kbc1212004@yahoo.com hath wroth:
>I plan to set up a PC on the thrid floor of my house. The cable
>modem/router is located in the basement, perhaps about four feet
>underground The distance between it and the upstairs room is about
>thirty feet horizontal,, thirty feet vertical away. This is a brick
>house build c. 1940..
Brick floors? My guess is wood floors and lath and plaster ceilings.
The wood is no problem, but the lath and plaster is difficult to
penetrate with 2.4GHz RF, especially if the lath has been reworked and
reinforced with chicken wire backing.
>The router is a Linksys WRT54GL router v1.1 flashed to DD-WRT .
Unless I missed something, the WRT54GL was only released in the 1.0
hardware version. However, the much much older WRT54G had a v1.1
version (I have one). There's a huge difference inside, especially in
amount of RAM, the RF section quality, and the processor clock speed.
See table at:
<http://en.wikipedia.org/wiki/WRT54G>
<http://www.linksysinfo.org/forums/showthread.php?t=47124>
and then check your serial number tag for the correct model number or
hardware version.
>Using a notebook with a Hawkings HWC54G card (not a Hi-Gain model),
There's no high-gain model. However, there is a high-power model.
>I am getting about -45dBm from a position one floor up, and about -72dBm
>two flights up.
Ok, the math is simple. -45dBmi is a very good strong signal. -72dBm
is marginal, but should work at very slow speeds. The difference is
27dB, which you have to make up in either path improvement or antenna
gain. A 24dBi gain antenna looks like and is about the size of an
outdoor barbeque grill. It will probably do the trick but I don't
think you would consider that very practical. It doesn't really
matter what style of antenna you select, anything over about 20dBi of
gain is going to be big and ugly. You could also distribute the gain
between the laptop and the access point with perhaps 13dBi gain at
both ends. That would look less obnoxious, but would require fixed
positioning of both ends. In other words, the laptop would not be
very portable. Similarly, pointing the antenna upwards from the
basement would probably all operation only in one part of the house.
Incidentally, your apparently loss through a single floor is:
-27dB / 3 = -9dB per floor.
>The Linksys router antennas were positioned
>perpendicular to the desired locations.
That will help one or two dB. However, you need about 27dB more gain,
which is not going to happen by juggling antennas, or even replacing
them with relatively small aftermarket antennas.
>Increasing the transmit power
>from 28 to 48ma improved the signal just two-three dBm, as indicated
>by posts in this forum.
That's 28 to 48mw (milliwatts). That's an increase of about 3dB,
which isn't going to do much for a 27dB shortfall. Increasing just
the transmit power only increases the gain in one direction. The
other remains the same. If you installed a kilowatt fire belching
power sucking amplifier on one end, it still won't work, because you
couldn't hear the return signals. When increasing transmit power, it
has to be done at BOTH ends of a link to be effective.
>While the signal on the third floor is rated poor/bad by the Hawkings
>ultility, and generally gets just two bars from Windows; the signal is
>steady, I didn't have any particular delays in getting webpages or
>downloads, and I got Speedtest scores of about 6mbits/sec, or about
>65% of wired speed.
Nice. I would have guessed that going through 3 wood or lath and
plaster floors would not have worked as well. It appears that you're
close and don't really need the full 27dB of gain. Perhaps a few dB
would be useful. As a rule of thumb, throughput doubles for every 6dB
of gain. The typical 8 to 10dBi patch, panel, or biquad antenna,
pointed upwards, should help. It will never give you a -45dBm signal
as you were getting at one floor, but it might make it work a bit
better.
>Given that the signal needs to get to a fixed point, it would seem
>that a directional antenna would be the best answer.
Maybe. Please note that there are alternatives to wireless, such as
running data over the CATV coax, phone line networking, power line
networks, and fiber optic links. These are particularly applicable
for fixed locations, which seems to be your situation. References on
request.
>Having seen the
>discussions about transmit power and the need to have both ends
>transmitting at about the same power for it to do much good, what's
>the case with antennas?
There is no need to have identical gain, type, or configuration
antennas on both ends of a link. Antennas are bi-directional and
redirect the signal identically in both directions. (Note: antennas
do not amplify the signal. They redirect it).
>If, for instance, I had a 10dBm directional
>antenna attached to the router, and a 10 dBm antenna upstairs, would I
>end up with double the improvement (assuming all was aligned
>correctly), or would I have just whatever improvement in each
>direction that came from each antenna?
Double, sorta. See my previous back of envelope calculations. If you
need 20dB of antenna gain, you can put it all at one end of the link,
or you can distribute it in any ratio between each end of the link as
long as the total gain is about 20dB. The real issue is antenna size.
A single 20dBi gain antenna is rather large. Two 10dBi antennas are
much smaller (and cheaper). You could even build your own. Another
benfit to distributing the gain is that the beamwidth is wider with
lower gain antennas, making aiming and alignment far less critical.
>I guess my real question is: Do these hi-gain antennas improve both
>transmission and reception, or just transmission?
They improve signal in both directions.
>Assuming that DIY is not an option, what would be good but fairly
>inexpensive (i.e. not much more than $50) manufactured directional
>antennas that would be compatible with the Linksys router?
Any of the lower gain antennas on this page:
<http://www.fab-corp.com/home.php?cat=255>
Be sure to look at the physical size of these antennas. The higher
gain antennas are probably far too large to be practical.
>Does the
>use of adapters (i.e.SMA to TNC) hurt performance?
Adapters are not a problem. Someone did my favorite test of stringing
all the adapters he could find together and measuring the loss. See:
<http://802.11junk.com/jeffl/antennas/connector-loss/index.html>
The loss was about the same as an equivalent length of coax cable.
However, I've seen far too many mechanically crude adapters, that look
like they would fall apart without much provocation. If it looks like
junk, it probably is and should be avoided.
>If appropriate,
>I'd also like to know about good antennas for a PCI card and notebook
>cards that can handle a good antenna, too.
PCI cards have RP-SMA connectors, which make adding an external
antenna easy. The problem is that they usually come with tiny thin
RG-316 coax cable, that's quite lossy. The antenna gain makes up for
the loss, but the net overall system gain is usually nominal for the
typical desktop vertical. I suggest a panel, patch, or biquad antenna
pointed at the basement with no more than about 6ft of the tiny coax.
Otherwise, go to larger coax, such as LMR-240.
Notebook cards are a problem was many do not have antenna connectors.
I don't think your Hawkings card has one. It may be best to either
get a PCMCIA card with an external connector:
<http://www.buffalotech.com/products/wireless/wireless-g-125-high-speed/wireless-g-125-high-speed-notebook-adapter/>
There are pigtail adapters from this connector to the more common SMA,
TNC, or N connectors.
>While I may do file transfers down the road, in all likelihood, this
>connection will be almost entirely for a plain vanilla Internet
>connection, no video streaming or P2Ping or torrents, big file
>downloads will probably be as strenuous as it will get. While I'd
>like to get the signal better, maybe I'm gilding the lily?
>
>Finally, is there anything important I've left out?
>
>Any answers or links would be deeply appreciated. Thanks!
--
Jeff Liebermann
jeffl@comix.santa-cruz.ca.us
150 Felker St #D
http://www.LearnByDestroying.com
Santa Cruz CA 95060
http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
High gain antennas at both ends? 
Linear Mode
