AN EXPERIMENTAL PATCH ANTENNA
months ago, after about an 18-year absence from amateur radio, I
decided to fire up my old IC22A on 2 metres and build some antennas.
I was fortunate enough to be introduced to the 2100 net by Ken,
VK3HKR and I found that the net was a very interesting forum which
discussed anything from painting to heavy technical discussions on
how a discone worked.
of the regular topics of discussion was working satellites. I
realised at this point if I was to be able to join in with these
activities, I needed to upgrade my equipment. So I lashed out and
bought second-hand IC271A and IC471A all mode transceivers.
building up various antennas and achieving some limited success
working UO14, FO20 and FO29 I came to the conclusion that to operate
with these birds, especially FO20 and FO29, you really needed some
sort of tracking equipment. As my Emotator Rotator was in bits
(Having pulled it apart 18 years ago to replace at pot) and my also
my QTH being a rented property, I didn't think that the landlord
would appreciate the erection of a tower in the backyard!
one of the 2100 net technical discussions one night, Robert VK3KRB
brought up the topic of 'patch' antennas and how patch arrays
could be electronically steered. The topic of 'patch' antennas
generally drew a blank from every body else, including myself and we
all decided that we didn't understand how they worked and these
antennas were really used at microwave frequencies and didn’t
really have any application for VHF and UHF,
had heard that there were a couple of high orbiting satellites
namely AO10 and
The big advantage with these satellites is that their footprint at
times covers half the world and the prospect that you could work the
States and Europe on VHF/UHF was quite exciting. The other advantage
was these satellites moved relatively slowly compared to the low
orbiting satellites which meant you could point your antennas at
them and be able to leave them there for a considerable period of
time without the need to track them across the sky.
then decided to concentrate on AO40 and managed to convert a 2.1 GHz
down converter to 2.4 GHz, down converting the downlink on AO40 to 2
Metres. After a lot of persistence, late one Saturday night using
this converter with its grid reflector, I managed to hear the middle
beacon of AO 40. It was only one S point above the noise. I could
also hear some weak sideband signals and some Morse code. In the
next few days I managed to transmit into AO 40 on 70 cm and heard
myself and couldn't believe the enormous time delay which seemed
to be at least 1 second.
first contact on AO 40 was to Charlie, VR2XMT in Hong Kong and
second was to Scott, NX7U in the States. After emailing Scott, and
having given him a signal report of 1*1, his report on my signal was
5*6. At this point I realised that my downlink was not working well.
I was over at Peters QTH, VK3DI, I spotted a large 1.5 m microwave
dish lying outside in his back yard, which Peter kindly loaned me. I
worked out that if I used this dish for the downlink, the additional
gain taking into account circular polarisation, I should achieve an
additional 10 db.
measuring up the dish to ascertain the focal length etc, I
calculated that I would need a 1.86 turn helical to optimally
illuminate this dish. The problem with this is that a 1.86 turn
helical would not perform as calculated as the general formulas for
helical antennas were only applicable for a helical of a number of
turns (probably 3 to 4 turns minimum). Any way I built a 2 turn
helical and this immediately lifted the AO 40 beacon from 1 S point
above the noise to 3 to 4 S points above the noise. I could also
hear a dramatic increase in the readability of other stations.
pleased with this I managed to work a number European stations and
even the UK.
was not convinced that the helical was illuminating the dish as I
thought I wasn’t achieving a 10-db increase over the grid antenna.
looking at the web for some clues, I found a commercially advertised
2.4 GHz dish which had a, you guessed it, a “patch” feed. The
radiation pattern was ideal and the axial ratio, which is a measure
of the antennas circular polarisation, was near perfect. (A 2 turn
helical does not have a good axial ratio).
