Is GR false?


On 8 Oct 1998, Chris Hillman wrote:


 Subject : Re: Re[9]: No such thing as BHs?
 From :	"C. Hillman" hillman@math.washington.edu
 Date :	Thu, 8 Oct 1998


The geodesic equations for the Schwarzschild coordinates for the Schwarzschild
vacuum are readily computed by the Lagrangian method from

   L(t,r,theta,phi, t*, r*, theta*, phi*, s)

   = (1-2m/r) t*^2 - r*^2/(1-2m/r) - r^2 (theta*^2 + sin^2 theta phi*^2)

 where * denotes differentiation with respect to the parameter s.  They are

  t** + 2m/(r-2m) t* r* = 0

  r** + m(r-2m)/r^3 t*^2 - m/(r(r-2m)) r*^2 + (r-2m) theta*^2
              - (r-2m) sin^2 theta phi*^2 = 0

  theta** + 2/r r* theta* - sin 2 theta theta* phi* = 0

  phi** + 2/r r* phi* -2 cot theta theta* phi* = 0

 For theta = pi/2 (orbit in the equatorial plane) and r = R (circular
 orbit) all but the third equation drops out, and after dividing by a
 common factor it becomes simply

    m/R^3 t*^2 - phi*^2 = 0

 or (d phi/dt)^2 = m/R^3 which is, of course, Kepler's Law for circular
 orbits, which remains valid for gtr (Not so for elliptical orbits, of
 course!). Thus, constant speed according to the Kepler law is the
 "projection" of a geodesic in the spacetime, which appears as a perfect
 helix in Schwarzschild coordinates.

 Now, we haven't checked that the tangent vectors to this putative world
 line are in fact timelike.  Plugging dphi^2 = m/R^3 dt^2, r = R, and theta
 = pi/2 into the line element we find

   ds^2 = (1-2m/R) dt^2 - R^2 m/R^3 dt^2

        = (1-3m/R) dt^2

 or ds/dt = sqrt{1-3m/R} when R  3 m.   In other words, when R  3m, the
 parametrized curve

    t(s) = s/sqrt(1-3m/R)
    r(s) = R
    theta(s) = pi/2
    phi(s) = sqrt(m/R^3) s/sqrt(1- 3m/R))

 is a timelike geodesic parametrized by arc length.  For r = 3m, this
 represents a null geodesic (world line of a photon in a circular
 orbit--- i.e. a perfect helix in spacetime--- this is not a closed
 timelike curve!). For r  3m, this is a spacelike geodesic.

 We still haven't checked whether these circular orbits are stable under
 small perturbations (see below for a surprise!).

Note that everything up to this point has been a study of circular orbits
of test particles around an isolated nonrotating massive body (analyzed in
the exterior static Schwarzschild coordinate chart).


 It is interesting to compare the time kept by a satellite orbiting in a
 circular orbit R_o with a clock on the surface of the Earth, which we
 idealize as a nonrotating ball with radius R_g  2 m, and with exterior
 geometry given by the Schwarzschild vacuum.  Then

   ds_o/dt = sqrt(1 - 3m/R_o)

   ds_g/dt = sqrt(1-2m/R_g)

 so

    ds_o   sqrt(1-3m/R_s)
    ---- = --------------
    ds_g   sqrt(1-3m/R_g)

 This is increasing, approaches sqrt(R_g-3m)/sqrt(R_g-2m) for R_o - R_g,
 approaches 1/sqrt(1-3m/R_g) for R_o - infty, and equals one for

    R_o = (3/2) R_g

 Thus, a satellite in low orbit moves so rapidly (according to Kepler's
 law) that the kinematical time dilation (plus the gravitational time
 dilation at radius R_o  R_g) of the rapidly moving satellite clock
 overwhelms the gravitational time dilation of the ground clock (at the
 radius R_g), so that it runs slow relative to the ground clock. Satellites
 with very high circular orbits are moving slowly and suffer little
 kinematical or gravitational red shift, so they run fast relative to
 ground clocks.  At R_o = (3/2) R_g, the competing str and gtr effects
 exactly balance, so that the clocks run at the same rate.


Chris Hillman

It is not true that the competing str and gtr effects exactly balance at R_o = (3/2) R_g. Indeed, the SR time dilation factor is sqrt(1-(GM/rc^2)*r/(r+H)), and the GR factor is given by sqrt(1+(GM/rc^2)*2H/(r+H)), where G is the universal gravitational constant, M the mass and r the radius of the Earth, and H the altitude of the satellite orbiting the Earth. Note that 1) sqrt(1 - (GM/rc^2)*r/(r+H)) or sqrt(1 - v^2/c^2) is the rate of the orbiting clock relative to a _stationary_ clock ** at the same height H **. (cf. Paul B. Andersen) 2) sqrt(1+(GM/rc^2)*2H/(r+H)) is the rate of a _stationary_ clock at the height H compared to a stationary clock at the surface. From 1) and 2), it is clear that the rate of the orbiting clock relative to a stationary clock at the surface of the Earth is given by the product of the SR clock slowing factor and the GR factor, i.e. sqrt((1 - (GM/rc^2)*r/(r+H) * (1+(GM/rc^2)*2H/(r+H)) = sqrt((1 - (GM/rc^2)*r/(r+H) + (GM/rc^2)*2H/(r+H) - (GM/rc^2)^2* 2Hr/(r+H)^2)) = sqrt((1 + (GM/rc^2)*(2H-r)/(r+H) - (GM/rc^2)^2 * 2Hr/(r+H)^2)) The GR solution is limited to sqrt((1 + (GM/rc^2)*(2H-r)/(r+H)), which indeed becomes 1 for 2H=r, because it neglects the higher order term (GM/rc^2)^2 * 2Hr/(r+H)^2. Iow, GRT gives an approximate solution. In order to obtain the correct altitude H at which the SR and GR effects exactly balance, one has to solve the equation (GM/rc^2)*(2H-r)/(r+H) - (GM/rc^2)^2 * 2Hr/(r+H)^2) = 0 Then, H = ((Rs-r)+ sqrt((Rs-r)^2+8*r^2))/4, where Rs is the Schwarzschild radius. It is only when Rs=0 (!), that the value of H is r/2, as predicted by GRT. Marcel Luttgens