Design help printed (crossed)-dipole

[Tom48]

hi,

im trying to design a printed (microstrip) dipole.
Im currently having a hard time finding some good information on how to design such a printed dipole antenna.

I would really appreciate some help/hints in form of links/books/equations that can help me to design such an antenna element. (or maybe tell me what some of the essential steps are to design/compute it)

Also I am quite unsure when to use a balun feeding network and when it is not necessary.

Thanks a lot in advance for your help! :)

 

[berkeman]

hi,

im trying to design a printed (microstrip) dipole.
Im currently having a hard time finding some good information on how to design such a printed dipole antenna.

I would really appreciate some help/hints in form of links/books/equations that can help me to design such an antenna element. (or maybe tell me what some of the essential steps are to design/compute it)

Also I am quite unsure when to use a balun feeding network and when it is not necessary.

Thanks a lot in advance for your help! :)

Can you say what frequency you are tuning this dipole for? What material will it be printed on, and how thick is it? What RF TX amp are you using, and how is it connected with your RX circuit and the antenna feedline?

EDIT -- And when you say "crossed dipole", do you intend to have both polarizations fed by the same TX/RX circuit, or will you have separate TX/RX circuits for each polarization?

 

[Tom48]

hi berkeman, thanks a lot for your fast response!

Frequency is 867MHz (RFID) and its planned to use a FR-4 substrate (epsilon 4.3).
Material thickness is about 1.55mm (but might be increased due to radiation efficiency by the multiple of 1.55mm)
For the start i would like to use coaxial probe feeds. (50Ohm)

Im not sure about the RX/TX amp yet but i would like to have both polarizations (vertical and horizontal) fed, using a crossed dipole.
In case of a crossed dipole the vertical polarization and the horizontal polarization should be evaluated separately respectively.

Motivation:
Lets say we have a RFID-Tag and we know its polarization is a linear vertical one. Now I am going to use the crossed dipole antenna we are talking about to check the polarization of the received signal sent by the RFID-Tag to the antenna.
As we know the response (sent signal by the RFID-Tag) is expected to be vertically polarized.
The idea was that in this case the vertical (dipole) feed of the antenna is going to receive "a much stronger signal" than the horizontal feed. (if both feeds are NOT add up but evaluated separately, maybe check the ratio as well)
So that way i could assume this is no multipath. As for a multipath a reflection might have occurred somewhere and the reflection most likely would have changed the signals polarization, causing a "weaker signal" on my vertical feed of the antenna.

Some hints/help on how to design such a printed (crossed)-dipole antenna would be great.
So far i really could not find any good information about it (how to compute the width, length [except each "arm" of the dipole should be approx. a quarter wave length as far as i know]), especially when do i need a balun feeding network etc.

Thank you so much!

 

[berkeman]

I think trying to lay out a crossed dipole pattern may be a bit difficult -- the feedlines will interfere with the pattern of the other polarization antenna.

Instead, I'd look at doing something more like an "L" shaped layout, with the end of each dipole at the corner of the L, and the feedlines going into the midpoints on the outsides of each leg of the L. That should minimize the interference of each antenna with the other antenna's feedline. There may be a better way that other posters can suggest.

I do think you will need baluns or some equivalent balun structures to match your unbalanced 50 Ohm feedlines to the dipoles' balanced Zo. BTW, there are some radio chips that already have provision for two antennas, for antenna spatial diversity to combat multi-path issues. In RX mode, some can automatically select which antenna is receiving the strongest signal, and focus the processing on that antenna channel. Check out the candidate radio chip datasheets to see if they automatically support 2 antennas.

 

[tech99]

I think trying to lay out a crossed dipole pattern may be a bit difficult -- the feedlines will interfere with the pattern of the other polarization antenna.

Instead, I'd look at doing something more like an "L" shaped layout, with the end of each dipole at the corner of the L, and the feedlines going into the midpoints on the outsides of each leg of the L. That should minimize the interference of each antenna with the other antenna's feedline. There may be a better way that other posters can suggest.

