About two years ago we began testing 60 GHz equipment for PtMP deployments.
During out testing phase we settled on max distance of 300m for residential client deployments. Any further and we had too many drops (we see a fair amount of heavy rain in the summer and snow in the winter).
So these deployments were for client close to our tower locations, and in a roof top deployment in a suburban neighborhood.
This past year we have had quite a few odd issue, and as such we rarely upgrade the 60 GHz equipment as it has been prone to 'breaking' the connections.
In the rooftop deployment subdivision we are using a WAP60AP and LHG60 at the clients. We have had this location deployed since late spring with little issue running version .
This past Thursday night we had a reasonable snow fall with temps cold enough that the snow accumulated (3-5 inches).
In the morning of Friday the six clients connected to the WAP despite all indicating good alignment RSSI all low 50s. Basically nothing appearing different then the day they we installed the units were dropping continuously. The longest they would stay connected was a few minutes, with many dropping every 20-30s. Performance was all over the place, latency going between 0-timeout, massive tx-packet errors. The units when they dropped also took much longer to reconnect then usual.
The units and AP were all running 6.45.9, we proceeded to test all the channels available to us, we with some difficulty (they would rarely stay connected long enough to upload the new version) upgraded all the units and the AP to 6.48.1. The problem persisted on v6.48.1 we then attempted to lock the tx-sector in (disable beamforming) this did appear to improve things, but disconnects were still occurring to all clients every 3-5 minutes. We did a truck roll and replaced the AP unit, we had out techs brush snow off the roof where the AP was located (AP is ~6ft above peak), and brush snow off a couple of the clients roofs. There was no snow or ice build up on the AP or any of the LHGs. Still the problem persisted. The techs were about to begin swapping the LHG60 units to 5 GHz based equipment to get them back online.
Taking a shot in the dark I decided to downgrade the AP from 6.48.1 to various versions, all the versions had the exact same problem.
That was until I decided to try the oldest compatible version 6.44.6, and low and behold all the units connected instantly, MCS rates stabilized, latency was a solid <1ms and dropouts were not occurring.
We rebooted the AP a few times, and every time all the clients connected very fast, performance was great, and no dropouts.
It has been about 32 hours since the downgrade and I just double checked and we have not had a single one of the clients have a dropout.
All the clients are still running v6.48.1, but the AP is running v6.44.6.
I was glad I stumbled upon the solution, but the solution is concerning as all the other versions I attempted to use all had major dropout and tx error problems.
I just wanted to share this info with other in case it helps.
FYI: We have other deployments using the WAP60Gx3 AP and they had problems with individual clients dropping a lot that was fixed after we upgraded to v6.48 and also set the mdmg-fix=yes
I mention this as we have locations that are working and have problems that appeared to be resolved with v6.48.x and this rooftop location where anything other then v6.44.6 appears to be unusable.