3000' wireless bridge performance

I have been using a pair of DLink DWL-900+'s configured as wireless
bridges with converted TV "downverter" wire grid dishes to connect
with my "host" 2950' distant for internet service. Best performance
is about 1Mbps. My host has up graded its bandwidth to 2Mbps but my
bridge won't support it. It has been suggested that my antenna is
about 15Db gain. My question is, is it reasonable and practical to
get better bandwidth out of my system or should I just be content with
what I have? My only options, I think, are to increase antenna gain,
or increase RF power. The D Links have been occassionally been
problematic... and stop working and / or won't reset. Any suggestions
of alternative equipment or hearing about experiences others are
having would be greatly appreciated.

"Paul Yarnall" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed) om...
> I have been using a pair of DLink DWL-900+'s configured as wireless
> bridges with converted TV "downverter" wire grid dishes to connect
> with my "host" 2950' distant for internet service. Best performance
> is about 1Mbps. My host has up graded its bandwidth to 2Mbps but my
> bridge won't support it. It has been suggested that my antenna is
> about 15Db gain. My question is, is it reasonable and practical to
> get better bandwidth out of my system or should I just be content with
> what I have? My only options, I think, are to increase antenna gain,
> or increase RF power. The D Links have been occassionally been
> problematic... and stop working and / or won't reset. Any suggestions
> of alternative equipment or hearing about experiences others are
> having would be greatly appreciated.

Who suggested your antenna is about 15 dBi gain, and on what did they base
that suggestion? If the suggestion is based on the surface area of the
reflector, then it's likely that the 15 dBi figure represents the maximum
possible gain from the reflector if it's optimally fed. What are you using
to feed the antenna? To optimally feed the antenna, the feed should have a
3 dB beamwidth angle that exactly illuminates the edges of the reflector
from the focus and the sides of the feed beam should have a steep drop-off.
If you're using a cantenna to feed the reflector, the opening of the can
should be at the focus.

The DLink DWL-900+ is a little underpowered, but with a good antenna and
short transmission line of high quality it should be more than adequate for
full speed at this distance. What is the length of your transmission line
and what kind of line are you using? The DWL-900+ should be positioned AT
THE ANTENNA to reduce transmission line length. Going straight from the
900+ to the antenna with 2' of LMR195 should give you good results.

Commercial antennas with 24dBi gain are inexpensive and readily available.
Even if your antenna has 15 dBi of gain, 9 dBi additional gain would
increase available signal at both ends of the link by 700%!

Go here and click on 'Antennas Parabolic', then scroll down to the 24 dBi
antenna for $69:
http://www.fab-corp.com/index.htm

....or here for $65:
http://www.hyperlinktech.com/web/hg2424g.php

On advantage of this antenna is the large increase in signal. Other
advantages are reduced interference from 'off-beam' sources and increased
security because your signal is only available to others in line with the
ISP. This also means that aiming the antenna will be a little more
critical, but at 3000' it shouldn't be too difficult.

Don W.

"Paul Yarnall" <(E-Mail Removed)> wrote...
> I have been using a pair of DLink DWL-900+'s configured as wireless
> bridges with converted TV "downverter" wire grid dishes to connect
> with my "host" 2950' distant for internet service. Best performance
> is about 1Mbps. My host has up graded its bandwidth to 2Mbps but my
> bridge won't support it. It has been suggested that my antenna is
> about 15Db gain. My question is, is it reasonable and practical to
> get better bandwidth out of my system or should I just be content with
> what I have? My only options, I think, are to increase antenna gain,
> or increase RF power. The D Links have been occassionally been
> problematic... and stop working and / or won't reset. Any suggestions
> of alternative equipment or hearing about experiences others are
> having would be greatly appreciated.

TeeVee antenna's are designed to be terminated at 300 ohm, the 900+ expects
50 ohm. A good portion of the signal is being lost over the wrong
termination impedance and more than likely high loss cable between the
two...

The TeeVee antenna (UHF) has a frequency that tops out at 800 mhz, the 900+
expects 2.4 ghz. More than likely a good portion of the signal is going
through the antenna due to the higher wavelength.
The 900+ can be configured for upto 22 mbps service, with a good antenna and
signal...

When you get the right cable and antenna, be sure to configure the 900+ to
use the RIGHT or external antenna connection and not the LEFT or internal
antenna or DIVERSITY setup...

