Page 466 - Maxwell House

446 Chapter 9

affair), build your computer model in AUTOCAD, CATIA, or another design tool. That
makes your model universal and acceptable by most commercial EM software packages.
11. You probably found that the numerical simulation ran too long or too big for your
computer. If so, break a big, overwhelming project into smaller, manageable parts. Look
back at Section 5.4 of Chapter 5 and decompose your model in Lego style. For example,
you can design and optimize the conical horn separately using the data from steps 2 - 4.
We are not going to collaborate further.
12. In the same manner, the separate Ka-band antenna of smaller diameter can be designed. If
you are lazy enough to do so, ask yourself the “million-dollar question” - “Do the system
require two dishes or could a single bigger diameter reflector operate simultaneously on
Ku- and Ka-band and meet all the specs in Table 9.1?” If you have some ideas, visit your
manager and meet a customer for discussion.
13. Do not trust the results of numerical simulations explicitly, use any available opportunity
to revise and cross-check their correctness. The intuition and common sense are the front
lines of your defense.
Finally, after several steps of validation and confirmation as Figure 9.1.5 demonstrates, you
came to the point of reviewing the conceptual model
and correcting the mathematical model by increasing
Physical Object the model complexity to better match the specs.
Suppose you were lucky and reached in the right spirit
(green smiley in Figure 9.1.2) the mechanical engineer
is responsible for antenna manufacturing. His/her
reaction might be the same size smile meaning full
acceptance of proposed design. But more likely, it will
be said that your design includes some non-existent,
inaccessible, or too expensive materials. Besides, you
Figure 9.1.5 Computer model must change this and that to make your antenna
validation and confirmation producible. Getting this extensive list of remarks, you
returned to your computer and models. In the worst
case, the conclusion might be disastrous: the proposed design is not producible in the available
mechanical shops. So you could start almost from the beginning or speak to another more
creative mechanical engineer. Hopefully, you were careful and discussed your project with
mechanical engineers several times before finishing it and avoided such unpleasant situation.
To end, you have gotten the green light with or without some reservations and sometime later
received your antenna ready for test and final validation.
All being well, the test demonstrates several minor discrepancies that can be corrected by
adjusting merely your final model. It means that after passing the vibration and shock,
environment and all other required tests without surprises the antenna is ready to customer
delivery. The manager recognizes you as a gifted engineer and starts discussing a new even
more challenging project.

It is important to note that an appropriate numerical solver has to be chosen for each EM field
problem. That requires lots of attention to details and needs the correct numerical method based
on the problem at hand. Several issues should be taken into accounts such as the required
accuracy in modeling complex geometries, computer memory requirements, and computational
time.