harvSki
Premium
join:2004-03-09
Suffolk, UK
| pop goes my fresh AP
In the continuing saga of problems (»Protecting AP / Omni) I'm having with one Access Point / Omni combination where the access point keeps going bad, this time I caught the event "on film" as it were.
The image shows a gap from 22:30 (10:30pm) to midnight when it went bad an no clients could connect (I had complaints). The information is coming from SNMP on a client of the AP, the continuous grey line at the top of the graphs is another AP/Client in a different village for comparison.
This is costing me a fortune now  with the only consolation at the moment that the clients are still seeing the AP OK(ish)
Any thoughts? |
|
superdog
I Need A Drink
Premium,MVM
join:2001-07-13
Lebanon, PA
| Can You make a drwing of this Pop(network layout), right down to what every wire is touching(even grounds)etc.?
--
»www.wavecrazy.net Join WISPA today! »www.wispa.org/ |
|
harvSki
Premium
join:2004-03-09
Suffolk, UK
| reply to harvSki
OK, I've managed to knock this diagram up in my lunch break, I've also discovered I can zoom into the graph on PRTG and notice that the client device where this data is coming from saw a signal drop just before losing contact with the AP (Osbridge). I also checked the uptime and no devices stopped running at all and the MikorTik in the diagram also stayed in contact with the wired side of the AP too.
remember the blue/grey line at the top of the graph is for ref from another system in another village. |
|
superdog
I Need A Drink
Premium,MVM
join:2001-07-13
Lebanon, PA
| OK, what other devices are on the roof?. Do any of Your CAT5 cables run very close to any electrical wires. Is the OSbridge AP in a plastic enclosure by any chance?
--
»www.wavecrazy.net Join WISPA today! »www.wispa.org/ |
|
harvSki
Premium
join:2004-03-09
Suffolk, UK
| reply to harvSki
On the roof there is - no power, but a taxi firm has a PMR omni about 3 metres long not sure what frequency but it could be 854Mhz as it has that written on the omni, seems odd though as I would have thought it would be VHF / UHF
The Osbridge is in a plastic box mounted onto the aluminium pole the omni is on. I 95% sure that the metal bracket that the Osbridge has only bolts onto the plastic of the box (no back plate)
The WR54G is in a custom water-tight plastic box with a metal back plate which is all connected up to the pole with the parabolic via a metal bracket.
Non of the antenna on the roof are earthed, neither mine nor the PMR.
I have 3 other osbridges mounted in a very similar way that have not shown this problem, they have all been in use longer too, (nearly 2 years)
It has been suggested to me that it could be an SWR problem on the omni as it is the only one of that type I have. I've orderd another one from my usual source with more believable gain figures too. »www.stelladoradus.com 10dbi 2.4gh omni. |
|
superdog
I Need A Drink
Premium,MVM
join:2001-07-13
Lebanon, PA
| said by harvSki  :The Osbridge is in a plastic box mounted onto the aluminium pole the omni is on. I 95% sure that the metal bracket that the Osbridge has only bolts onto the plastic of the box (no back plate) As learned by Caleb and Hal at Deliberant, You need to ground that AP. Wind over plastic creates static electricity, and I am willing to bet that this is where Your problem lies.:D
--
»www.wavecrazy.net Join WISPA today! »www.wispa.org/ |
|
harvSki
Premium
join:2004-03-09
Suffolk, UK
| reply to harvSki
Thanks superdog, that does seem somehow plausible, here are the weather readings taken from about 20Km away, there is some wind about yesterday but not a huge amount and thinks did change around 22:00.
This osbridge is probably the most exposed one I have so I maybe protect it (or a new one) by putting it lower on the pole so it is not directly in the air stream.... that is how most the others are mounted and they have never given me any problems.... |
|
wolfcreek
join:2003-12-02
Pagosa Springs, CO
| reply to harvSki
We have seen static damge even on non windy days. During certain snow or rain events it seems that the falling precipitation can cause a static buildup in the atmosphere. We were not losing radios but the switches after the radio. Soon as we put on a Cat5 surge suppresor and all was fixed. |
|
lutful
Premium
join:2005-06-16
Ottawa, ON
| reply to harvSki
said by harvSki :a signal drop just before losing contact with the AP (Osbridge). Several high power CPEs with 24dB grid very close to the omni could be saturating the receiver - many CMOS chipsets will fail with just 1 milliwatt total power input.
You really need to measure the maximum 2.4Ghz RF power that is being "seen" at the LMR400 end of the 15dB omni.
Replacing APs after APs will cost you more and more  |
|
harvSki
Premium
join:2004-03-09
Suffolk, UK
| I will try and measure the RF at the end of the LMR400, the 24db grid is 5 metres away from the Omni and all the clients are ETSI standard 20dbi EIRP max and hundreds of meters away.
What has me really puzzled is the hour and a half gap in the data, I didn't lose contact with the access point at this time but the clients disconnected. If it was static does this mean it lasted 1.5hrs and then was gone? |
|
lutful
Premium
join:2005-06-16
Ottawa, ON
| said by harvSki :What has me really puzzled is the hour and a half gap in the data, I didn't lose contact with the access point at this time but the clients disconnected. If it was static does this mean it lasted 1.5hrs and then was gone? Our experience with static is that LAN port dies earlier than RF port. Also, static discharge is momentary.
However, if the AP was exposed to more than 1mW RF input even for a short period it is possible for all clients to lose connection since the driver could have locked up.
IMHO, the "more than 1mW RF input" theory needs to be confirmed or eliminated at your leisure. |
|
harvSki
Premium
join:2004-03-09
Suffolk, UK
| reply to harvSki
UPDATE
Last Saturday I took a look at the graph and saw this graph, as you can see at about 4am the interference stopped and the radio system returned to normal 
I have now also taken steps to put some static earthing on to the access point to protect it that way too.
I am writing this 4 days since the problem disappeared and still none the wiser what was happening, although it is nice that the AP had not died after all.
Thanks for all your help
Harvey |
|
harvSki
Premium
join:2004-03-09
Suffolk, UK | reply to harvSki
and the Taxi radio antenna says 84.5 not 845! |
|
Semaphore
Premium
join:2003-11-18
Arnprior On. | Any chance it's the omni, or connectors, and not the AP that caused this? Have you tried cranking the power down by 3-5 db when this interference is happening ? It may be a standing wave ?
--
www.gozoom.ca |
|
John Galt
Forward, March
Premium
join:2004-09-30
Happy Camp | reply to harvSki
Did it start raining?
--
A is A |
|
harvSki
Premium
join:2004-03-09
Suffolk, UK
| reply to harvSki
The power coming out from the current AP is supposed to be 10db and I cannot turn it down, the AP that "died" was adjustable and by the time I pulled it out it was set to max power.... After logging all this data I'm now not sure if the original AP died at all and if I get a chance I am going to put it back in for 30mins and see what it does.
The weather on the day things recovered was v light wind, -1C and dark clear skies, sunrise is about 8:00am I think and the signal improved at 4am. |
|
