared with others
- information accumulated over time
- information does not have to be sent in QSO - saves time and
  effort in contests etc.
- information can be 'posted' when suitable

I'd be happy to do the necessary programming. Tell me if you think it is a good idea?

On 8-Jan-1998 G3SEK wrote:
You don't need any other information from the other station in order to collect data about polarization spreading - it's a receive-only measurement.

As for a paper on polarization reporting... well, after the reception the idea got from most people on here, I think it might wait until the next century!

Anyhow, let's talk about it first.

On ??-Jan-1998 G4SWX wrote:
I agree with what Darrel's saying, it is exactly what I said to both you and Graham on the phone except:

The non-uniformities in the ionosphere will give rise to the generation of elliptical polarisation of both right hand and left hand. The circular components of these will tend to reduce the signal as detected by a plane wave antenna. The reduction in extreme cases, 50% RHCP and 50% LHCP would be many dB if not quite infinite.

The most interesting case is the reception of echoes, the signal passing back down the same path in the ionosphere which it went up. It looks to me from the maths that the elliptical(as opposed to the plane wave rotational term) component changes sign on the return path. The effect of this is that unlike the plane term where where the rotations add, the circular term reverses. The effect of this term being to cancel out the circular component.

This being the case one's own echoes will always be stronger and more plane polarised than echoes received by anybody else. As another passage through a different path in the ionosphere is very unlikely to have the same birefringence. Hence Leif's observations that his echoes are predominantly plane polarised seem to be backed up by the theory.

BTW some of the GOES satellites have stabilised plane polarised sinals which allow accurate real-time measurement of Faraday (only one path which is unlikely to be the same as the moon)

On 11-Jan-1998 AA7FV & G3SYS wrote:
The thread on polarization changes of EME signals travelling through the ionosphere motivated me to open some of my old textbooks, and to do a little web surfing.  Maybe some of the following notes and web addresses  could be of interest to others, although I am sure most of the points are already familiar to many.

  Most of the textbooks have a statement along the lines of "A plane polarized wave incident on the ionosphere will in general become elliptically polarized."   The details are defined by the Appleton-Hartree equation, dating back to 1932 or earlier.  I came across a readable description of this on the web, at: with the title:
"Magneto-ionic Theory and the Appleton-Hartree Equation."

  The details of the polarization change depend on the frequency, the total electron density and magnetic field integrated along the line of site, and the direction of the magnetic field with respect to the direction of propagation and the plane of polarization.   Any plane wave can be considered as the vector sum of 2 or more plane waves with different polarization angles.  For example a 100% vertically polarized wave can be considered as the sum of two different linearly polarized waves in phase, each containing half the power, one with the plane of polarization 45 degrees clockwise of vertical and the other 45 degrees anti-clockwise of vertical.  The earth's magnetic field makes it easier for electrons of the ionosphere to move in some
directions than others, and this causes a slightly different refractive index, and so different propagation velocities, for the linearly
polarized waves at different polarization angles.  This in turn causes a relative phase shift between the different components of the wave, which in general will result in elliptical polarization.  This is the main factor responsible for the change of polarization.  These parameters all appear in the Appleton-Hartree equation.

  Circular polarization is a special case of elliptical polarization; sometimes the wave and ionospheric parameters will inevitably resultin circular polarization.  When a circularly polarized wave is received by a linearly polarized antenna, the received signal
strength is completely unaffected by rotation of the plane of polarization of the antenna, although the polarization mismatch gives
a constant 3 dB loss.  Under these circumstances, switching the receive antenna to the correct circular polarization would gain 3 dB, but switching to the opposite sense of circular would lose the signal altogether.  Does anyone have the capability of making this test on EME signals?

  Another factor affecting the strength of VHF and UHF signals passing through the ionosphere is known as scintillation.  In effect, blobs of enhanced electron density act as lenses;  they can enhance the signal, by focusing energy to a particular spot on the earth's surface, or they can reduce signal strength by redirecting the energy that should have arrived at a given spot to somewhere else.  In reality it's a little more complicated, involving diffraction as well as refraction, and is variable with time.   The phenomenon was first noticed in 1946 by Hey, Parsons and Philips in some early radio astronomy work at 64 MHz. An analogy is the observed optical twinkling of stars seen through an unstable atmosphere.

   The web is full of references to the topic, but I stumbled across a nice summary of ionospheric scintillation with literature
references at: called  "Mid-latitude Ionospheric Scintillations."
This includes a link to some interesting abstracts from the professional literature: .
Ionospheric scintillation is very much more pronounced at VHF and lower frequencies, but one abstract here mentions ionospheric scintillation measured at 1.7 GHz, with a 2.3 dB peak-to-peak variation in signal strength, and another describes measurements of ionospheric scintillation during a magnetic storm at frequencies as high as 11.5 GHz.

  Ionospheric scintillation can affect EME signals, but there is also interplanetary scintillation, which affects waves passing through the solar wind within our solar system, and even interstellar scintillation which can affect reception of very distant radio
astronomical sources.  Not yet a problem for radio amateur communication.

