The PRINGLES CAN antenna is quite popular for 802.11 applications. The
devivation of this design appears to be from the COFFEE CAN antenna that
is popular for feeting homebrew parabolic dishes in the Ham Radio world.
It got it's name since the 1" Coffee Can is about the correct
dimensions for this antenna. It was popularised by the book Building
Community Networks.
According to the Satellite Experimenter's Handbook by Martin Davidoff,
K2UBC, the correct dimensions for this antenna are
Diameter
3 / 4 * λ
Length
1.0 - 1.5 λ
Feedpoint from end
3 / 8 * λ
Feedpoint Length
1 / 4 * λ
Of course other dimensions will work, and may work well. Unfortunately
they are likely not to work as well. Despite appearances, this antenna
is linearly polarised. At other dimensions though there could be a
tendancy for the antenna to change polarisation, or decrease in gain
as the antenna attempts to transmit left and right circular polarisation
at the same time.
More Analysis from the NoCat List
James Roe [jroe@jamesroe.com] analysed the performance of the pringles
antenna and posted the result on nocat@pez.oreillynet.com
I'm new to the list and I'm not sure this is the one that was interested
in the coffee can antenna (I'm traveling and away from my home
computer archives), but I've been dusting off microwave skills I walked
away from a long time ago and I've come up with some info that may
be helpful.
For the frequencies of interest near 2400 MHz, the minimum can diameter is
something like 74 mm (say, 3 in) and the maximum is about 90 mm (say, 3.5
in) before higher order modes set up in the waveguide which can mess
up the radiated pattern. Note that larger cans WILL
radiate, but the results will not repeat too well from example
to example due to small differences in feed element, etc.
A 'bare' can of 90 mm diameter will have a maximum gain of about 6.5
dBi, but this can be increased considerably by forming a 'horn' that would
flare up from the 90 mm diameter to a larger one. There are optimal flare
angles and total lengths depending upon the starting and ending diameters,
and I have calculated a couple that might be of interest.
Can Dia
Final Dia
Length
Gain
90 mm
192 mm
14.9 mm
11 dBi
90 mm
305 mm
141 mm
15 dBi
This last is about a foot in diameter but just less than 6 in long. The
horn can be made by making a paper template to cut out an arc of
copper sheet that can be rolled up and soldered to make the horn. The
template can be constructed by drawing two concentric arcs of
length 218 deg with one radius at 74.3 mm and the larger at 252 mm.
This idea can be pushed on up in gain but the horns get pretty big such
that a parabolic reflector probably would serve better - a 20 dBi
horn would be about 21 inches in diameter and 25 inches long.
Q:What is the optimum length for the can?
Can LENGTH is really not important as long as it isn't too short (say less
that a couple of wavelengths). Higher order modes will be launched at the
feed, but will die out quickly if the can diameter is too small for them
to propagate. The 1/4, or 3/4 (or any odd multiple of 1/4) wavelength
refers to the placement of the feed element from the back, conducting wall
of the can (short circuits at these distances appear to be an open circuit
to the feed - hence not there).
On Horn Waveguides
The purpose of the horn is to match the
characteristic impedance of the circular waveguide to that of free space
(the air). The impedance of free space is a constant, but the
characteristic impedance of the circular waveguide depends
upon the frequency and the diameter, so different diameter cans will
need differently sized horns to achieve the same result (ie, gain). The
final diameter of the horn will usually be the
same for a given gain, but the angle of the cone and overall length will
be different to get the impedance match right. I know of no
simple geometrical procedure to calculate these parameters.
The effect of a mis-match of the impedances is to cause reflections (on
both the transmit and receive sides) which means less power transmitted or
received - effectively a loss of gain. The difference between
an 'optimal' design and one slapped together may not be too great, but the
effort to construct is the same - so why not squeeze the design for all
it's worth?
And yes, the horn must be soldered to the end of the can for a
good electrical connection.
Jim Roe
Seattle Wireless have better antenna family on their WWW site - The Cookie
Cantenna. This antenna is actually correctly sized. Check it out