I do think you will need baluns or some equivalent balun structures to match your unbalanced 50 Ohm feedlines to the dipoles' balanced Zo. BTW, there are some radio chips that already have provision for two antennas, for antenna spatial diversity to combat multi-path issues. In RX mode, some can automatically select which antenna is receiving the strongest signal, and focus the processing on that antenna channel. Check out the candidate radio chip datasheets to see if they automatically support 2 antennas.

I agree with Berkeman that it will be difficult to lay out the feeds for the crossed dipole configuration. However, I feel that the proposed L structure will introduce excessive coupling between the two dipoles, as the ends are close to each other and the layout is not symmetrical.
How about having separate dipoles, spaced apart by, say, half a wavelength?

 

[berkeman]

However, I feel that the proposed L structure will introduce excessive coupling between the two dipoles, as the ends are close to each other and the layout is not symmetrical.

Yes good point. @Tom48 -- Do you have room on your product to use two external antennas that can swivel? That way you can just use standard hardware and tell your users to set up the two antennas tilted opposite ways at 45 degrees (to offset them by 90 degrees).

http://g02.a.alicdn.com/kf/HTB1z1f_...um-body-Wifi-Router-VPN-intelligent-smart.jpg

 

[Tom48]

Thank you so much for the input, both of you!
Unfortunately it has to be a printed antenna. (no external one)
So do you have some ideas about printed ones?
I would at least like to try out to model a crossed dipole and simulate it in order to check how strong the coupling/cross-polarization is going to be actually so
i would still be very happy about some hints on how to design such a printed dipole. :)

How about having separate dipoles, spaced apart by, say, half a wavelength?

Do you mean one vertically polarized dipole and half a wavelength apart the second dipole that is horizontally polarized?
If so, i already wanted to try that out but then there's the problem that I am still not sure how to model a dipole in the first place...
So i would really appreciate some help on designing printed dipoles. (i really could not find any info on the net or books so far that could actually help me...)

thank you so much for your time!

 

[Tom.G]

The Amateur Radio group, ARRL, has some links for antenna software.
http://www.arrl.org/antenna-modeling

Also Google returns many hits.
https://www.google.com/?gws_rd=ssl#q=antenna+simulator+free

At least some sites have downloadable manuals for the software.

 

[tech99]

Thank you so much for the input, both of you!
Unfortunately it has to be a printed antenna. (no external one)
So do you have some ideas about printed ones?
I would at least like to try out to model a crossed dipole and simulate it in order to check how strong the coupling/cross-polarization is going to be actually so
i would still be very happy about some hints on how to design such a printed dipole. :)Do you mean one vertically polarized dipole and half a wavelength apart the second dipole that is horizontally polarized?
If so, i already wanted to try that out but then there's the problem that I am still not sure how to model a dipole in the first place...
So i would really appreciate some help on designing printed dipoles. (i really could not find any info on the net or books so far that could actually help me...)

thank you so much for your time!

This calculator will enable you to find the wavelength along the microstrip line forming the antenna. Then you can make the dipole, say, 45% of this length. Do a test and then scale things up or down slightly for the real dipole to obtain the correct frequency.
http://www.microwaves101.com/calculators/873-wavelength-calculator

 

[berkeman]

Unfortunately it has to be a printed antenna. (no external one)

Is the plane of your PCB going to be orthogonal to the direction to the RFID tag? In the picture I posted, the antenna patterns for the two vertical antennas will be horizontal. If you use a horizontal PCB antenna assembly, the two patterns will be vertical donuts, so they will be directional in the horizontal plane.

Can you provide a sketch of your reader and the position of the tag?