    Finally, since I don't personally have the sensitivity to be able to make useful measurements on real EME signals, I've found it
interesting to monitor some VHF and UHF transmissions from some geostationary satellites.   In particular, for those in the USA:

  GOES-8 and GOES-9 have time code transmissions from geostationary orbit, on 468.8375 and 468.825 MHz.  You can find more information on .  The transmissions have a well-defined and very stable carrier which is easy to monitor.

  There is a list of satellites with essentially continuous beacons on frequencies of 150 MHz and below at: .

   I've found signals from GOES 2 and GOES 3, nominally on 136.380 MHz, particularly useful; in practice the 2 satellites differ in frequency by about 800 Hz, which puts both signals conveniently within a single soundblaster DSP passband.  The signals are unmodulated, apart from variations caused by the spacecraft spin, which often gives a 6-Hz split in the spectrum that helps in identification.

  Apologies for this note having become so long, but I hope something in it may be relevant to some of the observed EME propagation effects.

On 16-Jan-1998 N1BWT wrote:
There was a recent flurry of messages regarding causes of fading on EME signals.

The December issue of IEEE Antennas and Propagation Magazine had an article addressing this: "10 Years of Radio-Scintillation Observations", by Jules Aarons of the Center for Space Physics at Boston University.

the article address fading from 54 MHz to 4 GHz for satellite transmissions and radio astronomy, paths similar to EME.  it reports fading which varies with frequency**-1.5 [-1.5 power], and proportional to solar flux, with peaks around noon and midnite.

a few numbers I picked out:

250 MHz: 25 dB fades in years of high solar flux, 3 to 5 dB in years of  low solar flux

1.6 GHz: 10 dB fades in years of high solar flux

4 GHz: 4 dB fades

the implication for GPS is being studied, but the author points out that the proposed LEO satellites have probably not considered this problem.

since this article is probably difficult to get for most of the worldwide EME population, if there is any interest, I would be willing to mail a copy to the first ham in each hemisphere who is willing to make an additional copy for the next interested party, and so on...

On 23-Jan-1998 GM4JJJ wrote:
Here is a copy of the latest 'Satgen' from John GM4IHJ which follows up on some of the discussions on this group:-

>From: John Branegan <>
>Subject: Satgen 461
>MIME-Version: 1.0
>Satgen461 Space Propagation Pt1   by GM4IHJ  (BID SGEN461) 1998-01-24
>Most satellite and EME operators know, there is a deal of difference between, the simple descriptions in amateur radio handbooks of the theory of what happens to a signal in space or travelling through the ionosphere, and, what actually arrives at the ground station receiver.
>Signals can be disturbed by Faraday effect, Sporadic E, Field aligned effects, Aurora, Scintillation, the Weather, and the Atmosphere.
>Taking Faraday effect first . Does a signal coming through the ionosphere simply have its phase rotated such that, EG vertical
>polarisation at the transmitter appears as a signal with something other than vertical polarisation at the ground station antenna ? No it does not , as a simple experiment listening to a signal from space with a ground antenna which can be turned to present varying angles of polarisation, soon reveals.  But you have to be careful. A Sat over Latitude X with its antenna pointing verticallly at the ground beneath does not have the same vertical as a ground station antenna aligned verticaly at Latitude Y. Many satellite operators ( particularly geo weather sats and TVRO sats ops, adjust the antenna at the satellite to match say the vertical at the centre of the ground service area of a spot beam . So under a spot beam the theoretical alignment ought not to be far out. But with a wide hemisphere beam on say a geosat you can be 60 degrees out at a station at latitude 60 degrees. So be careful even as you start simple one way experiments.
>On TV signals at 10GHz Faraday rotation is not a problem. But at UHF frequencies it is becoming a problem and by VHF frequencies it is a serious problem on satellite signals which come from linearly polarised antennas and also from those that come from supposedly circularly polarised sat antennas which produce varying polarisation the further you are off the sat antenna central axis.
>Worse still when you take the trouble to set up an experiment targeted on a geosat fixed in your sky sending linear polarisation , and you receive it on an antenna which can be turned to align it to the strongest signal ie correct for the transmitted polarisation plus Faraday rotation, you invariably get a shock . When you discover that instead of the 20 dB difference in signal expected between correct polarisation match and, 90 degrees away at total mismatch, you only get about 6 dBs of difference, because whatever it started out as , a linear signal coming through the ionosphere is inevitably elliptically
>polarised when you receive it at the ground. A phenomenon fully reported nearly 70 years ago by Appleton and Hartree.
>So if this simple test already reveals that a one way trip through the ionosphere produces elliptical polarisation, the results of a two way trip to a satellite or to the Moon will be very different from simple theory. Hence the move in most professional systems to the much more complex circular polarisation mode, where, if helical antennas are used instead of XY or spaced dipoles to get the circularity , signals will be almost completely unaffected by Faraday rotation at any frequency above 136 MHz. Below 136 MHz antennas producing circular polarisation are too bulky and rarely produce reasonable circularity.
>Chronological List title - Space Propagation Pt1
>Suggested Index references - Faraday : Propagation : Space Experiments
>John Branegan

On 23-Jan-1998 WB4APR wrote:
On Fri, 23 Jan 1998, David G.L. Anderson wrote:

> >ionosphere produces elliptical polarisation, the results of a two way
> >trip to a satellite or to the Moon will be very different from simple
> >theory. Hence the move in most professional systems to the much more
> >complex circular polarisation mode, where, if helical antennas are
> >used..   singals will be almost completely unaffected by Faraday
> >rotation at any frequency above 136 MHz.

