Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Limit for acceptable collimation errors?

collimation equipment
  • Please log in to reply
45 replies to this topic

#1 ronnekex

ronnekex

    Sputnik

  • -----
  • topic starter
  • Posts: 39
  • Joined: 16 Nov 2008

Posted 10 December 2018 - 03:53 PM

I have recently tested about 10 binoculars in the price range under 140$ with 8x-10x magnification and 40-50 mm aperture (full report will come soon). Vertical collimation was good on all of them. However, I was  surprised by how many of them had large horizontal collimation errors:

  •   6 of them had horizontal errors between 2° and 3°.
  •   1 of them had a horizontal error of 1.3° (convergent)
  •   3 of them had an errors of 0.6° or less.

Is this normal? What is an acceptable limit (in this price range) for collimation errors on binoculars that will be used for night sky observations? I know this may vary between individuals, but I am looking for a limit that would be acceptable for, say 60 to 80 % of all observers.

 

What I believe to know
In general, large collimation errors will cause problems for the brain with assembling the image correctly. When the collimation error is moderate and the observed subject contains many clear details and patterns, our brain's center of vision will usually be able to assemble the immage correctly. However, many astronomical objects and star fields contain relatively few details, and I therefore expect the tollerable level of collimation errors to be smaller for astro use than for most other uses.

 

Collimation errors may also cause physical strain to the eyes, since they are not used to being pointed in two different directions

 

I made a search on the forum and found the following posts by EdZ's, refering to a book by J. W. Seyfried, see https://www.cloudyni...ted/#entry68300. When converting the limts given by EdZ into apparent angles (those seen in the eyepieces) I get the approximate limits listed below:

  • Maximum 0.6° vertical error (one eye up, the other down)
  • Maximum 1.4° horizontal convergence (eyes crossed over)
  • Maximum 0.9° horizontal divergence (eyes spread apart)

However, EdZ also writes that he himself can hardly tolerate one half of the errors above. Also, I suspect that the tollerance is lower when whatching a sparse star field then when whatching daytime objects with lots of clear patterns.
 

An easy way to measure binocular collimation errors
Collimation error for a binocular means that the direction towards the image appears different for the left and the right eye. This is illustrated in figure a) below, where the image seen by the left and the right eye are shown in red and cyan, respectively.

collimation.jpg

 

Collimation errors should not be confused with deviating fields of view, which is illustrated in figure b). In this case the objects in the image appear in the same direction for both eyes, so they are easilly combined into one single image by the brain. The non-overlapping areas on each side contribute to a wider field of view, which is fine, allthough we have only one-eyed sight to these regions. A horizontal deviation in FOV of 5-10 % of the field diameter is not uncommon, and will usually not cause problems for the observer.

 

Figures c) and d) above illustrate how the collimation can be tested. The binoculars should be placed on a stand or on a steady surface where their direction can be adjusted while looking through them. Point the binoculars towards a clearly visible horizontal or vertical line at least 30 m away. The top or side of a large house may be suitable.

 

Now move your eyes 30 to 50 cm behind the eyepieces and locate the line through one of them. (Here, you may need to adjust focus to see a well defined exit pupil). Fine adjust the direction of the binocular until the observed line is alligned inside and outside the eyepiece, as illustrated in figure c) and (roughly) for the right eyepiece in figure d).

 

Now, look through the other eyepiece. If the line is not alligned inside and outside this eyepiece, then there is a collimation error in the direction perpendicular to the line. The magnitude of the error equals the apparent angle between the inside and outside line segments. This angle can be found by measuring the apparent separation of the lines with a ruler that is placed 57 cm from your eyes. At this distance, 10 mm on the ruler spans an angle of 1°.

 

The example in the figure shows a perfect vertical collimation, while the horizontal collimation deviates by 2°.

 

Erlend


Edited by ronnekex, 11 December 2018 - 03:52 PM.

  • djeber2, Stellarfire and Nate1701 like this

#2 photoracer18

photoracer18

    Fly Me to the Moon

  • *****
  • Posts: 5,043
  • Joined: 02 Sep 2013
  • Loc: Martinsburg, WV

Posted 10 December 2018 - 04:01 PM

My rule of thumb was always if my head hurt looking thru them they were too far out. So I only buy ones I can test or top brands. I think I have 5 pairs give or take one. All are either Vixen, Pentax, or Nikon (one older top Celestron version, the 10x42 Nobles). Not that they are the best or the only ones. It just worked out that way. I worked for a Celestron/Meade/Pentax/Zeiss dealer for a time so I got to compare them a lot. I also got to see a good number of used bad ones. If they were worth it we sent them to a specialist (retired or passed away now) and got them recollimated. Never tried fixing them myself although I did everything else.

