At 20 meter a RB500 ( came with 2.9.11, now 2.9.20 ) in mikrotik outdoor box. ( with 3xCM9 radio )
At 21 meter a18dbi panel 5hz for backbone
At 23 meters 1 omni 9dbi 5Ghz and a omni 12dbi 2,4Ghz mounted to pole ( diameter 50mm)
first time ;
the tower was powered up the 2,4Ghz radio worked for a couple of days and died on wireless side.
replaced radio, measured the cables , nothing found
Second time ;
After a couple of days again the 2,4Ghz radio died on wireless side.
replaced radio, measured the cables , replaced pigtail, placed radio in other slot on router board.
Third time ;
After a couple of hours checked, and yes again the 2,4 radio died on wireless side
replaced radio, replaced antenna, replaced cable, grounded the tower and mikrotik box.
Fourth time ;
Within a hour, again the radio on 2,4Ghz died on wireless side.
I’m out of ideas , how can give me some hints
Could it be that the 9db 5Ghz blows away the 2,4 radio ?, as the Omni’s are next to each other ?
Could it be a software issue ? witch disables the wireless side ?
It is a common problem that the pigtail u.fl connector sleeve can touch a capacitor and short the radio – and it dissappears from the PCI bus. Check this.
Also if the board really overheats, usually the PCI bus dissappears first. Maybe add some ventilation holes if you think it is getting really hot. Or replace the RB500 board to see if there is simply a defect with that board. If you are overclocking to 330MHz, then you should go back to 266MHz.
The CM9 cards are original Wistron , last one i put in was a clone 5212
I use these types in all my towers , never had any prblem with clone or original cards, but to be sure it isnt a bad series of CM9’s i placed a older 5212 the last time wich i used in my laptop so i knew it was a good one.
i will check if pigtailsleeve touches the circuit, but don’t think its the problem as the cards works further fine, only the wireless part dies.
( i checked dead cards in antother system to be sure )
Could overclocking ( the board is indeed running at 330Mhz ) damage wireless part of a radio card ?
I think i may go for the last option you mentioned tully, replace the whole board.
only strange thing is that other two cards are running fine for days now,
Well think CM9 number 5 has died on wireless, still works a little (very weak signals on a scan from all AP’s in range )
before i put in number 5 i changed position of omni 5Ghz to a lower point in the tower, so 2,4ghz and 5ghz omni are not in opposit of eachother anymore, also i discharched myself 3 times to the tower just to be suree
What I was getting at with the grounding is the possibility that with more than one ground, you can have a differential between your grounds, which can cause voltage and equipment failure. Try to isolate all of your grounds to one point only (this may require you isolate your ap from the tower and it’s ground ). With this many fried cards, you either have a bad board (which is very unlikely with this kind of problem), a bad power supply and/or ethernet/power run, or you’ve got a ground loop. I bet it’s your ground. Does this board with radios installed work fine when tested on the bench?
found a quote to put in this topic, could be useful to anyone
Another point, MT boards have isolated onboard power supplies (very good practice) which means earth-potential-differences indoor to outdoor aren’t a problem. Not true on some other boards, where you can get “earth loops” or worse if you ground the outdoor earth/antenna - which you should do of course.
The CM9 cards are original Wistron , last one i put in was a clone 5212
I use these types in all my towers , never had any prblem with clone or original cards, but to be sure it isnt a bad series of CM9’s i placed a older 5212 the last time wich i used in my laptop so i knew it was a good one.
i will check if pigtailsleeve touches the circuit, but don’t think its the problem as the cards works further fine, only the wireless part dies.
( i checked dead cards in antother system to be sure )
Could overclocking ( the board is indeed running at 330Mhz ) damage wireless part of a radio card ?
I think i may go for the last option you mentioned tully, replace the whole board.
only strange thing is that other two cards are running fine for days now,
well we replaced RB500 with doughterboard and the broken CM9
hope it goes well now. ( running for couple of hours now )
also we measured with a microAmpere meter if there are any strange voltage streams, and INDEED there was some between tower and ground.
dont know if it did broke the CM9 several times, but now we have isolated everything from the tower, and measured between antenneground and tower/ground to check isolation.
from now on , we ISOLATE antenna and check ALWAYS with micro-amp/volt meter.
COAX/connectors we always check with Ohm megger, also a nice way to see if moisture has came into the connector.
will put the RB500 board on the bench for testing do, keep you updated.
But I thing that you should check your power supply, I bet that it`s the main cause which produces you that problem.
In the WISP where I work, recently we have a very similar situation. We have a RB532 (MT 2.9.18 ) with the daughtboard and on it we put two 400mW 2,4GHz cards and a CM9 card (3 cards altogether).
So, we have a problem with our power supply (24V. - 2,5A), because it drops only 17V instead of 24V. So, the RouterBoard reset every time it reach the login prompt. Finaly, we change the power supply and could login correctly, but we found that the configuration of the 2,4GHz cards had been lost and we could not see the CM9 card in the wireless submenu.
I thing that the firmware of the CM9 card probably blow up because of the continous “auto-resets”, but I don`t have any idea why the routerboard lost the configuration of the other cards.
Right now Im doing test with two routerboards, one have MT 2.9.4 and the another have 2.9.24 The older one, lost the wireless configuration, the other one not. But I dìdnt do any test with the CM9 cards, I don`t want to loose another one card! jejejje