Helix's are fine between two fixed sites since one can use RHCP and the other LHCP, but helix's are not a good idea for random QSO's because you need to be able to switch circularity depending on who you are talking to.

Two RHCP's or two LHCP's will be cross polarized due to the circulartiy reversal on reflection from the moon.  A station transmitting RHCP has to switch to LHCP to hear his own echos.  I know, I have tried it with 100 watts at 2.2 GHz and the echos are undetectible unless I switch circularity between transmit and receive.

On 25-Jan-1998 SM5BSZ wrote:
It seems like I am the only one on this list who thinks ionospheric absorbtion is a rare phenomenon on 144MHz. The "normal" ionosphere as shown in my first posting with data from a conference proceedings is obviously not accepted at all by the EME community.

I find the situation remarkable, and I will try to explain why in this maybe somewhat lengthy posting. As a starting point, take the copy below of an E-mail from Darrel Emerson, AA7FV & G3SYS that arrived from G3SEK through G4SWX while moon net was down. (Maybe you all could share the full discussion with the rest of us ? - here, or at the EA6VQ site)

Darrel shows here in a very pedagogic way how we can understand the pure cases when the propagation direction is parallel or perpendicular to the magnetic field.

**************** From Darrel Emerson Jan 18 1998 ******************
  A plane polarized wave can ALWAYS be considered as the sum of a RHCP and a LHCP wave, with 50% of the power in each of the orthogonal pure circular polarizations.   Similarly, adding two orthogonal circularly polarized waves, of equal amplitude, always results in a 100% linearly polarized wave.  This, as such, is nothing to do with any ionosphere.  If the RHCP and LHCP are in phase when both of their rotating vectors are horizontal, then the result of the combination is a 100% linearly polarized wave, horizontally polarized.  If they are 180 degrees out of phase, the result is a 100% linearly polarized wave, vertically polarized.  If the RHCP and LHCP waves are 90 degrees out of phase with each other, the result is a 100% linearly polarized wave, at 45 degrees to the horizontal. Adding RHCP and LHCP waves of equal amplitude always gives a 100% linearly polarized wave, with position angle equal to half the phase difference between the RHCP and LHCP terms.  So, if the 50% RHCP is added to the 50% LHCP, if their relative phases are right there will be 100% coupling to (e.g.) a horizontal linearly polarized receive antenna, while if their phases are wrong (180 degrees different in this example, resulting in a vertical linear plane wave) there will be zero coupling to a horizontal linearly polarized antenna.  It's the relative phases of the RHCP and LHCP that count, not the fact that there is a RHCP and LHCP wave in the first place.   I think this is what John meant by "The reduction in extreme cases ..."   but I want to make sure I'm not misunderstanding him.

  Faraday rotation with propagation along a magnetic field line arises because, after you've decomposed an incident plane wave
into it's orthgonal RHCP and LHCP components (you can do this before it gets to the ionosphere) then the refractive index for RHCP is different from that of the LHCP.   In effect, depending on the sign of the field, the electrons try to move in partial circles which are in the same sense as the rotating vector of one of the orthogonal circular polarizations, in which case they have more influence on the wave, while the electrons move in circles opposite to the sense of rotation of the opposite circular polarization, and so interact with that circular term less.  This gives the different refractive index of the ortghogonal circular terms, which results in a different relative phase as the RHCP and LHCP terms emerge from the ionosphere, which then results in a rotation of the linear plane of polarization.  In this ideal case "linear in gives linear out" but with the plane rotated.  For this ideal case of a linearly polarized wave propagating exactly along a magnetic field line, the wave never becomes elliptically polarized at all.

  The textbooks talk about the "ordinary" and the "extraordinary" ray.  For propagation along magnetic field lines (the longitudinal case) the ordinary and extraordinary rays are orthogonal circularly polarized components.  For propagation perpendicular to the magnetic field lines (the transverse case) the ordinary and extraordinary terms are linearly polarized components, with the E-field respectively parallel to the magnetic field, and perpendicular to the magnetic field.  Of course, depending on the relative phase and amplitude of the two linear terms, the result of the combination may be anything - linearly, circularly, or in general elliptically polarized.

  For propagation along the magnetic field lines, it's convenient to decompose the wave into the sum of orthogonal circular terms. The change in polarization comes from the change in relative phase of the two orthogonal terms.  With propagation across the field lines, it's more convenient to decompose the wave into orthogonal linear terms, one with the E field parallel to, the other with the E field perpendicular to, the magnetic field.  This is true even if the incident wave is 100% circularly polarized (i.e. orthogonal linear terms 90 degrees out of phase.)