Edited by photoracer18, 10 December 2018 - 04:03 PM.


#3 TOMDEY

TOMDEY

    Voyager 1

  • *****
  • Posts: 11,631
  • Joined: 10 Feb 2014
  • Loc: Springwater, NY

Posted 10 December 2018 - 05:57 PM

Some of the literature allows up to 2.7 deg horizontal mismatch (almost entirely convergent)...  Almost makes me think some shops intentionally bias their builds on the convergent side... by a Lot?!  But truly-comfortable numbers are more like 3.4 arc-min vertical, 3.4 horizontal divergent, 8.6 arc-min convergent.  I'm (admittedly) guessing that in that low price range, maybe they are building them to the loosest tolerances found in the literature? You are using an instrument that does not depend upon the user's own fusion judgment?  Some instruments get around that issue, with creative pupil-sharing displays, etc.

 

Ummm... You are using a Collimator, right?  I'm not trying to be a jerk; but worked at B&L, back where they were still a true optics company.  They, of course, had walk-up collimators, seemed all over the place.

 

A biggish telescope, with a reticule at focus, makes a Great collimator!   Tom  



#4 Rich V.

Rich V.

    Cosmos

  • *****
  • Posts: 8,954
  • Joined: 02 Jan 2005
  • Loc: Lake Tahoe area, Nevada

Posted 10 December 2018 - 06:10 PM

Good post, Erlend.  Collimation is something we hear a lot about in the Binoculars forum; sadly, perhaps more than we should.  Some binoculars are infamous because of their reputation for poor collimation.

 

Regarding the standards from EdZ you posted:

 

 

I made a search on the forum and found the following posts by EdZ's, refering to a book by J. W. Seyfried, see https://www.cloudyni...ted/#entry68300. When converting the limts given by EdZ apparent angles (those seen in the eyepieces) I get the approximate limits listed below:

 

Maximum 0.6° vertical error (one eye up, the other down)
Maximum 1.4° horizontal convergence (eyes crossed over)
Maximum 0.9° horizontal divergence (eyes spread apart)

 

However, EdZ also writes that he himself can hardly tolerate one half of the errors above. Also, I suspect that the tollerance is lower when whatching a sparse star field then when whatching daytime objects with lots of clear patterns.

 

 

In a 2005 thread, forum member Claudio posted his revised collimation standards that may be more appropriate than the above standards:

 

 

The tolerated apparent alignment errors (= referred to image space, i.e. the images seen through the eyepieces, i.e. magnified by the binocular under test) I have more often found are:

 

Vertical step: 15 arc minutes
Divergence: 20 arc minutes
Convergence: 60 arc minutes

 

But Bill, Cory and anyone who overhauls seriously binoculars would stay well inside these limits. “Tolerances are for those who can’t get it right”, isn’t that so, Bill?

 

A reason more to adopt apparent alignment errors, i.e. data referred to the image space: they are what the eye see through the eyepieces, so they indicate (or SHOULD INDICATE, but let me say again that they are too permissive) the threshold that says what eyes sense. So, in your 16x the axes step of 3’ doesn’t give me an immediate, graphic idea of how much the misalignment is detected by eyes, on the contrary the resultant apparent step (= the step detected on the image space = 3’ x 16 = 48’) make me instantly think that your glass would benefit a lot from an overhaul of Bill or Cory. That vertical step at 100 m is about 9 cm, negligible at most for a 5x, but not for a 16x.

 

The data reported by Seyfried are too “tolerant”. Moreover,they give the same maximum axes errors for too wide ranges of magnification. For instance, in those tables a 5x and a 10x can have the same allowed axes errors, which cannot be.

 

Another point made by Claudio explains why some standards for divergence and convergence have values that appear reversed compared to others:

 

 

When we find tolerance data and they allow a greater error in divergence, certainly those data are referred to the axes.

When tolerance data allow a greater error in convergence, they are certainly referred to the images seen through the eyepieces.

