Hi Guys

When adding new links to your network, what is the best way to go about doing it?

Below is what I do, please share your thoughts and at the end I will edit my post with a set of "best practice" tips which we can sticky.

This is for a ptp link

1) Align both antennas
2) Try and get signal to as close as -55 as possible:
if > -55 then decreas TX power
if < -55 then increase.
Changing channels

3) try and get CCQ as close to 100/100 by a combination of the following:
Lowering data rate
Turning on/off adaptive noise immunity

4) try and get p-throughput as high as possible by:
Increasing data rates until CCQ drops
Turning on/off nstreme
Turning on/off adaptive noise immunity

5) Monitoring disconnects
If too frequent, then increase hw-retries in increments until max 15

I'm sure there are better steps than mine, if so please post below

Awesome this helps, anymore advice on wireless would be highly appreciated!

I might be wrong tho, so would appreciate if somebody could give their advice too.

well sir as far as i can see you have the best method....

I think you forgot the point 0) Planning
Planning a wireless link with Radio Mobile does for me the 90% of the duty before putting it on production. If Radio Mobile is correctly configured, you can find which signal level you should expect on that link and aim for it.
I'm not deploying any backbone ptp link with a bad fresnel zone.. No matter what signal i might get. I've a perfectly working -74 link and a average link with -57 which has some obstacles on the lower fresnel zone.

Hi Medianet

I have radio mobile installed but I must admit that I do not use it properly to plan links. I basically use it to check if there is LOS, either yes or no. I do not take fresnel zones etc into account.

How important is the fresnal zones etc? I agree, and sometimes noticed that I get better performace with a -70 compared to a -54 which probably explains my lack of understanding regarding fresnal zones.

Do you use specific antenna patterns for each link or do you use a default antenna design?

I use a directive antenna pattern 23dbi (usually i cover all links with a 23dbi)
Here the patterns.
Tune your Network properties to fit your requirements also.

Thanks will give it a look

See attached radio mobile path profile...it shows LOS but it is quite an unstable link...what do I need to look at, from the radio mobile path profile, to determine in advance if the link will not be good?

Attached is my network properties settings...

Anything you recommend I change?

Thanks

Look at the white lines just after the source point that are obstructed by the mountain (looks the coast of the mountain tbh). That's your issue. If there's some small tree on that coast it will make your connection fluctuate too..
I don't know what is the best solution to fix this for you... I would suggest to move the Source point a little forward towards the destination point. But might be impossible.
Your conf looks fine.

Thanks alot, that makes sense...Maybee its time to look for new highsite

How do you calculate expected signal level?

Just a note - don't increase the tx power above the default. This will eventually fry your card

Your antenna height can (not just freznel zone) be critical to a good stable link. Once you have created your link in path loss, move the antenna height up and down at each end and watch your signal strength change. You will notice that the is a "sweet spot" where the signal is peaked and does not change much for small height changes.

Now look at the second picture and you will not that the signal is now around -59 to -60, also note the lower antenna height at one end, note the differences on the where the signal strength lines now end. In a nut shell the first sample would likely not be a good stable link as it can "fall" in to a trough as the link conditions change and the line shifts due to atmospheric and ground conditions.

There is also a quick and dirty spread sheet that can be down loaded from the Proxim site that will do basic link calcs and provide you with a percentage up time of the link which is not provided in Radio mobile.

thanks 16mc, your sweet spot shows a signal of -77, is that the option that you are currently using?

Please see my attachment ( link4), I get about a -64 but the throughput is only about 2mbps FD before I get packet loss. Have I identified the sweet spots correctly?

Regarding link3, I get about 8mbps FD before packet loss ( 2xMIMO with SR7), would I get even better performace if the profile did not "dip"

thanks for the info

A bit of clarification. The First one with the sweet spot marked is a poor link at -77 as the sweet spot is not at the end of the link as in graphic # 2 where the signal is -60ish. So to move the sweet spot to your antenna you need to lower (or raise ) one of the antennas to get a graphic show in the second pic.

Your graph looks ok as there is no dips or bumps at the end, like in my 1st graph.

From what I have read 802.11N has some issues at present with going more the a few KM. I have seen a few posts elsewhere indicating that Ubquiti's Rocket gear has/had similar issues.

Oh so the sweetspot needs to be at the end of the profile...Is it irrelevent if there are bumps/dips half wah through the profile?

No, the bumps are normal, and will be different for every link and antenna height combination. You just want to avoid a dip at the end (at your antenna). You also want to avoid a sharp(steep) bump at the receiving antenna as it can cause fade problems when it "rolls" of the peak due to changes on the link(atmospheric etc.).

Got it thanks...Did you place your GPS directly underneath your antenna to get a pinpoint accurate reading?

At the moment I use google earth but apparantly it's not that accurate

I use both. Google earth is handy for getting positions but only if you have super Hi res of the area, if not it is a bit useless. If you have any terrain or buildings to deal with a good gps is well worth the cost as you can travel the path and find out if an obstruction is on the "line" or far enough to the side that it will not affect your link. There are many sources on for DEM(digital elevation Maps) that are available for free but in many case the resolution is to low, particularly for places out side of the US. We purchased some high res DEM for our area as I was not happy with the free coverage available.

Would buildings affect the radio mobile range profiles? ie would buildings shift the 'sweet spot'?

I think I am going to try the GPS solution next and compare it to google earth, if it really is that bad then I might consider purchasing some DEM aswell...thanks for the idea.

Is it quite expensive to get terrain data that has the buildings taken into account aswell?

Building and "clutter"(trees etc) data tends to be very expensive, and you really only need it for doing distribution as a any microwave p2p link should be LOS when ever possible or they can be come less predicable and reliable.

If you have a building you can calculate the Fresnel zone size at that location, then using the GPS you can tell if the the building is going to be a problem.... if it is to close you can have strange issues due to reflections and refraction....so yes it can be effected by buildings but the the software and data to do the calc are quite pricey, typical ranging from $2k to $20k just for the software! So a good gps and short trip are your best friend.

16mc

How far up and down are you taking about moving the antennas. Are we talking a few feet or yards?