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: Using an MFJ-269 to measure wires?  (Read 3439 times)
AB1JX
Member

Posts: 15




Ignore
« on: October 25, 2012, 10:08:26 PM »

I want to build a 4 element collinear (horizontal, for 48.85 MHz, not vertical) from a 1960 ARRL antenna book.  I thought I could use my MFJ-269 to check the lengths of the 1/4 wave phasing stubs and 1/2 wave elements, but from the manual it's not clear just how.

The phasing stubs will be 1/4 wave pieces of zip cord, probably coiled up.  I don't know the velocity factor.  The 1/2 wave elements will be about 10 foot long pieces of plain bare wire.  Measuring the complete antenna isn't going to tell me much because it's a 4 element collinear with 4 half wave straight sections and 2 phasing stubs.  I need to cut the pieces to length before I put them together.  The 48.85 MHz is a locally quiet frequency for doing radio astronomy.  Aperture counts here, which is why I want to use a collinear.

Thanks,

  Alan, ab1jx

Logged
WB6BYU
Member

Posts: 13578




Ignore
« Reply #1 on: October 25, 2012, 11:10:27 PM »

Quote from: AB1JX

The phasing stubs will be 1/4 wave pieces of zip cord, probably coiled up...



I strongly recommend that you reconsider both of those aspects:  zip cord will have
relatively high losses, and coiling up any sort of unshielded transmission line is rarely
a good idea.


There are several methods of measuring the stubs, generally relying on the fact that
a quarter wave stub open at one end will look like a short circuit at the other, and
vice versa.  For example, you can connect an open quarter wave stub across the
SWR analyzer and adjust the length for the minimum R value with zero reactance
(assuming that the meter is accurate at such high SWRs.)

If you only have an SWR measurement, then you can put a short circuited
stub across a 50 ohm resistor and adjust the length for minimum SWR at the
desired frequency.


The good news, however, is that the stub length isn't particularly critical for
the radiation pattern of the antenna.  I ran a model and it didn't make a lot of
difference even when I changed the stub length by two feet.  You probably
can get close than that with a nominal velocity factor.  (Just don't try to coil
it up, however:  the turns couple to each other and the performance is
unpredictable.)

You can also use some of the measured data available for zip cord, even though
there can be a lot of variation.  VK1OD includes "ZIP 105" as the last entry in
the transmission line table of his handy calculator here:

http://vk1od.net/calc/tl/tllc.php

One way to calculate a quarter wave line is by terminating the far end in a large
value such as 10,000 ohms and adjusting the length until the calculator gives you
a low resistance with near-zero reactance.
Logged
STAYVERTICAL
Member

Posts: 875




Ignore
« Reply #2 on: October 25, 2012, 11:35:45 PM »

First, a caveat - I have never actually done anything like that before, but to get you thinking on one way of doing it - here is my contribution.


QUARTER WAVE STUBS (unknown velocity factor):

Any feedline will have a reflection of its terminating impedance at half wave multiples (with velocity factor already taken into account).
At quarter wave multiples it will have an anti-node or node compared to what is at the end.

So, lay out enough feedline so that you are sure it is more than a half wavelength at 48Mhz and short the end.
Start low in frequency and sweep with the MFJ watching the impedance until it reads a short.
This is the frequency for which the feedline length you have laid out is a half wavelength.
Calculate what the free space wavelength length would be for this frequency (300/f (Mhz) assuming metres are used)
Then measure the length of feedline you actually have laid out.
If you divide the actual length multiplied by 2, by the free space wavelength you will get a value between 0 and 1 , this is the velocity factor.

Since you now know the velocity factor you can make your stubs.
Calculate the free space wavelength for your frequency of interest.
Multiply this value by the velocity factor.
Divide that result by 4 for quarter wave stubs.

HALF WAVE ELEMENTS:

To calculate the 1/2 wave element lengths you could use the well know 468/f (Mhz) , giving the answer in feet.
However it will not be resonant exactly because it has to be in its final location before trimming.
I would calculate the length (468/f), make it a bit longer, raise it to what height it will be, put some coax on it, and sweep frequency with the MFJ269 until you find resonance.
This will be when the resistive component is around 70 ohms and the reactance is 0.
Then, assuming you cut the element longer, just trim and repeat until you have the frequency you want.
Measure this final element length and it will be repeated for your other elements.

