Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1]   Go Down
  Print  
Author Topic: Pfeiffer (Maltese cross) Antenna - Good or Bad?  (Read 2507 times)
G7DMQ
Member

Posts: 40




Ignore
« on: May 08, 2013, 05:02:38 PM »

I'm interested in these because I would like a rotatable antenna for 20m, but don't have the space (or friendly enough neighbors) for a beam or cubic quad - and was looking for a more compact alternative.

I've read Pfeiffer's paper about the antennas - which (perhaps obviously) is very enthusiastic about them.  But most of what I've read (mostly on QRZ) has been negative - many suggesting that it's the feed line that is transmitting and it would work better without the antenna at all!

I don't have any simulation software - nor know how to use it.  I'm more practical, build it & see if it works then modify to make better (or worse).  But I'd like an opinion, is this a turkey? - before I start cutting up hundreds of feet of wire!

Thanks in advance,

Si
Logged
WB6BYU
Member

Posts: 13567




Ignore
« Reply #1 on: May 08, 2013, 06:47:32 PM »

Quote from: G7DMQ

...I don't have any simulation software - nor know how to use it...



And this is not the ideal candidate for a first attempt at modeling!  I used 32 wires
with a total of 300 segments, and the adjustments for fine-tuning the reflector are not
fun.  And that was using the 4-sided 12m design rather than the 8-sided 20m one.
Boom length is about 4 1/2 feet, roughly according to his formula.  (If anyone else
is going to try modeling it, I strongly recommend using a program that supports
equations, as the whole antenna is defined by about 3 dimensions.  That would
make tuning experiments much simpler!)

It really isn't that bad of an antenna, though some of the statements made in the
original article are rather simplistic given the current state of the art for quad design.

My model has about 5dBi gain and 10dB F/B ratio.  Either might be improved slightly
with further reflector tuning, but I'm down to the point where the differences are
probably too small to reproduce reliably in an HF antenna.

By comparison a standard 2-element W4RNL 2-element quad has about 7dBi gain
and at least 20dB F/B ratio, using a 6 foot boom.  As part of the process I had the
Pfeiffer driven element in front of a conventional reflector at the wider spacing,
and it achieved about 6dBi.

On the other hand, a Moxon might be another option - my model has 6dBi gain.

The original quad gives a good match with a quarter wave of 75 ohm line, while
the Moxon had a direct 50 ohm feed.  The Pfeiffer quad when properly adjusted
can have a direct 50 ohm feed as well.  (I haven't finished playing with that part
of the modeling yet.)


So, yes, it seems to work about as any other such shrunken antenna would.
Is it better than some of the alternative shapes?  Not necessarily.  It is rather
quirky to adjust, since his instructions say to adjust the inner ends of all 4
loading segments, but if you fixed the top and bottom and just moved the side
ones in and out it would be easier.

The 20m version would lose somewhat more due to the additional folding -
perhaps an extra 1dB in gain, but that is only a guess:  that model requires
even more wires and segments, and I don't have time for that at the moment.
Logged
WB6BYU
Member

Posts: 13567




Ignore
« Reply #2 on: May 08, 2013, 08:48:34 PM »

I did some further experiments with a 20m 4-armed version.  One of the main limitations
on performance is due to trying to use too short of a spacing between the elements.  This
is consistent with the findings of experiments by W4RNL and others:  you can get reasonable
performance from a small beam with elements about 2/3 of regular size, and passable results
at 1/2 size, but you can't do so if you also shrink the boom length.

His formula gave something like a 7' boom on 20m, but the model really started looking good
when I increased it to 10'.  There is a lot of interaction among the dimensions for the two
elements and the boom length, so make sure you have a good starting point before building.
That may be why he liked to use a gamma match - he didn't have to worry about the
feedpoint impedance.

