It does not matter waht you think, it is a matter of playing with every settings to see what works and what does not.
Not the best wording. Perhaps what you really meant was:
We need to take a logical process of elimination approach. Please try xyz, that way we can eliminate abc.
"It does not matter what you think" is arrogant and confrontational
"it is a matter of playing..." implies no logic or professionality
Quote:
another thing is that i am using a 24dbi directional to a 15dbi omni.
So you've got 200mW gear at either end? You're talking about:
23dB + 24dB = 47dB and
23dB + 15dB = 38dB
The law says you cant radiate more than 4W (36d so welcome to illegality. Try knocking back your transmit power, you may be interfering with yourself through multipath
(BTW the above calculations are what's called a link budget, you're supposed to take off your losses as well such as connector/cable/attenuator loss which would bring the 15dB down a few dB to under 36dB, I was stressing a point, the 24dB antenna end would still be illegal though)
24dB's are made because FCC regulations allow it if the transmitter power is lower, that way you can balance it all out to get a higher EIRP. Note that this is the FCC's regulation (America) not the MED's regulation (New Zealand)
For your example: linksys + 24dBi you'll noticed that I defined the link budget a bit further in my previous post. Remember to take off your losses, so the linksys + 24dBi comes to 39dB? Let's assume connectors and the cabling takes off a conservative 3dB, now you're down to 36dB.
Tx in (dBm) + Antenna (dBi) - Loss (d = link budget
SO
15 + 24 - 3 = 36dB. You follow?
Let's give this a more practical application to NZ's regulations:
Let's say you have a device that does 17dB, and a 19dBi antenna. Your losses are 3dB (cable, connector insertion loss)
17 + 19 - 3 = 33dB. Great.
Now let's say the connection isnt quite good enough, you're on the fringe of the range. +3dB signifies a doubling of power right? That could sure help.. You've got two variables there, the Tx and the antenna. Let's say we want to try a higher rated antenna:
With a 24dB antenna it becomes
17 + 24 - 3 = 38dB. Hmm. But you could drop it down to 36dB by introducing 2dB of loss through a longer cable or an attenuator. That way you can get 36dB perfectly and bob's your uncle.
After flooding you with that, can you see why they make 24dB antennas now?
What is multipath? I'll only lead the horse to the water:
open up www.google.com and type in "wifi multipath" and take a look at the results
(not that I dont want to help, I've just been at work for 8 hours now and my OOS is flaring up. Easiest way to describe it is to imagine a ghosted image on your TV. That's multipath. In the wifi context, rebounded data is arriving after it has already arrived directly, this confuses the receiver. Much like ghosting on a tv blurs the image, making it less than desirable)
i was wondering this:
how far would you say the antenna should be mounted above a metal structure
on the 1 mile link mentioned above, i have that 24 about 1 ft off a metal bin and a 16dbi about 4 ft off a metal silo. but between that there is absolutly nothing it is perfect los, no trees or anything.
in the above setup, what kindof multipath issues do i have?
from what i read of the link you posted it said that 4w was the max for omndirectional. i was using grid antennas
You read wrongly, you were reading about the FCC (eg American) regulations. NZ's are summed up:
New Zealand 1W any mod, any usage - 4W freq hop & dig mod only
As wifi is freqency hopping technology, it has a 4W EIRP, which is irrespective of antenna type.
i was wondering this:
how far would you say the antenna should be mounted above a metal structure
on the 1 mile link mentioned above, i have that 24 about 1 ft off a metal bin and a 16dbi about 4 ft off a metal silo. but between that there is absolutly nothing it is perfect los, no trees or anything.
in the above setup, what kindof multipath issues do i have?
thankyou for your help so far too
I'm sorry, I'm from a stubborn metric upbringing. I can get the gist of some of it, but you'll have to talk to me in terms of metres and kilometres.
Your concern about metal structures is related to the Farraday effect, however it really shouldnt be a problem. As for multipath, signal can bounce off a lot of things.. it could skim the earth's surface for example and shoot back up.
Grid antennas have a remarkable coverage pattern, they're supposed to be quite tight but I can pick up AP's behind my one, one that is 45 degrees off and about 2km away for example!
My 19dB is also picking up a cafe's network on the other side of wellington harbour, at least 10km away
mighty short link you've got there, what you've got is completely overkill. The easiest way to imagine it is two people face to face screaming at each other when all they need to do is talk
Please you two, just talk, and you ma'am, you need to ditch the hero and get yourself a zero, like me!
