How to increase BW of microstrip antenna?

[deki]

I'm using the DN007 antenna: http://www.ti.com/lit/an/swra351a/swra351a.pdf and I've managed to get the same loss/response in CST as what they have in their data sheet, however I'm wondering how I can increase the BW? It passes through ~2.4GHz, but I'd like to increase that to include about +-100MHz. My research so far has brought up adding in a patch antenna on top, though I'm unsure how to do this? Any ideas?

 

[tech99]

I'm using the DN007 antenna: http://www.ti.com/lit/an/swra351a/swra351a.pdf and I've managed to get the same loss/response in CST as what they have in their data sheet, however I'm wondering how I can increase the BW? It passes through ~2.4GHz, but I'd like to increase that to include about +-100MHz. My research so far has brought up adding in a patch antenna on top, though I'm unsure how to do this? Any ideas?

The antenna covers 2.2 to 2.65 GHz with a return loss of 10 dB, which seems to meet your requirement. It looks from the response that some reactance compensation has already been done; notice a small square capacitor at the feed point. My basic feeling is that to extend the bandwidth much further might be problematic.

 

[deki]

Actually I was going to ask about that but forgot! So the 280 MHz BW means that it will work for the range of 2.2-2.65 GHz? What I don't understand is why my CST simulation (and the data sheet response) shows it passing through ~2.4 GHz more so? Then the loss decreases progressively on either side. I think what I'm thinking of is a filter though. I'm still a novice when it comes to antenna design/response (EE student).

 

[tech99]

Actually I was going to ask about that but forgot! So the 280 MHz BW means that it will work for the range of 2.2-2.65 GHz? What I don't understand is why my CST simulation (and the data sheet response) shows it passing through ~2.4 GHz more so? Then the loss decreases progressively on either side. I think what I'm thinking of is a filter though. I'm still a novice when it comes to antenna design/response (EE student).

I think you might be confusing Loss for Return Loss. The latter relates to the impedance matching - the higher the number the better. For instance, if half the transmitter power is reflected at the antenna (due to mismatch), then the reflected signal has experienced a loss of 3dB. This is the Return Loss.

 

[JakeBrodskyPE]

My ham radio mentors from decades ago warned me about misunderstanding this subject. Your confusion is very similar to my confusion back then, so I am going to convey those concerns to you.

First, please note that you are discussing bandwidth and impedance match on this antenna. You are not discussing antenna gain. In general, if there is a dielectric anywhere near the antenna, you will probably lose signal. Dielectrics, particularly circuit board materials, start to have significant losses as frequencies increase toward the microwave regions. Printed circuit board antennas are usually poor performers for transmitting purposes.

Second, note the wavelength of this signal. A half wave in free space should be slightly less than 6 cm in length. Less than a 1/4 wave antenna tends to have isotropic gain at best.

Third, a well matched antenna does not necessarily radiate well. As my mentors pointed out, a dummy load may have a nearly perfect match, but it doesn't radiate well at all. It is even possible that a badly mismatched antenna might radiate quite well. That said, for high speed digital modulation, reflections caused by mismatched antennas can lead to intra- and inter-symbol distortion. Perhaps you may be old enough to remember analog TVs with mismatched antennas that lead to an image with shadows. That's what a mismatch can do to higher bandwidth signals.

As you shrink antennas, and as you try to keep your radiation efficiency up (compared to an isotropic radiator) your bandwidth will decrease. Very small antennas, if they radiate well, will have correspondingly narrow bandwidths.

So, if you're trying to improve performance and bandwidth, you will probably need to get much closer to a half wavelength in free space. Try to keep other things at least a half wave away from that antenna. There is much more to discuss about this subject. Antennas are the target of more ignorance and snake-oil salesmanship than most other fields I know. However, the study of antenna and transmission line theory can be very rewarding...

Jake Brodsky, PE
Amateur Radio Station AB3A

 

[JakeBrodskyPE]

One further note: Having reviewed the app note swru120b I have to wonder what they were drinking when they wrote up the gain figures for this antenna. There is no reference to isotropic radiators or dipoles. There is just a bland "gain" figure. They don't say how they measured it. I think it's someone''s misconception of what antenna gain is.

 

[deki]

Thanks for the detailed response Jake. So going by your post, the Reflection response they have on page 11 isn't the same as my response from CST:
http://i62.tinypic.com/42k36.jpg <-- so that is totally unrelated to BW?

 

[JakeBrodskyPE]

Deki, when you throw around terms like bandwidth, you should include a context so that people know what you're talking about. In this case, I suspect you're discussing bandwidth of the 10 db Return Loss match. Having a return loss of 10 dB or greater is a reasonable match for most purposes (unless you're using modulation that requires 20 dB or more of dynamic range).

The picture of S₁₁ in the application note may have been measured. The note doesn't say much about the data source for the graph. They don't say why this bandwidth is resistant to dielectric effects, they just claim it as if it were so. In practice, many dielectrics have permittivity of 2 or thereabout, so it is wouldn't have a huge effect. But the quality of the match of the antenna doesn't have much to do with that.

Beware of excessively wide bandwidth claims. Some designs, such as fractal antennas, are continuously loaded with capacity stubs in a fractal formation. These achieve about the best broad impedance match bandwidth you can get for a small antenna. However, actual radiation performance is still awful. Small antennas (less than 1/4 wave) are poor radiators. You wouldn't have enough of the radiating surface of the RF cycle to efficiently couple your energy to free space.

Usually, when discussing small antennas and wide bandwidth, the signal power is being dissipated as resistive heat. It does not help to radiate a stronger signal.

 

[deki]

Thanks for the response Jake. I'm going to try and design my own PCB antenna.