Just as a word of warning, I'm neither an RF expert nor pretend to be one. I spent most of my days in the university ditching classes, hanging out at coffee shops, and attending dance rehearsals. I spent the next decade and a half pretending to work at my various jobs. I'm the furthest thing you can get from an PhD'd RF engineer. But if you're willing to accept all that and are still willing to believe me, then read on:

I made a very interesting discovery recently while I was trying to improve the range of my 2.4 GHz wireless boards. I was getting limited range out of the on-board SMD antennas that I was using and was doing testing on multiple boards to see if it was due to the variance of the discrete RF components or if it was an inherent property of the system.

The funny thing was that in my sample of about five boards, one board constantly demonstrated a high received signal and transmitter output, almost twice that of the other boards. I was keeping detailed notes of the component values on all the boards so I copied the exact same component values on another board but couldn’t duplicate the results. The complete matching circuit was identical and I even replaced the radio and the balun to see if those were causing the discrepancy.

I spent about two days investigating this and trying to get to the bottom of why that one board was constantly outperforming the rest of them. I finally found the solution and it was staring at me in the face.

On all of my radio boards, I have two antenna options. One of them is for the onboard SMD antenna and the other is for the right angle SMA connector which can interface to external antennas. You can choose between the options by moving a capacitor to either the SMD antenna path or the SMA path. The board that was constantly outperforming the others had the capacitor in the SMA path rather than the SMD path. The strange thing was that there was no external antenna connected to the SMA.

I began investigating it and found that the SMA connector was actually acting as the RF antenna. I had to say that I was surprised that the SMA was able to radiate at all, but the biggest  surprise was how efficiently it was radiating. It was actually outperforming the 2.4 GHz SMD antenna.

Now I was asking myself how this could happen. The 2.4 GHz SMD antenna was designed and optimized for these types of applications, but it was getting trounced by a lil’ old SMA connector. You have to remember that I had tuned the matching circuit for 50 ohms using my network analyzer and these test boards were all able to achieve close to 99% efficiency as measured at the SMA. Hence, I could guarantee that the SMA RF path was very efficient.

A problem arises when you need to measure the SMD antenna path. Since the SMD antenna path doesn’t have a connector, you need to solder an SMA to it and then measure the return loss. No electrical delays are required when measuring the return loss since you’re just looking at the ratio of transmit energy to reflected energy. However it’s hard to get very accurate readings when doing this because the measurements will change based on where you choose your ground, the orientation of the SMA you hacked on to it, etc. So tuning a circuit with a hacked connector is a lot like fumbling your way through the dark.

Since I’m able to guarantee the matching circuit up to the point where the RF path diverges to the SMA or SMD antenna, the main difference between the two paths would be the difference between the trace lengths at the divergence. In other words, since trace length is just an inductor at high frequencies, the difference would theoretically be the same as adding an equivalent inductor into the RF path. Since I matched the trace lengths for the two paths, that inductor would be extremely small, much less than 0.1 nH according to my board characteristics.

So my conclusion is that the SMD antennas that I’m using are most likely poorly matched to 50 ohms which is why they seem to not radiate efficiently. But more importantly, I discovered that the right angle SMAs are able to radiate efficiently and it looks like its because the total length of the center conductor of the SMA is approximately 30 mm. The 1/4 wavelength of 2.4 GHz is approximately 31 mm which would explain this phenomenon.

By now, you’re probably tired of all this RF talk and want to know what my main point is. The main point is that by using the right angle SMA as a 2.4 GHz antenna, you can achieve a decent, if not good, radiation efficiency. Its more than adequate for short range wireless and I’ve been able to transmit successfully inside a hallway (read: hostile RF environment) at up to 40m with a couple of dB to spare on my Rx signal strength.

If you’re not satisfied with the performance, you can attach an external omni-directional or high gain antenna to the SMA without needing to change a thing. Hence, this interesting discovery allows people to have multiple antenna options at 2.4 GHz based on the form factor and range they want without the need for having two separate matching circuits for an SMD antenna and an external antenna. If you don’t know the importance of this, try ordering an XBee with both an SMD antenna and external SMA connector.

