On Sat, 07 Apr 2007 14:06:19 -0700, Tom kingman wrote in :
Why not use a reflector instead (thereby not taking risks like voiding any warranty)? See wikis below for links to info on how to easily build a cheap reflector that will actually work better than many much more expensive antennas.
Sure. I take apart everything I can get my hands on.
Fine. Time for "Learn by Destroying(tm)".
Hint: If its not easy or obvious, that's because the designers don't want you taking it apart.
Always go to the FCC ID site first. See photos at:
The radio (Atheros chipset) is under the "T" shaped metal shield on the right. The metal plate at the edge of the circuit board to the right of the shield area is the actual antenna. The antennas are PIFA type which will have a convenient 50 ohm feed point.
However, you're going to have a big problem. The box has *THREE* antennas. Two for receive (diversity) and one for xmit. See block diagram at:
In order to attach an external antenna, you'll need to either impliment a T-R switch (not recommended) or attach at least *TWO* external antennas. I suggest you consider this carefully before you blunder onwards.
Verify the FCC ID number from the serial number sticker to make sure it's the same as what you have.
As for how to take it apart, you're on your own. Please consider the sound of breaking plastic as an indication that you're doing it wrong.
Because the 2700HG has *THREE* antennas inside (2 receive, one xmit). A single corner or parabolic reflector just isn't going to work if there has to be 3 different points of focus.
A flat plate reflector will work yielding 3dB of gain, which is better than nothing. Find a random piece of very flat aluminium sheet metal and position it 1/4 wavelength behind the approximate plane of the antennas. Play with the spacing for maximum gain.
Sir, can you point me to some instructions on how to find those fcc photos? I gather you must start with the fcc ID number but I haven't stumbled yet into how to use that number to get anything useful from the FCC.
There's photos of the product, block diagrams, and other info, all of interest. I've looked but have found no basic instructions from the FCC site on how to use it.
Good question. The problem with the FCC ID search page is that the search is broken. I haven't figured out exactly how it's broken, but it often fails to find a known valid FCC ID, even when inscribed completely. I assume you want to find something where you do not know the FCC ID and only have the manufacturer and the model number. That's a common problem for everyone which is aggravated by various vendors not bothering to supply the FCC ID number on any of their literature or web pages.
You can sometimes get lucky with a Google search using: FCC equipment IDs example search: "fcc B4Z-34009-PIR" Remember to put the word "fcc" before the equipment ID. Even the example doesn't work, but it's worth a try.
Otherwise, start at:
which will eventually lead you to:
There are really only 3 items of interest on this page that need to be found. The "Grantee Code" which is the first 3 characters of the FCC ID, and the "Product Code" which is the rest of the line. The last is the "Show XX records at a time" at the bottom of the page. Change that to 100.
If you don't know the 3 character Grantee Code, then go to:
For example, plugging in Cisco (for Linksys) will yield prefixes of: LDK Cisco PKW Cisco-Linksys Q8G Cisco-Linksys Q87 Cisco-Linksys Most of the older Linksys stuff is under PKW and the newer stuff under Q87. I think Q8G is dead.
Now, we go back to the search page and try PKW for the "Grantee Code".
Don't forget to set the "Show xx records..." to 100.
That yields 227 hits for Q87 (Linksys). Go down the FCC ID column until you find the model number of the product. In most cases, the model number will be obvious, but not always. A few vendors use really obscure FCC ID numbers, which will require a model by model search for the correct FCC ID. You will also find more than one hit for a given model number.
Now, here's the part where the FCC ID search absolutely sucks. You can't do a partial search by model number. For example, if you use Q87 for the "Grantee Code" but insert just "WRT54G" for part of the remaining "Product Code", the search finds nothing. The help page shows that it should do a left justified substring search as in: * Wild Card Right - When the search is performed the characters entered into this field will be compared to the database using a wild card on the right. For example, if the letter "A" is entered then all records that start with the letter "A" in this field will be returned (ABC, Apple, Able...). However, it usually doesn't not work unless you know the tricks.
One trick is to put a "-" in front of the search. That's because the "-" is significant and part of the FCC ID number. The problem is that not every company uses the "-" so this trick doesn't work for everyone. Using Q87 for the "Grantee Code" and "-WRT54G" for the "Product Code", yields 32 hits. However, because of the left justified nature of the search, manufacturers that imbed the model number in the product code prefixed with some random garbage, cannot be found in this manner.
The rest is just a matter of looking down the list and reducing the number of hits to manageable size. There are a few other tricks but this will do for now. I'll also spare you my opinion of the FCC and it's general inability to do anything correctly.
That's exactly the (unofficial) reaction I received when I first complained. Same with the screwed up ULS system or whatever it's called this week. I've had to deal with the FCC for about the last 40 years and find both the attitude, quality, and responsiveness to be seriously lacking. However, you're correct. Better free and broken, than not at all. Were the FCC to seriously review my opinion of the search page, and decide that it's worth fixing, I'm certain the entire system will degenerate into a bureaucratic quagmire resulting in limited access or total shutdown based on some "national security" excuse along with the traditional cost over-run and 5 year beta test period. Of course, the manufacturers will want the site shut down completely for confidentiality reasons.
The only reason the site exists in the first place is that the FCC is required by law to "make available" under some ammendment to the Freedom of Information Act, any documentation that is likely to be requested more than a few times. To avoid the FOIA document request hassle, they just dumped it on the internet. Ah... foundit. See
As a minimal request, I've asked the FCC twice to include the "-" trick in the help file (to the left of the "product code" box) and to not touch the search engine for fear of breaking something. No response.
The FCC "registration" site is the CORES web site. If you want to do business with the FCC, you need to register and get a number. Any semblance to Big Brother is strictly coincidental.
The license info is under the ULS (Universal Licensing System) at:
Typing in my name "Jeff Liebermann" into the licensee search page results in the usual nothing found.
However, if you go to the CORES (Commission Registration System) page:
and search by just my last name, you'll eventually find my FRN (FCC Registration Number) as: 0003838828 You can then go back to the ULS page, insert the FRN number, and display my current and previous ham license data.
However, if you want to display my FCC license info, you have to the FRN page, and copy down my name exactly as it's scribbled: Liebermann, Jeff I Then paste that into the ULS search page to display all 3 of my assorted licenses.
As I previously mumbled, you don't stand a chance using the FCC search pages without knowing the tricks.
On the other foot, the page I use the most is the ULS Advanced License Search at:
which works amazingly well and doesn't require much trickery.
However, on page 19, it shows that it drops quite a bit at the band edges for 802.11b. Yech. 20.1dBm (100mw). There's something wrong here. No way does any access point I've seen act like that unless there's something broken in the 2.4GHz BPF (band pass filters). However, if I go to Page 30, the 802.11g power output is a consistent
26dBm (400mw) across the entire band. Ok, got it. 802.11b is tx power is weird.