Discussion: Antenna Diversity

[Rayengine]

“fading null”

Hi,

Any good articles best describe the antenna diverity?
If we assume using two antennae
(1) How these two antenna oriented?

(2) In what distance should be separated ?

Any idea welcomed!

Rayengine

 

[flatulent]

orientation

All of this depends on the situation. In the case of wireless LAN, each of the two antennas should be oriented to make the received signal largest in the case that each one is the only one. For instance, if the transmitter is vertical polarity, the receiver should be vertical polarity. They should be a quarter wave apart. This insures that if one is in a multipath null, the other one will not.

In the case of long distance communication, they should be 20 wavelengths apart. The orientatin depends upon the path from the TX to the RX. If tthere is Faraday rotation, one antenna should be vertical polarity, the other horizontal. If the main problem is fading, both should be the same as the TX.

 

[viterbi]

As a continuation to the above discussion,
What is the benifit of using cross polarisation antennas at the Tx and Rx stations? I saw same the the GSM transreceivers using the cross polarisation antennas and they don't incorporate any diversity. I would like to have a detailed description. Any articles are warmly welcomes

 

[flatulent]

cross polarization

You may have seen a case of circular polarization. The usual cross diversity is used when going through long distances of the atmosphere where the polarization can be rotated. For short multipath environments, the circular polarization is an advantage because reflections reverse the rotation. This way the receiver can get rid of some of the multipath signals by the rotation discrimination. The diversity is made by trying both rotation directions and selecting the one that has the best signal. To get circular, one antenna is fed at a 90 degree delay from the other. To reverse, the 90 degree delay is fed to the other antenna.

Another thing that may be happening is that people rarely position their portable radios with the antenna perfectly vertical when they use them. The two (H and V) antennas may be to pick the one that is the most vertical or to take advantage of the multipath reflections being less severe on one than the other.

 

[Rayengine]

Thanks for your valuable experience!

One more question, if it necessary to made the shape of antennae similar to ensure a good performance.

Any papers - related to Fading with respect to Antenna Diversity.

 

[flatulent]

radiation polarity only

Once the radio waves leave the antenna they only have the properties of amplitude and polarization. Any antenna shape that produces the polarization you want will do and the amplitude is controlled by the power into the antenna and the antenna gain.

There are two main diversity research areas. In the old days it was HF through the ionosphere. More recently it has been VHF and UHF in multipath situations. The first had just a few multipath components, the latter has many. I do not know of the papers on the subject. Most books on mobile communications will have information o the latter. Look for the key words “Raleigh fading” when you do a search.

 

[dd22]

What is cross polarization? My understanding to GSM polarization is that GSM does not need polarization diversity. The reason is simple: most of GSM mobile use linearly polarized antenna, such as dipole or similar. The antenna polarization for base station should be the same as mobile. People talk about antenna diversity, most of time, it is referred to space diversity. If the direction of antenna is very good, we can achieve some benefit in frequency effiency. The polarization diversity are rare for mobile communications, it may help for some terrian situation.

 

[flatulent]

hand held units

Hand held units are rarely operated exactly vertical or horizontal. Most mobiles I have seen have antennas tilted 30-45 degrees from vertical. I am just guessing what the cell site is doing.

Cross polarization is what makes zero coupling. In linear it is 90 degrees rotated. In circular it is the opposite rotation.

In some countries FM broadcast is circular because home antennas are horizontal and mobiles are vertical. You can take advantage of this in home antennas by using circular and getting rid of some multipath.

Space diversity is very easy at UHF. It is frequently included in one antenna radome. It is expensive on HF requiring separations of 10 - 100 wavelengths which requires a lot of land.

 

[Rayengine]

Now I play with the digital phone design using Frequency Hopping (slow hopped 100Hz), I have two question?

(1) Someone tell me that antenna diversity is useless in frequency hopping system! Is it correct ?

(2) How to verify the performance of antenna diversity compared with no antenna diversity because someone claimed that it the system using antenna diversity has lower sensitivity in open field test ?

Best Rgs
Rayengine

 

[ms1234]

If there are two Rx ckt before making decision which one is better, I think diversity is useful. It depends on how you design the system.
As for lower sensitivity in open field test, I have similar experience when I test wireless lan. I think the effect that need diversity antenna not happened when test.
Diversity antenna switch ckt loss and periodic switching between 2 antennas to decide witch antenna is good maybe lower the packet error rate when test (This kind of test result is wrong. It should be test without switching antenna during the test).

 

[flatulent]

two things

The switch in the antenna circuit has loss which will reduce the sensitivity of the receiver.

