Chipset Choice (p2)

Trying to decide between two radio bridges for client mode. Your opinion?

Both radios are from same vendor.

Model X has conexant / intersil Prism 2.5 chipset but has had occasional firmware and other complaints from users. Said by one company rep. to be better in range than model Y. 200mw transmit output, 95dBm receive at

1Mbs

Model Y has Realtek RTL8186 SoC chipset and is new to market, firmware either new or recently revised. 400mw transmit power, 92dBm receive at

1Mbs.

Both have 8dbi integrated panel antennas 35 degree horizontal and vertical spread.

My main AP is 2/5 mile away, near line of sight, 90 degrees to the side of the primary vector of the signal which is pointed in a downward tilt to cover an immediate two block area. My line of sight window is about

45 degrees vertical by 100 degrees horizontal, so my guess is I am capturing only the portion of the signal at right angle to it and within that window. Several other AP choices at further distance to be determined if I ever get it mast mounted with a rotator.

So far I have had to forget an 85mw radio that could only get a stable connection about 20% of the time with a small whip antenna (5dbi?), due my guess to some waving tree branches that partly obstruct. What is your best guess as to which model, X or Y I should try next? In talking to one mfg. rep. he states firmware is better on model Y, but this is a new model with a fairly new chipset. I also need something that will be compatible with a wide range of OS's and utility software, which speaks better for model X (prism). What do you think?

Reply to
JDavidson
Loading thread data ...

JDavidson hath wroth:

I'm not going to offer any opinions on the quality of the devices without knowing the names and models. However, I can offer some clues as to whether your antenna scheme is going to work. Methinks not.

An 8dBi panel antenna has a -3dB beamwidth of about 60 degrees. At 90 degrees from the center line, the pattern appears to be about -40dB down. That might work if your coverage area was close enough.

I can't decode what you're doing. I have no idea what a "main AP" is or what manner of hardware it implied. I don't even know what the selection of bridge radios has to do with this AP. Is this a seperate question or part of the bridge?

For fun, lets see if 8dBi antennas will even work with your prospective bridge radios. See the FAQ at: |

formatting link
in numbers for the Model X bridge, I get: TX power = +23 dBm TX coax loss = 1 dB (mostly internal connectors) TX ant gain = 8dBi Distance = 0.4 miles RX ant gain = 8dBi RX coax loss = 1 dBi RX sens = -95 dBm (at 1 Mbits/sec) Fade margin = unknown Plugging into:
formatting link
get a fade margin of 35.8dB. No problem as anything over 20dB fade margin will work well.

If you really only want 1Mbit/sec connection for about a 300kbit/sec thruput, you can have it. Incidentally, I believe the -95dBm sensitivity at 1Mbit/sec is baloney. It should be around -89dBm.

I think you'll find that higher speeds are more useful. Using my numbers from the above URL (because you didn't specify the maker and model so I can't look them up), the fastest connection speed possible, for a 20dB fade margin, will be about 24 Mbits/sec for a thruput of about 12 Mbits/sec. Good enough.

Run the numbers yourself for Model Y for practice.

Please note that the above calculations are the best case. It only gets worse. You casually mumbled something about tree branches waving in the wind, which implies that you have a tree in the way. I suggest you calculate your Fresnel Zone at:

formatting link
should have 0.8*Fresnel zone clearance at midpoint. For 0.4 miles, you need a 12 foot radius about the center line clearance, which includes trees, buildings, and the ground.

Incidentally, the stock rubber ducky is not 5dBi gain. It's about

2dBi at best.
Reply to
Jeff Liebermann

Jeff Liebermann hath wroth:

Oh, you're using the bridge radios as client radios. I thought you were building a point to point bridge with two of them. That make my point to point link cals kinda worthless. Substitute the real numbers for your access point and try again. However, since you're well outside the major lobe of the antenna on one end, the antenna gain should be from the pattern graphs for your angle of depression. That's going to be far less than the maximum antenna gain.

Incidentally, I have the same problem at home. I live on the side of a hill. The clients are all downhill from me at about -45 degrees. I have to use multiple antennas to cover them all.

My points about the fresnel zone still apply.

Reply to
Jeff Liebermann

I sent you a detailed answer to your previous to this reply, but it was either cancelled or does not appear on my servers, so you probably did not receive it; not sure you'll get this one either.

When you are trying to receive an AP signal approx 70-90 degrees off it's central vector (straight ahead from a sector antenna), is there a method for estimating the amount of signal strength loss when using a panel antenna to receive? Can you give a simplified explaination of how this is done using those antenna signal spread graphs I see all the time?

Also, once these outdoor radios are on a mast, if they screw up and you have to reset them, then you must have remote reset, or else you have to climb the pole so to speak?

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

Reply to
JDavidson

JDavidson hath wroth:

Nothing appeared on the sbcglobal.net and newsguy.com. I can check others but those are the one's I read.

Very easy. Let's take a typcial real sector antenna pattern:

formatting link
is a 2.4GHz 120 degree 12.2dBi gain. Horizontal pattern is at:
formatting link
degrees off the center line shows a gain of -5dB below the maximum gain (outer) circle. Therefore, it has the equivalent of a: 12.2 - 5 = 7.2dBi gain. Similarly, at 90 degrees off axis, the gain is -9dB below maximum or: 12.2 - 9 = 3.2dbi gain. Incidentally, the 120 degree horizontal beamwidth is based upon the -3dB points.

Note that the horizontal gain pattern on this sector antenna is very different from the vertical gain pattern. It's quite broad on the horizontal (120 degrees), but very narrow vertically (14 degrees). That means it won't tolerate much in the way of differences in elevation between the end points of a link. That's also why sector antennas and panels usually are mounted with some down-tilt.

Well, the easiest way to do a reset is to cycle the power. Unless you have a solar powered radio, with a gel battery on top of the pole, you should have access to the power source, especially with PoE. It should be easy to power cycle one of those.

If the radio is remotely located, a drive to the mountain radio site is no fun, especially in winter. I have a few pager receivers and VHF/UHF radios setup to trigger a relay or solid state switch to cycle the power on mountain top installs. Of course, the one and only remote weather station where I do NOT have a remote reset system is the one than seems to hang once a week. Grrrrr....

Reply to
Jeff Liebermann

Thanks much, Jeff. I am glad your on this group. The post musta been cancelled by some rogue A**hole. It appeared and then disappeared, even from the server it was posted to. I have a pretty good idea what I'm up against now, thanks to you. I will post later if I get stuck.

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

Reply to
JDavidson

On Fri, 7 Jul 2006 19:20:36 +0000 (UTC), JDavidson wrote in :

I very much doubt that -- most newservers don't honor cancels, and it doesn't appear on any of the major services. More likely some issue or problem with your news provider. Usenet propagation is far from perfect

-- a small but significant percentage of posts simply get lost.

Reply to
John Navas

JDavidson hath wroth:

Nope. It's not a cancelbot. You're apparently posting through aioe.org which declares it's policy on control messages as:

formatting link
Aioe.org server doesn?t accept neither Control Messages nor Cancel Requests.

As John memtioned, some news messages simply disappear. I switch back an forth between Newsguy and SBCglobal.net news servers. It's not unusual for me to see messages on one, that don't appear on the other. It's also not unusual to have a few messages magically appear many days after they were posted. Also, some of my postings have disappeared, only to be listed as spam by Newsguy.

Reply to
Jeff Liebermann

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.