You only need to shout to one another when you're a considerable distance apart. What you may be having is the digital equivalent of ringing/bleeding ears
Though that still doesnt explain why it wasnt happening with the .b gear :? :?
well all i had at the time of setup was a 16 and some 24s i could have gone 24 to 24 but instead i went 16 to 24 figuring it would be overkill like you said.
Sorry about the "It does not matter what you think" Not the best advice.
Are you able to get pings through the Wireless if you have both units next to each other in the same room?
The other thing to do is to setup an AP at one end and measure the signal at the other end using Netstumbler or a Wireless Client of some type using the antenna.
i flashed the firmware, and when i went up there to reset, my antenna was off. i readjusted it and now i can connect again.
still for some reason when i try to ping in ad-hoc it will not ping through, but when i connect it to an ap right next to(which is using an omni) i can ping everything.
i did 1000 pings and lost 4 out of 1000, this seem good?
the problem with many AP's is that when they say "bridge" they really mean point to point bridge (eg the other end has to be setup the same), not point to multipoint bridge (which is what we want)
If it were a case of just point to point, then it'd be easy, but because you want to do point to mulitpoint off your omni, life gets a bit more fun.
This is where a dedicated bridge device helps - they're transparent, just a set and forget process.
I wonder if a Linksys WET54G could be modded up to 200mW like the old WET11 v1's could.. hmmm
[internet, router, rest of network] to [ptp bridge] wireless [ptp bridge] to
[switch] to [802.11b AP] wireless ((clients))
all the hardware is in place for this to work, i just cant get my 2 ptp bridges to connect, so right now i bypassed that and am connected tot he .b AP just so that something is going now.
if only i got that ptp bridge setup working i think i would be good.
okay the way I understand it is that on each you have to set them up as a point to point bridge, and the setup will ask you for a remote MAC
Let's say AP1 's MAC is 00-0F-20-FE-16-D7
and AP2's is 00-0F-20-FE-16-D8
On AP1, you'd setup the remote MAC address as 00-0F-20-FE-16-D8 (eg AP2's MAC) and
On AP2, you'd setup the remote MAC address as 00-0F-20-FE-16-D7 (eg AP1's MAC)
ya i tried that, still says it connects good, but no ping throughs. since i reaimed the antenna, it now says i have 81% even though that is a meaningless number. setting the bssid as the other points mac does not be seeming to work.
how many aps in the area would be too many for a specific channel?
I'm inclined now to say bugger it... setup the end with the omni as an AP and the end with the directional as a point to multipoint bridge client. drop the output power at both ends right down and bring them up
the other thing to do is to setup one end with netstumbler and let us know what kind of -dBm rating you're getting... you're aiming for -60dBm or better generally...
ia m working on it again, because i got it up as a ptmp to an ap(ap has a directional on it)
the snr is ~55db. when i use a 15dbm radio. so on the senaos 21 mw it would be about 60db snr.
the signal is ~40db and the noise is ~100db.
this should be the perfect setup.
now it only puts through about 500kbps (kilobits) of bandwidth, this is absolutely unacceptable for what i am doing.
it is so annoying!!
when i plugged in my laptop to the one end (as opposed to the senao(it was running . i had no packet loss out of 1000.
i also tested the ethernet cable going into the client bridge and it is just fine as well.
i have completely exhausted everything i know about wireless. all i have left to do i shoot that shit that senao calls equipment. id go and just setup computers, but the winters here are cold, and i dont want to worry about hd failrues or the like. i have to get this stuff working. but i cannot no matter what i do. it is the most annoying thing ever. anyone have any idea at all on this??
Originally posted by variable@Jan 20 2005, 07:05 PM for some reason my senao 3054 ap will randomly drop the connection, or it will just have massive packet loss.
this did not happen before when i was using senao 2511 .b equipment, i was just wondering if this was a common problem with senao equipment or what.
another thing is that i am using a 24dbi directional to a 15dbi omni.
thanks for anyhelp.
[snapback]4381[/snapback]
It could be the fact that you are using an external entenna. This issue is resolved by upgrading the firmware to the latest edition. Try Netgates site for that stuff
senao 3054 cb3 plus deluxe problems 
so welcome to illegality. Try knocking back your transmit power, you may be interfering with yourself through multipath
Linear Mode