Incidentally, I checked Google and couldn’t find any references to using a right angle SMA connector as an antenna. Thus, I will name this antenna the “WTF Antenna”. It accurately reflects my feelings when I first saw this phenomenon.

And for those interested in the exact right angle SMA connector I’m using, it’s part number 07259 at 4UCON . However I’ve tested this out with a different manufacturer’s right angle SMA and another right angle RP-SMA connector and have seen similar results. Don’t take my word for it. Try it out for yourself :)

P.S. If this type of antenna does indeed already have a name, then please let me know...

PPS: Here's a pic of one of the boards I'm using to test. None of them have an SMD antenna and they are all communicating very well:

TrackBack URI for this entry

Comments (18)

Subscribe to this comment's feed

...
written by Bill Owens, February 10, 2010

Are you sure about that measurement? The 1/4 wave is about 31 mm, but the center conductor on the SMA is probably more like 3 mm, isn't it?

report abuse

vote down

vote up

Votes: +0

...
written by Akiba, February 10, 2010

If you check the spec sheet in the link at the bottom, the total length of the center conductor of the SMA is 14.2mm (plus) 9.63mm (plus) 4mm (bottom pin) so I guess 28mm. The main arm of the inverted F antenna is about 26 mm so the SMA might be acting similarly to that. The main thing is that I'm getting much better performance out of a bare SMA than I am out of my SMD antennas which is totally perplexing.

report abuse

vote down

vote up

Votes: +0

...
written by Bill Owens, February 10, 2010

Ah, I see, it's a female connector. But that means the entire center conductor is inside the shell, and it seems as though that ought to shield it reasonably well. Definitely a weird phenomenon. I don't suppose the trace leading up to the SMA is anything close to 31 mm?

report abuse

vote down

vote up

Votes: +0

...
written by Akiba, February 10, 2010

Just measured the trace. It's approximately 4 mm from the capacitor lead. Actually I realized that the center conductor is less than the 28mm since the mechanical measurements is also including the shell. I suspect that the center conductor is acting like a crooked F-antenna. And I agree, its totally weird, but really cool

Incidentally, I checked the RSSI in different orientations as well and the directionality is pretty consistent so I don't think its adding a lot of directional gain. But if omni-directionality is required, you can just throw on a whip antenna.

report abuse

vote down

vote up

Votes: +0

...
written by Greg Peek, February 10, 2010

I have seen this effect on 2.4GHz WiFi. The SMA's radiate, but I wouldn't call them a "good" antenna, or even really an antenna at all, just a parasitic radiator. Unlikely to be even on the order of an inverted F for efficiency.

You should look into inverted F design for projects that don't need the best range. An antenna for the cost of a bit of extra copper clad board - cheaper and more efficient than an SMA

I agree that your chip antenna is likely very poorly matched. Perhaps you are soldering it in backwards Too bad you don't have a way to measure the SWR.

Have you tried sticking a 1/4 wave length of bare wire in that SMA connector? That should be a better, very cheap antenna.

report abuse

vote down

vote up

Votes: +0

...
written by Akiba, February 10, 2010

Just redid the test with the SMD antennas soldered in backwards. They're still able to transmit but the signal is much lower than the SMAs. I'm seeing a difference of up to 30 dB on the received signal strength.

Also just compared the Atmel RavenUSB sticks versus the naked SMA and the SMA had better received signal strength for the same distance by about 12 to 18 dB. In each case, I'm using the exact same software.

So even if its just parasitics, them's some damn good parasitics.

But seriously, the main point of this is that if its possible to use the external antenna connector itself as the antenna, then it presents multiple antenna options to the user. You can then easily switch from no antenna to omni-directional or high gain directional antenna rather than switching out boards.

report abuse

vote down

vote up

Votes: +0

...
written by Akiba, February 10, 2010

Oops, just checked the AT86RF230 datasheet and I'm reading the Energy Detect register (PHY_ED_LEVEL). The resolution of the energy detect register is in increments of 1 dB and the RSSI register is in increments of 3 dB. The ED register gets updated with the received strength after each frame and is just a reading of the RSSI register averaged over 8 symbol periods.