NOTES:
When calculating the velocity factor, you don't have to use a short at the end of the cable - a 10 ohm non-inductive resistor may be easier for the MFJ to handle.
Just remember this 10 ohms will be repeated at half wave intervals, with the velocity factor already taken into account.
Theoretically, if you make it longer than necessary, measure the frequency where the 10 ohms repeats and then trim the stub, you will eventually get to 48 Mhz with no calculations required.
Then halve this half wave stub value to get a quarter wave stub length.

I am sure you will get some better replies, but this is to get you started.

73 - Rob
« Last Edit: October 25, 2012, 11:40:31 PM by STAYVERTICAL » Logged
W8JI
Member

Posts: 9296


WWW

Ignore
« Reply #3 on: October 26, 2012, 11:55:05 AM »

I want to build a 4 element collinear (horizontal, for 48.85 MHz, not vertical) from a 1960 ARRL antenna book.  I thought I could use my MFJ-269 to check the lengths of the 1/4 wave phasing stubs and 1/2 wave elements, but from the manual it's not clear just how.

The manual should be clear. All you do is leave the far end of the cable or stub open and tune the analyzer for minimum Z. That will be the frequency where it is an odd quarter wave long. If you make the stub for 49 MHz, the dip will be at 49 MHz. If it dips high, it is too short. If impedance dips too low in frequency, the stub is too long.

To check the half wave, do the same and dip at half frequency or short it and dip at full frequency.

If you are trying to check the element length, you would have to do it like a dipole.... or do it by formula. 

Quote
The phasing stubs will be 1/4 wave pieces of zip cord, probably coiled up.  I don't know the velocity factor.
 

That's a VERY bad idea. I'd use twin lead or ladder line at a minimum, and preferably an air insulated stub.

Quote
The 1/2 wave elements will be about 10 foot long pieces of plain bare wire.  Measuring the complete antenna isn't going to tell me much because it's a 4 element collinear with 4 half wave straight sections and 2 phasing stubs.  I need to cut the pieces to length before I put them together.  The 48.85 MHz is a locally quiet frequency for doing radio astronomy.
 

Is there a link to the antenna?

Quote
Aperture counts here, which is why I want to use a collinear.

Physical size is NOT related directly to capture area except in very special cases of standard high-efficiency antennas. Some virtually lossless perfect dish antennas, for example, have a known size to capture area relationship.

Capture area relates directly to GAIN and frequency. As such, it really has nothing to do with physical size with nearly all real antennas.

If both antennas are on the same frequency, an antenna ten miles long with 3 dB gain has less capture area than an antenna three inches long with 4 dB gain.

http://www.w8ji.com/capture_area_ae_effective_aperture.htm



Logged
AB1JX
Member

Posts: 15




Ignore
« Reply #4 on: October 26, 2012, 08:21:31 PM »

Thanks for all the replies.  While it's still on my clipboard here's a picture of one: http://wavelab.homestead.com/east_ant_48.jpg  The article itself was in a 1960 book (paper).

The air insulated phasing stubs are the classic way to do it but there's a picture at http://astro.u-strasbg.fr/~koppen/RJove/antenna.html of using twin lead stuffed into a cylinder to be less obtrusive.  The wavelab picture uses air insulated.

Since it's only for receiving and I don't have to worry about SWR I'll probably just cut it from formulas and put it up, but I thought it should be possible to check the electrical lengths more accurately with the MFJ-269.  I can make the phasing stubs out of coax and look up the velocity factor that way, the elements are just bare wire so I don't have to worry about those.  I can coil up the coax to keep it out of the snow.  I plan to put this about 5 feet above ground (I'm listening to the sky and being low helps cut man-made noise) so I don't want dangling phasing stubs.

I'll check MFJ's site for a newer PDF but the one I've got doesn't cover how to cut a plain piece of wire that's going to be part of something more complicated to resonance.  I suppose I could just stick the end in the coax connector, suspend it about the right height and measure that way.  It also doesn't talk about making 1/4 wave feedline stubs except maybe in the "distance to fault" section.  And of course end feeding a halfwave isn't commonly done.

