Agh I'm going insane

Can anyone answer this? physics.stackexchange.com/questions/100575/…

Law of equipartition considers all modes of kinetic energy and potential energy due to vibration. But what about the potential energy due to inter-molecular attraction? Equipartition says nothing about that. Then, how can we say that total internal energy for, say, mono-atomic gas is $\frac{3}{2}{nRT}$, when we haven't considered the potential energy due to molecular attraction?

@ManishEarth ?

I haven't visited the new questions tab for a while...

Revisiting every page now, I find something strange

At the first, it starts with question "asked 2 days ago"

then, it goes "asked Feb 19 at..."

then, "asked Dec 13 '13 at..."

"Jan 23"

"Oct 7 '13"

and soon, it goes into "Jun 19 '12", and "May 24 '12"

^^^ to the older questions... o_O

At the end of the page, it says "Aug. 24 '10" <--- Woah!!!!! O_O

After all, I'm at page 3, and things go exponentially (???)

@ShuklaSannidhya we ignore that

IIRC, it's usually not much

remember, equipartition is used for gases

yeah, if it is used for gases, and if we assume that they're ideal, then there'll be no attraction. right?

@ManishEarth But I've seen equipartition being used for liquids and solids too. Don't we say internal energy of crystalline solid is $3nRT$?

Also, molar heat capacity of metal is derived using equipartition...

@ShuklaSannidhya I think we only look at the parameters that can contain changing energy

Wait no

Not sure

Ah, that is encoded in the vibration term. Equipartition is usually for oscillators, and the intramolecular attraction provides oscilaltors for us to analyse. So we analyse them

oh... so the potential energy of oscillator = potential energy due to attraction?

@ManishEarth ping

not sure

@ShuklaSannidhya I have a blog post that might be of interest on that issue

In this case, I think the intermolecular potential is nearly independent of particle position for a gas (except when two particles get close to each other) so you can treat it as negligible. But technically, yes, it would have an effect.

In other news, I'm wondering if we should have a banner about safety issues

@DavidZ Have you written any post about how research is?

@DavidZ A banner saying something along the lines of "We're not safety professionals, so please take every precaution suggested with a large grain of salt"?

Hey can someone explain to me a situation where I could use spherical coordinates to calculate an electric field that isn't along a cartesian axis?

I mean, most of the time I've seen E-field problems they reduce to something along, say, the z axis due to symmetry, but that makes the spherical coordinates to just have a cos(theta) or something, and reduce...

@KyleKanos something like that, yeah

But it would also remind people posting comments or answers not to give safety recommendations.

@jinawee what do you mean?

@DavidZ do you have an example?

Of what?

An efield you can calculate in spherical coordinates that doesn't just result in something like a zhat field?

Rings of charge can be summed over, but you end up just taking a cosine times the R, what if you have asymmetric distributions?

Like of a point charge?

I'm not sure I see what you're asking

I know, I wasn't very clear...

In Cartesian coordinates you can calculate E-fields by integrating component wise, like from a semi infinite wire- you can take the x and z coordinates separately.

But in Spherical coordinates, I feel like all they are ever good for is, when presented with large symmetries, like rings of charge, eliminating the symmetric parts.

Is there a case where you can take something like a cube, or something asymmetric, and calculate the field using spherical, not cartesian?

Probably not a cube, because that's bad right?

You can, but there's often not much point in doing so

@DavidZ I mean how is a typical research day? Do you just sit and think, talk with other researchers, do calculations, etc.

What's an example, though? I can't think of any case where it would be possible- the theta and phi coords just seem like junk to me.

@jinawee oh, no I haven't written anything about that. People get plenty of that from other blogs.

@Anthony you might be able to do it with a cube, I'm not sure.

You could do it with an infinitesimal electric dipole

Can you explain?

I don't follow what you mean by the dipole.

A charge distribution with one positive charge and one negative charge of equal magnitude, the same distance away from the origin in opposite directions

And that could be done in spherical?

Yeah, it's a pretty standard calculation. Wikipedia has the formula, for instance.

Mind if I ask a question?

(I know I'm contravening the don't ask to ask, but politeness calls)

@Alyosha anything goes in chat :-)

(pretty much)

Excellent!

Is there a physical justification for assuming that, in a metal ring, the temperature is a seperable function of position and time?

@DavidZ but wait this is only for two particles

I get that you could just sum spherical ones

but then with an integral, you're summing over some coordinate...

I need to sleep now, I may post it on main tomorrow.

@Anthony still not sure I understand. But I have to run now. Maybe we can continue this later.

Alright...

@dmckee you there?

More or less.

Can you give me an example where we can calculate an efield in spherical coordinates, but where it isn't a matter of "recognizing symmetries" to get rid of theta/phi dependencies?

Like for a wire extending from 0 to infinity, we can calculate the field of that in cartesian, how can we do that in spherical?

Or rather, I guess polar.

@dmckee?

@KyleKanos So Hotspot analysis is pretty awesome, but I am playing with the hardware stuff now and it gives me a different set of hotspots

But I don't know what to do with the results now because it is telling me that _int_malloc and malloc_consolidate are the chokepoints in my code, both are in libc