“patch” feed had come back to haunt me.
referring the matter to the 2100 net technical forum, Robert, VK3KRB
emailed me some URLs on articles on “patch” feeds. One of the
articles was a very detailed description of all the antennas on AO
40 written by the people who built the bird. Low and behold, the
antenna used for 70cm was an array of 6 'patch' antennas. Whilst
it didn’t give a technical description of how they worked, there
was very detailed information on dimensions, feed points, impedances
and how to connect to create circular polarisation.
only technical description referred to a patch antenna being
equivalent to a couple of slot antennas. I thought it was time to
re-familarise myself with the principles of a slot antenna, so I dug
out and old textbook, which dealt with slot antennas and after
skimming over a lot of maths the final analysis, is relatively
slot antenna is simply a slot cut in a large sheet of metal. However
if you were to make a dipole out of the piece of metal, which was
cut out to create the slot, there is an interesting relationship
between that dipole and the slot created.
two antennas are complimentary and almost exact duels, in other
words everything is opposite i.e.
The centre feed point impedance of a dipole is minimum
at the centre corresponding to a current maximum and impedance
maximum at the ends corresponding to a voltage maximum.
The centre feed point impedance of a slot is a maximum
across the centre of the slot corresponding to a voltage maximum,
and impedance minimum at the ends corresponding to a current
The impedance of a centre fed resonant dipole in free
space varies from about 60 to 70 ohms (depending how fat it is) and
the impedance of a centre fed resonant slot varies from about 590 to
510 ohms (depending on how wide the slot is). The fatter the dipole
the lower the impedance, the wider the slot the higher the
A horizontal dipole creates horizontal polarisation
whereas a horizontal slot creates vertical polarisation.
how does this relate to the 'patch' antenna?
having to do further reading on the subject and not knowing whether
this is strictly correct, the analogy I see is that a “patch”
antenna is the unbalanced form of slot antenna as the quarter wave
ground plane antenna is an unbalanced form of a dipole.
“patch” antenna has 2 vertical slots and two horizontal slots
created by the space between the edges of the patch and the ground
plane. The horizontal slots radiate vertical polarisation and the
vertical slots radiate horizontal polarisation. The patch is
grounded in the centre so the feed point impedance varies from 0
Ohms at the centre to somewhere near 300 Ohms at the centre of the
patch edges (assuming my analogy is correct). So the “patch”
antenna is like 2 separate antenna sets in one (although I am not
sure whether you could feed each antenna set with a different
both polarisations available lends this antenna to creating either
right or left hand circular polarisation by connecting the
horizontal and vertical slots together with a quarter wavelength of
coaxial cable to provide the 90 degrees phase shift. This is similar
to how crossed yagis are connected to create circular polarisation.
I was ready to built one of these for 2.4 GHz using the dimensions
quoted however after further discussions with the 2100 net, Dave,
VK3AAD suggested I build one for 70 cm to make sure it works before
building the 2.4GHz version
is strongly that prior to building on of these antennas you:
Read the article on the AO40 70 cm 'patch'
Adhere closely to the dimensions given.
ground plane I used in the prototype was a piece of aluminium sheet
600 * 600 mm. The size of this is not critical and this can be
reduced somewhat. I believe that the ground plane should be at least
100 mm larger than the patch (providing 50 mm extension of ground
plane on each side of the patch)
patch is made of 1 mm brass sheet and measures 315*315 mm. I decided
to use brass sheet as this would make it easy to solder a connection
to (There is no reason why you couldn’t use aluminium sheet). This
was found to resonate at 438 MHz. This is slightly smaller than
quoted in the Amsat article as I found that the original size of
0.47 wavelengths (325*325mm) resonated somewhere below 430 MHz.
centre of the patch is grounded and spaced from the ground plane a
distance of 12mm, by using a 3/16 stainless steel bolt with two
stainless steel locknuts used as spacers. The patch is supported at
the corners at a height of 12 mm using ¼ inch nylon bolts with two
nuts used as spacers. The support points in my prototype are 40 mm
from each corner along the diagonals of the patch.
the “patch” antenna is achieved by running the coax from behind
the antenna, through a hole in the ground plane, connecting the
outer to the ground plane and the inner to the feed point on the
patch. As is the case generally when working with UHF, all lead
lengths should be kept to a minimum. The outer braid of the coax
should be continued through the ground plane and cut just short of
the connection to the patch.