One thing that bugs me about using it is that there are lots of little false peaks and dips so you have to scan in the right direction, and don't believe every one you see. You have to pretty much know what to expect before you start.

I had modeled this so I made a copy and changed the lengths of the phasing sections about a foot, and sure enough it made very little difference in either the pattern or gain.  Maybe that's a limitation of the "method of moments" modellers, I was using NEC2.  I didn't try setting up a feed for each element and changing the phases on the elements.  Maybe it isn't calculating the effective delay in going though the phasing stubs for some reason. Changing them both by 20% or more should make some difference.

  Alan
Logged
W8JI
Member

Posts: 9296


WWW

Ignore
« Reply #5 on: October 26, 2012, 08:55:53 PM »

Thanks for all the replies.  While it's still on my clipboard here's a picture of one: http://wavelab.homestead.com/east_ant_48.jpg  The article itself was in a 1960 book (paper).

The air insulated phasing stubs are the classic way to do it but there's a picture at http://astro.u-strasbg.fr/~koppen/RJove/antenna.html of using twin lead stuffed into a cylinder to be less obtrusive.  The wavelab picture uses air insulated.

Since it's only for receiving and I don't have to worry about SWR I'll probably just cut it from formulas and put it up, but I thought it should be possible to check the electrical lengths more accurately with the MFJ-269.  I can make the phasing stubs out of coax and look up the velocity factor that way, the elements are just bare wire so I don't have to worry about those.  I can coil up the coax to keep it out of the snow.  I plan to put this about 5 feet above ground (I'm listening to the sky and being low helps cut man-made noise) so I don't want dangling phasing stubs.

The stubs act like they have infinite SWR, so they have to be very low loss or they will dissipate significant power. Coax will  also have significant common mode currents, and that will ruin pattern. Use air insulated parallel conductor line.

Quote
I'll check MFJ's site for a newer PDF but the one I've got doesn't cover how to cut a plain piece of wire that's going to be part of something more complicated to resonance.  I suppose I could just stick the end in the coax connector, suspend it about the right height and measure that way.  It also doesn't talk about making 1/4 wave feedline stubs except maybe in the "distance to fault" section.  And of course end feeding a halfwave isn't commonly done.

Measure the wire by formula and it will be close enough. Feed a sample piece as a dipole and trim it to length, and that would also work.

The stubs are measured using DTF mode.


Quote
One thing that bugs me about using it is that there are lots of little false peaks and dips so you have to scan in the right direction, and don't believe every one you see. You have to pretty much know what to expect before you start.

I'm not sure what that means.

Code:
I had modeled this so I made a copy and changed the lengths of the phasing sections about a foot, and sure enough it made very little difference in either the pattern or gain.  Maybe that's a limitation of the "method of moments" modellers, I was using NEC2.  I didn't try setting up a feed for each element and changing the phases on the elements.  Maybe it isn't calculating the effective delay in going though the phasing stubs for some reason. Changing them both by 20% or more should make some difference.

All you have is three half waves in-phase.  The gain will be about 4 dBd. Capture area will be just a little over  twice that of a regular dipole, provided it has about 4 dB gain over a dipole.

Any antenna you installed that had 4 dB gain on the same band would have the same effective aperture, or capture area. A two element Yagi would actually have a bit more capture area.

Personally, if it were me, I'd build a collinear array with two dipoles fed with coax spaced 1/4 wave apart at the center,  or I'd build a broadside array with two parallel coax fed dipoles spaced 5/8th wave apart. I'd lay a reflector screen below the dipoles.

Those are two much easier ways to do the same thing, and would have much more usable bandwidth.
Logged
WB6BYU
Member

Posts: 13578




Ignore
« Reply #6 on: October 26, 2012, 09:35:22 PM »

Quote from: AB1JX
...
The air insulated phasing stubs are the classic way to do it but there's a picture at http://astro.u-strasbg.fr/~koppen/RJove/antenna.html of using twin lead stuffed into a cylinder to be less obtrusive.  The wavelab picture uses air insulated.



Just because someone describes an antenna on the internet doesn't mean it is an
optimized design.  Besides, if changing the stub length by a foot or two doesn't affect
the pattern much, your chances that a non-optimal solution will appear to work may
be pretty good, especially if you never check the antenna to see that is is actually
working the way you think it is.



