Femtime – mismatch for 2 different tlists

Hi,

The system I solve is a coupled system (resistive heating) --> higher temperature --> higher current --> higher temperature and so on.

The system I look at has sharp temperature changes in a very short time (few ps). The overall simulation length however, is 100ns.

Therefore, I would like to split the calculation into a coarse tlist and a short one. I use COMSOL 3.5a with Matlab interface.

fem.sol=femtime(fem, ...
'init',fem0.sol, ...
'solcomp',{'T','sim_V1_p_i','V','Vconstra_g1_emdc'}, ...
'outcomp',{'T','sim_V1_p_i','V','Vconstra_g1_emdc'}, ...
'blocksize','auto', ...
'tlist',[colon((kappa+tau-1)*timeStep,timeStep/1E1,(kappa+tau)*timeStep)], ...
'tout','tlist', ...
'linsolver','spooles');

--------------------------------
coarse step
0ps:5ps:50ps
and for the fine one
0ps:0.1ps:10ps

The problem is that the solutions seems to kind of forget the old solution besides that it is fed back with fem0.sol to have a good starting value and I have no clue why.

If I use for the coarse step
0ps:5ps:50ps
and for the fine one
0ps:0.05ps:50ps

it works nicely meaning there the feedback fem0.sol works, however the step of 50ps is too coarse to accommodate the modified conductivity so I have to use 10ps and update the conductivity.
I am quite puzzled why COMSOL is struggling to use two different time steps.

If anyone could help me on that it would be great.

Many thanks and best regards,
Toby