Re: Back-to-back antennas
John Navas
<http://wireless.navas.us/wiki/Wi-Fi_How_To#Extend_Wi-Fi_through_a_brick_wall>
Dave M. writes:
>My first reaction to using wired back to back antennas to get around a
>brick wall is that it won't work. There is too much loss. I admit I
>have not done the calculations for this particular instance, but I have
>in other instances and found it always is too lossy. I am interested to
>know if anyone has really made this work. Maybe I should put some
>numbers down here.
>+15 dBm transmit power, 28 mW, WRT54G stock
>+7 dB transmit antenna gain, optional high gain antenna
>(+ 22 dBm ERP, exceeds 200 mW by 2 dB, may be an infraction)
>-54 dB free space loss at 2.4 GHz and 5 meters
>+8 dB receive antenna gain, could get a higher gain antenna
>-1 dB coax between antennas and mismatch losses
>+8 dB second transmit antenna gain, could get a higher gain antenna
>-60 dB free space loss at 2.4 GHz and 10 meters
>+2 dB second receive antenna gain, estimated notebook or PCI card type
>-90 dB, Net Path Loss, sum of all gains and losses
>-75 dBm receive power at receiver, transmit power + Net Path Loss
>This system needs about 20 dB more receive power to be feasible. If one
>could change the back to back antennas to 18 dB antennas, you might
>make it.
>Still, the two end terminals here could not be farther apart than the
>two distances shown, a total of 15 meters. Free space loss at 15 meters
>is 64 dB, compared to 70 dB for the 18 dB gain back to back antennas if
>you had them. So if you don't have good isolation from other extraneous
>reflections and refractions, you may get self-interference, maybe
>picket fencing, that is problematic. Oh Oh, the wikipedia reference
>does not explain the proper nature or origin of picket fencing.
>A 2x2.5 ft passive repeater, properly aligned, should have a gain of
>about 50 dB, which is 16 dB better than the two 18 dBi back to back
>antennas. That 50 dB gain is just what I cranked out here in a
>spreadsheet, I would believe it when I see it. It assumes a zero degree
>angle between the incoming and outgoing path, which is obviously wrong.
>Need to correct by cos(alpha), the incident angle. :(
>Other considerations: Use of such a high gain antenna facing past the
>wall means high directivity, so you won't get a wide coverage angle.
>Use of a reflector has the same problem, only more directive due to
>higher gain. And if you really are within 5 meters of the wall on the
>inside, better just move the router the 5 meters!
>Now why am I writing this???? I should be going to bed!
John Navas
At first glance, it's sorta correct but very disorganized and has
some math errors. Also, the +2dBi antenna gain on the "notebook or
PCI card" is way optimistic. 0dBm at best.
However, the wiki is for answers, not questions. I suggest you ask
him to clean it up so that it's definitive instructions on how to use
a passive repeater (a.k.a. back to back antennas). Ask the questions
in the newsgroup, then post the consensus in the FAQ.
Revised number using most of his numbers and only correcting coax
losses and
Transmitter: +15dBm
Transmit coax loss: -1dB (about 0.5 meters with connectors)
Transmit antenna: +7dBi
Free space loss: -54dB at 5 meters
Receive antenna: +7dBi
Rx and tx coax loss: -1dB (about 0.5 meters with connetors)
2nd transmit antenna: +8dBi
Free space loss: -60dB at 10 meters
Rx antenna gain 0dBi (internal laptop antenna)
Delivered RX power ??dB
Doing the arithmetic, I get -79dBm which is low but usable to about
24 Mbits/sec. See my chart at:
<http://802.11junk.com/jeffl/rx-sens/receiver%20sensitivity.htm>