 

Example: let's imagine that our binocular is not collimated. When pointed to a far electricity tower, this target is in the centre of view for the left eye, while for the right eye it is on the same plane (no vertical step) but right from the centre. We are used to say that there is some divergence, which is correct when referred to the target. If we want to be more exhaustive, we have to say “ images diverge, thus axes converge”.

The same for the vertical step. For instance, if we detect a lower image through the right eyepiece, then it means that the right axis points higher than the left one.

 

Some texts express tolerances referred to the images, others to the axes. As explained, just by observing which is the wider tolerance between convergence and divergence we can understand what they are referred to.

 

Food for thought, anyhow...

 

Rich

 


  • Jon Isaacs, Mark9473, ronnekex and 2 others like this

#5 SMark

SMark

    Fly Me to the Moon

  • *****
  • Posts: 5,232
  • Joined: 29 Aug 2011
  • Loc: Clay County, AL USA

Posted 10 December 2018 - 08:59 PM

Another element that I will always have to deal with (and I'm sure others will too...) is that my eyes are poorly aligned to begin with. So I have a much greater tolerance in one direction, and much less tolerance in the other. gottahurt.gif


  • Jon Isaacs, Philip Levine, MeridianStarGazer and 1 other like this

#6 FrankL

FrankL

    Viking 1

  • -----
  • Posts: 818
  • Joined: 30 Jul 2009
  • Loc: Winnipeg, Manitoba Canada

Posted 10 December 2018 - 09:39 PM

See Section 3.4 in this link: https://sites.google...n-of-binoculars


  • Philip Levine likes this

#7 TOMDEY

TOMDEY

    Voyager 1

  • *****
  • Posts: 11,631
  • Joined: 10 Feb 2014
  • Loc: Springwater, NY

Posted 10 December 2018 - 09:52 PM

Another element that I will always have to deal with (and I'm sure others will too...) is that my eyes are poorly aligned to begin with. So I have a much greater tolerance in one direction, and much less tolerance in the other. gottahurt.gif

Great point! We all have some personal "prism"... as the eye doctors call it. And THAT eats up some of our abailable margin! I have exactly one bino that allows "on the fly" adjustment of the vertical and horizonal... The result is Amazing Comfort! Those extra adjustments take some getting used to, but essential when using binos at 200X.



#8 MartinPond

MartinPond

    Fly Me to the Moon

  • *****
  • Posts: 6,530
  • Joined: 16 Sep 2014

Posted 11 December 2018 - 10:16 AM

-------------------

"6 of them had horizontal errors between 2° and 3°."

--------------------

 

Something doesn't seem right.   

First, to clarify:  2-3 degrees out the barrel, or in the view?

 

I think I have only seen one showroom binocular out of dozens

  that had disturbing alignment.

 

I get the IPD distance settled in very well,

 and then look at a distant object with the eyepieces 3-4 inches from my eyes.

That easily multiplies the error while narrowing the view.

Anything that is minor or non-existent that way is impossible to detect

   with the eyes right up to the eyepieces.

Setting the IPD wrong often causes trouble.


  • Philip Levine and MeridianStarGazer like this

#9 Rich V.

Rich V.

    Cosmos

  • *****
  • Posts: 8,954
  • Joined: 02 Jan 2005
  • Loc: Lake Tahoe area, Nevada

Posted 11 December 2018 - 11:41 AM

-------------------

"6 of them had horizontal errors between 2° and 3°."

--------------------

 

Something doesn't seem right.   

First, to clarify:  2-3 degrees out the barrel, or in the view?

 

I read the OP's statement as saying he had converted EdZ's angular tolerances that were originally listed in arc minutes to what was seen in the eyepieces's magnified (apparent) image space. So his 2° error in the image space at 10x, let's say, would be a 12' angular error between the bino's axes.  Still way out from the Bureau of Naval Personnel's 1966 standards of the time.

 

Rich


  • Jon Isaacs and ronnekex like this

#10 MartinPond

MartinPond

    Fly Me to the Moon

  • *****
  • Posts: 6,530
  • Joined: 16 Sep 2014

Posted 11 December 2018 - 02:48 PM

--------------------

So his 2° error in the image space at 10x, let's say, would be a 12' angular error between the bino's axes.  Still way out from the Bureau of Naval Personnel's 1966 standards of the time.

--------------------

 

OK...that's a bit more realistic.

So, for example, out of a 60 degree view, 2 degrees.  

Or..... 3.3% of the field.