Quote

Since it's only for receiving and I don't have to worry about SWR I'll probably just cut it from formulas and put it up, but I thought it should be possible to check the electrical lengths more accurately with the MFJ-269.



It is possible, as several people have described.  This may require a bit of creativity
as to how you connect the twinlead to the connector on the SWR analyzer.  

Here is one way:  take a 47 ohm resistor and bend the leads so one goes in the
center pin of the connector and the other lead touches the shell.  Measure the
impedance using the MFJ-269.  Now connect one end of the twinlead across
the resistor - one lead on each side of it.  Adjust the length of the twinlead
until you get X = 0, or at least some reading close to the original measurement.
If the far end of the twinlead is open circuited and it isn't laying on a dielectric
material that will change the velocity factor, you should have a 1/4 wavelength
of feedline.

But it doesn't make much difference in the antenna performance.



Quote

I'll check MFJ's site for a newer PDF but the one I've got doesn't cover how to cut a plain piece of wire that's going to be part of something more complicated to resonance.


You aren't cutting a plain piece of wire - you are cutting a length of feedline.
You need to use both wires of the feedline in your measurements, just as you do when
you connect it to an antenna.


Quote

One thing that bugs me about using it is that there are lots of little false peaks and dips...



What parameter are you looking at when you see these?  How do you have the twinlead
connected to the analyzer?





Quote

I had modeled this so I made a copy and changed the lengths of the phasing sections about a foot, and sure enough it made very little difference in either the pattern or gain.  Maybe that's a limitation of the "method of moments" modellers, I was using NEC2.



Nothing to do with NEC or other modeling software - that's the way the antenna works.

Here's why:

First, I'm assuming you are using 4 half wave elements, so the overall antenna length is
about 40'.  If you had a 30' overall length with 3 elements the impedance at the center
would be low instead of high (and might be an adequate match with a 4 : 1 balun.)  But
except for the center section and feedpoint, the principles I'm describing apply to either
design.

Each side of the antenna has 3/2 of wire:  the inner wire, the outer wire, and the total
length of wire in the stub.  There will be a point of maximum current in the middle of each
of these sections, and the wire closest to those points will account for most of the radiation
from the antenna.  In this case, because the stub wires are folded, the radiation cancels
from the two conductors at that point, so only the 4 horizontal half wave wires contribute
to the antenna pattern.

The phase of the current reverses in each half wavelength.  If you string up a full wave
wire with no phasing stub, the two adjacent half wave portions are out of phase, which
cancels radiation broadside to the antenna (because you are picking up equal and opposite
fields from both sides of the antenna.)  By adding a phasing stub, the currents in the
two horizontal wires on each side are now in phase, so they add instead of cancelling.

Ignoring the matter of the feedpoint impedance, the pattern depends on the fact that
the currents are in phase in all the wires.  If you change the length of the phasing stub
this shifts the relative locations of the current maximum in the wire between the feedpoint
and the stub.  But until this moves far enough to cause significant out-of-phase
radiation from these wires, the actual radiation pattern won't change much.  If you make
the simplifying assumption that the current is only significant in the middle half of each
half wave section, then you can shift the current maximum by 1/8 wavelength along the
wire before any significant out-of-phase radiation occurs.  (That is what starts breaking
your pattern up into lobes.)  

It doesn't matter whether you model the antenna using NEC2, NEC4, or if you build it and
try it out - you'll get the same results because that is how the antenna behaves.


At 5' above ground, a quarter wave of 300 ohm twinlead feeding the center of the antenna
should give you a reasonable match to coax cable.  (My model suggests an impedance around
1500 ohms at resonance.)  Or use a quarter wave of 600 ohm line to a 4 : 1 balun.
Logged
N8CMQ
Member

Posts: 403




Ignore
« Reply #7 on: October 26, 2012, 09:45:09 PM »

I like the design, except, I would use two half wave elements,
one on each side of the center fed half wave, but you can do it your way too.
I would also use ladder line for the phase stubs for a transmitting antenna, rather than zip cord.
Twin lead is mentioned, which is OK for low power.
Do not stuff it in a tube, let it hang down. If it is too near the ground, the antenna is too low.
More cores on the coax to choke the RF from flowing on the outside of the coax, or several loops of coax will work too.
A 1:1 balun using 75 ohm cable would be a better match, but 50 ohm will work, and you wouldn't need the loops or cores.
If you go with looping the coax, the number of turns is determined by the frequency of operation.
Height above ground is important! 3/8 wave would be my minimum height.
This design is also used as a vertical antenna, and is very good!

