At Cadence Design Systems’ subsidiary, Tality Corp.,
prototype testing exposed large recirculation zones within a customer’s
telecommunications system. The unexpected results came after flow analyses,
provided by an outside consultant using CFdesign, had failed to identify
any potential problems.
Having spent thousands of dollars and weeks getting a wrong answer, Tality
needed a quick way to recoup if it were to meet its tight delivery schedule.
A magazine article pointed Tality toward Coolit, and after a rapid ramp-up
and brief trial, engineers attacked the problem, identifying the flow patterns,
pinpointing stagnation zones and then modeling baffle arrangements that would fix
the problem.
A single CFdesign analysis for only one slice through the chassis
took approximately 1 to 1-1/2 days to build
and mesh, and another 24 hours to run. Coolit required only two hours total:
one hour to model the entire system and one hour to run the calculations.
And follow-on prototype testing verified that the Coolit predictions did,
indeed, match the real life results.
After having been bitten by its initial FEA chassis predictions, engineers
decided it would be prudent to revisit the board level analyses. Here, Coolit
predicted that the boards, though close to their limits, would perform as
required. Again, Tality verified the Coolit predictions against
measurements; the four boards matched actual results to within 2%, 5%, 7%, and 10%,
respectively (engineers believe the latter number reflects an erroneous
power input assumption provided by the customer).
|