 

I seem to remember Bill Cook saying the Naval Standard

   was very small and the Russian Std.  was more like what people tolerate.

If I were setting a US Navy standard for re-collimating

binoculars leaving a repair facility, I suppose  that would be possible,

once you hooked everything up and tweeked to the standard.

 

 

I tried some 8x30s, 7x35s, and 10x50s from 1950----1980 Japan

 Once the IPD is set, I can't get the image to break up, even 

 a whopping 6 inches from my eyes.

 

I have moderately-priced Carson 8x32s ...made in China.  Very nice, in use.

I can get an image split at 3 inches from the eyes.

Scaling the 'stagger distance' to the view, seems like they are actually about

2% of the field off..... about 0.17 degrees.  Not irritating, that I can tell. 

Not sure if alignment is harder for roofs, or if China assembly is looser.

 

I'm still an order of magnitude from the above, though..


Edited by MartinPond, 11 December 2018 - 02:49 PM.


#11 MartinPond

MartinPond

    Fly Me to the Moon

  • *****
  • Posts: 6,530
  • Joined: 16 Sep 2014

Posted 11 December 2018 - 02:57 PM

Of course, if you are looking at trees nearby,

   as in an illustration above, you should expect a deviation...

 

If your eyes are 3 inches apart, and the trees are 100ft away,

 that's :    0.14 degrees difference into your barrels,

    and:   1.4 degrees of your view (!!) at 10x

  Even 300ft away, that' s a whopper.

 

Looking at one point source has the same trouble,

  unless it's 1/2 mile away or more.....



#12 Rich V.

Rich V.

    Cosmos

  • *****
  • Posts: 8,954
  • Joined: 02 Jan 2005
  • Loc: Lake Tahoe area, Nevada

Posted 11 December 2018 - 04:35 PM

Erlend's post also points to another "collimation" issue we don't talk about a lot but it happens.  

 

Collimation errors should not be confused with deviating fields of view, which is illustrated in figure b). In this case the objects in the image appear in the same direction for both eyes, so they are easilly combined into one single image by the brain. The non-overlapping areas on each side contribute to a wider field of view, which is fine, allthough we have only one-eyed sight to these regions. A horizontal deviation in FOV of 5-10 % of the field diameter is not uncommon, and will usually not cause problems for the observer.

 

This is something I personally don't care for one bit: non-concentric FOVs.  To me, it's good reason not to attempt "eyeball" aligning of your binos by trying to match the field stops with target objects rather than merging the objects within the FOV.  You could end up with a considerably misaligned bino while thinking you've "nailed it".  wink.gif

 

As long as these non-concentric FOVs share merged images, I suppose the bino can get a "pass" since there's no eye strain involved or double images.  I still don't like seeing the half-bright "arcs" at the edges of the field caused by FOVs that don't match, though.  I remember Mr. Bill wrestling with this issue when he built his BinoBox BT years ago. Tiny misalignments add up and they're hard to get to the bottom of in your home workshop.

 

Last year when I reviewed three 16x70s, the brand new APM MS ED, though impressed greatly by them otherwise, had this annoying misalignment of fields; the similar Lunts to a lesser degree.  The images through the bino were merged just fine but I didn't like the wall-eyed appearance of the diverging FOVs.  Perhaps I'm too fussy but it just bugged me. They ended going to our much-appreciated friend, Cory Suddarth and being thoroughly looked over and properly collimated as many of us have come to expect from him .  They came back with perfectly relaxed views to my eyes; collimated to a high standard and with nice, matching FOVs as well.  Why couldn't they have come from the factory that way in the first place?  shrug.gif

 

Rich


  • ihf likes this

#13 ronnekex

ronnekex

    Sputnik

  • -----
  • topic starter
  • Posts: 39
  • Joined: 16 Nov 2008

Posted 11 December 2018 - 07:18 PM

This thread has gathered many posts discussing standards for collimation tollerances, which a appreciate!
I guess, however, that most of these standards are adapted for landscape observations. It would be nice if more people can quantify, from their own experience, how large collimation errors you can tollerate without problems when whatching the sky?

 

 

TOMDEY wrote:

"But truly-comfortable numbers are more like 3.4 arc-min vertical, 3.4 horizontal divergent, 8.6 arc-min convergent."