 

[Tom48]

This calculator will enable you to find the wavelength along the microstrip line forming the antenna. Then you can make the dipole, say, 45% of this length. Do a test and then scale things up or down slightly for the real dipole to obtain the correct frequency.
http://www.microwaves101.com/calculators/873-wavelength-calculator

ok, thank you.
I know how to compute the "dipole arm length", but clueless about the width (just match the width to 50ohm?) and the feeding.
Do i need a balun, transformer? Maybe a simple sketch of a dipole would help, so you could tell me how you obtain the values of the particular part of the antenna.

Is the plane of your PCB going to be orthogonal to the direction to the RFID tag? In the picture I posted, the antenna patterns for the two vertical antennas will be horizontal. If you use a horizontal PCB antenna assembly, the two patterns will be vertical donuts, so they will be directional in the horizontal plane.

Can you provide a sketch of your reader and the position of the tag?

thank you!
Im not quite sure about it yet, for now i would like to have a dual polarized (linear) printed antenna so i can test how reflection will affect the polarization of a given signal. (already designed a rectangular microstric patch antenna, with 2 feeds placed orthogonally to each other, but would like to have another dual linearly polarized antenna for comparison)

thanks a lot!

 

[Tom.G]

Take a look at:
http://www.microwavejournal.com/art...-wlan-and-wireless-communication-applications

Google returns over 300 000 hits:
https://www.google.com/?gws_rd=ssl#q=printed+cross+dipole+antenna
Hint: also try the search without the word "cross"

 

[Tom48]

ok, thank you!
Im going to do some more research then.

In case someone still has some useful info regarding printed dipoles, please do not hesitate to post it here.

Thanks! :)

 

[Baluncore]

You do not need a balun to match a printed antenna because the width of the antenna strip can be adjusted to match the transmission line. A quarter wavelength long transmission line transformer could be used on the PCB. It would have an impedance that was the geometric mean of the transmission line and the antenna. Z_qw = √ ( Z₁ * Z₂ )

If the RFID tags are linear polarised, and you do not know the orientation of the tag, you could use a single circular polarised antenna that will receive any linear polarisation.

It is best to avoid a ground plane behind a printed antenna because an equal and opposite current flows in the ground plane that will cancel much of the radiated signal.

 

[tech99]

You do not need a balun to match a printed antenna because the width of the antenna strip can be adjusted to match the transmission line. A quarter wavelength long transmission line transformer could be used on the PCB. It would have an impedance that was the geometric mean of the transmission line and the antenna. Z_qw = √ ( Z₁ * Z₂ )

If the RFID tags are linear polarised, and you do not know the orientation of the tag, you could use a single circular polarised antenna that will receive any linear polarisation.

It is best to avoid a ground plane behind a printed antenna because an equal and opposite current flows in the ground plane that will cancel much of the radiated signal.

I have not made a microstrip dipole, but my suggestion is not to have a metal substrate behind it. If you use the folded dipole shape you can use a 300 ohm balanced line to feed it. Then make a 300 to 50 balun just in the strip line. Maybe a quarter wave stub with a tapping point would do this. Really, unless you can find an existing design, you have a deceptively difficult task, which would tax even an experienced engineer.

 

[Tom48]

So i managed to model a design in cst for my desired frequency. (printed dipole)
The obtained values so far seem to be pretty good.
Radiation pattern looks promising as well:
https://postimg.org/image/cwoehaskj/

I used this design (with via balun feed):
http://www.qsl.net/va3iul/Antenna/P...inted_Dipole_with_Microstrip_Balun_2.4GHz.gif

So far so good but now a new problem arises.
I would like to orthogonally add a second dipole to the one i already created so i get a kind of crossed dipole, like also can be seen in the
right bottom corner in this picture: http://www.qsl.net/va3iul/Antenna/P...inted_Dipole_with_Microstrip_Balun_2.4GHz.gif
The problem now is that i can not get a proper dipole radiation pattern out of the design, because the big ground plane seems to prevent the radiation here.
Is it even possible to get a good dipole radiation pattern out of this "crossed dipole" design with such a huge gnd plane?
As such a design seems to be used already i assume it is possible.
Do you have any suggestions what to try out in order to get a useful design out of it?

thank you!