[mems-modsim] Coventorware RAM (Memory) problems >> Unexpected Error failures

[Mehmet Yilmaz]

Hello all,
I am trying to simulate a comb-actuator device. The geometry of the device
is roughly:
*******************************************************************
Springs:
1000um length, 10um width, 10um depth >> Folded beam springs

Comb-actutor geometry:
Finger width: 5um,
finger length: 40um;
Finger depth: 10um,
initial overlap of comb-fingers: 10um
Distance between two neighboring comb-fingers for the primary capacitance
 2um,
Initial paracitic capacitance distance:30um

I have 200 comb-finger pairs
*********************************************************************

My computer is:

Dell XPS 600, Pentium4 CPU3.80GHz, and 2GB of RAM.

And I increase the solver memory from MemMech Advanced Setting to 5000 MB
for my CoSolve Simulation.



Although I can simulate good results with small and simple structures that
have analytical solutions to compare (Such as a clamped beam actuated with
an electrode under the free end of the beam), I have NEVER obtained good
results for my device geometry.



I can NOT obtain Electrostatic Force Convergence while reducing the mesh
size by refining edges and refining the structure globally. My device is
approximately composed of 50000 Manhattan Bricks before I start educing the
mesh size by refining edges and refining the structure globally. At first,
Coventor says that it needs approximately 350MB of RAM for the device
WITHOUT reducing the mesh size by refining edges and refining the structure
globally.



Then, gradually, I do the steps given below:



refine edges >> 1, refine globally >> No

refine edges >> 1, refine globally >> 1

refine edges >> 1, refine globally >> 2



And,

refine edges >> 1, refine globally >> 1  >>>>> gives results, but they are
quite different than the 1st results, so I assume that, there is no
convergence yet.



After that, I use,

refine edges >> 1, refine globally >> 2 >>>>> simulations give an error as
'unexpected error occured, log file may contain useful informatin'. But
there is no useful information about the reason of the failure. Or at least
I can not see the useful information!



Is there somebody who faced the same problem?

What should I do to get rid of that 'unexpected error'?



My mesh size is approximately 4um in x-direction x 4um in y-direction x 5um
in z-direction globally.



Best regards,



Mehmet Yilmaz