Non-geodesic circular orbit? [closed]

From N. Straumann, General Relativity

Exercise 4.9: Calculate the radial acceleration for a non-geodesic circular orbit in the Schwarzschild spacetime. Show that this becomes positive for $r>3GM$. This counter-intuitive result has led to lots of discussions.

This is one of those problems where I have absolutely no clue what to do. Since it says non-geodesic, I can't use any of the usual equations. I don't know what equation to solve. Maybe I solve $\nabla_{\dot\gamma}\dot\gamma=f$ with $f$ some force that makes $\gamma$ non-geodesic. But I don't know where to go from there if that's the way to do it.

Also any specific links to discussions?

Any help would be greatly appreciated.

EDIT: So I tried solving $\nabla_u u=f$ with the constraints $\theta=\pi/2$, $u^\theta=0$ and $u^r=0$. lionelbrits has explained I must also add $\dot u^r=0$ to my list. This all leads to $$(r_S-2Ar)(u^\varphi)^2+\frac{r_S}{r^2}=f^r$$ ($A=1-r_S/r$, notation is standard Schwarzschild) The problem with this is that the $u^\varphi$ term is negative for $r>3m$. So somewhere a sign got screwed up and for the life of me I don't know where it is. A decent documentation of my work: http://www.texpaste.com/n/a6upfhqo, http://www.texpaste.com/n/dugoxg4a.

closed as off-topic by ACuriousMind♦, Brandon Enright, Kyle Kanos, John Rennie, JamalSDec 6 '14 at 13:29

This question appears to be off-topic. The users who voted to close gave this specific reason:

• "Homework-like questions should ask about a specific physics concept and show some effort to work through the problem. We want our questions to be useful to the broader community, and to future users. See our meta site for more guidance on how to edit your question to make it better" – ACuriousMind, Brandon Enright, Kyle Kanos, John Rennie, JamalS
If this question can be reworded to fit the rules in the help center, please edit the question.

• You can find a general path with $\dot r = 0, \theta = \pi/2$. Then, yes, you just calculate that acceleration term. – Jerry Schirmer Dec 5 '14 at 21:07
• correct. Any curve should have unit length. – Jerry Schirmer Dec 5 '14 at 21:33
• @Jerry Schirmer What do I do about the $2\Gamma^r_{tt}(u^t)^2$ term? I'm not sure what $u^t$ is in a non-geodesic prescription. ⟨u,u⟩=−1 still holds, right? – Ryan Unger Dec 5 '14 at 21:33
• @JerrySchirmer What I have so far: texpaste.com/n/a6upfhqo. I feel like I should be able to express $(u^t)^2$ in terms of $(u^\varphi)^2$ using $\langle u,u\rangle=-1$, right? – Ryan Unger Dec 5 '14 at 21:47
• Yes, you solve for $u^t$ in terms of $u^\varphi$ using the norm of the velocity vector being -1. Basically this just means that you have suitably chosen an affine parameter. – lionelbrits Dec 5 '14 at 22:32

The equivalence principle tells us that we can evaluate $\nabla_u u$ in a co-moving reference frame and that for geodesics we should find no acceleration (to the occoupants of an elevator in free-fall, the contents seem to be experiencing no acceleration). Therefore, if we evaluate this when we are not along a geodesic (elevator sitting on earth), we find that it is not zero. Because it is a vector, if it is non-zero in one frame, it must be non-zero in another. In other words, yes, $f^r$ is what you have to calculate. The ingredient that you are missing is that $r=\mathrm{const}$ for a circular orbit implies that $\dot{u}^r = 0$. This is not a local thing, it is simply because you are forcing the orbit to be circular.

• So by doing this in texpaste.com/n/dugoxg4a I get $0=-r_S(u^\varphi)^2-\frac{r_S}{r^2}+2Ar(u^\varphi)^2+f^r$. I think I screwed up a sign because when I rearrange to get $\frac{r_S}{r^2}+(u^\varphi)^2(r_S-2Ar)=f^r$, it is actually negative for $r>3m$. Any idea what's up? – Ryan Unger Dec 6 '14 at 13:23
• Is that a covariant/contravariant issue? – lionelbrits Dec 6 '14 at 13:43
• You haven't defined $A$ yet. – lionelbrits Dec 6 '14 at 14:03
• I don't think so. $r_S-2Ar=6m-2r$. Obviously this is negative when it should be positive. On the other hand this is not a simple overall sign flip because the term $r_S/r^2$ is positive when it should be. So $f^r$ can either be positive or negative depending on $(u^\varphi)^2$ and $m$ it seems. – Ryan Unger Dec 6 '14 at 14:08
• $A$ is defined in the OP and in the links in the above discussion. – Ryan Unger Dec 6 '14 at 14:12