Derivation of Hyperbolic Representation from Lorentz/Minkowski equations in SR

In summary: Just like how the square root of 9 is 3 and the reciprocal of 9 is 1/9.Fourth line: This is just an algebraic identity. (A-B)^2=A^2-2AB+B^2.Fifth line: 1) The result from the fourth line. 2) Definition of \gamma. 3) A different way of writing 1-v^2. 4) \gamma^2=1/(1-v^2) is a well-known identity. You can verify it by substituting v^2=1-1
  • #1
stevmg
696
3
This is a carryover from a previous thread:

https://www.physicsforums.com/showpost.php?p=2875138&postcount=68

Sports Fans:

I am familiar with the Minkowski equations and the Lorentz transformations in one or two dimensions:

A) In algebraic form
(1) t2 - x2 = t'2 - x'2

(2) t' = [itex]\gamma[/itex](t - xv/c2)
(3) x' = [itex]\gamma[/itex](x - vt)

Now, by using equations (2) and (3) and doing the right substituions, one gets equation (1) and by starhaus:

starthaus said:
The standard interpretationfollows correctly from the INVARIANCE of the Minkowski metric as follows:

[tex] (c d \tau)^2=(cdt)^2-dx^2[/tex]

I am sure I have demonstarted this before.

Thus [itex]\tau^2[/itex] is invariant which is saying the same thing, no matter where on the hyperbola your x and t are, wherever you go on that same hyperbolic curve, [itex]t^2 - x^2 = \tau^2[/itex]

Notice I am using the timelike format.

Now, the equivalent Lorentz equations in hyperbolic format are:

B) In hyperbolic form.
(4) t' = -xsinh [itex]\vartheta[/itex] + tcosh [itex]\vartheta[/itex]
(5) x' = xcosh [itex]\vartheta[/itex] - tsinh [itex]\vartheta[/itex]

where

(6) sinh [itex]\vartheta[/itex] = [itex](v/c)\gamma [/itex]
(7) cosh [itex]\vartheta[/itex] = [itex]\gamma[/itex]

and where

(8) [itex]\vartheta[/itex] = tanh-1(v/c)

Now, I have tried, tried and tried and I cannot get from A) to B)

In the figure posted we have the t-axis vertical and the x-axis horizontal and this is a time-like representation:

In this equation that starthaus presented:

c2t2 - x2 = [itex]\tau^2[/itex]
[itex]\tau [/itex] should be the t intercept at x = 0

any point on the hyperbola should be represented by the parametric equations of

x = [itex]\tau[/itex]sinh [itex]\phi[/itex] and
t = [itex]\tau[/itex]cosh [itex]\phi[/itex]

where tanh-1 [itex]\phi[/itex] = x/t for any point (x, t) on the hyperbola

How do I make the connection?
 

Attachments

  • Hyperbolic Equation.jpg
    Hyperbolic Equation.jpg
    31.7 KB · Views: 530
Physics news on Phys.org
  • #2
stevmg said:
Now, I have tried, tried and tried and I cannot get from A) to B)
Hey, didn't I tell you that once already? :biggrin:

https://www.physicsforums.com/showthread.php?p=2678572

(Start at "This is what I wrote in my notes:". After the first sentence, you can skip the stuff about adding velocities, and continue at "The definition of rapidity...".)
 
  • #3
You have explicitly stated c in some equations, but not in others. Here are the same equations with all the c's explicitly stated:
...
A) In algebraic form
(1) (ct)2 - x2 = (ct')2 - x'2

(2) t' = [itex]\gamma[/itex](t - xv/c2)
(3) x' = [itex]\gamma[/itex](x - vt)
...
B) In hyperbolic form.

(4) t' = t cosh [itex]\vartheta[/itex] - (x/c) sinh [itex]\vartheta[/itex]
(5) x' = x cosh [itex]\vartheta[/itex] - ct sinh [itex]\vartheta[/itex]
...
where

(6) sinh [itex]\vartheta[/itex] = [itex](v/c)\gamma [/itex]
(7) cosh [itex]\vartheta[/itex] = [itex]\gamma[/itex]
...
Now starting from the Lorentz transformation for time:

[tex]t' = \gamma(t-xv/c) = t(\gamma) - (x/c)(v/c)\gamma = t\, cosh\, \vartheta - (x/c)\, sinh\, \vartheta [/tex]

which is (2) --> (4)

and starting from the Lorentz transformation for distance:

[tex]x' = \gamma(x-vt) = x(\gamma) - (vt)\gamma = x(\gamma) - ct(v/c)\gamma = x\, cosh\, \vartheta - ct\, sinh\, \vartheta [/tex]

which is (3) --> (5).

Now I am sure you are looking for something much deeper than that, but I thought we might as well get the algebraic relationship between the hyperbolic and regular expressions established in a (hopefully, barring typos) consistent way first.
 
Last edited:
  • #4
stevmg said:
any point on the hyperbola should be represented by the parametric equations of

x = [itex]\tau[/itex]sinh [itex]\phi[/itex] and
t = [itex]\tau[/itex]cosh [itex]\phi[/itex]

where tanh-1 [itex]\phi[/itex] = x/t for any point (x, t) on the hyperbola

I think that should be:

any point on the hyperbola should be represented by the parametric equations of

x = c[itex] \tau[/itex] sinh [itex]\phi[/itex] and
ct = c[itex] \tau[/itex] cosh [itex]\phi[/itex]

where tanh [itex]\phi[/itex] = x/(ct) = v/c = [itex]\beta[/itex] for any point (x, ct) on the hyperbola.

