Rocket M9's losing associations

[dougncbp]

Not sure if this section of the forum is still active, but I thought it might be worthwhile to post here....

We have a couple of Rocket M9's on a T-Mobile tower. They work well to get into areas with heavy foliage. On July 5th, we had a huge rainstorm area come through that area, and a combination of lightning and rain downed the site. The climber actually found water inside the enclosure on the bottom of the AM-9M13 sector antenna that one of the M9's reside in, so the rain was probably torrential and sideways at the peak of the storm. Dried all that out, and thought we were in the clear.

We had to replace every radio on the tower, save for a Mikrotik Dyna Dish. In addition to the two M9's, this included a Rocket M2 AP and a Rocket Prism 5AC that we use for a backhaul connection. We also replaced the Netonix WS-8150-AC switch, even though all the Ethernet cables there had recently been replaced with the "flooded" type to prevent water migration. But the key difference is, the Rocket M2 AP was replaced with one of the new Rocket 2AC units that have a lot more capacity for the legacy subscriber units.

Since that replacement, the M9 AP's have been regularly losing all station associations. I set up a watchdog timer on both of them to automatically reboot if a particular customer can't be pinged. That has activated multiple times on both of the units, and each time all stations come back with good signal like they always had. So it's not interference of any kind that I can see. The units are back-to-back, and have good isolation. Nose levels are -89 and -104.

The only thing I can think is that there might be some interaction with the new 2AC AP, even though they are in entirely different bands. Would be interested to know what others think about this problem. The ping watchdogs are doing what I need them to do, but it's a last-choice solution.

Thanks,

Doug N

[sirhc]

If I remember correctly the 900MHZ airMAX radios are down converted 2.4 GHz so they "could" possibly see noise from the 2GHz radios.

What antennas do you have on the 2.4 GHz Rockets?

[dougncbp]

In this case, it's one of the omni-directional types, the AMO-2G13. Both of the M9's are in pretty close proximity to that omni. We have deployed these 2AC units in two other places, but this is the only place we have the M9's. We are running 10MHz bandwidth on everything, so unless there's a lot more spread on the 2AC, what can I do to mitigate this?

[sirhc]

https://www.rfarmor.com/omni-kits/uos2x.html

https://www.rfarmor.com/sector-kits/usk1209.html

[dougncbp]

Yep, we actually have two of those already for the 900 sectors. Got them in trade from another WISP for something they wanted. But the cost to deploy them would be a little steep, since those are rather large antennas and would require bringing them to the ground to deploy the armor. So we have held off on using them.

The M9's are running on channels 5 and 7. There are only 7 channels available at 10MHz bandwidth, since there is no overlap. I would like to know the conversion rate that they use to go from 2.4GHz channels to 900MHz, so that I could put the 2AC radio on a channel as far away from those frequencies as possible. Since it is also constrained to 10MHz spectrum use, it seems to me that would help a great deal.

[sakita]

The 915 band runs from 902 to 928 MHZ for a total of 26MHz... a quick check on one of our Loco M900 radios shows that with a 10 MHz channel width the frequencies are 907, 909, 912, 914, 917, 919, and 922. So, to make two radios non-overlapping at 10MHz means they would have to pretty much be at opposite ends of the band: for example, 907 and 922.

As for the conversion rate, I don't know what it is or what channels to avoid. Something else I might curious about though, if the radios are really close together, is what frequency the microprocessor is at and if the 915 band is near a harmonic of it?

In any case, the shields are a big help compared to the stock plastic radio housings and antennas.

[lligetfa]

In any case, the shields are a big help compared to the stock plastic radio housings and antennas.

Shields may help with intermediate (out of band) interference but does squat for in-band colo interference. IMHO, UBNT M9 doesn't colo worth shit. You pretty much need GPS sync to colo 900 MHz.

[sakita]

...and it certainly doesn't hurt to make sure your own channels aren't overlapping

[dougncbp]

Thanks to everyone who replied to my original question. We did deploy the RF Armor shields on the 900MHz sector antennas, and the problem is greatly reduced, but not gone completely. I also lowered the power by a couple of dB on the new Rocket 2AC unit with no ill effects. The channels on the M900 are also separated as far as possible. 909 (904-914) and 922 (917-927) are the center frequencies and (frequency bands). But both units only reboot a couple of times a week now when the watchdog timers kick in. And of course, this is the season for tropospheric ducting, so I am chalking some of this up to external interference.

Word of advice, though: If you apply the shields to sector antennas that have been up on a tower for a few years, plan on fighting the bolts for the mounting brackets quite a bit. I had to stand on the antenna hung over the edge of a pickup truck tailgate while the other tech was underneath the whole thing to start the bolts in. The internal nuts on the sectors float inside a small space, and getting them to line up right AND put enough pressure on the bracket with the armor behind it is tricky. It adds just enough thickness that it's an issue if the antenna has "swelled" a bit and doesn't fit into the shield assembly just right. Slightly longer bolts would be helpful.

Doug

[sirhc]

Keep in mind this is the 900 MHz sector kits, the 2.4/3/5 GHz kits are easy to install but the 900 is a beast, never was an easy install.