search

CNS-OIST / STEPS4ModelRelease

Validation and performance models for STEPS4
0 stars 0 forks source link

Rallpack1: Add z_min as a membrane #13

Closed jlallouette closed 2 years ago

jlallouette commented 2 years ago

This PR is intended for debugging an issue in which the addition of a second membrane to rallpack1 leads to a difference in the results between STEPS 3 and STEPS 4. It could be related to the caburst issue.

jlallouette commented 2 years ago

Using mesh/axon_cube_L1000um_D866nm_1135tets.msh:

download download

iahepburn commented 2 years ago

is this with the bugfixed STEPS 4 code, Jules?

jlallouette commented 2 years ago

Yes, it is with katta/fix_inputs_v_i_on_ghosts

jlallouette commented 2 years ago

STEPS 3 goes up when z_min is added as a membrane: download

download

iahepburn commented 2 years ago

oh- weird. So STEPS 3 has the issue perhaps. Both are moving actually but STEPS 3 moves more. Impossible to say which one actually has the issue here without investigating further.

We have, fundamentally, Q=CV here. We increase C (by adding more membrane), Q is constant which means to my mind V should decrease a bit (relative to the 1 membrane model). So I'm confused.

jlallouette commented 2 years ago

Without leak, STEPS 3 and STEPS 4 are nearly identical:

download download

jlallouette commented 2 years ago

Unitary leak conductance was slightly lower for STEPS 3 with 2 membranes. After fixing this STEPS 3 and STEPS 4 are identical. I'm closing the PR.