(not tanh-1)
 
  • #5
Fredrik said:
Hey, didn't I tell you that once already? :biggrin:

https://www.physicsforums.com/showthread.php?p=2678572

(Start at "This is what I wrote in my notes:". After the first sentence, you can skip the stuff about adding velocities, and continue at "The definition of rapidity...".)

Fredrik -

That post: https://www.physicsforums.com/showpost.php?p=2678572&postcount=36 is way beyond me. This may disqualify me from going further but that is the case.

I haven't a clue what you wrote. Might as well be speaking Martian or some other alien tongue.

Doc
stevmg
 
  • #6
yuiop said:
I think that should be:

A) In algebraic form
(1) c2t2 - x2 = c2t'2 - x'2

(2) ct' = (ct - xv/c2)
(3) x' = (x - vct)

(4) t' = -xsinh [itex]\vartheta[/itex] + ctcosh [itex]\vartheta[/itex]
(5) x' = xcosh [itex]\vartheta[/itex] - ctsinh [itex]\vartheta[/itex]

where

(6) sinh [itex]\vartheta[/itex] = [itex](v/c)\gamma [/itex]
(7) cosh [itex]\vartheta[/itex] = [itex]\gamma[/itex]

and where

(8) tanh [itex]\vartheta[/itex] = (v/c), or [itex]\vartheta[/itex] = tanh-1 (v/c). The arctanh of v/c is [itex]\vartheta[/itex]

Is that better?

Now, get me going as if I cannot further progress than this, there is NO POINT in you wasting your time in trying to explain it to someone with subnormal intelligence (in physics) such as me.

I wish I could get those damn quotes within quotes as you have done.

I will try but it will take thirteen forevers with edits and re-edits to try and I guaranatee you no success.

Go to post #7 below
 
Last edited:
  • #7
Continuation from post #6 above,

I still can't get from A) to B) so my error isn't withing the typos but within the brain!

Doc
(by the way, when you address me in a post you can and should call me "Doc" as that's what everyone in the Air Force used to call me (better than some of the other phrases that I can't repeat here on PF.)
stevmg
 
  • #8
stevmg said:
Fredrik -

That post: https://www.physicsforums.com/showpost.php?p=2678572&postcount=36 is way beyond me. This may disqualify me from going further but that is the case.

I haven't a clue what you wrote. Might as well be speaking Martian or some other alien tongue.

Doc
stevmg
I didn't expect you to understand all of it, but I'm surprised that you that you claim to understand nothing of it. This is the essential part:

Fredrik said:
[tex]v^2=\tanh^2\phi=\frac{\sinh^2\phi}{\cosh^2\phi}=\frac{\cosh^2\phi-1}{\cosh^2\phi}=1-\frac{1}{\cosh^2\phi}[/tex]

[tex]\frac{1}{\cosh^2\phi}=1-v^2=\frac{1}{\gamma^2}[/tex]

[tex]\cosh\phi=\gamma[/tex]

[tex]\sinh^2\phi=\cosh^2-1=\gamma^2-1=\frac{1}{1-v^2}-\frac{1-v^2}{1-v^2}=\frac{v^2}{1-v^2}=\gamma^2v^2[/tex]

[tex]\sinh\phi=\gamma v[/tex]
What part of that is impossible for you to understand? The third and fifth lines are the results that take you from A) to B).

Explanation of the equalities:

First line: 1) Definition of rapidity ([itex]\tanh\phi=v[/itex]). 2) Definition of tanh. 3) The identity [itex]\cosh^2x-\sinh^2x=1[/itex], which is easy to derive from the definitions of sinh and cosh. 4) For all real numbers A,B with A non-zero, we have (A+B)/A=1+B/A.

Second line: 1) The result from the first line. 2) Definition of [itex]\gamma[/itex].

Third line: 1) The square root of the result in the second line.

Fourth line: 1) The identity [itex]\cosh^2x-\sinh^2x=1[/itex] again. ([itex]\cosh^2[/itex] should of course be [itex]\cosh^2\phi[/itex]. I left out the variable by accident). 2) The result from the third line. 3) Definition of [itex]\gamma[/itex] and the fact that 1=A/A for all non-zero real numbers A. 4) For all real numbers A,B,C with C non-zero, we have A/C+B/C=(A+B)/C. 5) Definition of [itex]\gamma[/itex].

Fifth line: 1) The square root of the result in the fourth line.
 
Last edited:
  • #9
Fredrik said:
I didn't expect you to understand all of it, but I'm surprised that you that you claim to understand nothing of it. This is the essential part:


What part of that is impossible for you to understand? The third and fifth lines are the results that take you from A) to B)

Explanation of the equalities:

First line: 1) Definition of rapidity ([itex]\tanh\phi=v[/itex]). 2) Definition of tanh. 3) The identity [itex]\cosh^2x-\sinh^2x=1[/itex], which is easy to derive from the definitions of sinh and cosh. 4) For all real numbers A,B with A non-zero, we have (A+B)/A=1+B/A.

Second line: 1) The result from the first line. 2) Definition of [itex]\gamma[/itex].

Third line: 1) The square root of the result in the second line.

Fourth line: 1) The identity [itex]\cosh^2x-\sinh^2x=1[/itex] again. 2) The result from the third line. 3) Definition of [itex]\gamma[/itex] and the fact that 1=A/A for all non-zero real numbers A. 4) For all real numbers A,B,C with C non-zero, we have A/C+B/C=(A+B)/C. 5) Definition of [itex]\gamma[/itex].

Fifth line: 1) The square root of the result in the fourth line.

I'll download it to paper and with my calculus book that goes into hyperbolic functions and their derivatives I will pour over it until it sticks in my brain.

