Unbalanced force in periodic boundary settings

Hi all,

I have been experimenting with "PeriTriaxController" engine for the past few months and I am getting curious results which I was hoping you can help me understand better.

1) When friction is not zero, the unbalanced force seems to be relatively high for samples that are prepared and compacted with PeriTriaxController. Even if I prescribe zero strain rate for all the strain components (sample at rest), the unbalance hardly goes below 0.005 after a day. (This is for a 5000 particles, 2D simulation).

For comparison, with "TriaxialStressController" I get the unbalanced force to decrease below 1e-4 in a span of couple of hours.
Is there a sound explanation for this behaviour?

(I understand that 1e-4 may be too conservative for most of the purposes, but for my type of studies, I need this range of equilibrium.)

2) Interestingly, I can get very small unbalanced forces with frictionless materials. I can even reach unbalanced force of 1e-10 in a couple of hours. But, intriguingly, as soon as I introduce friction to the contacts via "setContactFriction", the unbalanced force jumps to ~1e-2, regardless of the type of control on the boundary.

This is confusing since typically increasing friction should enhance the stability. Again, this is not the case with "TriaxialStressController".

I did not include examples since I was hoping you are already aware of this issue. I can certainly prepare small examples if needs be.

Thanks!