I'll try to summarize the extreme, special cases, assuming that the incident wave is 100% linearly polarized.:

(A) Propagation parallel to the magnetic field lines.
(1) Ordinary/Extraordinary waves are RHCP and LHCP (or vice versa).

(2) There will be a change in the relative phase of the RHCP and  LHCP components. This will introduce a pure rotation of the
  linear plane of polarization.  No ellipticity is introduced.

(3) Changes in relative amplitude (e.g. by absorption of the extraordinary ray) WOULD introduce ellipticity, but this is not   normally important at VHF/UHF wavelengths (but can be sometimes.)

(4) Rotation of the plane of polarization is cumulative.  On the way back from the moon, the total rotation of the plane of
  polarization is doubled.

(B) Propagation pendicular to the magnetic field lines.
(1) Ordinary/Extraordinary waves are respectively linear polarizations with the E field parallel to the magnetic field, and with the E field perpendicular to the magnetic field.

(2) There will be change in relative phase of these two orthogonal linear terms, because of the different refractive index of the ordinary and extraordinary waves.  Consider some special cases:
  (i) If the incident wave is polarized with the E field parallel to the magnetic field, there is no extraordinary ray, and nothing changes in the polarization.  No polarization rotation, no ellipticity introduced.
  (ii) If the incident wave is polarized with the E field perpendicular to the magnetic field, it becomes 100% the extraordinary ray.  Again, there will be no change in polarization.  No polarization rotation and no ellipticity introduced.
  (iii) If the incident wave is polarized with the E field at 45 degrees to the magnetic field, then half the power can be considered as a plane wave parallel to the magnetic field, and half perpendicular to the magnetic field.  The two terms propagate with different velocities, so there is a relative phase shift.
     (a) If this phase shift is odd multiples of (pi/2), the result is 100% circularly polarized radiation, either RHC or LHC depending on how many (pi/2) there are in the phase shift.
     (b) If this phase shift is odd multiples of (pi), the result is a 100% linearly polarized wave, but rotated in plane of polarization by 90 degrees.
     (c) If this phase shift is even multiples of pi, obviously nothing changes.

(3) Change in relative amplitude of the extraordinary and ordinary ray, caused by differential absorption, would in general cause some rotation of the incident plane of polarization, but would not in itself introduce ellipticity into the wave.  This case is probably not important.

(4) The relative phase shift between the extraordinary and ordinary ray is cumulative.   There will be twice as much differential phase shift after the return journey from the moon.
  However, there are some special cases.
    (a) If the one-way phase shift is pi/2, then on return from the moon it becomes pi.  So, what landed on the moon as 100%
     circular, from the single pi/2 shift, returns as 100% linear but rotated in plane of polarization by 90 degrees. (The reflection from the moon changes RHC into LHC as well; I've lost track of whether this adds or subtracts to the rotation of plane of polarization, or even has any effect.)
   (b) If the one-way phase shift is any multiple n of pi, then on return this becomes 2.n.pi, so the wave returns with the same
     linear polarization it left.
********************** End of Darrel's text ************************

<<<<<<<<<<<<<< Theory section (may contain errors) >>>>>>>>>>>>>>>>>>
The general case is more difficult to visualise. One has to split the radiation into two orthogonal elliptic polarisations. The end result will however be similar to what we obtain by replacing a passage through the real ionosphere with some angle alfa between the magnetic field and the propagation direction by a passage through two different ionospheric layers of thicknesses cos(alfa) and sin(alfa) in which the magnetic field is parallelled respectively orthogonal to the magnetic field.

Once we have quantitatively correct refractive indices for the extreme cases I think it will be obvious what will happen in the general case.

The reference: which Darrel gave in a previous posting gives the equations, but I do not know what numbers to put in so I can not make make calculations directly from equ (2.16). It is however possible to compare equ (2.26) to (2.24 and 2.25). These equations are for the pure cases described above in Darrels text, and they are for frequencies above 8MHz.

(A) Propagation parallel to the magnetic field lines.
(1) Ordinary/Extraordinary waves are RHCP and LHCP (or vice versa). For frequencies above 8MHz:

n*n = 1 - X / (1+Y)  Ordinary (?) equ (2.26)

n*n = 1 - X / (1-Y)  Extraordinary (?) equ (2.26)

(B) Propagation pendicular to the magnetic field lines.
(1) Ordinary/Extraordinary waves are respectively linear polarizations with the E field parallel to the magnetic field,
  and with the E field perpendicular to the magnetic field.

n*n = 1 - X*(1-X) / (1-X-Y*Y/2+Y*Y/2)  Ordinary equ (2.24)

n*n = 1 - X*(1-X) / (1-X-Y*Y/2-Y*Y/2)  Extraordinary equ (2.25)

X is the square of the ratio of the signal frequency to the plasma frequency. The plasma frequency is the limit frequency for a radio wave to be reflected by the ionosphere at 90 degrees so we can safely assume that X is not much above 0.01 at 144MHz. The interesting quantity is the difference between the refractive indices for the two waves in the two cases.
Y is the electron gyro frequency divided by the signal frequency, so Y is also a small number.

