WiFi antenna comparison using T23 and Broadcom BCM94306

[rkawakami]

Since this question has come up several times before ("which antenna is the best"), and me being the inquisitive engineering type, I decided to start up an experiment. Sourced from eBay, these are the antennas that I have just purchased:

Cisco WIFI miniPCI antenna 23cm long
Wireless & bluetooth 2.4-2.5GHZ mini PCI Antenna/Aerial
Pair Internal Antenna For Laptop Wireless Mini PCI Card
New 2 x Antenna pigtail for MiniPCI wireless card U.FL

Note: Since these listings will only be alive for the next 6 months or so, I have saved them as Adobe .PDF files and will link to them later.

Here's the plan: Using a T23 with one of my usual Broadcom BCM94306MP 802.11b/g cards, I'm going to hook up each one of the antennas and measure the signal strength on the opposite side of my house from the wireless access point (Netgear ME102 802.11b). I will be using NetStumbler 0.4.0 to do this since it has a nice signal vs. time graphing capability. Each antenna will be located outside of the laptop on some sort of stand which I've yet to build. I'm thinking about using some wooden clothespins to hold onto each antenna, at the same height as the built-in antennas in the T23 lid. A single factory T23 antenna will be used as the control reference. This will hopefully compare the effectiveness of each antenna (i.e., the "gain" of each antenna design). Note that these add-on antennas will not have the proper mounting that they were designed for (in antenna design parlance, a "ground plane"). However, this is how most people are going to use them anyway. I will also place the antennas outside the T23 laptop near the end of the battery bay to simulate the installation of the antenna inside at that location. Sorry, but I'm not going to rip open a T23 four times and install each antenna there . This should give me some idea on how the antennas perform when they are aligned in that manner (again, in antenna terminology, vertical vs. horizontal polarization).

Any questions or comments?

[DarkMark]

Definitely sounds interesting.

I'm going to follow this, for one because it's nice to know what works best, and because iI would like to upgrade one ore two of my laptops with internal wireless in the near future, and those antennas aren't even that expensive.

So I'll be glad to here how good they do.

Thumbs up:

[BillP]

A wavelength is only 4.9" at 2.4 GHz. You might find considerable signal strength variation with small changes in antenna position. The position of your body with respect to the antenna may also have an effect.

It's extra work, but I'd suggest repeating the test sequence on at least three different days to check for consistency. Try to replicate the setup as closely as possible on each day's test. Hopefully the results will not vary all over the map.

[rkawakami]

All of the antennas that I have seen so far appear to be 1/4 wave designs (approx. 1.25"). The only problem that I can see is trying to compare a built-in T23 antenna with the free-standing versions that I bought. The T23 antenna is going to have a blind spot because of the lid. My only solution so far is to "aim" the antenna opening in the direction of my access point. I'm also going to have to determine which antennas are the main and aux and then use only one of them for my measurements. Hopefully a simple Faraday cage of aluminum foil will reveal that difference.

[rkawakami]

The last of the four sets of WiFi antennas just came in a few minutes ago. Wasn't too bad; it was bought 8 days ago and came from Taiwan. Two of the others were also from overseas; China and Hong Kong (both received yesterday), the fourth one from Florida (received in three days). I will start the signal strength measurements within a couple of days and set up a web site with the results.

[DarkMark]

Ahh, glad to hear from you again in this thread. Like I already mentioned I'm quite interested in the results. I installed XP tablet edition on my Transnote today, and adding internal wifi would be great.

Thanks for the links also, they arent realy that expensive and shipping is ok too, even to germany.

[carbon_unit]

I bought a set like these http://cgi.ebay.com/ws/eBayISAPI.dll?Vi ... 0156546798
The leads are awfully long and seem to have a lot of loss. I tried them in the lid on my 570 and the LCD cover blocked quite a bit of signal. I put them under the palm rest and it gets better signal than under the LCD cover.
I like the look of these better http://cgi.ebay.com/ws/eBayISAPI.dll?Vi ... 0148705611

[LPC]

Any update on this project Ray? I am thinking of getting these you are testing.
New 2 x Antenna pigtail for MiniPCI wireless card U.FL

By the discription (pigtail) did it come with what I would need to hook up an external antenna and rount the other antenna under the keyboard?

