What's different on 47 GHz EME ?

adapted from the essay from DJ7FJ Josef Fehrenbach:
What's different on 10 GHz EME
to
What's different on 47 GHz EME

by PA0EHG

1. Introduction

The differences on 47 GHz EME, compared with the well known conditions using 2m or 70 EME are pointed out in this essay.
On 47 GHz EME earth atmosphere is also in important factor due to attenuation in the atmosphere. Effects for higher temperature and humidity causing extra attenuation and therefore the elevation used is an important element in working 47 GHz EME.
There are severaol effects of the very small beamwidth and the change in characteristic of reflection at the moon.
Also, the received moon noise causes a reduction of sensitivity in the receiver.

2. Path loss - RADAR equation

Usually, the path loss is calculated using the RADAR equation.
It is given by:
Pr =
(s.Pt.Gt.Gr.l²)/((4.p)^3.d^4)
Pt :   transmitted power
Gt :   gain of the transmitter antenna
Gr:    gain of the receiver antenna
l :    free space wave, length
d :    distance between transmitter/receiver and target
s :    cross section of the target

In this basic form of the RADAR equation, the following units have to be used:
- power in 'Watt' or 'mWatt'
- antenna gain in absolut terms
- wavelength and distance both in the same unit e.g. 'm'
- cross section in m
²,

The effective cross section a of a target is defined in the unit of an area.
Using a sphere with a well conducting surface and a circumference of more than 10
l,
it can be calculated as:
s = r².p = d².(p/4)

If the surface of the sphere is only poor or non conducting, e.g. the moon, the cross section has to be multiplied with a corresponding reflection coefficient.

The effective cross section of the moon can therefore be calculated as:
- diameter  =     3.500.000 m
- reflection coefficient
=   6.5%   =  0.065

s = (3.5.10 ^6m)².(p/4).0,065 = 6,25.10^11 m²

So, the moon has the same characteristic of reflection as a well conducting sphere with a cross section of 6,25. 10^11m²
If the target isn't a sphere, equivalent values for s  have to be measured or calculated.
For practical use, in dB, the RADAR equation can also be written as:

Pr = P(t) + G(t) + G(r) + 10 log s (m²) - 20 log f (MHz) - 40 log d (km) - 103,4

Calculating the path loss instead of the received power, the values P(t) and the antenna gains G(t) and G(r) have, to be fixed to the factor 1, equal to 0 dB.
With the moon, considering an average in distance, the path loss on 47 GHz can be calculated as 302 dB.

Attention:
The accuracy of the RADAR equation is only sufficient considering the following conditions:
a) The beam of the transmitter/receiver antenna has to cover the target several times.
b) It is not allowed, that the target is formed by combining spreaded single reflectors.
On 47 GHz EME, both restricting conditions are broken.

3. Antenna beams

Observed from earth, the moon covers an average cross section of 0.5º width.
- in perigee with a distance of 356,400 km
Þ 33'32"
- in apogee with a distance of 406,700 km
Þ 29'14"

On 47 GHz a dish of 0,2mtr has a 3 dB angle of 2º. Therefore, a 0.8 mtr dish has a 3 dB angle of 1º and a 0.9 mtr dish only 0.5º.
Larger dishes cover only a section of the moon.

4. Reflectors

On lower frequencies, e.g. 2 m, usually 50% of the backscattered signal is reflected from a small central area on the moon.
RADAR echoes on 400 MHz already show, that not only the center, but also the other areas up to the rim cause a part of the reflection.

With RADAR echoes in the X-band, this effect is increased

Therefore, on 47 GHz the whole visible area are more or less producing reflections. Approximately, the moon surface is a combination of a huge, number of single reflectors. So, the echo is scattered and contains parts with different travel times.
References for this thesis are:
b) Doppler smear on 47 GHz QSO's

5. Doppler and dopplersmear

The basic doppler on 47 GHz can be more than 90 kHz. As the doppler can be calculated in advance by different programs, this shouldn't be a problem. The dopplersmear is a SHF/EHF specific problem on EME. Containing spectral parts from both sides of the center frequency, the CW-signal in the RX isn't a clear single tone. It sounds a little bit similar to rain-scatter signals, although the rain-scatter signals might be from a rougher type. Using signal processing care has to be taken about the covered bandwidth. As a reason for the signalsmear, the libration of the moon was assumed several times in previous applications. The authors opinion is, that the main part of the smear is caused by the doppler of different moon areas formed by the rotations of the earth below the moon.

