Hello,
I work for a small fiber-optic ISP and we use MikroTik (hEX s, hAP-ac, RB4011, etc.) as the customer premise equipment. We are seeing some problems pop up after prolonged use, usually around 2-3 years or so, to where several customers are reporting drops, inconsistent or just plain poor service on our older (3 years or so) batch of routers where we utilize the SFP port as the fiber uplink port. We have tried extensive troubleshooting, but nothing that we do seems to work. There are no errors that give any indication as to what the problem is, no errors on any of the interfaces, etc. It does not seem to be a wireless issue, as the issues are not exclusive to wireless, per multiple customer reports. Several/all devices appear to be affected when there is a problem. We have installed the problem routers in other places with reports of similar issues, too. When we swap out equipment for new, the problems disappear. Nevertheless, it is becoming a real (expensive) issue, as we cannot dependably reinstall this failed equipment. We believe this to be a heat issue with the transceiver being the culprit. We are using fs.com 10k transceivers and they get quite hot. This can also be seen from the temperature log of the device. I have seen the temperature rise near or above 70 deg Celsius, which I did not think was even possible, given that is not within operating specifications. Usually, the light dBm levels are higher than (or below, depending on how you look at it) -8, sometimes as high as -4 or even -3, which seems really quite high. We have tried attenuators and there is potentially some promise to those; but, it seems once it gets past a certain point of no return, the devices appear to be not consistently functional. We have tried running tests back at the shop and the CPE seems to work okay, but it could very well be that we are not testing over long enough period of time or running the right tests. I am asking if anyone knows what the operating light levels should be, as I do not see anything posted. While I think attenuators could (and probably do) help, I don't have hard evidence to back this up. I guess I am asking if anyone has experienced something similar, knows a potential solution, or can shed some light on this issue. Any help is appreciated. Thanks.