If the AP you are connecting with doesn't support D-LINK stuff directly, ie.
that it is too, then also turn off the "4X Mode"... Both of the above are in
the ADVANCED and PERFORMANCE settings of the 900+, if you are running the
later versions of firmware...

"Darrel Toepfer" <(E-Mail Removed)> wrote in message news:<nmc0b.14634$(E-Mail Removed)> ...
> "Paul Yarnall" <(E-Mail Removed)> wrote...
> > I have been using a pair of DLink DWL-900+'s configured as wireless
> > bridges with converted TV "downverter" wire grid dishes to connect
> > with my "host" 2950' distant for internet service. Best performance
> > is about 1Mbps. My host has up graded its bandwidth to 2Mbps but my
> > bridge won't support it. It has been suggested that my antenna is
> > about 15Db gain. My question is, is it reasonable and practical to
> > get better bandwidth out of my system or should I just be content with
> > what I have? My only options, I think, are to increase antenna gain,
> > or increase RF power. The D Links have been occassionally been
> > problematic... and stop working and / or won't reset. Any suggestions
> > of alternative equipment or hearing about experiences others are
> > having would be greatly appreciated.
>
> TeeVee antenna's are designed to be terminated at 300 ohm, the 900+ expects
> 50 ohm. A good portion of the signal is being lost over the wrong
> termination impedance and more than likely high loss cable between the
> two...
>
> The TeeVee antenna (UHF) has a frequency that tops out at 800 mhz, the 900+
> expects 2.4 ghz. More than likely a good portion of the signal is going
> through the antenna due to the higher wavelength.
> The 900+ can be configured for upto 22 mbps service, with a good antenna and
> signal...
>
> When you get the right cable and antenna, be sure to configure the 900+ to
> use the RIGHT or external antenna connection and not the LEFT or internal
> antenna or DIVERSITY setup...
>
> If the AP you are connecting with doesn't support D-LINK stuff directly, ie.
> that it is too, then also turn off the "4X Mode"... Both of the above are in
> the ADVANCED and PERFORMANCE settings of the 900+, if you are running the
> later versions of firmware...

Thanks Guys for your feedback. I did modify the "TV" antenna head
from its original form. I was advised that the dipole on the circuit
board of the convertor was in fact intended for the 2.4 gig range, and
the circuitry on the board converted the signal into one of the TV VHF
frequencies, so what I did was cut the traces from the board that
connected the dipole to the rest of the stuff and soldered on the
special hi frequency low loss coax to the dipole traces. The
reflector is 14x23"... looks just like the units you referred me to
and at least based on its size should be around 19Db assuming my head
is functioning well, but that is questionable. The coax on one unit
is about 4' long, the other is about 12' long. What I hear you both
saying is that with a better antenna I should be able to get the full
capability of the DLinks bandwidth (which is more than I need)over a
3000' distance. I didn't want to waste more time & $ for a marginal
gain in performance... the antennas you mentioned are affordable.

----- Original Email Message -----
From: "Paul Yarnall" <(E-Mail Removed)>
To: "Don W." <(E-Mail Removed)>; "Darrel"
Sent: Tuesday, August 19, 2003 5:55 PM
Subject: 3000' bridge

Hi Don & Darrel,

I hope I am not breaking any etiquette by emailing you guys directly... if
so I apologize, just let me know... I am new to the user group
thing. Anyhow I have attached a photo of the "downverter" modified to use
only the dipole portion of the circuit to which I attached the low loss
coax. I have no way to measure the SWR or gain of the antenna but there
was an attempt to match the 50 ohm and frequency. I am not using the
latest firmware apparently because there are no antenna choices. The
reflector is 14x23"... looks just like the units you referred me to and at
least based on its size should be around 19Db assuming my head is
functioning well, but that is questionable. The coax on one unit is about
4' long, the other is about 12' long. What I hear you both saying is that
with a better antenna I should be able to get the full capability of the
DLinks bandwidth (which is more than I need) over a 3000' distance. I
didn't want to waste more time & $ for a marginal gain in
performance... the antennas you mentioned are affordable. My box only has
the one stubby antenna and one of the reasons I chose the 900 was that the
antenna was removable so I could connect the remote antennas. My bridge
has been in service for over a year now. When I started I could not find
affordable antennas so I made my own. If you guys think that buying the
ready to go 24 Db antennas should do the trick, that is the way I will go.

