On site now.
The AP's are up and running and have been for 22days, which eliminates the AP's having lost power. The error logs dont contain anything, however the time and date are incorrect (defaulted to 2001) despoite me setting them abpout a month ago, but on only the link in question!
The antennas are perfectly alligned, but I havent managed to investigate wherther water is getting in to the column as weather if great.
There was a furthr drop out on Monday for 2 hours even though the weather was dry and warm.
The camera on the link closest to the control centre has been fine all the time.
I will attach the screen shots when I figure out how to!!!
The links are -35 (8 jitter) and -37 (1 jitter) the second is -36 (1 jitter) and -38 (2 jitter).
I'd like to hear what Dave has to say about this but...
Try turning your power down. At -36 dBm these devices are screaming at each other. I'd like to see the signal somewhere in the mid to high -50's at best.
Thanks for your reply, however would that and the bad weather work against the link to cause the drop outs?
The three AP's that are in the same link are more or less in a perfect line to each other, so AP three talks to AP 2, but also clearly has AP 1 in its line of sight (200M away). It is possible that the rain has 'scrambled' their signals to AP 1 with AP 1 just dropping them all as AP 1 is the only one of the three that hasnt had any issues so far?
I think you need to stop looking at the weather thing, unless the RF or Ethernet connectors are getting wet. It isn't a signal fading problem with the rain.
I assume that you have the two links of different channels; make sure they are far away from each other on the spectrum. Not channel 36 and 40 because, especially at that high of power, they will interfere with each other.
How many devices are there? I thought this was two separate links, but is the middle AP a repeater? I'm guessing not since the two endpoints are so close.
If you look at it from a link budget standpoint, you could have - 60dBm and still do great even if it is pouring rain. Having such a high signal just doesn't sit well with me, although I have no facts to back up why.
we have two seperate 'daisy chain' links for CCTV. The one that is having the isse has three cameras on along the high street, hence the need for 3 AP's. The final AP goes to another building then to the police station. The other link is a few hundred meters away and had no issues, although this too has some short links of only 70M but no other AP diretly in its line of sight.
Thge reason I keep coming back to the weather is that they were fine for a couple of weeks until we had bad rain and hailstone.
Yes, I agree with Grant.
Without going into details here [ I'll put a post up some other time ], having too much power [ especially on a relatively short link ] can be just as much of a problem as not enough power.
When we send an 802.11 signal, it has strange little "things" or "wings" "attached" to the "normal" part of the signal.
These "things" are called "side-bands".
When we increase power on a transmitter, a phenomenon called "spectral regrowth" can occur, and these "wings" can become larger in amplitude compared with the "good" part of the signal.
They can result in many problems:
1. Interference with adjacent channels
2. Increase in noise floor level
3. Saturation of receiver front-end [ bit like screaming at the top of your lungs into someone's ear-hole. Your ear/brain would saturate and distort the information ]. This can also occur even if the "wings" are within limits by the way.
More is not always better.
Spectral re-growth is highly complex and mathematical, but there are "limits" to keep it in check. These limits can be seen diagramatically in the "masks" that we see in the CCNA study books [ and in tech docs ]. Provided we stay within these masks, the spectral re-growth problems will likely be contained. Big problem is that you either need to know how to apply these limits on a spectrum analyzer screen ot have a software mask built in. Few people do this.
Just as a point of interest, how are the cameras being powered ?
Here's a new twist: I think the problem might have to do with revenge from the ghosts of all the hoodies and other ASBO associated pond scum who have been caught on tape by these cameras in the past !!!!
The cameras are powered by mains at each CCTV column. It had crossed my mind about someone 'disturbing' them, but where they are think the local ASBO's are a bit lacking in the thought department...
Although I can see the point about the full power (which I will correct next week) I am still at a loss how the link failed during bad weather aftr two weeks, then had an issue in good weatger as well. s mentioned, there are no reorts in the error log. The only annomily is that the time ad date isn't correct despite me doing at point of install.
By the way [ for CWNA folks ], you can see examples of transmit spectral masks on:Page 207 of the new Sybex CWNA book, page 182 of the old one.
Well, finally got back up with the cherry picker and found that the gys hadn't glanded or grommeted the cable entry points on the two columns that were experiencing issues in bad weather. I have now weather sealed the entry points and it has started to rain here so lets hope we get a good down fall (besides taking the edge on the heat!) to see if the results are different. If all is well this now I will adjust the power next.
Thanks for your help so far