I suspect these strict tollerances are refering to the angles "in the field", and not to the apparent (magnified) angles that you see in the eyepiece. Actually, confusion between in-the-field angles magnified apparent angles seems to be common in these discussioons. (For the eye 3.4 arc-minutes is a very small distance. In darkness with wide open pupils it is probably close to the limit of what we can resolve.)

 

 

Rich V. sited Claudio:

"Vertical step: 15 arc minutes  

Divergence: 20 arc minutes
Convergence: 60 arc minutes"

I found even more discussions by Claudio on various standards here: https://www.cloudyni...ested/?p=323377

 

 

Rich V. also wrote:

"Another point made by Claudio explains why some standards for divergence and convergence have values that appear reversed compared to others: (...)"

This is interresting! The definition of what is convergence and what is divergence depends on what you consider as the "forward" direction; either the direction you are looking in or the direction in which the beam is propagating. If you need to cross your eyes to allign the images seen in the eyepiece (what I call convergent collimation), then the rays emerging out of the two eyepieces from a common object are actually divergent! 

 

 

FrankL linked to a text by Rafael Chamón Cobos:

"See Section 3.4 in this link: https://sites.google.com/site/rchamon/home/sun-images-method-for-collimation-of-binoculars"

Actually, Raphael has switched the meaning of divergence and convergence compared to EdZ, and his tolerances are therefore larger for "divergence" for "convergence". (Raphael's definition of divergence/convergence becomes clear from the discussion below Figure 10.)

 

 

Rich V. discussed deviating FOVs:

"This is something I personally don't care for one bit: non-concentric FOVs.

(...) Last year when I reviewed three 16x70s, the brand new APM MS ED, though impressed greatly by them otherwise, had this annoying misalignment of fields; the similar Lunts to a lesser degree.  The images through the bino were merged just fine but I didn't like the wall-eyed appearance of the diverging FOVs."

I hadn't thought about this, but in cases with little detail in the field it may be a problem that the eyes automatically allignes on the field stops of the two telescopes rather on a few stars inside the field. This can cause irritation, I guess.

 

Erlend


Edited by ronnekex, 11 December 2018 - 07:28 PM.


#14 ronnekex

ronnekex

    Sputnik

  • -----
  • topic starter
  • Posts: 39
  • Joined: 16 Nov 2008

Posted 13 December 2018 - 04:33 PM

Here is a potential argument why convergent collimation errors may NOT be a problem: I am very comfortable with reading a book 50 cm away from my eyes. In that situation my eyes are convergent by 8°. This should be similar to looking through a binocular with 8° apparent convergent collimation error... 

 

So, I suspect that apparent convergent collimation errors of 2 to 3 degrees will not cause problems. I can believe that we are much more sensitive to to vertical errors and diverging horizontal errors, though. 

 

Am I missing something important here?

 

Erlend



#15 Rich V.

Rich V.

    Cosmos

  • *****
  • Posts: 8,954
  • Joined: 02 Jan 2005
  • Loc: Lake Tahoe area, Nevada

Posted 13 December 2018 - 04:52 PM

Yes, our eye's ability to converge to merge at close distances gives us quite a bit of leeway.  If you're going to settle for any degree of binocular misalignment, that's the way to err. 

 

Still, preferably, staring through binoculars into infinity with fully relaxed eyes looking straight ahead is best in the long run for most of us.  I'm sure that's the basis of the standards in the first place.

 

Rich



#16 MartinPond

MartinPond

    Fly Me to the Moon

  • *****
  • Posts: 6,530
  • Joined: 16 Sep 2014

Posted 13 December 2018 - 08:54 PM

Here is a potential argument why convergent collimation errors may NOT be a problem: I am very comfortable with reading a book 50 cm away from my eyes. In that situation my eyes are convergent by 8°. This should be similar to looking through a binocular with 8° apparent convergent collimation error... 

 

So, I suspect that apparent convergent collimation errors of 2 to 3 degrees will not cause problems. I can believe that we are much more sensitive to to vertical errors and diverging horizontal errors, though. 

 

Am I missing something important here?

 

Erlend

 

What's missing is the power...

At 10x, a certain deviation outside will cause 10x the variance at your eyes.

 

Nevertheless, you instinct applies, even at 10x the sensitivity..

 

Just substititute  the example I gave of the image difference nearby,

    and consider your example in terms of the variation somebody actually

    sees and expects for stereo vision.

