Focusing Issues

[DOUG66]

Advertisement (gone after registration)

I do not have a source but as Pete has explained, the laser beam is now passing through

the telescope and being magnified and the exit from the viewfinder is larger than the 1.25mm

exit hole of the laser. I can however give you an example: the old 135mm Elmarit which has

attached telescopic specs cat. no. 11829. The magnification of the specs is 1.5 which does

two things it allowe the 90mm frame lines in the viewfinder to act as 135mm frame lines and

it increases the effective base of the rangefinder to enable more accurate focusing.

Focusing with a laser is exploiting the same principal.

Doug.

[Guest mc_k]

It seems clear to me. In 'normal' usage light travels through the viewfinder and into your eye and is magnified (diminished) by a factor of 0.68 owing to the optics of the viewfinder.

 

But when shining the laser through the viewfinder onto a subject the light is travelling in the opposite direction and the viewfinder's optics will magnify the laser by 1/0.68 = 1.47.

...

 

this is clear, but wasn't my question.

 

But I get it, the sensitivity has only to do with the scaling of the distance between the two images, whether seen through the viewfinder or projected by the laser. And the exit hole of the laser has no effect on the sensitivity of the setup, any more than the size of an ink dot you focus on affects the sensitivity.

[Guest mc_k]

...

 

o.k., so you would need to add a better than 2X magnifier to compete with the laser in this aspect. How do you take advantage of this? Can you test focus at, say, distances of 12 ft. or more? What would you use for a target? What is the limiting factor at a distance: lens resolution, target, ability to focus?

[DOUG66]

Hi McKeogh K,

On some of the first tests that I carried out M.T. (private communication)

pointed out to me that the sensor plane should be maintained at the 30 degrees

to the chart and rotating the camera altered this angle and introduced an error.

However the error is different depending on which side of the chart is furthest

from the camera, as the camera is always rotated anti- clockwise. I described

this wrongly as an apparent change in depth of focus, it is really a reduction

in error. The micrometer stage is the better solution keeping the angle constant

in the horizontal set-up. The vertical set-up is much easier.

 

Your other point regarding testing at greater distances. I am working on this.

The first step is the small laser alignment device which clips into the flash shoe.

This is shown in the two photos I posted already in this thread. The idea is to

align the laser beam exactly above the axis of the lens and use this to maintain

the required 30 degree angle. I have been testing this device today. The first

problem is adjusting the lateral alignment of the laser to bring it exactly over

the lens axis. As you will appreciate the precision required in the making of

this device is extremely high. I have discovered that it also requires an angular

adjustment as the alignment of the flash shoe though good, for the distances required

is not good enough. So it is back to the drawing board.

[Guest mc_k]

...with the RHS side of the chart furthest from the camera. This is

on account of the paralax shift of the range finder of the Leica.

This artificially reduces the aparant depth of focus of the lens.

Sloping the chart the other way increases the depth of focus.

This is due to the slight angular offset of the sensor plane as

the camera is focussed on near objects...

 

o.k. I get this but why not explicitly say it minimizes an error

[Guest mc_k]

sorry, didn't see your post! Thanks, I appreciate the info.

[Guest mc_k]

Advertisement (gone after registration)

apparently, if you try to shoot a ruler/chart at x degrees (to the lens axis) and you are off by +-y degrees, the percent error in your result is

 

[1 - (cos x) / (cos (x+-y)] X 100

 

E.g. if you are off by 5 degrees your result will be off by 4-6 % (30 deg. target) or 8-10 % (45 deg. target) or 13-18 % (60 deg. target).

 

So you can see that how you slope the target makes a little bit of a difference, and your angle makes even more of a difference.

[farnz]

apparently, if you try to shoot a ruler/chart at x degrees (to the lens axis) and you are off by +-y degrees, the percent error in your result is

 

[1 - (cos x) / (cos (x+-y)] X 100

 

E.g. if you are off by 5 degrees your result will be off by 4-6 % (30 deg. target) or 8-10 % (45 deg. target) or 13-18 % (60 deg. target).

 

So you can see that how you slope the target makes a little bit of a difference, and your angle makes even more of a difference.

Should that be [1 - {(cos x) / (cos (x+-y)}] X 100?

 

Pete.

[Guest mc_k]

Should that be [1 - {(cos x) / (cos (x+-y)}] X 100?

 

Pete.

 

It's worse than that, I have an extra " ( ". It should be

 

[1 - (cos x) / cos (x+-y)] X 100

 

(And you could also add your own absolute value sign around the whole thing.)

 

In other words, if you shoot a ruler/chart at a "bad" angle, the bad result times the cosine of the bad angle = the true result times the cosine of the true angle. This you can get from a diagram with the two lens axes, the two targets sloped in the different directions, and the two critical focus planes.

 

The takeaway for me is that if you are a little sloppy, there is an error just from setting up the target.

[Nicoleica]

OK. I think I understand the equation now. If you don't set it up perfectly, you'll absolutely get more errors 'cos of this and 'cos of that.

[Guest mc_k]

OK. I think I understand the equation now. If you don't set it up perfectly, you'll absolutely get more errors 'cos of this and 'cos of that.

 

o.k., a laser wit by all sines

[Guest mc_k]

I guess that covers errors from setting up the target and from focusing the target--how about from reading the target? Does anyone else use the moire? I've seen this mentioned in a couple of places.

[DOUG66]

Hi,

I think I might have come up with a method of setting up the test chart accurately

which I will publish here when I get it checked out. I have been busy in the workshop

making the components. One of the components requires a extremely accurate

30 deg angle.. I have just measured this component and find that the angle is 30.017 deg.

an error of 0.06%. However when I put this figure into the formula it gives an error of

0.02%. Since errors are additive , how can an error in a component result in a smaller

total error in the overall test. Am I missing something ?

Doug.

[Guest mc_k]

If you're cutting out paper right triangles, and making a mess of one of the other angles, if we use your numbers (30 vs. 30. 017 degrees) the percent error in the hypotenuse is less than the percent error in the angle.

 

If I've made a mistake, it should be easy enough for someone to supply the correct formula. Note the angle in the formula is w.r.t. the lens axis.

[jaapv]

Ummm. I use the method: I take photographs. If I don't like the focussing results I take a close-up of a glass of wine before drinking it - if the wine is blurred I send the lens in. Then I drink the wine. Tastes much better than cosines.

 

The moral being: The owner of the lens only needs to notice whether something is wrong. Even if he is mistaken he needs somebody to look at it for his peace of mind. The utmost precision is needed by the technician only.

[Nicoleica]

I use a similar method to Jaap. Except that I drink the wine before taking the photograph. If the image of the subject is more blurred than it appears to be with my eyes, then I find that a further glass of wine normally restores the balance.

[bill]

...is this thread for real?

 

Regards,

 

Bill

[SJP]

...is this thread for real?

 

Regards,

 

Bill

sure, exept the following equation is more convenient to remember obviously:

 

[1 - {(cos x) / (cos (x+-y)}] X 100 = [{tan(x)sin(y)+cos(y)-1}/{cos(y) + tan(x)sin(y)}] X 100

[jaapv]

I use a similar method to Jaap. Except that I drink the wine before taking the photograph. If the image of the subject is more blurred than it appears to be with my eyes, then I find that a further glass of wine normally restores the balance.

Ah yes - the double blur method. Quite valid, although I find it works best using Otard VSOP.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[Guest mc_k]

...The utmost precision is needed by the technician only.

 

well it all depends on what you are shooting and what you are trying to accomplish, doesn't it.