Einstein's elevator equivalence gedanken is rubbish.

[crawler]

Einstein's elevator gedanken for the bending of a light beam doesn't work if u examine it closely.
A beam of light crossing an accelerating elevator will appear to curve. However if we consider individual photons in the beam then every photon must remain parallel to its initial alignment at all times whilst crossing the elevator.
Think of the beam as a straight line of arrows in deep outer space.
If we apply that elevator beam to light passing the Sun, then that light beam will curve towards the Sun as it approaches the Sun, due to gravity. Think of this curve as being a simple ballistic curve, its not quite correct but nearnuff for today. But, that inward curving beam must then curve away from the Sun as it departs. Think, each photon/arrow continues at speed c km/s, measured along the arrow direction, at all times. Plus (or minus) there is a radial component of "fall" due to acceleration.

The inwardly curved approach & the outwardly curved departure result in an S curve. At a great distance the beam will be moving in a straight line which is parallel to the line of its initial traject. And the beam will never cross the centerline of the Sun, no matter how far u follow it.
Here the acceleration towards the Sun is equivalent to the acceleration of say millions of elevators moving radially away from the Sun. The beam crosses each elevator. And, whilst crossing, each photon/arrow maintains its initial straight line aim & speed (as seen by an outside observer). So, we have converted thems elevator gedankens to give an equivalent gravitational trajectory. The wording here is clumsy, but i think everyone gets the drift.

In that S-beam scenario an Einsteinian Ring, ie where the light of a faraway galaxy might be seen as a ring around a very massive galaxy, is impossible.

If the massive galaxy is very massive, such that it bends the light of the faraway galaxy a huge amount, then some of the light can eventually almost reach the centerline of that massive galaxy, but, it will never reach the centerline, & eventually its trajectory will be parallel to the centerline.

The individual beams will form a severe S. If we consider that there are millions of such severe S-beams all around the massive galaxy then these will almost join near the centerline, & will form a small ring of beams. A distant observer would not be able to see all of the ring at one time, he/she would only be able to see a small part at any one location. It would not look like a ring, it would look like a small patch of light. And that patch would appear to be very near the centerline of the massive galaxy. The observer would not see an Einsteinian Ring around the massive galaxy.

The above is based on Einstein's elevator gedanken. But we know that we can see lots of Einsteinian Rings, hence the Einsteinian elevator gedanken is not correct.

And where it fails is of course because photons passing the Sun do not maintain their initial aim, their aim gradually follows their actual trajectory. Hence the photon beam doesn't describe an S, the trajectory of the approach to the Sun is similar to the trajectory of the departure, & some beams can/do cross the centerline of the Sun at a great distance.

Conclusions: Einstein's elevator gedanken for the bending of a light beam is typical Einsteinian krapp. And GTR is krapp.

[crawler]

An improved version of OP.
Einstein's elevator gedanken for the bending of a light beam doesn't work if we examine it closely.

LIE No1. The first problem is that according to Einstein's silly elevator gedanken the bending of light passing the Sun would be 0.87 arcsec not Einstein's claimed 1.75 arcsec. The silly elevator gedanken gives us no more than the equivalent of the well known ballistic bending calculated by Soldner. The theoretical 1.75 arcsec is only obtained if including theoretical assumptions not existing in or near a theoretical elevator. This was obvious to me from the first. And i see that Paul Marmet pointed it out in 1999 -- Incompatibility between Einstein's General Relativity and the Principle of Equivalence.

Einsteinian apologists sob that its not a simple matter to synchronize clocks near the ceiling of an elevator with clocks near the floor, & they wail that that difficulty adds another 0.87 arcsec. Yes folks, we sure are in the Einsteinian Dark Age of science, but the times they are a-changin'.
Anyhow, ignoring the weeping pleas of Einsteinian apologists (easily done)(i wont bother to waste my breath on this today), according to a proper application of Einstein's elevator gedanken we only get a half of the well known 1.75 arcsec of bending near the Sun. Hence Einstein's elevator gedanken is false (re the bending of light).

GOOF No2. It gets worse. Marmet missed seeing another problem for Einstein. Marmet made the same mistake as Einstein. They both looked at a beam or a ray of light. Me myself i looked at the individual photons in the beam, & the photons told me that Einstein lied (again). But i think that it wasn’t a lie, it was a goof, they didn’t have photons in the oldendays.

I usually think of a photon as being a cylinder. I know that some fellows don’t like to give a photon a size or shape, so let's describe photons as being arrows. Hence a beam of light is a straight line of arrows (for an outside observer)(in deep outer space here).
If we look at individual photons in the beam crossing the elevator then every photon/arrow must remain parallel to its initial alignment at all times. After all, that there beam appears dead straight for an outside observer, at all times.