To adjust the SWR, first, do the center fed dipole, then add the half wave element cut to the same length as the dipole.
use a stub longer than 1/4 wave and adjust the SWR to minimum with the shorting bar.
If you add the second half wave and stub, do it now, just like the first half wave section.

Being that you want this as an astronomy antenna, and receive only, you can install it at 1/4 wave over a good ground plane and have a major lobe going straight up.
When I say a good ground plane, I mean a series of parallel wires about 4 inches apart and slightly longer than the radiating element of the antenna.
The more ground plane, the better the results. It becomes like a reflector on a beam antenna.

In order to have the stubs not in contact with the ground, you can use an insulated rod and counter balance the stubs horizontal to straight up.
I would go straight up if there is a concern for icing.
Good luck!
« Last Edit: October 26, 2012, 10:31:18 PM by N8CMQ » Logged

N8CMQ   Jeff
AB1JX
Member

Posts: 15




Ignore
« Reply #8 on: October 27, 2012, 09:13:37 PM »

The collinear in the picture is a 4 element, two half waves each side of the center.  They're nicely versatile, you can make a 3 element where the center is a halfwave dipole, two elements both halfwaves, lopsided where you have unequal numbers each side of center, or longer than 4 elements.  This is from an ARRL antenna book, 1960 edition.  The picture is of Hans Michlmayr's antenna in Australia: http://wavelab.homestead.com/index.html (vk6zt)

The other thing about them is that they're in one line, which keeps things simpler since there'll be two of them about 220 feet apart in parallel as an interferometer.  Each antenna goes in a north-south line, and as the earth rotates signals from distant radio sources hit the two antennas with phase angles that are constantly changing, which is what you measure.

Impedances vary depending on the configuration but I've had good luck lately winding baluns on toroids out of old computer power supplies so I can match just about anything to my RG6.  This 4 element should have an impedance around 781 ohms, so a 10.438 impedance or 42:13 turns ratio is what I plan to use.

I know about the reflector wires, but I'm not sure I can get them out of the way of the lawnmower.  Stapling them down and letting them grow in maybe.

The plain piece of wire I'm referring to is each of the halfwave elements, which will be #13 galvanized electric fence wire about 10 feet long.  Maybe a coupling loop around one might work like with a dip meter, but I'll probably just measure with a tape measure.

I haven't seen any of the false dips on this antenna because I haven't gotten that far, but I've used it enough to know there isn't going to be just one.  I seem to get dips on harmonics of the oscillator frequency, that sort of thing.

Radio astronomy types have strange notions of antenna apertures, some probably outdated by ham standards.  Another is measuring signal power as degrees Kelvin and calibrating from noise generators.  From what I can figure there's no direct microvolt or db equivalent because the receiver bandwidth is also involved.  It's a matter of how much energy you can capture.  Systems are calibrated from the antenna to the receiver, with the result that they can read out apparent temperatures of distant objects. I'm new at it and frankly trying to stay away from that area.  All I know is that I can't detect the sun yet, which should be easy.

OK, dangling open wire stubs it is.
Logged
W8JI
Member

Posts: 9296


WWW

Ignore
« Reply #9 on: October 28, 2012, 03:31:18 AM »

The collinear in the picture is a 4 element, two half waves each side of the center.  They're nicely versatile, you can make a 3 element where the center is a halfwave dipole, two elements both halfwaves, lopsided where you have unequal numbers each side of center, or longer than 4 elements.  This is from an ARRL antenna book, 1960 edition.  The picture is of Hans Michlmayr's antenna in Australia: http://wavelab.homestead.com/index.html (vk6zt)

With four half waves you would have a very high center impedance making it difficult to feed. That's why I assumed it was three half waves.