In case A the difference between the squares is:

X/(1-Y) - X/(1+Y)  =  2*X*Y (approx)

In case B the difference is

X*(1-X)/(1-X-Y*Y/2-Y*Y/2) - X*(1-X)/(1-X-Y*Y/2+Y*Y/2) = X*Y*Y (approx)

Faraday rotation in the longitudinal case is 2/Y times stronger than conversion from linear to circular in the transverse case.

Note that the refractive index is a real quantity here. The imaginary part which is associated with collisions is a measure of attenuation while the real part is a measure of the wave speed.

From equ (2.16) it seems highly unlikely that the small difference in attenuation between the ordinary and the extraordinary wave will have any effect at all on the polarisation.
<<<<<<<<<<<<<<<<<<<<<< End of theory section >>>>>>>>>>>>>>>>>>>>>>>>>>

The equipment I use since more than three years uses a computer to correlate the signal from the two parts of my cross yagi antenna. This evaluation gives the relative amplitudes AND THE PHASE of the two signals. The algorithm is reasonably efficient and the computer always presents the polarisation plane and the "circularity", a number that goes from 100 for a circularly polarised wave to zero for a linearly polarised wave.

The receive system is implemented in a dedicated hardware (TMS320C25 + 80186) and it is inconvenient to program, so the algorithm uses a preselected polarisation to find the signal and tune the filters to it before an analysis of the polarisation becomes possible. Under normal circumstances I use circular as the preselected polarisation because then the system will lock properly to all signals except the opposite circular which no one uses.

When the system finds a signal, the polarisation indicator gradually changes from "circularity" 100 to very near 0. For reasonably strong signals this is a very fast process, but for weak ones it may take some time.

I can assure you all that EME signals retain their linear polarisation under current solar conditions (as they have been during the last few years). I have checked this hundreds of times, and it is part of my normal QSO procedure to measure the polarisation of the other station in order to know which polarisation I should use myself. I have seen perfect circular polarisation from HB9JAW but everyone else is accurately linear always.

Of course I can not know what will happen during sun spot maximum, but from theory I think one should not expect conversion from linear to elliptic to cause losses of signal at 144 or higher except during unusual circumstances.

<<<<<<<<<<<<<<     The controversy:   >>>>>>>>>>>>>>
G3SEK and G3LTF (Jan.7) both say that polarisation smearing is often at -10dB on 432MHz.
GM4IHJ (Jan.23)  If you compare best and worst linear polarisation on GEOSAT you will find 6dB, not the expected 20dB.

My answer is definitely no - everyone else seems to say yes!!!

We all read now and then about "polarity lockout". Such a thing would be extremely rare if polarisation smearing due to ionospheric effects was the normal situation.

Situations when a whole continent has a polarity lockout for hours indicate that the Faraday rotation is very small - otherwise it can not be the same for so long and over so large areas. Polarisation smearing can not happen if the ionisation is so weak that the Faraday rotation is small.
<<<<<<<<<<<<<<<<<<<                   >>>>>>>>>>>>>>>>

<<<<<<<<<<<<<<<  Scintillation >>>>>>>>>>>>>>>>>>
Scintillation is surely a phenomenon that happens in the ionosphere at VHF. The phenomenon is analogue to the twinkling of the stars and it is a phenomenon that causes QSB, but it does not change the average power level of the received signal.

I am not sure the EME path is affected by scintillation because of the size of the moon. Stars twinkle, but planets do not - they are too big!!!

The QSB caused by scintillation (if it contributes at all) will add to the libration fading caused by the rocking motion of the moon and perhaps change the character of the QSB slightly although I personally do not think this is a significant effect.

If scintillation was the problem I think reports of bad conditions would look quite different from what we are used to see!
<<<<<<<<<<<<<<<<<              >>>>>>>>>>>>>>>>>>>>>

For those of you who still think elliptical polarisation causes loss of signal on EME I suggest a stereo system. On 70cm where elliptic is often received for unknown reasons a system like the one I am using will give one or two dB better signal than a conventional system because the computer (or your brains if you listen in stereo) has no prejudice - it adapts to the polarisation actually present.

My guess is that a cross yagi with a stereo system will often give a 3dB improvement on 50MHz EME, and probably also on terrestrial modes on that band.

My new receive system, which is not complete yet, will run in a PC computer with a Soundblaster 16. It listens in stereo at 20kHz bandwidth and allows fast finding of all CQ callers - and their polarisations. When it runs properly I will put the software on my internet site as an EXE file and as source code for those of you who like to use it. Having a better general view of the band makes operating more fun!

On 25-Jan-1998 K5GW wrote:
I will have to throw my lot in with SM5BSZ on the absortion issue but for a different reason:

I make frequent sun and radio star noise measurements on my 2 meter and 70cm systems. Both systems are large enough to get very reliable and repeatable noise readings. If there were an absorbtion phenomen taking place, I would immediately see lower noise readings on these systems. I have *NEVER* seen a time during the past 5 years and hundreds of readings that the sun or radio star noise was below the expected and calculated value. There of course have been times of elevated sun noise readings. For this reason, I must conclude that signal loss is a polarity issue only or at least mostly.