While a 10' boom is capable of up to 6dBi, that comes with a fairly narrow bandwidth and
poor F/B ratio.  By relaxing the reflector tuning a bit we can get perhaps 5.5dBi and 15+ dB
F/B, with SWR better than 1.5 : 1 from 14.2 to 14.35 MHz.  That looks like a reasonable small
beam - still might be a bit tricky to adjust, but by measuring the resonant frequencies of the
two elements beforehand (they likely won't follow the handbook formulas) it shouldn't be too
difficult to get reasonable performance.  That's another problem with shrunken antennas - the
construction tolerances can get impractical to maintain in the real world if you try to squeeze
too much gain out of a small package.  This antenna extends about 7' on each side of the
boom.

By comparison, these gains are dBi in free space, so a dipole is 2.1dBi.  The 5dBi version is
quickly approaching the part where the useful gain over a dipole disappears as it gets
smaller, especially at shorter boom lengths that don't give you very good F/B ratio either.


In the end it has to be your decision where to go with this or some other option, partly
depending on your preferences for the trade-off between F/B and gain, mast height,
etc.  Other options to consider are the HEX-beam and other various sorts of shortened
yagis and quad elements.  One other thing to consider if you have a tall mast is extending
the quad element vertically while making it narrower in width:  I have one model using
about 16' spacing that is only 13' wide but 23' tall, with a pretty flat SWR across the
20m band in spite of also achieving over 7dBi of gain.
Logged
WB6BYU
Member

Posts: 13567




Ignore
« Reply #3 on: May 08, 2013, 10:03:10 PM »

Sorry to keep posting:  I'm supposed to be cleaning up the house, but its much
more fun to play with the modeling program...


At 12' boom length it is easier to get a wider SWR curve and good F/B:  20dB or
better.  But this also points up some of the disadvantages of compact beams.
Here are the numbers for gain (dBi), F/B ratio, and SWR across the band:

14.0   4.8    1   5.8 : 1
14.1   6.3    8   2.1 : 1
14.2   5.7   30  1.1 : 1
14.3   4.8   12  1.3 : 1

While we can get usable SWR across the whole phone band, the F/B and gain
peak in different places in the band; F/B is poor at the low end and not all
that good at the high end, and while the gain curve is well centered, the gain
drops off quite a bit at the band edges.  It has pretty good performance over
about 100kHz, but at least is better than a dipole across the band.
Logged
G7DMQ
Member

Posts: 40




Ignore
« Reply #4 on: May 09, 2013, 12:19:03 AM »

Thanks Dale - that's a fantastic & comprehensive response!

I can see that some modelling software might be on my Christmas List!

Out of interest, do any of the apps allow you to import an antenna from a CAD drawing - just taking the line segments as wires?  I'm a design Engineer by trade and use Solidworks all day.  Drawing even complex antennas is easy & quick (Solidworks allows parametric, equation driven drawings) whereas the drawing interface in the modelling software I've seen is a bit clunky!

Are there any decent books / resources to get me going with modelling?  It finally struck me last night writing the post that not using software is daft!

Si
Logged
KD2CJJ
Member

Posts: 369




Ignore
« Reply #5 on: May 09, 2013, 05:52:28 AM »

Interesting antenna - I was too looking for a small beam that worked better than a dipole and had TV antenna look due to neighbors, etc.

I ended up with a TGM MQ26 mini beam.  Its a quad for the reflector and end loaded dipole...  They have larger versions which give you more gain and F/B. The only issue is that is has very narrow bandwidth but a tuner cleans that up with no issues.  I have compared it to my 20m full wave end fed and a fan dipole at the same heights and it is about 1s  above the fan dipole and some times more than 2 on the end fed...  Its also 1 - 2s quieter than both - I suspect due to the directivity of having a beam (less than 1s F/B though)

Here is a pic off the side of my house.. Good luck in your search!

http://qrz.com/db/KD2CJJ




I'm interested in these because I would like a rotatable antenna for 20m, but don't have the space (or friendly enough neighbors) for a beam or cubic quad - and was looking for a more compact alternative.

I've read Pfeiffer's paper about the antennas - which (perhaps obviously) is very enthusiastic about them.  But most of what I've read (mostly on QRZ) has been negative - many suggesting that it's the feed line that is transmitting and it would work better without the antenna at all!