50-Ohm feed point is located at 0.078 wavelengths from the centre of
the patch along the line joining the centre of the patch and the
middle of one of the patch edges. Feeding the patch at this point
will excite the slot closest to the feed point and the other
to the other pair of slots is done in exactly the same manner.
create circular polarisation, with a linear antenna array we require
Two identical antennas polarised at 90 degrees to each
A phase shift in the feed to the antennas of 90
degrees so that one of the antennas is either leading or lagging the
other by 90 degrees depending on whether you want
Left Hand (LHCP) or Right Hand Circular Polarisation (RHCP).
concept of creating circular polarisation with 2 linear antennas can
be quite confusing especially when it comes to determining the which
connection creates RHCP as opposed to LHPC.
are a number of important points regarding circular polarisation:
The direction of rotation of circular polarisation is
the direction of rotation of the electric field vector in the
direction of propagation.
The polarisation rotation is as viewed from behind the
RHCP is the most the common.
In the case of an axial mode helical, the direction of
rotation is the same as the winding direction.
When using linear antennas to create circular polarisation, impedance
matching is vital. A quarter wave piece of coax is a 'dangerous' length to work with because any mismatch here will
produce impedance transformation leading unequal power being
delivered each linear antenna.
For crossed yagis (one at 90 degrees to the other) feed the
90 degree one first, with a quarter wave connection to the 0 degree
one for RHCP. Vice versa for LHCP.
For 'patch' antennas feed the horizontal slots first,
with a quarter wave connection to 90-degree connection to the
vertical slots for RHCP. Vice versa for LHCP.
When the antenna is driving a dish the polarisation is
reversed i.e. for RHCP drive the dish with LHCP.
Another simple way to change from RHCP to LHCP is to add a
half wavelength of coax to the existing quarter wavelength of coax.
All lengths are electrical lengths and the velocity factor of
the coax must be taken into account together with the electrical
lengths of connectors etc.
the patch antenna can be tapped anywhere between the centre and the
outside edge, the 50 Ohm points on the vertical and horizontal slots
can be joined by quarter wave length of 50 ohm coax. This provides
the 90-degree phase shift for circular polarisation.
paralleling the 50-ohm points, the impedance at those points now
becomes 25 ohms. By now tapping the 100-ohm point, the impedance at
this point now becomes 50 ohms after connecting the 50-ohm points
with the quarter wavelength of 50-ohm coax. The 100-ohm point is
found at 0.115 wavelengths from the centre.
my prototype, I actually used N type connectors to tap the 50 and
100-ohm points so that I could do various measurements. As a result,
when taking into account the electrical length of the connectors and
connection to the feed point, I had to add either a half or full
wavelength of coaxial cable to the quarter wavelength because the
resultant piece of coax to make up the electrical difference was too
short and would not reach the connectors!!
I tested the antenna connected for vertical polarisation. The VSWR
at 438 MHz was less than 1.1 : 1 as driving the antenna with 25
watts from my IC 471A, there was virtually no reflected power as
read on my Bird Wattmeter on the 25 watt range.
the antenna at a height of 6 metres at my QTH in Bentleigh, all the
local repeaters (RMU,RHF,RMM,RSE,RGL) were full scale on receive.
Repeaters RAD at Mitcham was about strength 5 and RPU at Arthur’s
Seat was strength 2.
did some further tests with Rod, VK3DQF in Beaumaris and determined
there was good cross polarisation rejection and using Rods 70 cm
helical there was about a 3db increase going from linear to circular
real test was a couple of nights later when I decided to see if I
could get a signal into AO40. A 5*3 signal report from Rene, DJ1KM
in Hanover, Germany completely exceeded my expectations. Not bad
when AO40 was more than 60,000 kms away during this QSO.
am so impressed with the performance of this very compact antenna; I
am now designing an array of 4 patches to be used for further
satellite work. I have also build a patch feed for 2.4 GHz to feed
the 1.5 metre dish, but haven't fully tested this.
antenna article Part#2 Patch