As a side note, my 70cm system is equipped with a very fast and capable home made polarity rotator (uses 1/3 horse power motor). The rotator changes polarity at the rate of about 8 degrees per second. I can echo test while rotating clockwise or counterclockwise to cause my echo to change polarity plus and minus about 20 degrees relative to the transmitted polarity. By measuring the echo strength and plotting it on the polarity loss chart (45gegrees=-3db, 90 degrees=-20db, etc) I can determine the local echo polarity rotation pretty closely. I can tell you that on *MOST* occasions over the past two years that my local echo polarity rotation has been in the 70 to 110 degree range. If this is true for other stations, and I bet it is, then testing for echoes has been pretty depressing over the past several months. As an estimate, my own echo strength has been typically 10 db or more below optimum. I mention this only because an inexperienced operator could draw the conclusion that his station has a serious deffect since echoes are always weaker than expected. Let me emphasize this is on 70cm and bears no relation to the 2 meter case.

On 26-Jan-1998 F/G8MBI wrote:
A new angle.? (pun intended)

the polarity discussion is fascinating...

It seems that quite a few non eme'ers who are here and subscribe presumably through real interest in these type of issues, and are actually professionals in various industries, appear to understand many of the effects and have clear math proofs, well beyond my humble math/observational capabilites..

has anyone actually approached these folks to help with defence documents against the guys that want to use low VHF for commercial LEO's....?? the ones that still persist in attacking the 144 band...??

Do these folks know these effects also..?...should they be told.?..and more importantly could a sensibly constructed paper presented to the various licensing administrations throw a further spanner in the works..??

even if not conclusive it might assist sew some seeds of doubt about suitability of low vhf, both with those that propose the system and those that license it..the more smoke/fog the better..??

On 28-Jan-1998 ZS6AXT wrote:
I noticed that signals from METEOSAT (1695 MHz) are stronger after the Sunset and also after the rain. Similar effect I noticed also on 1296 MHz.So there if obviously some other attenuation, caused maybe by dust in the atmosphere and other reasons. So that it looks that we have more types of attenuation especially on the higher frequencies than the "standard" theoretical types !! Meteosat signal here is marginal and any change from good level is immediately noticeable, as noise on the picture.

On 30-Jan-1998 AA7FV wrote:
To: John Regnault
Cc: Ian White, G3SEK; Darrel Emerson
From: Darrel Emerson on Fri, Jan 30, 1998 3:21 am
Subject: Another word on ellipticity

Hi John,
   It was good to talk to you a couple of days ago.

   If I remember our telephone discussion correctly, you were taking the differential phase shift derived from transverse-to-magnetic-field propagation of RHC and LHC terms, responsible for transverse Faraday rotation, and suggesting that the differential phase shift for the linearly polarised transverse ordinary and extraordinary rays  (i.e. linearly polarised with E-field parallel to the magnetic field and normal to the magnetic field,) would be similar.  [That's the longest sentence I've written for some time.]

   I don't think that's quite right.  Just taking formulas from Kraus, and plugging in "typical" ionospheric values, at a frequency of 100 MHz I get the following values for the differential refractive indices:

(1) Longitudinal-to-magnetic-field propagation, difference in refractive index for RHC and LHC terms (this is responsible for effectively all of the Faraday rotation) is approximately 10^-4.

(2) Transverse-to-magnetic-field propagation, difference in refractive index for RHC and LHC terms, which is responsible for Faraday rotation in transverse propagation:   approximately 8*10^-7, let's say 10^-6.  I.e. transverse Faraday rotation is down about a hundred on longitudinal Faraday rotation.  This agrees exactly with Kraus (p. 143). This also agrees reasonably with the graph of typical longitudinal and transverse Faraday rotation from your book that you faxed me a couple of days ago.

(3)  Transverse propagation, difference in refractive index for linearly polarised wave with E-field parallel to, and linear wave with E-field normal to, the magnetic field.   This is approximately 6.5*10^-9, or let's say about 10^-8.

   Those values are for 100 MHz.  The ratios between the factors in the 3 cases increase as frequency increases.

   Summarising: the differential refractive index for orthogonal circular terms propagating transverse to the magnetic field is about 100:1 down on the differential refractive index for orthogonal circular propagating along the magnetic field.  The differential refractive index between orthogonal linear terms of transverse propagation, which would be responsible for introducing ellipticity, is about 100:1 down further on that. I think you assumed differential refractive index for transverse propagation orthogonal linears would be about the same as for transverse orthogonal circulars, not another factor of 100 down.

  There are basically two ways of converting a linearly polarised wave into elliptically polarised.  A linearly polarized wave can always be decomposed into either (a) two equal RHC and LHC terms, or (b) two equal orthogonal linear terms, each with their E-field at 45 degrees to the the original wave. 