Examples:  Conditions:    - The test stations are in the northern hemisphere.
- moon in perigee or apogee

-   a) the moon is for the test station in culmination.
There is no variation in distance to the center and the northern and southern parts of the moon
Þ no basic doppler occurs.
Caused by the rotation of the earth, the test station travels to the left rim of the moon and simultaneously away from the right rim.  So, the left side produces a positive doppler and the right side produces a negative doppler with identical value.  The maximum frequency shift will be approx. ± 180 Hz.
The spectral distribution isn't exactly known to the author.  With wider antenna beams, it is supposed to be a cos
² function.

b) The station works at moonrise or moonset.
Here, the normal doppler has it's maximum and the relative variations in distance to the lateral rims are in a minimum.  Therefore, also the dopplersmear has its minimum.

Fig.  Nr.3 shows the relations between doppler and pass of the maximum frequency shift by the dopplersmear.

- c) QSO between stations in North America and Europe, both located around the same latitude.
Many variations of dopplersmear are possible.  If the moon has the same elevation for both stations, the smear is at minimum.

- d) The test station uses a very big antenna. e.g. 4.5 m dish (3dB beamwidth ~ 0. 1°)

The test station illuminates only parts of the moon.  As the antenna doesn't illuminate zones with a big relative motion, there is only a little dopplersmear, also when the moon is in culmination.

0H6DD has already seen this effect (see DUBUS 2/94).  He worked with a very big dish during previous tests and  found out, that almost no dopplersmear occurred.

6. Additional attenuation by the atmosphere

Fig.4 shows typical values of the Faraday-Rotation.

This figure says, that there is no considerable Faraday-Rotation on 47 GHz.   It is therefore not necessary to work with circular polarisation.  Using linear polarisation, you have to take care about the basic rotation for reflections at the moon.
The American recommendation, used by most of the active stations, seems to be a usable compromise.
It is called: - stations in North America work with horizontal polarisation
- stations in Europe work with vertical polarisation
Considering this, the rotation among American stations is kept low, among the European stations it is even lower.  Between America and Europe there are always approx. 90
°.  Currently, 47 GHz EME stations only exist in America, Japan and Europe, so you can contact every partner in a session without any modifications. For optimal performance as needed on 47 GHz it is advised to adjust polarisation for testing between JA and W

8. Antenna and feed

On mm bands the sun and the moon are no longer usable as point sources. When using large dishes on 47 GHz, larger than 1 meter, the feed cannot be optimized by simply adjusting for optimal sun noise.

For optimalisation of the feed focal point it is not enough to adjust the feed focal point at maximum solar or moon noise.
Read this document for an alternative method in adjusting the feed focal point

9. Noise sources

• The earth-noise should be minimised by optimising the antenna.  Nevertheless, this part of the ant -noise will always be 20 to 40K.

• The galactic background noise could be ignored, because it's lower than 1O K.

• The antenna noise temperature on 47 GHz is strongly dependant on antenna elevation and temperature and humidity.
At low elevation angles (below 25 degrees) the noise from atmospheric losses is the most important effect for 47 GHz EME

• On 47 GHz, the moon-noise, causes the second most important effect for the sensitivity of the system.  It can be seen as a noise source with a temperature of arround max. 28O K.  On one hand this is a positive effect, as you can use the moon-noise for the tracking of the antenna.  On the other hand, it limits the sensitivity, especially with larger antennas and therefore the noise factor of the pre-amp gets less important. Probably when using state of the art low noise amplifiers combined with cooling of the LNA the moon noise will become a limiting factor in sensitivity. It should be considered to use the lowest period of moon temperature for EME testing, a few days after new moon. Moon temperature will be at it's lowest level and this will bring an advantage of a few tenth of a dB.

• As today noise factors of lower than 3 dB are common, the preamp shouldn't cause any problems. The loss of the RX-feed should be minimised by using waveguide relays and placing the preamp direct at the feed.