Below, I have posted a part of a lecture that sort of proves it, too but this author just made a wild a-- guess which turns out to be true (equation 32) or [itex]\beta = v/c[/itex] which is just a WAG. Nothing like knowing the answer before you take the test. How do you think we doctors got through medical school? Certainly not on out of genius.

Lorentz-hyperbolictransformations.jpg
 
  • #10
I wouldn't describe (32) as a guess. It's the definition of rapidity. (The same one I'm using, except that I'm calling it [itex]\phi[/itex] instead of [itex]\beta[/itex]. I don't like calling it [itex]\beta[/itex], because [itex]\beta[/itex] is often defined as v/c). Of course, when it's presented this way, it looks like we were just lucky that the definition turned out to be useful.

Fredrik said:
Note that I'm using units such that c=1. Keeping that c around is like putting something sharp in one of your shoes before you go out running. It's a pain with no benefits.
I actually made the edit that added the quote above before I saw this below: :smile:
yuiop said:
LOL. Pesky c's. Missed a couple myself.
 
  • #11
LOL. Pesky c's. Missed a couple myself. Things are still not quite right.

stevmg said:
A) In algebraic form
(1) c2t2 - x2 = c2t'2 - x'2

(2) ct' = (ct - xv/c2)
(3) x' = (x - vct)

(4) t' = -xsinh [itex]\vartheta[/itex] + ctcosh [itex]\vartheta[/itex]
(5) x' = xcosh [itex]\vartheta[/itex] - ctsinh [itex]\vartheta[/itex]

where

(6) sinh [itex]\vartheta[/itex] = [itex](v/c)\gamma [/itex]
(7) cosh [itex]\vartheta[/itex] = [itex]\gamma[/itex]

and where

(8) tanh [itex]\vartheta[/itex] = (v/c), or [itex]\vartheta[/itex] = tanh-1 (v/c). The arctanh of v/c is [itex]\vartheta[/itex]

Is that better?

(2) ct' = (ct - xv/c2)

should be (2) ct' = [itex] \gamma [/itex](ct - xv/c)

(3) x' = (x - vct)

should be (3) x' = [itex] \gamma [/itex](x - vt)

(4) t' = -xsinh [itex]\vartheta[/itex] + ctcosh [itex]\vartheta[/itex]

should be (4) ct' = -x sinh [itex]\vartheta[/itex] + ct cosh [itex]\vartheta[/itex]

For some reason you dropped some gammas too, that I have put back in. You can probably go from there now.

(1),(5),(6),(7) and (8) are OK.

I have edited post #3 for the "c" I missed in equation (1) so all the equations in #3 should be exactly correct now, but they are presented slightly differently to the equations presented here.

The complete and hopefully finally correct set of equations should be:

(1) c2t2 - x2 = c2t'2 - x'2

(2) ct' = [itex] \gamma [/itex](ct - xv/c)
(3) x' = [itex] \gamma [/itex](x - vt)

(4) ct' = -x sinh [itex]\vartheta[/itex] + ct cosh [itex]\vartheta[/itex]
(5) x' = xcosh [itex]\vartheta[/itex] - ctsinh [itex]\vartheta[/itex]

(6) sinh [itex]\vartheta[/itex] = [itex](v/c)\gamma [/itex]
(7) cosh [itex]\vartheta[/itex] = [itex]\gamma[/itex]

(8) tanh [itex]\vartheta[/itex] = (v/c)

(9) x = c[itex] \tau[/itex] sinh [itex]\phi[/itex]
(10) ct = c[itex] \tau[/itex] cosh [itex]\phi[/itex]

(If they are not correct, PM me. We can't spend the whole thread correcting typos :tongue: )
 
Last edited:
  • #12
To yiuop and Fredrik -

Using the equations [itex](ct)^2 - x^2 = \tau^2[/itex] I KNOW there has to be a more direct and intuitive way to prove the hyperbolic assertions.

Using the timelike diagram where the ct axis is up and down and the x-axis ls the left and right, [itex] sinh \vartheta = x/\tau [/itex] and [itex] cosh \vartheta = ct/\tau [/itex] there has got to be a way of working [itex] v [/itex] and [itex] c [/itex] into this.

[itex] x = \tau sinh \vartheta [/itex] and [itex] t = \tau cosh \vartheta[/itex]

[itex] tanh \vartheta = x/t [/itex]

[itex] \vartheta = arctan x/t [/itex]

Now, let's get from [itex] x/t[/itex] to [itex] v/c [/itex]
 
Last edited:
  • #13
Everyone, go to post 12 above. All else is superfluous. I could not edit, correct and re-edit in time.

Doc
 
  • #14
stevmg said:
Using the equations [itex](ct)^2 - x^2 = \tau^2[/itex] I KNOW there has to be a more direct and intuitive way to prove the hyperbolic assertions.
This is the equation of a specific hyperbola in the diagram, so it's not a good starting point if you want to find a result that's valid for all events in the diagram.

stevmg said:
Everyone, go to post 12 above. All else is superfluous.
I would say that about post 8. I think you should start by writing down the definitions of sinh, cosh and tanh, and using them to prove that cosh2x-sinh2x=1 for all x. Then study every equality in post 8 carefully. I don't think there is a way to do it that is simpler than the way I did it.
 
  • #15
stevmg said:
To yiuop and Fredrik -

Using the equations [itex](ct)^2 - x^2 = \tau^2[/itex] I KNOW there has to be a more direct and intuitive way to prove the hyperbolic assertions.