[rkawakami]

Sorry for a delayed start to this project. I've been busy with other things. Experiment is scheduled for tomorrow (Saturday) when I have enough time and space on the kitchen table to lug out all of the necessary equipment. An antenna stand has be constructed, all of the ordered antennas are on hand and (hopefully) the physical arrangement of the laptop and routing of the wires determined.

The "pigtail" description that the eBay seller used in his auction is not really accurate. To me, a "pigtail" is an adapter which converts one type of antenna jack into another. For my 600X project, I used one which had a U.FL plug and RP-SMA jack. You would need something similar if you intend to connect an external antenna. One end has to be a U.FL plug since the MiniPCI wireless cards have U.FL jacks. The other end is up to you and the availability of "off-the-shelf" antennas. You can use SMA, BNC, MMCX or any other version based on those standards (i.e., reverse polarity (RP)). I chose RP-SMA since that is what fit my application and antennas using that style of connector are widely available.

[LPC]

Wonderful! I greatly anticipate the results

[rkawakami]

Spent the better part of Saturday afternoon and early evening gathering the results for the antenna comparison. Had an emergency "honey-do" project assigned to me by "The Boss" (reviving an Epson AIO for her employers) which slightly delayed my experiment. I have posted it at:

http://www.kawakami-ca.com/antenna_comparison/

It's still a work-in-progress but the raw data is there and some inferences can be made. I am part of the way done with explaining the results but still have several sections to complete. I should be able to complete the site sometime today (Sunday). Got to go to sleep....

[LPC]

Ray thanks for all the hard work!

By your graphs would I be safe to say that the
Cisco WIFI miniPCI antenna 23cm long
Pair Internal Antenna For Laptop Wireless Mini PCI Card

performed the best?

[rkawakami]

By your graphs would I be safe to say that the Cisco WIFI miniPCI antenna 23cm long Pair Internal Antenna For Laptop Wireless Mini PCI Card performed the best?

Yes, that's the information presented by the graphs. However, note that the "Cisco" antenna has the shortest coaxial lead of the entire group (9" long) which will severely limit the possible locations for it. I've seen antennas like this, or even shorter, being offered with the advice of just sticking them into the MiniPCI bay ("no disassembly required!!!"). That might be true with a laptop which has all plastic panels but with the Thinkpads, putting an antenna inside will not work too well unless you can find a spot which can "see" the outside world without being covered by any metal. For the T2x systems, that means under the keyboard bezel near the end of the battery bay and/or under the PCMCIA slot where the ventilation holes are.

P.S. Site is now complete. Well, at least Version 1.0 is. I may have some additional updates if I decide to try to modify one of the antennas and see if it helps improve the signal strength.

[LPC]

Wow...thanks for all the work again!

Question on the fourth antenna. You stated in your write up that you would cut off two feet of coax. I remember reading here that shorter antennas get better signal? 36"-46" seems to long for a T23 application. If I used this antenna could I shorten it to fit my needs and not damage it? I was under the impression that the lengths of antennas could not be modified.

Thanks,

[Harryc]

Looking at the results so far, it seems to me that the factory antenna is best because it is right up near the top performance and it fits inside the LCD bezel.

[rkawakami]

The "active" portion of the antenna is what receives/transmits the radio signal. That part is usually what's laid out on the small printed circuit board or metal frame of the antenna assembly. That length makes all of the difference in the world when considering the efficiency of the antenna design. The frequency being used for the radio dictates the length of the active antenna. The higher the frequency, the shorter the antenna needs to be. As mentioned earlier by BillP, the 802.11b/g radio system operates at 2.4Ghz and a full-wave antenna is about 5" long. A 802.11a antenna would be around 2.5" as that system operates at 5Ghz.

The length of coaxial cable that connects the antenna to the wireless card does matter somewhat in that the longer it is, the more signal loss occurs. Having the shortest possible "downlead" will minimize this signal loss. The problem with trying to shorten an existing antenna design is it can be hard to do properly unless you have the tool to peel back the coaxial cable (a good pair of wire strippers) and mark/remember which part of the cable goes where (the center conductor and shield).