In the graph is shown the great variation in moon temperature during a month between 208 and 275 Kelvin. This causes a variation in measured moon noise of almost 0.35 dB.
For a station with a noise figure of 5.7 dB typical moon noise will vary between 0.45 and 0.6 dB
For a station with a noise figure of 2.5 dB typical moon noise will vary between 1.07 and 1.4 dB

10. Moon tracking

Using programs, e.g. from VK3UM, it is possible to calculate the azimuth and elevation angle or the declination and rectaszension with a sufficient accuracy. Nevertheless, it is recommended to control the alignment of the antenna continuously and check tracking with the moon-noise.  Therefore, a broadband noise-RX in a parallel receiver section can be used.
Modern SDR receivers can assist in this like Spectravue by using the Continuum mode for monitoring moon noise level

11. Power density in an antenna beam

In order to have a closer look at the reflections on the moon, it is necessary to analyse the power density in the antenna beam.
The antenna used in the following tests is a parabolic dish with a standard pattern characteristic.  The first side lobe is 20 dB down and the second approx. 26 dB.  All the other side lobes are ignored.  There is no difference between E- and H-plane in the radiation pattern.
A numerical method is used to get the power density of the antenna.  So it is possible to avoid complicated mathematic methodes.

Fig.6 shows the radiation pattern and the equivalent distribution of power density.

Some explanations to this graph:

- The x-axis is the rotation angle.
- Only one half of the pattern is displayed.
- The factor 1 on the x-axis marks the 3 dB point or the half 3dB angle.
- The left y-axis shows values of relative power level of the radiation pattern.
- The right y-axis shows relative values of the transmitted power. They are related to an illuminated circle area that covers the double angle as
marked on the x-axis, observed from the antenna source

For example:

• How much power is in the antenna beam from the maximum to all 3 dB points?
Answer: Approx. 55% of the transmitted power.

• How much power is in the antenna beam up to the double of the 3 dB angle?
Answer: Approx. 90% of the transmitted power.

The result for the reception of moon noise shows Fig.7.

The non-linear gradient in this graph is caused by the radiation pattern of the antenna.

Examples:

• How much moon noise is received using a 0.9m dish?

• How much noise is received with a 0.22m dish?

A result of this investigation shows, that the noise temperature (noise figure) of the RX is not critical, working with a large dish.

12. Relation between antenna beam and received power

In this chapter is only taken care about the absolute received power.  The signal/noise ratio will be considered in the following chapters.
Fig. 8 shows the connection between antenna size and received power level, assuming a constant transmitted power.  For transmitting and receiving, the same antenna is used and a dish with 0.88 m in diameter is taken as a reference.

The diagram shows that with antenna diameters smaller than 0.44 mtr the RX-signal behaves as commonly assumed.  At diameters from 0.44 to 0.88 mtr appears an additional loss of 1.5 dB. The cause for this is the power density in the radiation pattern as the 3 dB points now approach the moon's rim.
With diameters from 0.88 to 1.2 mtr the signal increases again like it does with smaller antennas, but a basic loss of 1.5 dB has to be added.
At larger diameters, the curve bends, and by using dishes of more d= 1.7 mtr, the receiving level increases not with the double antenna gain.  It increases only with the single amount.

Explanation:
Using small dishes (relative wide antenna beams), only a small part of the beam illuminates the moon.  Considering Fig. 8, there is an almost linear relation between beam width and power level with antenna sizes.
Is the antenna beam approx. 0.5
°, the moon therefore fully covered, the power density on the moon is direct related to the radiation pattern of the antenna.
With large dishes (narrow beam widths), the beam illuminates only a spot at the moon.  So, almost all transmitted power reaches the moon (as with LASER), only reduced by the loss in the atmosphere.  The power reflected by the moon can not be increased by more antenna gain.  The losses of reflection are assumed to be the same, whether the diameter of the spot is 1000km or 100km.
The reflections on the moon surface are diffuse.  When the same antenna is used for transmitting and receiving always the same moon spot is illuminated or observed.
Using bigger dishes, the received power increases only with the single gain factor.  Now, more gain can only be made on the receiving side.  The moon always reflects the same sum of energy.  If the surface of the receiving antenna increases, this will have an effect on the power level.

What happens with the received power when stations use different antennas?
With antennas up to 1.3 m in diameter everything is O.K. If only one partner uses a bigger antenna (>1.3 mtr), nobody gets a higher RX-signal level. Only the Doppler smear decreases.