  In case (a) a difference in phase just rotates the linear wave you get by adding the RHC and LHC terms back together.  No ellipticity is introduced.  However, differential absorption, so that the RHC and LHC terms are no longer equal in amplitude, WILL create ellipticity.  If there is 10% (0.46 dB) differential absorption, then the result will be elliptical with an axial ratio of about 13 dB.  It will take 100% absorption of one of the RHC or LHC terms to give a 1:1 axial ratio (or 100% circular) polarization.

  In case (b) differential absorption creates no ellipticity, but differential phase shift, from the differential refractive index, will.
This is what my calculations so far are all about, and seem to indicate that ellipticity will in effect never be introduced at 144 MHz.

   I don't know whether the effect in (a), differential absorption of RHC and LHC terms, is strong enough to produce significant ellipticity at 144 MHz. More sums are needed!

    In other words, I just don't know.

    I've asked one of my former colleagues, who has written some classic and oft-quoted papers about interpretation of polarization measurements in radio astronomy, to help out here.  He owes me a favour.  I'll see what he comes up with.

On 6-Feb-1998 AA7FV wrote:
To: John Regnault
From: Darrel Emerson on Fri, Feb 6, 1998 10:52 pm
Subject: Revised calculations.

Hi John and Ian.
   I got some help from an astrophysicist friend of mine to check the calculations on "how far through the ionosphere does a wave become circularly polarised?"  He immediately found a problem.  (The friend is Kurt Weiler, K7BLT, who does research at the Naval Research Labs.)

   My earlier calculations were wrong by a large factor.  The problem was that the textbook I took the equations from had over-simplified the equations.  Going back to the more rigorous form of the equations, the numbers are very different.   Interestingly, in comparing different reference books on this, two of the well known standard texts had errors in at least one of the relevant equations.  Given the assumptions, I now have much more confidence in the following figures.  As before, it's a "worst case"  scenario, with propagation precisely perpendicular to the magnetic field lines, with a linearly polarised wave having it's E-vector at exactly 45 degrees to the magnetic field lines.

   John, your intuition, that the differential phase shift between the linearly polarised ordinary and extraordinary waves ought to be comparable to the differential phase shift between transverse RHC and LHC terms, was absolutely right.  The earlier message I sent you claiming there was a factor of about 2 orders of magnitude between the relative phase shifts, was wrong; it was based on the oversimplified equations.

One-way distance through a uniform ionosphere, propagation normal to the magnetic field, in which a linearly polarised wave, E-field 45 degrees to the magnetic field lines, will become elliptically polarised to the indicated degree.  AR is axial ratio of the polarization.

50 MHz 144 MHz 432 MHz
(AR=0 dB) 100 2800 76000 (km)

AR=6 dB 70 1700 45000

AR=10 dB 49 1000 30000

AR=20 dB 14 400 10000

   These are fairly extreme values;  Propagation along the field lines introduces no ellipticity (to a good approximation anyway).  If the E-field of the wave propagating across the magnetic field is polarised parallel to or perpendicular to the magnetic field, NO ellipticity is introduced.  If the propagation is not quite perpendicular to the magnetic field, then Faraday rotation will cause the major axis of the ellipse to rotate, in the normal way.  If the Faraday rotation is an exact multiple of half-turns, then the ellipticity will cancel out exactly.  This factor alone will limit how elliptical a wave can become, especially if Faraday rotation of several turns is involved.  You never get more ellipticity than can be built up in a quarter-turn of Faraday rotation.

   So, the general conclusion (assuming I've not screwed up again) is that at 432 MHz there will never be significant ellipticity introduced by the ionosphere.  At 144 MHz 20 dB axial ratios may be common, but 6 dB axial ratio will almost never happen (although maybe it could approach that at very low elevation angles exactly perpendicular to the magnetic field?) At 50 MHz, high ellipticity, even perfectly circular polarization, should be common.

   This doesn't take account of other extreme conditions, where major ionospheric disturbances may temporarily give extreme electron densities.

I believe the numbers should be representative of a "normal" ionosphere.

   Does this agree more with your practical experiences?  Unless another major screw-up is discovered, I'll post the numbers to the moon-net, with a short explanation.  I may put the details of the calculation on to a web page for anyone interested, and so folks can find more of my blunders.

      Any thoughts?

On 9-Feb-1998 G4SWX wrote:

This looks fine to me as it stands but I have some further worries about the way that we are going.

If you look at the "classic" derivations of the ionospheric refractive index in the first two lines you get a comment " this proof ignores collisions".

If you take the case for waves transverse to the magnetic field and add in collisions you end up with some additional complications. One work that I have read says " n^2 becomes complex  and its modulus never fades".............."it can be shown that the extraordinary mode wave is always attenuated more than the ordinary mode wave".

If this is the case then, dependant upon the particle input to the ionosphere and where the collisions take place. I had assumed that the D layer, where most of the classical absorption takes place has little effect upon the Faraday rotation and that most rotation was in the f layer. If there are collision effects in the e and f layers then the attenuation of the extraordinary wave will lead to greater elliptical polarisation that simple theory would predict.