Using the timelike diagram where the ct axis is up and down and the x-axis ls the left and right, [itex] sinh \vartheta = x/\tau [/itex] and [itex] cosh \vartheta = ct/\tau [/itex] there has got to be a way of working [itex] v [/itex] and [itex] c [/itex] into this.

[itex] x = \tau sinh \vartheta [/itex] and [itex] t = \tau cosh \vartheta[/itex]

[itex] tanh \vartheta = x/t [/itex]
[itex] \vartheta = arctan x/t [/itex]

Now, let's get from [itex] x/t[/itex] to [itex] v/c [/itex]
Perhaps we should start with two images:
268761_f520.jpg

271083_f520.jpg


Source: http://hubpages.com/hub/Hyperbolic-Functions
 
  • #16
Maybe this is a place to start (using c=1 as recommended by Fredrick :wink: ).

The hyperbolic identity (similar to the trigonometric identity) is defined as:

[tex]cosh^2(\phi) - sinh^2(\phi) = 1[/tex]

From this we can create:

[tex]\tau^2 cosh^2(\phi) - \tau^2sinh^2(\phi) = \tau^2[/tex]

We also know the Minkowski metric is:

[tex]t^2 -x^2 = \tau^2[/tex]

So we can say:

[tex]\tau^2 \, cosh^2(\phi) - \tau^2 \, sinh^2(\phi) = t^2 -x^2[/tex]

From the above equation it is easy to make the natural associations:

[tex]\tau \, cosh (\phi) = t [/tex]

[tex]\tau \, sinh(\phi) = x[/tex]

and:

[tex]v = \frac{x}{t} = \frac{\tau \, sinh(\phi)}{\tau \, cosh(\phi)} = tanh(\phi)[/tex]

Now we already know that gamma is equal to [itex]t/\tau[/itex] so from [itex]\tau \, cosh(\phi) = t [/itex] we can get [itex]cosh (\phi) = (t/ \tau) \rightarrow \, cosh(\phi) = \gamma [/itex] and from [itex]\tau \, sinh(\phi) = x[/itex] we get [itex]sinh(\phi) = (x/\tau) = \gamma(x/t) = \gamma v[/itex] which is the celerity or proper velocity (w). The mysterious quantity [itex]\phi[/itex] is defined as the hyperbolic velocity angle or rapidity and is obtained from the preceding equation as [itex]\phi = asinh(w)[/itex]. The rapidity can also be obtained in various other forms from any of the earlier equation that include [itex]\phi[/itex].
 
Last edited:
  • #17
stevmg said:
To yiuop and Fredrik -

Using the equations [itex](ct)^2 - x^2 = \tau^2[/itex] I KNOW there has to be a more direct and intuitive way to prove the hyperbolic assertions.

Using the timelike diagram where the ct axis is up and down and the x-axis ls the left and right, [itex] sinh \vartheta = x/\tau [/itex] and [itex] cosh \vartheta = ct/\tau [/itex] there has got to be a way of working [itex] v [/itex] and [itex] c [/itex] into this.

[itex] x = \tau sinh \vartheta [/itex] and [itex] t = \tau cosh \vartheta[/itex]

[itex] tanh \vartheta = x/t [/itex]

[itex] \vartheta = arctan x/t [/itex]

Now, let's get from [itex] x/t[/itex] to [itex] v/c [/itex]

Passionflower said:
Perhaps we should start with two images:
268761_f520.jpg

271083_f520.jpg


Source: http://hubpages.com/hub/Hyperbolic-Functions

Passionflower - I have seen those diagrams at hubpages before

How does that get one from
[itex] x = \tau sinh \vartheta [/itex] and [itex] t = \tau cosh \vartheta[/itex]

[itex] tanh \vartheta = x/t [/itex]

[itex] \vartheta = arctan x/t [/itex]

Now, let's get from [itex] x/t[/itex] to [itex] v/c [/itex]?

Make the connection for me. Right now all I see is "true - true - not related."

Doc
 
  • #18
yiuop, I again remove the quotes so that we have the text in "real time" and we can work with it, is that OK? You liked that as I used it before and also I tried your "quotes within quotes" and it didn't work for me, as nothing ever does.

The hyperbolic identity (similar to the trigonometric identity) is defined as:

[tex]cosh^2(\phi) - sinh^2(\phi) = 1[/tex] - OK, even though I am not being sarcastic, that's rocket science!

From this we can create:

[tex]\tau^2 cosh^2(\phi) - \tau^2sinh^2(\phi) = \tau^2[/tex] beautiful!

We also know the Minkowski metric is:

[tex]t^2 -x^2 = \tau^2[/tex] Thank God for starthaus!

So we can say:

[tex]\tau^2 \, cosh^2(\phi) - \tau^2 \, sinh^2(\phi) = t^2 -x^2[/tex] by algebraic substitution

From the above equation it is easy to make the natural associations:

Now, where does this come from?
[tex]\tau \, cosh (\phi) = t [/tex]

[tex]\tau \, sinh(\phi) = x[/tex]

Oh yes! I think I postulated that in my prior post but that was based on the geometry of [itex](ct)^2 - x^2 = \tau^2[/itex]. That's the basic geometric definitions of hyperbolic cosines and sines.

and:

[tex]v = \frac{x}{t} = \frac{\tau \, sinh(\phi)}{\tau \, cosh(\phi)} = tanh(\phi)[/tex]
By algebraic division and equality. Good, so far we are on the exact same page. Ge that, folks? [itex]v = x/t[/itex]

Now we already know that gamma is equal to [itex]t/\tau[/itex] How did we get this?