Explanation:
The OM with the relative small antenna illuminates the moon almost completely.  The OM with the big antenna only observes a spot at the moon.  If he now increases the antenna, his receiver level stays almost the same because the observed spot becomes smaller.  The reduction in the spot size is compensated by the bigger antenna size.
This effect is reciprocal. It's also valid when the OM with the big antenna is transmitting.  He only illuminates a spot at the moon; therefore the OM with the small antenna receives the same power level, independent from the spot size.
To simplify the calculation of the path loss in such combinations, the gain of the big antenna can be fixed to the gain of a 6 mtr dish.
If both partners use big antennas, the station with the smaller antenna produces the limiting factor.  To calculate the path loss, both antennas have to he set to the same size, equal to the dimensions of the small antenna.

Conclusion:
The effects of big dishes from institutes, e.g. VE3ONT at Algonquin Park, on 'normal' partners are the same as with dishes of only 1.3 m in diameter.

13.  Relation between antenna size and signal/noise ratio

When the antenna gain will be increased in lower amateur bands, the signal/noise ratio also increase.  This relation changes, when the moon noise starts to add a contribution to the basic noise that is worth mentioning.  On 47 GHz, system noise of approx.  200 K looks realistic.  It is, e.g., consisted of 180K from the preamp and the receiver and 30K external noise received by the antenna (earth, feedloss etc.). A system temperature of 200K seams to be possible without cooling and a temperature of 50K will be lower limit with cooling the pre amp.

On 47 GHz EME the received noise temperature of the moon has to be added to the system temperature.  Using 47 GHz, the moon has a max noise temperature of approx. 210 K, The received amount of moon noise depends on the size of the antenna, only big antennas (diameter > 1.3 mtr) receive the full noise power.  The smaller the antenna, the smaller the amount of noise power that is received.

Considering the antenna beams from chapter 12 the noise parts can be found in Fig.7.
Tab. 1 shows the relations between the antenna diameter, gain, beam width and the amount of noise received from the moon.
Of course, the received noise from the moon has an influence on the sensitivity of the system.  With a 0.44 mtr dish the overall noise temperature rises about 40K and a 0.88 mtr dish adds even 120K to the system noise.  Some examples have been worked out to have a closer look at this effect.  Therefore, dishes from 0.44 to 1.7 mtr and system noise (without moon noise) of 50K, 100 K and 150 K are used. The reflected signals from the moon were related to the according moon noise using a fixed RX-bandwidth.  In order to get an easier look at the relations, it will be assumed that the partners antenna beam is wider than the moon and therefore the reflected signals are from the same distribution (room and area) as the noise signal.

The meaning of the used factors are:

0 dB        - In the used bandwidth, the received noise from the moon has the same value as the reflected power.
3 dB        - The signal power is two times bigger than the noise power of the moon.
6 dB        - The signal has 4 times more power than the moon noise.
10 dB      - The signal is 10 times bigger than the moon noise.

Tab 2 and Fig 9 show the effects on the signal/noise ratio in the correct form (signal + noise)/noise.  The results are really interesting.

It can be seen at once, that already an antenna diameter of 0.88 mtr is the limit. Bigger antennas don't produce a better signal/noise ratio. On the other hand, the receiving loss using a 0.44 mtr dish is relative small.

14. View to the combination of antenna gain, noise, and power

Working with 20 W, a 0.66 mtr dish and a RX noise figure of 2 dB the own echoes can not be heard.  Using this combination as a reference, the following equipment and its effects are displayed:

• Increasing the power is the only effect, that directly effects the received power of the partner.

• Antennas bigger than 0.66 mtr hardly improve the own receiving quality, but they produce, as long as they are smaller than 1.8 mtr, a bigger RX-signal for the partner.

• Antennas smaller than 0.66 mtr can be useful as receiving, as long as a very good preamp is used.  But the power should be increased to produce sufficient receiving power at the partner.

• Improving the System noise to values lower than 3 dB only produces a positive effect, when small dishes are used.

• Antennas bigger than 1.3 mtr in diameter are not very useful.  The partner with the small antenna receives a signal like transmitted with a 1.3 mtr dish.  The OM with the big antenna doesn't receive a stronger signal as by using a 1.3 mtr dish.  Only the doppler smear decreases.

