Node B RF Form Factor and RF Performance

Given that operators may be asked to share access hardware and given that operators
have been allocated different RF carriers, it may also be necessary to produce small
form factor Node Bs capable of processing more than one × 5 MHz RF carrier—ideally
60 MHz, though this is at present unrealistic in terms of digital sampling techniques.
The two major design challenges for Node B products are transmit linearity, including
the ability to handle multiple downlink OVSF codes per user, and receive sensitivity,
including the ability to handle multiple uplink OVSF codes per user. We have said
that receive sensitivity can be improved by using electronic downtilt (reducing the
exposure of the Node B to visible interference) and multiuser detection where the
Node B uses the short codes embedded in each individual handset’s offered traffic
stream to cancel out unwanted interference energy. Multiuser detection is a longerterm
upgrade (rather like frequency hopping was in GSM in the early 1990s).

Receive sensitivity is also a product of how well the radio planning in the network
has been done and how well the Node B sites have been placed in relation to the
offered traffic. We address these issues in the next section.
As with handsets, RF power budgets can be reduced by increasing processor overhead.
For example, we can implement adaptive smart antennas on a Node B, which
will provide significant uplink and downlink gain (potentially 20 or 25 dB). This
reduces the amount of RF power needed on the downlink and RF power needed on the
uplink. However, if the processor power consumption involved (to support the many
MIPS of processing required) is high compared to the RF power saved, then very little
overall gain would have been achieved. You will just have spent a lot of money on
expensive DSPs.
As with handset design, DSPs can do much of the heavy lifting at bit level and symbol
level but run out of steam at chip level. There is also a need for substantial parallel
processing to support multiple users, each with multiple uplink and downlink OVSF
code streams. These factors presently determine existing Node B form factor and
functionality. The design objective has to be to balance good practical RF design with
judicious use of DSPs and ASICs to deliver power-efficient processor gain. The requirement,
as with GSM, is to keep power consumption for small Node Bs in the region of
tens of Watts. This means it is easy to install Node Bs indoors without greatly adding
to the landlord/hosting energy bill, and for outdoor applications, it provides the basis
for solar-powered or wind-powered base station/Node B implementation. 246