So, the beam of light crossing the elevator consists of say horizontal arrows. For an inside observer the beam appears to bend down (the elevator is accelerating up), but the arrows nonetheless remain horizontal (for the inside observer)(& for the outside observer).
Nextly we apply the elevator gedanken to a beam passing the Sun. Now, Einstein was happy to use one elevator, & he was happy to use one beam & one bend. That simple approach does indeed give 0.87 arcsec of bending, as per Marmet. But, lets look at our one beam crossing millions of elevators, each accelerating radially away from the Sun.

Each time a photon/arrow crosses an elevator its traject bends down, but the arrow retains its initial angle. No matter how many elevators the arrow crosses it retains its initial angle. And each arrow retains its velocity c km/s.
Now, if the photon/arrow was originally moving tangentially to the Sun as it passed (by definition), then we can draw a centerline passing throo the Sun parallel to that tangent.

A simple examination of the traject for an arrow shows that it can never cross the centerline. When or if the arrow eventually enters the last elevator, the elevator at or next to the centerline, the acceleration of the elevator will be parallel to the arrow. The arrow will never get to the far wall. Or if u like it gets to the far wall at infinity. But it can never cross the centerline.

So, the arrow traject bends towards the Sun on approach, in a ballistic way, & then is parallel to the Sun at closest approach, & after passing the Sun the traject must reverse, such that the arrow never reaches the centerline. Hence the traject follows an S kind of traject. The arrow at some time reaches a point of closest approach to the centerline, & then diverges away & leaves the centerline, & much later its traject becomes nearnuff parallel to its original traject, albeit displaced sideways towards the Sun.

Hence according to a proper application of Einstein's elevator gedanken we can never see an Einsteinian Ring. All we can see at any one time & place is a small part of a half-baked ring. However, we know that Einsteinian Rings exist. Hence Einstein's elevator gedanken is false (re the bending of light).

If the ray of light originates at a light-source on the centerline of the massive body, & if the massive body is super massive such that the S trajects of the beams/photons/arrows almost meet & touch the centerline on the far side, briefly, before diverging, & if the converged photons are somehow seen by an observer, then that observer will see a patch of light, not a ring, & that patch will appear to be at the centerline of the supper massive body.

[crawler]

Wiki says that Orest Khvolson predicted the Khvolson Halo in a paper in 1924.
Einstein thought of an Einstein Ring in 1912.
Einstein wrote a paper in 1936.

I dont know what logic Einstein used for his 1912 Ring.
I suspect that he used the bending of time, in which case he would have predicted 0.87 arcsec of bending of light passing the Sun.
In praps 1915 he might have added the bending of space, which added another 0.87 arcsec, giving the correct 1.75 arcsec found by Hipparcos in the modern era.

I dont know when Einstein first thort of his Elevator Gedanken , ie his Spacious Chest Gedanken. It might have been in 1915.
Anyhow, as i explained earlier, the Elevator Gedanken gives only 0.87 arcsec (the same as Khvolson i suppose)(the same as Soldner in 1901)(praps the same as Cavendish in 1786), & more importantly it doesnt predict rings at all, in fact the Elevator Gedanken predicts that gravitational rings are impossible (as i explained earlier).

[crawler]

In the past i have often used Excel to calculate the bending of light passing the Sun.
Lemmeseenow, i have done…..
(1) Ballistic bending, ie as per Soldner in 1801, Excel gave 0.87 arcsec.
(2) Time bending, ie as per Einstein in 1911, Excel gave 0.87 arcsec.
(3) Space bending, ie as per Einstein in 1915, Excel gave 0.87 arcsec.
(4) (2) plus (3) gives 1.75 arcsec, as per Einstein's 1915 GTR.
(5) This week i used Einstein's elevator gedanken, Excel gave 0.876 arcsec (but Excel lied).