is that because [itex]\gamma = 1/\sqrt{1 - v^2)}[/itex] We are assuming c = 1 for ease of calculations
= [itex]t/\sqrt{t^2 - x^2)}[/itex]
= [itex]t/\tau[/itex]

THERE! I was right! It is s-o-o-o nice to be such a genius!


so from [itex]\tau \, cosh(\phi) = t [/itex] we can get [itex]cosh (\phi) = (t/ \tau) \rightarrow \, cosh(\phi) = \gamma [/itex] and from [itex]\tau \, sinh(\phi) = x[/itex] we get [itex]sinh(\phi) = (x/\tau) = \gamma(x/t) = \gamma v[/itex] which is the celerity or proper velocity (w). The mysterious quantity [itex]\phi[/itex] is defined as the hyperbolic velocity angle or rapidity and I already knew this. starthaus got me going on this - a way of adding up velocities in hyperbolic functions directly is obtained from the preceding equation as [itex]\phi = asinh(w)[/itex]. The rapidity can also be obtained in various other forms from any of the earlier equation that include [itex]\phi[/itex].

Now, Passionflower, starthaus, DrGreg, et al... This is what I have been looking for!

You have to approach me like I'm four years old and assume I know less than nothing, because that's true. What is obvious to you and unspoken is totally oblivious to me. But guess what now? It isn't any more because I was brought from A [tex]\rightarrow[/tex] B and could understand every algebraic step on the way and it was intuitive. I AM teachable. Lot of work, but it can be done!!

You know something, if I ever got to learn this stuff I would be the best damn teacher in the world because I would know every nook and cranny and blind alley the students would go through and would know how to get them out of it. When I taught medicine, I was the same way. I read every pertinent text and made every mistake you could make so when my students and residents made mistakes I was already there, knew what they did wrong, why they did it wrong and got them going on the right track.

Such is the price of a learning disability, when you get it right, you really get it right.

yiuop, DrGreg, Passionflower, starthaus - thanks for your help. Couldn't have done it without you.

Doc
 
  • #19
yuiop said:
Maybe this is a place to start
Start doing what? It's not clear what you're trying to do in this post. Are you trying to explain how someone who hasn't yet done the calculations I did in #8 might have guessed that [itex]\tanh\phi=v[/itex] would be a useful definition?

yuiop said:
We also know the Minkowski metric is:

[tex]t^2 -x^2 = \tau^2[/tex]
This is just the equation of a hyperbola. If you're going to use it, you should explain why it's relevant. The Minkowski metric has the property that such hyperbolas are preserved by linear transformations that preserve the metric (i.e. Lorentz transformations), but is that relevant for what you're trying to do?
 
Last edited:
  • #20
I don't see how those diagrams could be helpful. Doc, what you want to prove is a simple algebraic fact. You just want to show that the definition [itex]\tanh\phi=v[/itex] implies [itex]\cosh\phi=\gamma[/itex] and [itex]\sinh\phi=\gamma v[/itex]. The easiest way to do that is to do exactly what I did in #8. I explained every equality in that post, so there's no reason for you to ignore it.
 
  • #21
Fredrik - not ignoring it. Have downloaded it onto paper and will pour over it.

Do not avoid your input into my future threads or postings. Trust me - all is appreciated and so is yours (most highly, I say.)

And thanks for calling me "Doc" - reminds me of the good old days when the troops called me that as I was their "father protector."

Doc
 
  • #22
Fredrik -

I have not ignored what you posted and have included it below in all germane posts regarding the subject from you. I will place the quotes in this post and then, in the following post (so that I can get PF to "swallow it" without kicking me out as too long) I will discuss it.

Fredrik said:
Stevemg, the Minkowski bilinear form g (I won't call it a scalar product, or an inner product, since it isn't positive definite) is defined by

[tex]g(x,y)=x^T\eta y[/tex]

I usually take

[tex]\eta=\begin{pmatrix}-1 & 0 & 0 & 0\\ 0 & 1 & 0 & 0\\0 & 0 & 1 & 0\\0 & 0 & 0 & 1\end{pmatrix}[/tex]

to be the definition of [itex]\eta[/itex], but you could choose the opposite sign if you want to. This wouldn't change any of the physics. It would just change a few signs here and there. The two different "norms" (they're not really norms, since they're not positive definite) are just different choices of what sign to use in the definition of [itex]\eta[/itex].

My definition above is more common in GR books, and the opposite sign is more common in QFT books. Some people prefer to write [tex]\langle x,y\rangle[/tex] instead of g(x,y).

Since you have taken an interest in the concept of rapidity, you may find the notes I made for myself useful. Note that I'm using units such that c=1. Keeping that c around is like putting something sharp in one of your shoes before you go out running. It's a pain with no benefits. This is what I wrote in my notes:

In special relativity, we define the rapidity [tex]\phi[/tex] of a particle moving with velocity v by [tex]\tanh\phi=v[/tex]. Note that the rapidity and the velocity are approximately the same when the velocity is small. To be more precise, in the limit [tex]\phi\rightarrow 0[/tex], we have [tex]v=\phi+O(\phi^2)[/tex].