Do you have any opinions on superior boxes as an alternative to the
900s? They are great when they are working, but I had to send two back
shortly after setting up the bridge because they would not reset and I
replaced another one last week that stopped working. I have bought 5 of
these things so far. (DLink replaced the two but I could not wait for
them... bridge was mission critical). Every time one goes down I have to
go to the other end to reprogram it for the new MAC address of the other
one. Bit of a hassle. The most recent failure I think was due to an RF
spike from a lightning strike which was not very close but directly in
sight of the failed unit antenna. Didn't hurt anything else on the network
so that is my theory.

Thanks again for you comments.... no one else in my area is doing anything
like this so there aren't many prospects for getting meaningful advice.

Paul J. Yarnall
Technical Resolutions
5711 Thomas Rd.
Canandaigua, NY 14424
(585) 394-8651
(585) 394-8375 FAX
---

I don't do email replies, comments need to stay here in the newsgroup due to
the fact that it may help someone else down the road who looks at the
archives of this newsgroup...

Chances are the units failed due to the severe mismatch on the antenna. I'm
guessing at this, but for that many units to fail, you have to look at what
is common to all of them, and that is your strange antenna...

Anything 14db or more should work fine, perhaps even something less, 24db
would be overkill...

As I stated before, look at your firmware to see that you are only using the
"RIGHT" antenna which is the external one...

> Reads like you have a 2.5ghz wireless cabletv antenna then, it would still

Isn't that an oxymoron? "wireless cable"

"David Taylor" <(E-Mail Removed)> wrote...
> > Reads like you have a 2.5ghz wireless cabletv antenna then, it would
still
>
> Isn't that an oxymoron? "wireless cable"
>
>

Actually its called MMDS, and is FCC licensed...

"David Taylor" <(E-Mail Removed)> wrote...
> > > Isn't that an oxymoron? "wireless cable"
> > >
> > >
> >
> > Actually its called MMDS, and is FCC licensed...
>
> I think you might have missed the humour!

I caught that, just missed what that type of service is called the first
time I mentioned it... Like the FCC has anything to say about what happens
in England... <g>

"Paul Yarnall" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed) om...
> "Darrel Toepfer" <(E-Mail Removed)> wrote in message
news:<hjA0b.165$(E-Mail Removed)>.. .
> > Reads like you have a 2.5ghz wireless cabletv antenna then, it would
still
> > be mismatched frequency wise and could very well have caused the
failures of
> > the radios, especially with the "downconverter board" in place.
Typically
> > those need a power inserter to feed power to the downconverter...
>
> Since it has been a year + since I got the antennas and did the mods I
> don't recall what the downverter was designed for... I thought it was
> 2.4 ghz, but maybe it is 2.5. Anyhow the "downverter" portion of the
> feed is completely disabled. The dipole antennas were "surgically"
> separated from the rest of the circuits and my coax attaches directly
> to the dipoles. I uploaded the newest firmware and noted the antenna
> choices. One of my problems with the Dlinks has been that I have
> managed to get "locked out" on several occassions, especially after
> the firmware changes and resets. Tech support got me back in by
> clearing out browser histories and security settings... be nice if
> those tricks were published somewhere... so my other unit that I
> thought might be blown is not. My spares are ready again. Sorry to
> ramble. I wanted to mention that today I added 1/4" screen to my
> reflector over the wire grid. The Dlinks now seem to support the
> "auto" speed setting and my last speed check is now 1724/323. That's
> good.
>
> It would be nice to have a way for measuring signal strength when
> setting up and aiming the dishes. How do you guys manage this task
> when you are bridging over many kilometers? I am doing eyeball
> sighting down the feed stalk at 3000' and with the beam angles on my
> antennas that is probably OK, but I would love to have a way to
> measure signal strength in software, or a cheap instrument. Any
> ideas?

If you like to build stuff, Analog Devices offers an evaluation board for
the AD8313 Logarithmic Amplifier/Detector for $100 with SMA connectors for
input and output. It will respond accurately to signals from a little
weaker than -70 dB up to about -10 dB with an analog output voltage from
about 0.5VDC to around 3.5VDC. The AD8313 will just barely drive a 1mA f.s.
meter movement, but you should buffer it through an op amp to drive the
meter. Or if you have a good multimeter you could connect that to the
output. Then the problem is how to keep your signal source transmitting!

Don W.