SIMULATION ||| Wireless
28 DE^ | Technology for Optimal Engineering Design August 2019 /// DigitalEngineering247.com
some of the nonlinear behavior of
the RF system.”
Exploring which trade-offs are
most effective in engineering and
cost requires tools and skills that
cut across those boundaries. “The
engineers need tools that allow them to consider some aspects
and incorporate more insight in the RF/analog part of the sys-
tem to do their work and know that they are doing it correctly,
more so than in the past,” Karnofsky says.An Interdisciplinary Approach
Some tools are emerging to help companies provide a more
holistic system outlook to engineers so that they can optimize
the entire design—not just the individual components. Math-
Works takes a simulation-to-test approach that provides a
toolbox for engineers that allows them to put together a single
model to help predict system performance.
“The RF engineer, for example, can give an accurate model
of what they are designing to the system architect or digital
designer, and they can incorporate that into their simulation as
they develop their part of the system,” Karnofsky says. “They
aren’t making flawed assumptions.”
Nebens offers MIMObit, a simulation tool that allows users
to assess the performance of a multiple input/multiple output
(MIMO) antenna system at the capacity and throughput level,
as well as the dynamic range of communication algorithms
with realistic antenna systems. It also helps simulate spectrum
sharing, dynamic spectrum access and system coexistence.
Nebens helps make a connection between antenna design
and capacity/throughput testing, according to Buris. “Right
now, antenna design specifications that people adhere to are
based on intermediate performance metrics, but those don’t
necessarily relate to capacity and throughput,” he says. “As a
result, they may pass these intermediate tests, but they don’t
pass the final certification testing. In our tools, we can predict
the performance at the throughput and certification test level
as they are designing the antenna systems.”
At MathWorks, Karnofsky says that the company has de-
veloped simplified interfaces so that an RF or antenna engi-
neer, for example, can reference a pared-down version of the
digital details of the standard. “They don’t need all of the de-
tails, just enough details that are relevant to their challenges
as an RF engineer,” Karnofsky says. “On the digital side, they
get every detail on the standard. But there is also a graphical
tool that simplifies that down to what an RF or antenna engi-
neer would need to use.”
COMSOL also recently released a COMSOL Multiphysics
tool so that RF designers can integrate electromagnetic simu-
lations with heat transfer, structural mechanics, fluid flow and
simulations. With this tool, designers also can model coupled
physics effects as they would occur in a real-world application.
At Keysight, Nichols says that the company emphasizes havingsolutions that can simulate the entire
system, which help reduce design is-
sues reaching the test phase. “It’s more
of an integrated approach instead of
measuring particular stimulus and
response approaches,” Nichols says.
“We have a network emulation system that cannot just make RF
measurements, but also make measurements to ensure the entire
system works properly.”
Keysight has also reduced the complexity of the measure-
ment process and engages in a lot of field training to help users
get the information they need quickly.
Buris says more companies will need to shift to this mul-
tidisciplinary approach to meet these evolving wireless stan-
dards challenges. “With this methodology, you can optimize
for the ultimate goal,” he says. “That gives you a product that
performs better, as well as being smaller and cheaper. The RF
and antenna people need to know more about communications,
and vice versa. All of the algorithms on the communication side
completely oversimplify the antenna systems.”
In some cases, companies outsource some element of the
design work to a specialist who already has the ability to bridge
those different design domains. “We get customers who have
done this type of design before, or maybe they do mechanical
design but not wireless,” Voler’s Maclay says. “In some cases
they don’t want to hire a team and train them, or they are
under deadline pressure.”Benefits of a Shared Simulation Environment
Having engineers collaborate or access a shared framework
also helps eliminate a common problem in wireless design:
over-engineering. When working on assumptions, for example,
designers sometimes overcompensate by taking more care than
necessary to avoid interference. That adds cost to the finished
product. With accurate simulation data from other depart-
ments, they can optimize the design more effectively.
“If you overdesign at the intermediate performance speci-
fication, it is impossible to get optimized performance and
throughput later,” Buris says.
This approach also saves time and additional costs by iden-
tifying problems prior to the physical testing phase, when it is
more expensive to fix any issues.
It also enables design optimization that can result in better
products. “A good example in 5G is MIMO technology, which
gives you the ability to steer and focus the signal,” Karnofsky
says. “That requires code design, digital, RF and antenna tech-
nologies, and they have to be done all together to optimize
across those domains. In addition to getting a cheaper, faster
process, you can get a better process through simulation.”
Implementing this type of approach takes time and train-
ing—a difficult problem in a busy organization. “How do you
get people up to speed on something new?” Karnofsky says.
“You identify pilot projects and try out new methods so people“This is not rocket science,
but it’s getting closer.”
— Roger Nichols,
Keysight TechnologyDE_0819_Wireless_Simulation_Albright.indd 28 7/11/19 1:48 PM