The concept of rapidity is useful because it's easier to "add" rapidities than velocities. If the velocity of frame B in frame A is v, and the velocity of a particle in frame B is v', then the velocity of the particle in frame A is

[tex]v\oplus v'=\frac{v+v'}{1+vv'}[/tex]

If the rapidity of frame B in frame A is [tex]\phi[/tex], and the rapidity of a particle in frame B is [tex]\phi'[/tex], then the rapidity of the particle in frame A is just [tex]\phi+\phi'[/tex]. This follows from the velocity addition formula above, the definition of rapidity, and the identity

[tex]\tanh(\phi+\phi')=\frac{\tanh\phi+\tanh\phi'}{1+\tanh\phi\ \tanh\phi'}[/tex]

The definition of rapidity also implies that

[tex]\cosh\phi=\gamma=\frac{dt}{d\tau}[/tex]

[tex]\sinh\phi=\gamma v=\frac{dx}{d\tau}[/tex]


Proof:

[tex]v^2=\tanh^2\phi=\frac{\sinh^2\phi}{\cosh^2\phi}=\frac{\cosh^2\phi-1}{\cosh^2\phi}=1-\frac{1}{\cosh^2\phi}[/tex]

[tex]\frac{1}{\cosh^2\phi}=1-v^2=\frac{1}{\gamma^2}[/tex]

[tex]\cosh\phi=\gamma[/tex]

[tex]\sinh^2\phi=\cosh^2-1=\gamma^2-1=\frac{1}{1-v^2}-\frac{1-v^2}{1-v^2}=\frac{v^2}{1-v^2}=\gamma^2v^2[/tex]

[tex]\sinh\phi=\gamma v[/tex]

[tex]\tau=\int\sqrt{dt^2-dx^2}=\int dt\sqrt{1-\dot x^2}=\int\frac{dt}{\gamma}[/tex]

[tex]\frac{d\tau}{dt}=\frac 1 \gamma[/tex]

[tex]\frac{dt}{d\tau}=\gamma=\cosh\phi[/tex]

[tex]\frac{dx}{d\tau}=\frac{dt}{d\tau}\frac{dx}{dt}=\gamma v=\sinh\phi[/tex]

These results enable us to express an arbitrary proper and orthochronous Lorentz transformation as a hyperbolic rotation by an "angle" equal to the rapidity.

[tex]\Lambda=\gamma\begin{pmatrix}1 & -v\\ -v & 1\end{pmatrix}=\begin{pmatrix}\cosh\phi & -\sinh\phi\\ -\sinh\phi & \cosh\phi\end{pmatrix}[/tex]

stevmg said:
Cyosis -

Would this be the correct representation with the hyperbolic functions?


[tex]
\begin{pmatrix}
ct' \\ x'
\end{pmatrix}

=

\begin{pmatrix}
\cosh \(arctanh(v/c) & -\sinh\(arctanh(v/c) \\
-\sinh\(arctanh(v/c) & \cosh\(arctanh(v/c)
\end{pmatrix}

\begin{pmatrix}
ct \\ x
\end{pmatrix}
[/tex]

If this is the case, then all one would have to do is look up the arctanh of v/c or ([tex]\beta[/tex]) and plug that into the cosh and sinh to get your matrix. That can be done on a scientific calculator directly.

Yes, but if you look at my last equation before the quote, I'm sure you'll see an easier way. :smile:

Fredrik said:
Explanation of the equalities:

First line: 1) Definition of rapidity ([itex]\tanh\phi=v[/itex]). 2) Definition of tanh. 3) The identity [itex]\cosh^2x-\sinh^2x=1[/itex], which is easy to derive from the definitions of sinh and cosh. 4) For all real numbers A,B with A non-zero, we have (A+B)/A=1+B/A.

Second line: 1) The result from the first line. 2) Definition of [itex]\gamma[/itex].

Third line: 1) The square root of the result in the second line.

Fourth line: 1) The identity [itex]\cosh^2x-\sinh^2x=1[/itex] again. ([itex]\cosh^2[/itex] should of course be [itex]\cosh^2\phi[/itex]. I left out the variable by accident). 2) The result from the third line. 3) Definition of [itex]\gamma[/itex] and the fact that 1=A/A for all non-zero real numbers A. 4) For all real numbers A,B,C with C non-zero, we have A/C+B/C=(A+B)/C. 5) Definition of [itex]\gamma[/itex].

Fifth line: 1) The square root of the result in the fourth line.

Go to post #23, https://www.physicsforums.com/showpost.php?p=2877552&postcount=23

Doc
 
  • #23
Continued from post #22 above, https://www.physicsforums.com/showpost.php?p=2877463&postcount=22

Fredrik-

I tried numerous times to answer you but PF would not let me continue with symbols, even with small posts as it would lock up and go no further. Very frustrating.

You have the posts I quoted above. I will write out what I was trying to say but "in prose" as the symbols do not work.

Basically we have the equation of a hyperbola tau^2 = t^2 - x^2 (for below, tau = SQRT(t^2 - x^2) with y intercept of +- tau at x = 0. There are no x-intercepts.

sinh phi = x/tau
cosh phi = t/tau, tanh phi = x/t

gamma = 1/SQRT(1 - v^2)

If v = x/t, tanh phi = x/t = v (this is "proven" and not assumed)

Then gamma = 1/SQRT(1 - x^2/t^2) = t/tau

Substituting back into the Lorentz transformations:

x' = xt/tau - v(t/tau)t = xcosh phi -tsinh phi
t' = (t/tau)t - (t/tau)xv = tcos phi -xsinh phi or the more familiar -xsinh phi + tcos phi which can be represented in matrix format which I cannot do as yet on PF.

This makes intuitive sense. For a given tau we have a hyperbola which represents all x's and t's which are different coordinates in different frames of reference but which "go back" to the same "proper time" or "tau." If one were totally inertial and not moving, x would = 0. If one were moving, for the initial FR where x does = 0, then any x and t on the hyperbola would represent the coordinates which would calculate back to that initial inertial frame where x = 0. It would take a speed of x/t = v to get to that point from x = 0 and t = tau.