I have a 1W frequency hopping cordless phone without diversity. It has worse performance than my old 0.1W or less direct spread one with divrsity as far as audio quality goes. This is because every where you stand at least one of the frequenies is in a multipath null and you and the person on the other end of the call hear a disturbing buzz sound as the frequency sequences. At very long ranges the buzzing increases because there are more multipath paths. I can receive a signal at least 10 times the distance but the buzzing prevenets voice communications. At short ranges within the room the two have identical good performance.

In packet networks there is always retransmission to replace the lost packets.

 

[MonkeyBusiness]

There are many articles on the subject.
You can search IEEE (https://ieeexplore.ieee.org/Xplore/DynWel.jsp) for 'antenna diversity gain', and you'll get many results. Unfortunately you need a subscription to download the actual articles.

 

[Rayengine]

I have consulted books about the separation of antennas, but there have two answer one says quater-wave and other say half-wave? Any one has information about fading null ? Please advise!

 

[flatulent]

nulls

The nulls are a half wave apart. If the antennas are a quarter wave apart, both of them cannot be in a null at the same time.

 

[yingyang]

antenna diversity

I do not fully agree with flatulent's comments on the separation....
It all depend on the frequencies of the two antenna. If they are not in their harmonics, a separation of more than 3*lambda by eight is sufficient....

However, if they are in the harmoinics, than flatulent is right...

 

[Darkseed]

I've seen valuable answers in this discussion but I think no one has explained why really diversity is good and how this affect to the good
distance between antennas.

The idea of diversity is getting two versions of the same signal (so,
incorrelated). This method is used usually, but not limited to as seen
along the topic, to fight against quick fading which cause is a changing environment near the receiver. This is usually modeled with Rayleight or
Rice models. In a Rayleight environment (Rice is similar), if
epsilon/lambda >= 1/4.13, where epsilon is the distance between
antennas, the correlation between the signals acquired in the differents
antennas is < 0.5. Then, it is considered they are incorrelated (this is the
origin of the lambda/4 measure defended by flatulent). More conservative
people use lambda/2. I have a nice graph about this (it is quite sinc
shaped) but my scanner is not working right now. :roll:

Hope that helps.

 

[flatulent]

it all depends

It all depends upon the environment. If you are inside an office environment with parallel walls, then there is huge multipath of almost equal signals which produce a sharp null pattern just like inside a waveguide. These nulls are a half wave apart.

The other extreme case is short wave recpetion across long distances. Here there is usually two paths from different layers (F1, F2, E) of the ionosphere. These paths have different lengths to different points on the earth. This makes their relative phase and amplitudes different at these points. The height of the ionosphere's layers are always changing with time which makes the signal level at any location vary or fade. The further apart two points are the less correlated are their fading.

An intermediate situation is land mobile in urban areas. There are less regular reflecting objects and fewer multipath paths. In the old days when lower frequencies were used (150 MHz) the driver of the vehicle would select a stopping point by slowing down and noting the fading pattern and then come to a full stop at a signal peak location.

 

[johnseven]

Hi,

Any good articles best describe the antenna diverity?
If we assume using two antennae
(1) How these two antenna oriented?

(2) In what distance should be separated ?

Any idea welcomed!

Rayengine

-----
RAYENGINE,
AS you've noticed there's a lot of diversity of ideas about antenna diversity.
From ITT's Reference for Radio Engineers:

I thas been shown that if two or more high-frequency radio channels are sufficiently separated in SPACE, FREQUENCY, ANGLE of arrival, TIME, or POLARIZATION, the fading on the various channels is more or less independent. Diversity systems make use of this fact to improve the overall performance, combining or selecting separate radio channels on a single high-frequency circuit.
Satisfactory diversity improvement can be obtained if the correlation coefficient of the fading on the various channels does not exceed about 0.6, and experiments have indicated that a frequency separation of the order of 400 Hertz gives satisfactory diversity performance on long high-frequency paths. (for FREQUENCY DIVERSITY)

SPACE DIVERSITY
Spacing between antennas at right angles to the direction of propagation should be about 10 WAVELENGTHS.
POLARIZATION DIVERSITY has been found to be about equivalent to space diversity in the high-frequency band.
TIME DIVERSITY
Measurements have indicated that times varying from 0.05 to 95 seconds may be necessary to obtain fading correlation coefficients as low as 0.6 in HF time-diversity systems.
ANGLE of ARRIVAL diversity requires the use of large antennas so as to obtain the required vertical directive characteristics. Differences in the angle of arrival of 2 degrees have been shown to give satisfactory diversity improvement on HF links.

Interesting topic
johnseven

 

[Rayengine]

In case of Polarization diversity, if I used two antennas which come out from same antennas port (i.e in the shape of XY), can it fullfill the requirement of polarization diveristy because of only antenna port. Any matching technique need to be noticed if we choose above configuration.

 

[egliu]

re

see the book of handbook of propagation effects for vehicular and personal mobile satellite system in ebook upload forrum