# CHARMM TIP3P

When I calculate the temperature of water in a region, the output
temperature was alway lower than the expected one.

[…]

I found it is because the presence of
‘fix regidwater water shake 0.0001 10 10 b 1 a 1’. If this
command was deleted, temperature of region ‘temp1’ was 300K.

Why? Are there something wrong with shake?

no. not at all. PEBCAC!
remember that fix shake takes away degrees of freedom.
a water molecule has 3N = 9 DOF, however you constrain
two bonds and an angle and thus have only 6 DOF.
i bet that makes exactly the difference. have you looked at
the internal compute that npt fix creates?

axel.

Thanks, axel.

But why the temperature for all of water is correct? Just for part of water, temperature is wrong?

2011/9/23 physics.hangyan.chen <[email protected]>:

On Fri, Sep 23, 2011 at 5:19 PM, physics.hangyan.chen

When I calculate the temperature of water in a region, the output
temperature was alway lower than the expected one.

[...]

I found it is because the presence of
'fix regidwater water shake 0.0001 10 10 b 1 a 1'. If this
command was deleted, temperature of region 'temp1' was 300K.

Why? Are there something wrong with shake?

no. not at all. PEBCAC!
remember that fix shake takes away degrees of freedom.
a water molecule has 3N = 9 DOF, however you constrain
two bonds and an angle and thus have only 6 DOF.
i bet that makes exactly the difference. have you looked at
the internal compute that npt fix creates?
axel.

Thanks, axel.

But why the temperature for all of water is correct? Just for part of water,
temperature is wrong?

i just explained it to you.
perhaps you need to clean your glasses?

axel.