3.5 GHz band and Fixed Wireless

The 3.5 GHz band has been an interesting journey. I did not have any experience with it until summer 2020 when starting the fixed wireless operation ElektraFi in Buna TX (2 hours east of Houston) in the pineywoods of SE Texas. Dense Everygreen pine forests with pockets of deciduous trees. Trees can be from 30 to 120 feet in height. open pasture land is common with walls of trees around them or within large tracts of land. Rural space, population groups in small towns and small settlements (anywhere from a coupe of houses to 10 to towns of several hundred homes)

Started with LTE 3GPP Airspan Airharmony 4200 then migrated to Airspeed 1030 within 4 months mainly due to hurricane damage 30 days into the startup.

Mainly due to stubborn trial and error, we got it to work well, providing 25 / 35 /50 Mbps internet service out to 5 miles with up to 30-foot antenna pipes or on taller roofs.

lessons learned
Vendors are almost useless at providing good operational guidance in making design decisions. Lots of trial-by-fire testing yielded what worked – very hard and cost time and money.

Examples – questions on loading of PCI/cells – we did not know that for 20 MHz channel sizes and planning for 64 to 256 QAM modulations (did not know until too late) that we should cap CPE capacity at around 30 and make sure all CPE’s were CAT12 or later models.

SMALL ops EPC servers, keep the seat count below 450, ideal size is about 350 max.

Signals and qualifications

RSRP of -105 minimum and SINR of 15 or higher to get to consistent 64 and 265 QAM modulation support and best use of PRB’s in the 20 MHz channel. Use the features properly with good signals and it works well. That mix of guidance would have been nice to have the details prior to building a site #1. #experience#startup#design

Fast forward to 2022 and putting in Tarana BN (SECTOR Radio) and RN (CPE)

Trees still matter, the Tarana spatial modulation and handling enables faster in the same environment of foliage. The RF digital drop point is sharper, thus blow by blow the Tarana solution has less range than 3GPP LTE in the same path geometry #testing#environment#digital#cbrs

Tarana needs the pathloss to be 149 dB or better and the SINR to be 10 or better for a reliable delivery of 100 to 200 down and 20 to 50 up. This will work well out to 8 miles in our forested environment with the RF highway path even with NLOS – totally depends on clutter and terrain. Tree foliage loss still obeys physics even with Tarana Star Trek Magic. #geometry

CBRS and SAS

The Tower sector radios (BN’s) are less impacted by DPA spectrum grant suspensions than the RN’s most of the time. By less impacted I mean they can move faster to non DPA spectrum channels than the RN’s can. I am hopeful that Tarana and the SAS providers can arrive at a faster move cycle to make DPA events smoother and minutes instead of hours long outages.