[lammps-users] Spring Tether Bug

Hi Steve,

I've read from lammps documentation site that there was a bug on fix
spring tether and the patch for it was released on August 1,2008 (I
haven't applied the patch yet). Our group has been using this command
since early this year (with the June 2007 version) to tether the center of
mass of a stiff polymer chain to 0.0 0.0 0.0 and we compared the average
angle between two consecutive bonds. We did a similar simulation without
tethering the chain. We got identical results hence we concluded that the
spring tether fix has no effect on the conformation of the chain.

So my question is, what exactly did the bug miscalculate?

(currently we are repeating the same test on the May 2008 version and will
be doing the same when we apply this patch)

Jan-Michael

Hi Steve,

I've read from lammps documentation site that there was a bug on fix
spring tether and the patch for it was released on August 1,2008 (I
haven't applied the patch yet). Our group has been using this command
since early this year (with the June 2007 version) to tether the center of
mass of a stiff polymer chain to 0.0 0.0 0.0 and we compared the average
angle between two consecutive bonds. We did a similar simulation without
tethering the chain. We got identical results hence we concluded that the
spring tether fix has no effect on the conformation of the chain.

So my question is, what exactly did the bug miscalculate?

jean-michael,

the force applied to the atoms in the tethered group was
repeatedly scaled by the total mass. so the first atom got
the correct force, each subsequent atom only a fraction.
it basically amounts to mostly fixing only one atom of
the group and having a weaker tethering force in total.
so if your results do not depend on the tethering force,
your runs should be still ok. the impact on the USER-SMD
package is _much_ worse. this is how the bug was found...

cheers,
   axel.

Axel's description is correct. I can't recall exactly, but
I think the bug got introduced in late 2007. So before
that it was correct, then it was incorrect for several months,
until it got corrected with the patch. Axel wrote his
fix smd starting from fix spring in that bad window, hence it had
the bug as well.

If you look in your older June 2007 code you should be able
to compare the changed lines (in the patch file) and see if
you had a correct or incorrect version.

Steve

Thanks Steve and Axel