I hope this makes sense and there are no assumptions other than what Minkowski himself projected that tau^2 = t^2 - x^2.

This is for timelike phenomena. Spacelike should be similar but the t's and x's would be reversed and the hyperboa would be "left and right" and a proper distance would be
distance^2 = x^s - t^2 and x = tau cosh phi while t = tau sinh phi.

I hope I am right.

Doc
 
Last edited:
  • #24
That was a freakin' bear!

By the way, what is this "[tex]\Lambda[/tex]" from the above post with the matrices? I've seen it before but haven't got a clue what it represents!

stevmg
Doc
 
  • #25
[itex]\Lambda[/itex] is a common symbol for the matrix form of the Lorentz transformation.
 
  • #26
[itex]\Lambda[/itex] is the uppercase version of the greek letter "lambda", and I used it the way jtbell described. Regarding your problem with "symbols", you're probably just experiencing the latex preview bug that's bothering all of us the same way. The only workaround is to refresh and resend after each preview. (I haven't read the rest of your post yet. Give me a few minutes to do that).
 
  • #27
stevmg said:
Basically we have the equation of a hyperbola tau^2 = t^2 - x^2 (for below, tau = SQRT(t^2 - x^2) with y intercept of +- tau at x = 0. There are no x-intercepts.
That's the equation of a hyperbola alright, but I don't know why you're considering a specific hyperbola.

stevmg said:
sinh phi = x/tau
cosh phi = t/tau, tanh phi = x/t
What are you doing here? Are these supposed to be definitions, calculations, or guesses?

stevmg said:
gamma = 1/SQRT(1 - v^2)
This is the standard definition of [itex]\gamma[/itex].

stevmg said:
If v = x/t, tanh phi = x/t = v (this is "proven" and not assumed)
It looks like you're just writing down the same equalities you wrote down earlier, so I don't see what you have proved, or how you did it.

stevmg said:
Then gamma = 1/SQRT(1 - x^2/t^2) = t/tau
Here it looks like you have used v=x/t, [itex]t^2-x^2=\tau^2[/itex], t>0 and [itex]\tau>0[/itex], but that means that your "proof" is only valid for the specific event (t,x) that lies at the t>0 intersection of the line t=x/v and the hyperbola [itex]t^2-x^2=\tau^2[/itex].
 
Last edited:
  • #28
I tried refreshing, resending, etc and the website would freeze and not let me go further. It wasn't just a mix-up of symbols as used to happen. Maybe it is my proxy server.

I used the general equation of a hyperbola. In this sense
t2 - x2 = t'2 - x'2 which is from Einstein's book on Relativity. This demonstrates the invariance of time2 - distance2. This should be proper time for a given family of points in (t, x). i.e., all the points on a hyperbola that satisfy a specific invariance as defined should all give the same [itex]\tau^2[/itex].

Oh well, enough for now. My brain cannot absorb any more today. As I have said before in prior posts, I have nobody local to talk to (that can give quick fixes) or go over scenarios with so it is a very stoccato way of learning and tedious.
 
Last edited:
  • #29
The reason I think you shouldn't be looking at specific hyperbolas, fancy diagrams, results from special relativity, etc., is pretty simple. All you want to do is to prove the following claim:

[tex]\text{If }\tanh\phi=v\text{ and }\gamma=\frac{1}{\sqrt{1-v^2}}\text{, then }\sinh\phi=\gamma v\text{ and }\cosh\phi=\gamma.[/tex]​

This requires absolutely no knowledge of special relativity; it doesn't involve physics in any way. In fact, it would be wrong to use any knowledge of physics when you do this.

You only need to use the identities

[tex]\tanh x=\frac{\sinh x}{\cosh x}[/tex]

[tex]\cosh^2x-\sinh^2x=1[/tex]

The first is just the definition of tanh, and the second is very easy to prove from the definitions of sinh and cosh. The definitions are

[tex]\sinh x=\frac{e^x-e^{-x}}{2}[/tex]

[tex]\cosh x=\frac{e^{x}+e^{-x}}{2}[/tex]
 
  • #30
Fredrik said:
The reason I think you shouldn't be looking at specific hyperbolas, fancy diagrams, results from special relativity, etc., is pretty simple. All you want to do is to prove the following claim:

[tex]\text{If }\tanh\phi=v\text{ and }\gamma=\frac{1}{\sqrt{1-v^2}}\text{, then }\sinh\phi=\gamma v\text{ and }\cosh\phi=\gamma.[/tex]​

This requires absolutely no knowledge of special relativity; it doesn't involve physics in any way. In fact, it would be wrong to use any knowledge of physics when you do this.

You only need to use the identities

[tex]\tanh x=\frac{\sinh x}{\cosh x}[/tex]

[tex]\cosh^2x-\sinh^2x=1[/tex]

The first is just the definition of tanh, and the second is very easy to prove from the definitions of sinh and cosh. The definitions are

[tex]\sinh x=\frac{e^x-e^{-x}}{2}[/tex]

[tex]\cosh x=\frac{e^{x}+e^{-x}}{2}[/tex]

Yea, I got that:

If one chooses [itex]tanh \phi = v[/itex] We will use v instead of [itex]\beta[/itex] because it is easier to write

then [itex] tanh \phi = sinh \phi/cosh \phi = v[/itex]
[itex] cosh^2 \phi - sinh^2 \phi = 1 [/itex]
[itex] 1 - tanh^2 \phi = 1/cosh^2 \phi [/itex]
[itex] 1 - v^2 = 1/cosh^2 \phi [/itex]
[itex] cosh \phi = 1/\sqrt{(1 - v^2)} = \gamma [/itex]
[itex] sinh^2 \phi = 1 - cosh^2 \phi [/itex], [itex] sinh^2 \phi = 1 - \gamma^2 = (1 - 1 + v^2)/(1 - v^2) [/itex]
[itex] sinh \phi = v/\sqrt{(1 - v^2)} = v\gamma [/itex]