Or is it that the more rigorous form of the equations takes account of collisions ?

What do you think ? am I heading in the wrong direction ?

On 9-Feb-1998 AA7FV wrote:
Hi John,
   No I think you're going in exactly the right direction!

   I believe that collisions, implying a differential absorption between the extraordinary and ordinary rays, can at least in principle generate ellipticity for the LONGITUDINAL propagation case.  If the extraordinary ray were 100% absorbed, you'd be left with 100% circular.  However, if anything, absorption on the TRANSVERSE propagation would actually REDUCE the amount of ellipticity introduced; if the extraordinary ray were 100% absorbed, no matter what the plane of polarization of the incident ray, you'd be back to 100% linear polarization parallel to the magnetic field with zero ellipticity.  (There's another approximation there, but I think it's reasonable.) The equations I used did not take collisions into account.

   I'll try to do some more sums including collisions.  By the way, as regards where most of the Faraday rotation takes place, it's just proportional to the integral of B.N.dl - i.e. the product of magnetic field and electron density at each point for longitudinal propagation, but proportional to the integral of (B^2).N.dl for the transverse case.  The transverse Faraday rotation decreases as 1/f^3, instead of the 1/f^2 of longitudinal Faraday rotation.  The transverse Faraday rotation is also weaker by roughly (1/f) than the longitudinal case, with f in MHz, so at 100 MHz and above can reasonably be ignored.

   Since I think the magnetic field varies much more gradually then the electron density, the rate of Faraday rotation will be highest where the electron density is highest.

          I'll get back to you on the collision question.

On 9-Feb-1998 G4SWX wrote:

The great debate continues..........
I had a number of doubts about how well Darrell's calculations fitted observed results. Despite comments from Leif, I still beleive that a lot of which we observe as "conditions" on 144MHz eme are not straight absorption but distortion of the polarisation of the signals. Of course the proof of the pudding will be in real results but I am hoping that this rather messy
analysis will further our understanding of what is going on.

On 31-May-1998 Darrel Emerson  wrote:
  I finally got round to tidying up (although not that much) the notes I'd been sending you both about the ionosphere introducing ellipticity into linearly polarized signals.  My postings to the moon-net seem to be bouncing back again, so I'm taking the liberty of telling you both directly - for what it's worth.

   I've put everything together on to a web page at:

  I've included the effect of collisions, as well as the more normal mechanism responsibility for introducing ellipticity.  The above page includes links to the relevant parts of my Mathcad worksheets, which includes all the relevant equations, constants and assumptions involved in the calculations.  Of course, this doesn't mean I've actually got anything right, but it should make it easier for anyone with   nothing better to do to see where I went wrong.

  As I'd mentioned before, the rough conclusion is that you don't get much ellipticity introduced at 144 MHz, much less at 432 MHz, in "typical" ionospheric conditions.  Things may be very different in abnormal conditions though.

About the "null" (Thanks SM5BSZ)

"The null to be filled" refers to a measurement of signal level versus the element angle. If you receive a signal that is purely linear in polarisation with a linearly polarised antenna, the signal amplitude will be exactly cos(P) where P means the angle between the polarisation plane and the plane of the antenna. When the angle P is exactly 90 degrees the signal disappears completely. That is the null.

In real life there are never exact nulls, the null is "filled" to some extent. When the null is filled to -6dB, the weakest signal one can find by twisting the polarisation angle is 6 dB below the maximum signal (90 degrees away)

The "null filling" happens because one or the other antenna radiates a little in the orthogonal plane with a phase that is 90 degrees out of phase with the main beam. That means that the polarisation contains a component of circular polarisation.

The null filling does not tell from where the circular component comes. It may be the transmit antenna, the receive antenna or the propagation channel.

An observation of a perfect null does not mean perfect linear polarisation - it can also mean that the antenna happens to be elliptically polarised exactly as much as the elliptical polarisation of the wave - but in opposite direction.

In principle (of course) both the reflection off the moon and the passage through the ionosphere cause some contribution from circular polarisation. What I am trying to make clear is that this is a very small effect on 144MHz - it does not effect EME communication at all except during extremely strong aurora, and only for signals passing the ionosphere in the auroral zone.

(Note: I'm not showing any E-Mail address here in order to avoid them from being collected by SpamBots. You can possibly find the E-Mail addresses of the above OM at QRZ.COM.)

This discussion is still not closed. If you have any opinions or additional related information you would like to be published here, just send me an E-Mail

Read other Moon-Net discussions

Share: Twitter   Facebook   Buzz  
Home page Site map Radio Sherlock search engine Join the Mailing List ! 

Select Language English EspaƱol

0 user comments

Add your public comment to this page. (Important: Only respectful comments directly related to the content of this page will be accepted here. Inquires and complains must be sent to the webmaster.)

Name / Callsign
E-mail (Will not appear online)
To prevent automated Bots form spamming, please enter the text you see in the image below in the appropriate input box. Your comment will only be submitted if the strings match. Please ensure that your browser supports and accepts cookies, or your comment cannot be verified correctly.