impedance transformer for a UWB antenna 1

[mariusnpr]

Who can give me some directions about designing an impedance transformer to use it between the waveguide and an ultra wide band antenna? The bandwidth is 400-4845 MHz

 

[VEBill]

What kind of waveguide are you using that covers 400 to 4845 MHz (1:12 ratio)? While we're on the subject of 1:12 ratio bandwidths, what sort of antenna?

 

[Higgler]

UWB antennas are coaxial fed, waveguide fed UWB antennas is a definite hmmn?

The antennas and propagation society through IEEE is the place to find UWB design info. That type antenna design is very challenging. So much so that reshaping of the pulsed energy sent through these antennas may require pre-distortion of the pulsed waveform.
DM Pozar won an award for his Antenna paper in 2004 on how to pre-distort the waveform for optimum pulse shape exiting a UWB antenna.

kch

 

[VEBill]

Waveguide at 400MHz would be fairly large and (IIRC) such waveguide wouldn't be appropriate to support reliable 4GHz transmission (something about the wave 'flopping over' into different modes because of the excessive dimensions...).

 

[Comcokid]

The connection of UWB or impulse signals to an antenna is complex.

I'll refer you to point #7 in an article Avoid the 'gotchas' of impulse UWB design in EE Times, Jan 10, 2005 pp 60. This article attributes the information to a book An Introduction to Ultra Wideband Communications which will be published in Feb 2005 by Prentice-Hall.

Point #7 warns the designer that classic wideband antennas may have excellent impedance matching but may create unacceptable time dispersion. What I would also like to point out is that classic wideband matching circuits may also create time dispersion - that is to say changing phase delay verses frequency which will smear your impulse. Beware.

 

[jchan]

I'm trying to design a microstrip(patch)antenna for indoor wireless application. The feeding technique is simply a microstrip line. My problem is that i tried to macth it with diffrent technique such as single stub and quarterwave however i'm only getting a bandwidth of 20-40 MHz for a return loss of -20dB or better is there any thechique that i could use to incease the bandwidth to around 80MHz. My substrate is RTduroid 5880 with a thickness of 3.175mm and a Er or 2.2. My patch size is around 40 to 40mm.

 

[Higgler]

jchan
Thin patches are inherently narrowband, key word = thin.
Patch antenna bandwidth is proportional to thickness (thicker = more bandwidth, 2x thicker= 2x bandwidth), you need a thicker dielectric if you want to stay with similar construction.
Also, bandwidth is proportional to dielectric constant. Foam or honeycomb is better than duroid 5880. Bandwidth is proportional to square root of dielectric constant.
kch

PS: seems a thread shift has occured.

 

[SolderingGunslinger]

The simplest way to think of waveguide is as an open-wire feedline with an infinite number of shorted quarter-wave stubs extending perpendicular to the plane of the wires in the feedline. A shorted quarter-wave stub looks like an open circuit at the design frequency.

As you move off the design frequency, the lengths of the infinite number of shorted stubs is no longer lambda/4 and losses build.

At 400 MHz, the waveguide would be 37.5 CM on the two widest sides. At 4.845 GHz, it would be 3.0959 CM.

Now there is a way to make the waveguide work on both 400 MHz and 4.845 GHz. The waveguide would be operated multi mode. It would need to be 37.5 X 3.0959 CM. Not practical, but it should work.

Waveguide is used at frequencies in the 400MHz range. High Power UHF TV uses waveguides, and they are big. Digital TV channel mask filters are built out of waveguide. Check out

http://www.dielectric.com

I remain,

The Old Soldering Gunslinger

 

[VEBill]

I believe that MariUSNPR was asking for continuous bandwidth from 400 MHz up to 4845 MHz; and everything in between. I don't think that waveguide (multimode) would provide a nice, level, frequency-independant response between those frequency limits.

It seems that Mari has wandered off anyway...

 

[jimvandamme]

You want easy & high bandwidth? Can't beat a bowtie.

 

[SolderingGunslinger]

It wasn't mentioned if this antenna is to be used for receiving, transmitting or both.

For a receive only antenna, try an active antenna. The element doesn't need to be at a resonant length for any of the frequencies desired, in fact, it really shouldn't be resonant at any of the frequencies desired.

You will need to design a broad band LNA, I doubt if a MMIC will have a low enough noise figure for your purposes.