I don't have any simulation software - nor know how to use it.  I'm more practical, build it & see if it works then modify to make better (or worse).  But I'd like an opinion, is this a turkey? - before I start cutting up hundreds of feet of wire!

Thanks in advance,

Si
Logged

73

Mike
KD2CJJ
G7DMQ
Member

Posts: 40




Ignore
« Reply #6 on: May 09, 2013, 10:11:21 AM »

had TV antenna look

Your TV antennas look a bit different to ours!  On the other hand, I guess the whole of the UK would fit in your back yard - so you need a BIG antenna! Wink

After a bit more reading (while I should have been working!), about the only compact antenna nobody has much bad to say about is a HexBeam - so maybe that's the way to go.  Decisions Decisions!

Si
Logged
WB6BYU
Member

Posts: 13567




Ignore
« Reply #7 on: May 09, 2013, 08:39:41 PM »

Quote from: G7DMQ

...about the only compact antenna nobody has much bad to say about is a HexBeam...




You can find people saying bad things about almost any antenna:  that doesn't mean it is
a bad one, or that it doesn't do exactly what it claims to do.  In this case the performance
claims are perhaps a bit better than modeling would indicate, but we also don't know what
type of quad it was being compared against or the accuracy with which differences could
be measured.

Basically it works about as well as such a shortened antenna can be expected to work.

Probably the best way to proceed is to determine the limiting dimensions (width and height)
that you have to work within and see what the options are in the space available.  Also give
some thought to what characteristics of a beam are most important:  Is it the reduced QRM
on receive due to the F/B ratio, or the strongest signal at the other end, or some trade-off
between the two?   How much of the 10m band do you regularly use where you want maximum
performance?  (Bandwidth probably isn't an issue for a CW operator, or one who uses
phone over a relatively small portion of the band.)

Once you have an idea of what factors are most important and what your limits are, then
we can look at which options best fit them.


Quote from: G7DMQ

Are there any decent books / resources to get me going with modelling?



The late W4RNL had an extensive website with a lot of information on modelling, but it
is now behind the AntenneX paywall.  It may be worth paying for access, or looking at
some of the compilations they have available for sale.  He also wrote an online training
class on the topic for ARRL, but I don't know if it is still available.

I got started with EZNEC by reading the tutorial, and about 1 1/2 chapters into it
I was off doing my own antennas.  You can download a free demo version (which limits
the size of models you can save) to get a sense of the behavior, and also what file
formats it supports for input and output.

Another option is 4NEC2, a very complete package that is available as freeware.  I find it
a bit more confusing to use, but then, I've never read any of the documentation on it, and,
given the number of features, that may be my fault more than that of the program.  It
includes options for defining antennas with variables and an optimizer, as well as quite a
number of demo files.


All the modeling programs have some limitations, and part of building a successful model
is knowing when to expect problems and what to do about them.  The NEC2 equations (used
in EZNEC and 4NEC2) require care with parallel wires (the segmentation has to match on each)
and most programs have problems when wires meet at very sharp angles because of how they
model intersections of wires.  MININEC is an older version designed for smaller computers that
couldn't run the original NEC program:  it has a number of differences, and its ground model
is famously poor for low horizontal antennas.  The newer NEC4 core is available as an option
in EZNEC and some other programs, but requires that one pay for a license to use it (as well
as getting permission from the US Government - apparently it is still categorized as a weapon.)

But I suspect that either 4NEC2 or EZNEC will be sufficient for most of your needs, and would
be my recommended places to start.


While there are some limitations to model accuracy, I'm still surprised sometimes how well
they can work.  I modeled a quad with a gamma match, and was surprised to see the
polarization rotated by about 45 degrees in the output plot.  Since I had several 2m quads
that used a gamma match, I immediately took one outside, connected it to a transmitter,
and probed the polarization with a yagi.  Sure enough, the polarization was rotated just
as EZNEC had suggested.  Changing to a different feed method for the same antenna
shifted it back to where I expected it to be.
Logged
Pages: [1]   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!