Thus the standard hyperbolic Lorentz transformations are proven.
[itex] x' = xcosh \phi - tsinh \phi [/itex]
[itex] t' = tcosh \phi - xsinh \phi [/itex]

[itex] x' = \gamma x - \gamma vt [/itex]
[itex] t' = \gamma t - \gamma xv [/itex]

Thus, choosing [itex]tanh \phi = v[/itex] is an educated guess which makes things work out. We started with the conclusion and worked backward through logically equivalent statements until we arrived at the original Lorentz transformations. This technique is oft used in mathematics but one must be careful so as to use a series of logically equivalent statements.

What I mean by logically equivalent statements is

A [itex]\Leftrightarrow[/itex] B
B [itex]\Leftrightarrow[/itex] C
C [itex]\Leftrightarrow[/itex] D
D [itex]\Leftrightarrow[/itex] E

In this case we start at D [itex]\Leftrightarrow[/itex] E and work backwards to A [itex]\Leftrightarrow[/itex] B

So, I was mistaken in looking for a hyperbola to match the hyperbolic functions.

According to yuiop on post # 11: https://www.physicsforums.com/showpost.php?p=2876438&postcount=11
one does exist but I probably read his post wrong.

stevmg
Doc
 
Last edited:
  • #31
Looks good.
 
  • #32
Thank you Fredrik.

I am saving this below so I can, in the future, write a damn matrix. I took out the quotes so I can see the symbols:

From Fredrik;2678572

[tex]\Lambda=\gamma\begin{pmatrix}1 & -v\\ -v & 1\end{pmatrix}=\begin{pmatrix}\cosh\phi & -\sinh\phi\\ -\sinh\phi & \cosh\phi\end{pmatrix}[/tex]
**************************************************************

I'm saving this so I can, in the future, write a square root with a long top bar as well as write fractions.

Fredrik;2877945

[tex]\text{If }\tanh\phi=v\text{ and }\gamma=\frac{1}{\sqrt{1-v^2}}\text{, then }\sinh\phi=\gamma v\text{ and }\cosh\phi=\gamma.[/tex][/indent]

[tex]\tanh x=\frac{\sinh x}{\cosh x}[/tex]

[tex]\sinh x=\frac{e^x-e^{-x}}{2}[/tex]

[tex]\cosh x=\frac{e^{x}+e^{-x}}{2}[/tex]
***************************************************************

Let's try it:

[tex]
\text{If }\tanh\phi=v\text{ and }\gamma=\frac{1}{\sqrt{1-v^2}}\text{, then }\sinh\phi=\gamma v\text{ and }\cosh\phi=\gamma.
[/tex]
***************************************************************
Fredrik-

Would this be proper notation?

[tex]\begin{pmatrix}x' \\ t'\end{pmatrix} = \Lambda \begin{pmatrix}x \\ t\end{pmatrix} [/tex]

[tex]\begin{bmatrix}x' \\ t'\end{bmatrix} = \Lambda \begin{bmatrix}x \\ t\end{bmatrix} [/tex]

Doc
 
Last edited:
  • #33
To anyone -

Would this be proper notation?

[tex]\begin{pmatrix}x' \\ t'\end{pmatrix} = \Lambda \begin{pmatrix}x \\ t\end{pmatrix} [/tex]

[tex]\begin{bmatrix}x' \\ t'\end{bmatrix} = \Lambda \begin{bmatrix}x \\ t\end{bmatrix} [/tex]

stevmg
 
  • #34
starthaus -

If one takes the Lorentz equatons:

[tex] x' = \gamma(x - vt)[/tex]
[tex] t' = \gamma(t - vx)[/tex] where v really is [tex] \beta [/tex]

and substitutes them into the Minkowski identity

[tex] c^2t'^2 - x'^2 = c^2t^2 - x^2 [/tex]

We wind up with an identity (either way) and a hyperbola is suggested

[tex]c^2t^2 - x^2 = a^2 [/tex]

This [itex] a^2 [/itex] looks like [itex] \tau^2 [/itex]. Is that so? It doesn't make sense that [itex] a [/itex] would be proper time.

You previously stated an equation [tex]c^2 d \tau^2 = c^2 dt^2 - dx^2 [/tex]

This is clearly not the same thing.

What does the [itex] a^2 [/itex] represent? Is there anyway of relating it to [itex] v [/itex] and [itex] \gamma [/tex]?
 
  • #35
Passionflower:

These two images you posted (post 15) represent hyperbolas.
268761_f520.jpg

271083_f520.jpg


Source: http://hubpages.com/hub/Hyperbolic-Functions[/QUOTE]

In your thinking what do they represent in SR? There's an invariance but how do we use that?

stevmg
 

Similar threads

  • Special and General Relativity
Replies
33
Views
2K
Replies
14
Views
975
  • Special and General Relativity
Replies
17
Views
1K
  • Special and General Relativity
Replies
2
Views
857
  • Special and General Relativity
Replies
5
Views
1K
  • Special and General Relativity
2
Replies
36
Views
3K
  • Special and General Relativity
Replies
7
Views
1K
  • Special and General Relativity
Replies
6
Views
1K
  • Special and General Relativity
Replies
29
Views
1K
  • Special and General Relativity
Replies
14
Views
1K
Back
Top