For transmit or transceive operations, it becomes a stickier situation, You might be able to design a conical monopole, biconical or discone with that kind of bandwidth, however there will be other problems such as beam tilt and group delay at certain frequencies.

A 17-element LPDA (Log Periodic Dipole Array) could be etched on a piece of double sided Printed circuit board which would cover that frequency range, however it would be directional and have around a 7 dBi gain. Try a Tau of .8 and a Sigma of .21. Be aware that the longest element will be 375 mm and the length of the array will be about 859 mm depending upon how you terminate or connectorize the antenna, so you will need a really BIG piece of double- side PCB.

I remain,

The Old Soldering Gunslinger

 

[mrkenneth]

SolderingGunslinger, may you explain the best way to design a receive-only antenna? I would like it to operate in 500 MHz to 3000 MHz, and have everything on a circuit board. How can it operate well without being at a resonant frequency?

Thank you in advance!

http://www.mrkenneth.com

 

[SolderingGunslinger]

A broad band active receive only antenna usually is a short (in terms of a quarter wavelength) vertical feeding the high z input of a FET amplifier. While you can use a self-resonant length, often the additional gain can cause strong local signals to overdrive the FET. A FET has an extremely high input impedence, so an impedence miss match doesn't cause as big of a problem as it otherwise would.

My first small active antenna used a nuvistor in a cathode follower circuit. Later I canibalized the antenna and used the amplifier on an active electrostatic shielded loop antenna for AM BCB DX...later yet, I took the same tube (with different components) and built an active shielded loop for 2-meter direction finding. The advent of good transistorized receivers for VHF work relegated that amp to the junque box.

A person could use a MMIC amplifier for a UHF-SHF broadband active angenna, however, often the noise figure of the MMIC isn't as good as could be acheived (by a long shot) with a FET amplifier.

Basically, the lack of gain is made up by the FET. If a low noise figure isn't necessary for your application, a MMIC might work for you.

You do not mention your application, however a skeleton feed horn (a V-Beam made from two triangular solid elements) designed for a Low Frequency cut-off below .5 GHz might work better than a snort active antenna, however it would not be on a circuit board...although it could be built from pieces of PCB.

See a copy of the ARRL antenna book for such an antenna.

I remain,

The Old Soldering Gunslinger

 

[mrkenneth]

Thanks for the detailed reply Gunslinger!

I am just trying to measure radiation levels (cellular phones, cordless phones, microwaves, etc.) with something like an AD8314 RF detector that produces a DC voltage in proportion to the voltage at the RF input.

The datasheet does not mention anything about the type of amplifier that is used. It does say that at frequencies above 1 GHz, the input impedance will drop to lower than 500 ohms.

For the antenna, I would like an antenna that can operate over the 500 MHz to 3000 MHz frequency range. I would like the antenna to use as little metal as possible, so reflections are reduced. I was thinking of either a equiangular spiral or a single-ended elliptical antenna. Do you think a balun should be used?

Thanks in advance!

http://www.mrkenneth.com

 

[SolderingGunslinger]

Without making a trip to the engineering library on campus to verify my (admittedly poor) memory, something in the back of my mind tells me that neither the equiangular spiral nor the single-ended elliptical antennas have balanced feeds. Do you propose to use a balanced feed on an unbalanced antenna?

If just impedance matching is needed, then an Un-Un) would be needed. I haven't ever built such a device, after all, why reinvent the wheel? Such devices are available from Mini Circuits

http://www.minicircuits.com/transf.shtml

I have always tried to use off-the-shelf materials in my designs whenever possible...after all, would you design a low frequency PLL using discrete parts? Why wind your own FR transformer.

Good luck with your project.

I remain,
The Old Soldering Gunslinger

 

[mrkenneth]

Thank you again.

The spiral antenna is just a dipole wrapped into a spiral, so it should be balanced. The single-ended elliptical antenna is not.

I have read about baluns on a PCB, but have not came across literature that describes the formulae and such.

I guess I can choose the unbalanced elliptical antenna and put the AD8314 directly on the antenna without any other connections except for a microstrip. The PCB would be larger and more expensive, but probably still less expensive and less difficult than having two boards.

Would the soldermask in front of the copper and tin on the PCB degrade the performance? I am planning to just use cheap FR4 dialetric.

Thanks in advance!

P.S. Are there any free (for personal use) software for simulating antennas?

http://www.mrkenneth.com