Still, it is a low gain antenna. Four elements with no spacing end-to-end is around 4.5dBd gain when perfectly optimized.

http://www.w8ji.com/stacking_broadside_collinear.htm


Quote
The other thing about them is that they're in one line, which keeps things simpler since there'll be two of them about 220 feet apart in parallel as an interferometer.  Each antenna goes in a north-south line, and as the earth rotates signals from distant radio sources hit the two antennas with phase angles that are constantly changing, which is what you measure.

I suppose the azimutal directivity isn't a concern then? The HPBW will likely be around 25 degrees if you build a 4 element correctly.


Quote
Impedances vary depending on the configuration but I've had good luck lately winding baluns on toroids out of old computer power supplies so I can match just about anything to my RG6.  This 4 element should have an impedance around 781 ohms, so a 10.438 impedance or 42:13 turns ratio is what I plan to use.

Transformer efficiency with sooo many turns on 50 MHz??? My bet is, not good. Smiley


Quote
Radio astronomy types have strange notions of antenna apertures, some probably outdated by ham standards.  Another is measuring signal power as degrees Kelvin and calibrating from noise generators.  From what I can figure there's no direct microvolt or db equivalent because the receiver bandwidth is also involved.  It's a matter of how much energy you can capture.  Systems are calibrated from the antenna to the receiver, with the result that they can read out apparent temperatures of distant objects. I'm new at it and frankly trying to stay away from that area.  All I know is that I can't detect the sun yet, which should be easy.

Well, since about 1900 or earlier we knew physical size doesn't mean capature area. Smiley Hams are just slow catching up.

The four element collinear still has about the same capture area as a two-element Yagi, because it has about the same gain.

I'm a little puzzled why a collinear is OK.  Is the goal to beam straight up at the zenith?

Logged
WB6BYU
Member

Posts: 13578




Ignore
« Reply #10 on: October 28, 2012, 08:13:00 PM »

If you want to dip a half wave wire to length, I'd just wrap the center of it
around the dip meter coil for 1 turn or so.  That should get you close enough.

Once you build the antenna you can adjust either the stub lengths or
the wire lengths for resonance.  It really doesn't matter which.

If you want shorter stubs to keep them out of way of the lawnmower you can
use 5/8 wave elements and 1/8 wave stubs to build a similar antenna:  basically
keeping a majority of the out of phase current in the stubs, but using the
wire where the current is low to increase the spacing between the elements.
You still get about the same gain for a given overall length, but you could use
3 elements with the middle one fed in the center for an easier impedance match.

My model suggests an overall length of 40' the stubs are 3' long (or a bit more)
and placed 12' in from each end, but those might be optimized a bit further.
Feedpoint impedance is about 125 ohms, much easier to match than the high
impedance of the original version.  Gain is half a dB higher than the original,
and the side lobes are slightly stronger.
Logged
AB1JX
Member

Posts: 15




Ignore
« Reply #11 on: October 28, 2012, 09:36:51 PM »

Beaming up at the zenith is fine for a beginner like me. There'll be some tilt to the earth with seasonal changes, but there are hundreds of radio sources in almost any direction, especially in the milky way.  Other than seasonal changes though I think a fixed antenna gets pretty much the same thing every day, so eventually you go to 1 GHz and up and use dishes that you can tilt.  You can actually map the sky by just moving from the south horizon to the zenith to the north horizon and waiting for the earth to rotate.  A narrow beam in declination (north-south) provides some selectivity.

Both MMana-Gal and NEC2 give this antenna 9 - 11 dbi gain, but I don't know how to get a "real" ground in NEC2 yet so I think that's figured with a ground plane.  Mmana's definitely using real ground, not perfect.  What I've got up on one side now is a fat dipole that's at http://ab1jx.webs.com/toys/jove/antennas/fat8/index.html but the other side's a fan dipole with less gain and I think the imbalance is causing trouble.  I've been pretty happy with it, it's just a little unorthodox to describe to someone and I don't have the materials to build a matching one for the east side.