An improvement in signal/noise ratio is only made when both partners use dishes with more than 1.3 mtr, under the condition, that both illuminate the same spot at the moon.

16.  Conclusions

•  Most amateurs at this time use dishes larger than 1.3 mtr, the beamwidth of these dishes are less than 0,3 degrees to 0,1 degree

•  Measuring moon noise with a large dish will not be influenced, or only very little, by the distance to the moon. Measuring at Perigee and Apogee will get almost the same moon noise levels. In both cases the beam hits only a small part of the moon. The received moon temperetaure remains the same and therefore the received moon noise will remain the same.

• Measuring moon noise with large dishes will not give any indication on the performance of the used dish

• Large dish moon noise level does not tell anything about gain or efficiency from the dish

• Measured moon noise is determined by losses due to antenna temperature, elevation angle related to atmospheric losses and overall noise figure of the receiver system used.

• Elevation, temperature and humidity are important factors responsible for large atmospheric losses, best to use high elevation at least above 20 degrees preferable higher, in early spring time with low temperatures and low humidity.

• Moon noise varies very strong related to the phase of the moon and the moons surface temperature

• Reflected signals from the moon are not a sharp beam

• For reflections (for QSO's) we need to use Perigee for minimum loss

• Measuring solar noise with large dishes will not give any indication on the performance of the used dish

• For QSO planning the best period will be with low moon noise, perigee, highest possible elevation on both sides, low temperature and low humidity and last but for sure not least low spreading

The VK3UM tool is not taking the effects as mentioned in this paper into account , if we calculate a single way EME contact between W5LAU and DL7YC we see the calculations as original by VK3UM giving the following results:

Both stations are in the calculations at 2.4 mtr dish and atmospheric attenuation is set at 3.6 dB
the received signal is -10.97 dB that is almost 10 dB different from practical result

Now we reduce antenna gain conform the model What's different on 47 GHz EME and we get:
Transmit antena gain reduced conform the statement above to 1.3 mtr diameter

Receive antenna gain reduced to 1.7 mtr diameter,
as stated above the gain should be for a 1.2 mtr dish 51.26-1.5 = 49.76 dB
increase from 1.2 mtr to 1.7 mtr lineair increase, for a 1.7 mtr dish 54.28 dB -1.5 = 52.78 dB
increase from 1.7 mtr to 2.4 mtr, increase half of normal increase, from 1.7 to 2.4 mtr dish is 3 dB increase
so 52.78 +1.5 = 54.28 dB which corresponds to a dish of 1.7 mtr

Moon noise calculated is very close to real measured values
Sun noise also seems quite close to real measured values, for a good sun noise calculation the sun flux has to be increased

In this table only look at the results from the home station

The dish from W5LUA is reduced to 1.3 mtr dish and the dish from DL7YC to 1.7 mtr, atmospheric attenuation set to 3.6 dB
received signal is -19.24 dB that is about 1 dB difference between practical results
The 1 dB difference can be explained by higher atmospheric losses and perhaps also the difference between Apogee and Perigee which might not be longer valid when using dishes larger than 1.3 mtr.

In this table only look at the results from the home station

In this last calculation I increased the distance of the moon to Apogee.
In this model it is not yet clear if the difference between Apogee and Perigee should be considered,
it should be confirmed by practical measurements

Up to now this is the closest calculation I can get giving also good results for moon noise and sun noise.
For S/N received this is only 0.5 dB different from the measured result.
This last difference probably is caused by uncertainty of atmospheric attenuation and perhaps also by effectiveness of the used dishes.

17. Conclusion

This essay shows some differences you have to deal when moon bouncing on 47 GHz.
The conditions on other EHF-bands are similar but atmospheric circumstances will bring big differences.
Some of the physical approaches are not in the highest scientific exactitude.  But the author is convinced that they are of sufficient accuracy for practical use.  Some QS0s and noise measurements with the moon prove this.

For 47 GHz EME this means that power output needed should be 20 Watt or more and for Receive side the dish should be larger than 1.8 mtr or the noise figure in Rx should be below 2.0 dB and losses between feed and pre amp minimized.

At power levels of 10 Watt the chance of a successful contact are reduced dramatically or even impossible. For testing with high elevation levels the atmospheric loss will be much lower and possible contacts could be possible.