[TowerTalk] 900MHz - Long and possibly OT

[Roger Parsons]

Please forgive my posting this question here - if anybody can suggest a more appropriate forum I would be grateful.

For several years I have had a 900MHz link between my home and remote stations. These are about 15km apart but are separated by VERY not-line-of-sight countryside. 

The link initially used Aironet BR1000 units with Hyperlink HA (5W) amplifiers and 14dB gain yagis at each end. The antenna at the home end was at 60' and at the remote end at 220', and the amplifiers were close to the antenna. This worked well (860kb) for several years until a wireless ISP set up shop on a tower directly in line with the path and about 2km from the remote site. This effectively destroyed the link as the Aironet units are very poor with co-channel interference, and reduced the throughput to a few tens of kb. This happened about a year ago.

I changed the radios to Ubiquiti XR9s/Mikrotik Routerboards which live much better with the interference, but found I also had to change the amplifiers to new (expensive) Hyperlink models which work with the modern radios. I was still only barely winning until I put the remote antenna up to 300' - that made a phenomenal difference and I then had around a 35dB margin, and was able to maintain a better than 1Mb link.

This worked extremely well from about August last year until a few weeks ago with literally no outages. (I don't think the same can be said for the ISP, but he wasn't prepared to cooperate with me in terms of frequency selection which would have helped both of us, and after all he is an unlicensed user...) I suddenly noticed that I was getting outages ranging from a few seconds up to several tens of minutes and the next day the link disappeared completely, and remained down for several weeks until three days ago. (During the 'up' periods the SNR was solid at 35dB in both directions.) I then observed some periods when the link connected for a few seconds, then for a few minutes, then for a few hours, until it has now been solid for 72 hours.

The weather here has not been conducive to extensive trouble-shooting and I was waiting for it to improve before doing anything much, but now I have nothing to look for! On the other hand I am well aware of the old tech's proverb "If you haven't found the fault you haven't fixed it".

The logs on both the radios state 'disconnected, extensive data loss' - I take that to mean complete loss of signal as it is the same fault logged when the power is removed from the opposite radio. I know that the radios were driving the amplifiers under the fault condition (although possibly at the wrong frequency), as the DC feed to the amplifier has an LED which flashes when the amplifier takes current. I checked that at both ends of the link (the check at the remote end involved a long uphill climb through several feet of snow!) I have tried resetting both ends several times which made no difference - and now it is working again they also reset back into a working state. It had been raining heavily for several days when the link failed, but it had also been raining heavily for several days when it reappeared!

My original plan had been to lower the (pneumatic) mast at this end and bridge the amplifier to see if that restored the link - it would have been poor but probably just about usable. Then bring the radio from the remote site here and check that I could establish a link locally. Then possibly swap the amplifier from this end for the one at the other end and so on.

I have clearly had an intermittent fault, but I cannot recall one which failed so completely for such a long period and then recovered so completely...... The possible culprits seem to me:

(1) An interference source which started up and has since stopped. Difficult to know what that would be - another radio system would hurt my throughput but I don't think would kill things stone dead.
(2) A faulty radio at either end. Possible, but I think it would have to be something like faulty modulation/demodulation or frequency. And it would have had to have mended itself all on its own!
(3) A faulty amplifier at either end. I think this is the most likely suspect, but again it would have had to self-repair.
(4) Faulty coax or antenna at either end. I think this is the least likely as water damage would be the probable cause and that is certainly not self repairing!
(5) Faulty power supplies. These checked good at both ends during the failures. Equipment has a UPS at each end - put those in as well as intruder detection after a local villain (unfortunately with a ham licence) broke into the remote site and spoiled/stole a lot of my work a few years ago.

I am at a lost to know how to proceed with troubleshooting - if I start playing with the equipment I am just as likely to keep it working as to break it so it can be repaired. On the other hand, I am certain that if I just hope it doesn't break again it will certainly do so in the middle of the 160m DX season...

Thank you for reading this far. Helpful suggestions would be appreciated.

73 Roger
VE3ZI