Calculix *FREQUENCY error

Hello All
I am doing some research for my thesis and am using Calculix to solve some of my models. I am having problems with the *FREQUENCY solver using SPOOLES under both CCX 1.8 from Convergent Solutions and CCX 1.7 on CAELinux. On some models I can solve in a desired frequency range on one in particular I can't. Below are my *STEP cards. The results from the static analysis are perfect and stored fine. The solver looks like it is solving for the modes and checking the memory usage I am sure it is because when I solve for a couple hundred modes it takes a couple hundred MB or more of memory. I looked around on the forum and web so see if someone else had this problem but I couldn't find anything. I can't load the model file since I haven't published my work yet. I have similar models that perform as expected though. I need to be able to solve in ranges because the system has to many modes to solve for at once because I just have enough ram or access to a 64bit machine. I tried running the CCX 1.8 executable from the calculix website on CAELinux but it didn't work. BASH is giving me an error. I can run the CCX 1.7 that came with CAELinux from BASH with no problem. Should I just compile version 1.8 myself? I tried using the SOLVE=ARPACK option but in both cases it tells me ARPACK is not linked. The docs say SPOOLES and ARPACK are used in the *FREQUENCY. CAELinux has the ARPACK libs but I am not sure how CCX on CAELinux was compiled. I have cross posted this on the Calculix Yahoo Group as well. Any suggestions would be appreciated. CAELinux and all the included packages are great, thanks for all the good work.
Thanks
Ben

*STEP
*STATIC, NLGEOM
*DLOAD
A,GRAV,9.810,0.,0.,-1.
B,GRAV,9.810,0.,0.,-1.
C,GRAV,9.810,0.,0.,-1.
D,GRAV,9.810,0.,0.,-1.
E,GRAV,9.810,0.,0.,-1.
F,GRAV,9.810,0.,0.,-1.
G,GRAV,9.810,0.,0.,-1.
H,GRAV,9.810,0.,0.,-1.
I,9.810,0.,0.,-1.
J,GRAV,9.810,0.,0.,-1.
K,GRAV,9.810,0.,0.,-1.
*CLOAD
5,2,53
5,3,-19


*NODE FILE, NSET=Nall
U
*NODE PRINT, NSET=Nall
U
*EL FILE, ELSET=Eall
S
*EL PRINT, ELSET=Eall
S
*END STEP
*STEP,PERTURBATION
*FREQUENCY
10,1.01,2
*NODE FILE
U
*END STEP

Hello Joël
There is no error message CCX reports that the job is finished so it doesn't crash either. I checked all the files for an error message but didn't see one. It seems like the results are not being written to the .frd file from the ram. I will try to build a version with debug support today and see what happens.
Thanks
Ben