The differences I was seeing were being multiplied by 3 because I had confused the ED and RSSI resolution. The difference between the backwards SMD antenna is actually 10 dB on the ED register and the difference between the Atmel Raven is 4 to 6 dB in the ED register. I'm using the same receiver in all cases and each of the boards is located in the same spot.

report abuse

vote down

vote up

Votes: +0

...
written by Bill Owens, February 10, 2010

I just read the datasheet for the chip antenna, and found it amusing that although they show SWR graphs and lots of physical specs, there's no antenna pattern. Considering that a non-inductive resistor has excellent SWR but not so great pattern, it makes me wonder

I know next to nothing about microstrips, and very little about GHz antenna design, but I wonder if you could replace the SMA connector with a short piece of bare wire, sticking straight out of the board, and just trim it for signal strength. Start at 35 mm or so and nip off a bit at a time? I have a vague recollection that 'fat' antenna elements have better SWR, so if you made it a short piece of brass tubing it might perform better and be robust enough for serious use.

report abuse

vote down

vote up

Votes: +1

...
written by Akiba, February 10, 2010

Ha ha ha...not sure how a piece of brass tubing sticking out a wireless dev board would look. I guess steampunk?

I think my point is that using the SMA as an antenna gives you more options because you can use it as a bare antenna or you can attach whip or gain antennas depending on the application. Its kind of cool because then you can have a very configurable WSN platform that can be used in the lab with no antennas for development purposes. And then when you want to do an actual application, like weather monitoring, you can slap on an outdoor directional antenna, etc. For a dev platform, its all about choices because if you can support different configs without having to re-tune the matching network, then things become so much easier.
I doubt this would actually get deployed in a real product but for rapid prototyping of WSN applications, I think its quite exciting

report abuse

vote down

vote up

Votes: +1

...
written by Brad, February 23, 2010

Nice Hack! If digikey finds out, they will charge more for connectors! I have observed the RAVENs being more directional along their edge than the circuit face. The LCDs occlude the antenna as well. I have not tested the USB sticks.

report abuse

vote down

vote up

Votes: +0

...
written by Vinod Ganesh, February 23, 2010

i am particularly happy with name you chose for this phenomenon

report abuse

vote down

vote up

Votes: +0

...
written by Akiba, February 23, 2010

Thanks. I kind of think the name has a nice ring to it m'self

report abuse

vote down

vote up

Votes: +0

at86rf212 boards getting ready
written by Vinod Ganesh, February 26, 2010

hi akiba,
i am getting ready a new module with same size as zigbit with UFL connector with AT86RF212 and ATMEGA1284P with a limited 25 numbers. How do we port the bitcloud to this design?
i want to be part of the opensource initiative.
regards
VinGan

report abuse

vote down

vote up

Votes: +0

...
written by Akiba, March 03, 2010

Unfortunately, you have a problem because BitCloud isn't open source. You'll have to write your own stack. You can use mine as a template and then add in (remove) the features you want. My stack is still under development (for eternity it seems) but all the layers are there and the frame formatting and mesh routing are working, as confirmed by my Daintree protocol analyzer.

report abuse

vote down

vote up

Votes: +0

...
written by Deepika, April 16, 2010

The blog contains quite a number of aspects as well as essential tips about WTF Antenna. I like this post , and the examples discussed have been remarkable as well.

report abuse

vote down

vote up

Votes: +0

...
written by Manpreet, April 23, 2010

Good tips about WTF Antenna. I like this post waiting for the next post.

report abuse

vote down

vote up

Votes: +0

consultant
written by colin s, April 29, 2010

I was looking over your schematics, and noticed that you are not grounding the center tap of the balun. Any reason for this choice? The Atmel datasheets show a ground, but I would have expected a greater output swing with an open circuit on the center tap, like you have.

report abuse

vote down

vote up

Votes: +0

...
written by Akiba, April 29, 2010

The balun is the Murata LDB182G4510C-110 which is part of the "C" series and has a different pinout than the "G" series. The only GND pin is pin 3 and the center pins 2,5 are NC. You can check out the specs here:

http://www.murata.com/products/catalog/pdf/k99/k99e_s0800.pdf

report abuse

vote down

vote up

Votes: +0

Show/Hide comment form

Name

Email

Website

Title

Comment

smaller | bigger

Subscribe via email (Registered users only)

I have read and agree to the Terms of Usage.

Add Comment