I just checked again: by MMANA-GAL a 3 element has 7.82 dbi gain, a 4 element has 9.21, I don't have space for longer than 4 elements.  End-end spacing used was 50 mm.  This is dbi not dbd.  I'm also sort of a fan of the coaxial collinear but those are about impossible to model.  Some people use a pair of dipoles, but usually with preamps.  LOFAR in the Netherlands uses a bunch of 6 meter inverted vees with preamps to cover 20 - 80 MHz. The multiple antennas aren't for gain, they're for different vantage points.
http://www.astron.nl/radio-observatory/astronomers/technical-information/antennae/antennae-description
And they're spread across Europe too, all linked together.

OK, I'll use a 13:4 turns ratio to try for better efficiency.  I think I can make each antenna out of 2 pieces of wire about 31 feet long with the bends to make the phasing stubs and threading the spacers on.

The WB6BYU post just came in while I was typing.  I'll play around with the 5/8 and 1/8 idea.
Logged
W8JI
Member

Posts: 9296


WWW

Ignore
« Reply #12 on: October 29, 2012, 05:51:55 AM »

Beaming up at the zenith is fine for a beginner like me. There'll be some tilt to the earth with seasonal changes, but there are hundreds of radio sources in almost any direction, especially in the milky way.  Other than seasonal changes though I think a fixed antenna gets pretty much the same thing every day, so eventually you go to 1 GHz and up and use dishes that you can tilt.  You can actually map the sky by just moving from the south horizon to the zenith to the north horizon and waiting for the earth to rotate.  A narrow beam in declination (north-south) provides some selectivity.

This subject is fascinating to me, but the antenna you posted is the only one I've seen.

Frankly, I wouldn't do it that way at all. I think it is a design fraught with potential problems for any possible small benefit in gain.  (Notice I won't use the myth or commonly abused "capture area".) The high feed impedance on ~50 MHz is a real issue, as is getting phase and current correct. 

Quote
Both MMana-Gal and NEC2 give this antenna 9 - 11 dbi gain, but I don't know how to get a "real" ground in NEC2 yet so I think that's figured with a ground plane.  Mmana's definitely using real ground, not perfect.  What I've got up on one side now is a fat dipole that's at http://ab1jx.webs.com/toys/jove/antennas/fat8/index.html but the other side's a fan dipole with less gain and I think the imbalance is causing trouble.  I've been pretty happy with it, it's just a little unorthodox to describe to someone and I don't have the materials to build a matching one for the east side.

A regular dipole at reasonable height over reasonable earth has 8-8.5 dBi gain. So what you are saying is you have 0.5-3 dB gain over a dipole. It is VERY easy to lose that much with design errors when dealing with stubs and matching systems.


Quote
I just checked again: by MMANA-GAL a 3 element has 7.82 dbi gain, a 4 element has 9.21, I don't have space for longer than 4 elements.  End-end spacing used was 50 mm.  This is dbi not dbd.  I'm also sort of a fan of the coaxial collinear but those are about impossible to model.  Some people use a pair of dipoles, but usually with preamps.  LOFAR in the Netherlands uses a bunch of 6 meter inverted vees with preamps to cover 20 - 80 MHz. The multiple antennas aren't for gain, they're for different vantage points.
http://www.astron.nl/radio-observatory/astronomers/technical-information/antennae/antennae-description
And they're spread across Europe too, all linked together.

A basic halfwave dipole at reasonable height over reasonable earth is 8-8.5 dBi. It would, of course, depend on ground losses.

I'd probably look at what I wanted to hear, and plan something around that.


Quote
OK, I'll use a 13:4 turns ratio to try for better efficiency.  I think I can make each antenna out of 2 pieces of wire about 31 feet long with the bends to make the phasing stubs and threading the spacers on.

The WB6BYU post just came in while I was typing.  I'll play around with the 5/8 and 1/8 idea.

It sounds like they really need a few good standard antenna designs with a reasonable matching systems. Frankly, for a single band antenna, I'd use a Q-section for matching and 1:1 choke balun. I'd stay away from transformers unless I knew the actual loss. You are using a primary-secondary transformer, this means the core will play a considerable role in efficiency. Think about this, on 160-20 meters I use a 2 turn to 5 turn transformer for 450 ohm Beverages to 75 ohm line.

You might consider a Q-section of the geometric mean of the antenna and transmission line impedances, and then just a simple 1:1 common mode choke with ferrites.
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!