This is how i did (5) in Excel.
(a) I used 9,022 elevators, the farthest was 3.98E14 km from the Sun, measured along a tangent, i call this XX km.
(b) The distance of the elevator from the center of the Sun is what i call the XY distance.
(c ) The distance from the center of the Sun to the tangent is a YY distance, & is a radius of the Sun.
(d) The nearest elevator was touching the Sun at the tangent, ie XX = 00 km, XY = 699,508 km, YY= 699,508 km.
(e ) Each elevator accelerated away from the center of the Sun, along XY, with g=G*m/r*r, where r is the XY distance to the center of the Sun. The XY angle & distance varies with each elevator.
(f) In each elevator the photons appeared to the inside observer to enter via a hole in the wall, the photon beam initially appearing to enter horizontally, ie parallel to the floor, & then the beam curved/bent down while crossing, & exited via a hole in the far wall a little lower than the entry.
(g) To a stationary outside observer the narrow photon beam (a single line of single photons) allways traveled along a tangent line drawn to the Sun, & each photon/arrow followed that there tangent at all times, & each photon/arrow remained parallel to that tangent.
(h) If in (f) the beam appears to enter horizontally then the elevator must be initially speeding towards the Sun. To the inside observer the photons will appear to be very short, & each short photon/arrow will appear to aim/lay/travel parallel to the floor, & they will appear to cross at much less than c km/s, ie at v km/s where v is c*Cos(XY), here (XY) being the angle tween XY & YY.
(i) For the farthest elevator the angle (XY) was 89.9999999 deg. For the last elevator XY was 0.00000000 deg.
(j) The vertical distance dropped by the beam while crossing the elevator was drop (km) = g*t*t/2. Here the vertical distance is the distance down the wall, the wall appearing to be vertical to the inside observer. This is an XY dimension for the outside observer.
(k) In (j) the time t (sec)(which varies) is the time to cross the elevator, & here t = w/v, where v is the apparent speed of light (km/s)(varies) measured parallel to the floor (see (h)), & w the width of the elevator (km)(which varies).
(l) But re (k) & (h), there is a simpler way to find t, t is the time taken to travel the XX distance tween elevators by a photon/arrow travelling at c km/s along XX. Its that simple.
(m) The width of every elevator varies. Each elevator has a width that spans the delta XX distance from the previous elevator to the next elevator. The total width of the 9,022 elevators equals the total distance XX to the Sun. The narrowest elevator is the nearest one, ie the elevator at the tangent point, & it is 121 km wide. The widest is the farthest & it is 3.59E14 km wide.

(n) My Excel page has 9028 rows & say 20 columns.

(o) I calculated the drop/width (km/km) for each elevator, & converted that to arcsec for each elevator. This arcsec is the naïve average slope of the beam. In fact the end of the beam has a greater angle, which happens to be double that there naïve average angle.
(p) I added these naïve arcsec's to get the total for the beam's journey to the Sun, & this added to 0.219 arcsec.
(q) However the real bend of the beam for each elevator crossing is double that there naïve calculation of drop/width, which thusly gives us 0.438 arcsec. The bend for the second half of the journey on the far side of the Sun is the same, hence the total bend for passing the Sun comes to 0.876 arcsec.
(r ) For interest, the total of the drops in the elevators for the journey to the Sun was 3,140 m (each drop is measured vertically in each elevator)(ie vertically according to the inside observer)(ie in the XY directions)(ie parallel to the walls), & this drop is equivalent to a total drop of 1.09 m measured in the YY direction. The YY direction is the one we want, ie it is the bending we are looking for. Alltho the drop is only of interest, what we all talk about is the bend mentioned in (q).

(s) In my Excel model the beam travels along a tangent to the Sun. A real beam acting in accordance with Einstein's elevator gedanken will after passing the Sun eventually bend the opposite way as it approaches the centerline of the Sun, & the beam will eventually be parallel to the centerline & almost touching the centerline, ie the beam will be parallel to the original direction. Remember, the photons/arrows retain that alignment/direction/aim at all times (according to Einstein's elevator gedanken). But the beam itself of course bends (& then bends back the other way)(an S bend).

(t) A better Excel model for the Einstein's elevator gedanken for an Einstein Ring would have the initial beam starting at a say star on the centerline of the Sun, the beam then grazing the Sun as it bends & passes, then bending further as it departs, then eventually when it gets close to the centerline the beam bends away from the centerline, & in the end the beam direction will be parallel to the initial direction.

(u) And there will be a large number of similar beams all around the Sun, all converging to near the centerline, & then all diverging. Hence an observer at the point of closest approach to the centerline would not see an Einstein Ring, he/she would see a small patch of light apparently coming from the centerline of the Sun.

(v) And when i said in (q) that the total bend comes to 0.876 arcsec, that might be true near Earth's orbit, but if u measure the bend at the point where the beam is closest to the centerline then the bend is say 0.000 arcsec (depending on definition). In any case once the beam starts to reverse then that there 0.876 arcsec at Earth's orbit starts to diminish, & then it falls to 0.000 arcsec, & then it falls to a negative arcsec, & eventually when the beam(s) is diverging it becomes difficult to say what the bend is or isn’t, the problem then being how in hell to define what the bend is or isn’t.

Anyhow this here analysis shows that Einstein's elevator gedanken was rubbish, koz it can't explain Einstein Rings. We need a good explanation for the bending of light, & Einstein's silly GTR is of no help here.
And a good explanation will have it that each individual photon bends along its own length, & its aim/alignment/direction is always the same as its trajectory. Actually that last specification only applies in a static aether, if there is an aetherwind then it must fail, but we wont worry about that today.