----> Looking at a bird 50 ft away with eyes 3 inches apart with perfectly aligned

        binoculars gives a variation left-to-right of  2.8 degrees between images/eyes.

 

Now......this actually does cause some double-vision sometimes (I know!),

    but a difference of image planes of 1.5-2 degrees can be easily handled

    ...at least, by me.  Relaxed and perfectly straight ahead is not how our eyes

    work from day to day....they are more flexible than that.

   So..I have real examples that suggest ~3% of a 60-degree field is fine,

     5% is not.     The width of the field is very important in this.

 

You were stating a practical example, and that inspired me to make a practical

   binocular-based example.

 

BTW, the nearby difference is why I think people looking for 8-10x and 

   'close focus' are asking for a splitting headache, and why I use a monocular

    in museums. 


Edited by MartinPond, 13 December 2018 - 08:58 PM.


#17 Philip Levine

Philip Levine

    Viking 1

  • *****
  • Posts: 813
  • Joined: 22 Mar 2007
  • Loc: near Boston, MA

Posted 14 December 2018 - 12:42 PM

Seasons Greetings All,

In the interest of sharing knowledge on collimation, I would like to add an excerpt from Bill Cook's Book: "Binoculars: Fallacy and Fact".  (note: permission was given by the author to use excerpt)

 

Happy Holidays All,

Phil Levine

 

 

 

 

 

 

Attached Thumbnails

  • collimation standards excerpt.png

  • Jawaid I. Abbasi and SMark like this

#18 MartinPond

MartinPond

    Fly Me to the Moon

  • *****
  • Posts: 6,530
  • Joined: 16 Sep 2014

Posted 14 December 2018 - 05:35 PM

Since the number of the "maximum theoretical collimation error"

is proportional to the inverse of the power, it looks like the numbers

refer to the error from the front of the binoculars.

 

At 6 power, then, you should exceed an 

  objective-objective convergence of 27.6 minutes.

 

If I look at a bird 33 feet away and my eyes are 3 inches apart,

   the path difference is 26 arc-minutes.

Trying this across the house onto font targets, I get no apparent strain

  or image effects...

 

If I step closer, to 24feet,

  (36 arc-seconds chart-difference,  3.6 degrees to my eyes)

   I can clearly see the two circles split.

  The targets still line up in my eyes, but if I

   look at a 3' x 3' whiteboard with a dime-sized black circle on it

   (a blackened dime),......the dime splits into two dimes sometimes,

   and there is a little eyestrain (the lack of detail makes it harder to align eyes)

 

So,....since the angles above are "pre-power" and out of the front,

   the standards in the chart seem slightly excessive, but not by much.

   So.....makes sense.   (was usuing 1955 USCO binocs with images that

    stay aligned when eyepieces are 7 inches from my eyes(!)) 

 

Now......the US Navy alignment standards stated there 

   are about 10 times tighter than, so they do seem very extreme.

If you are fixing a pair and have the machinery to see, I suppose you

   might as well tweek things to 2 arc-minutes, but you won't be able

   to tell in use until you get to near 20.



#19 ronnekex

ronnekex

    Sputnik

  • -----
  • topic starter
  • Posts: 39
  • Joined: 16 Nov 2008

Posted 14 December 2018 - 06:37 PM

Martin,

 

You wrote: "What's missing is the power...".   But  I was speaking about 2° to 3° convergent apparent collimation errors. This is what one see in the eyepiece after magnification, so the power is not "missing" here. These apparent angles correspond to watching an object that is between 2 and 1.3 m away without a binocular. I don't believe many would get an headache from that.  (If a 10x binocular had 2° to 3° collimation error before magnification it would have 20° to 30° apparent error seen in the eyepiece, which is horrible. However, this is not what I described in my example.)

 

Note: I have been speaking about apparent errors throughout all of this thread thread. Other posters have also contributed to pointing out the different between apparent (magnified) angles and angles "in-the-field" before magnification.

 

You also wrote: "So..I have real examples that suggest ~3% of a 60-degree field is fine,  5% is not."
This transforms into: 1.8° error is fine,  3° is not. 

 

In your last post you commented about Bill Cook's table (thanks a lot Philipe for posting this!): "Since the number of the "maximum theoretical collimation error" is proportional to the inverse of the power, it looks like the numbers refer to the error from the front of the binoculars." 
I agree in that. When converting Bill Cook's values into apparent angles we get the following requirements:

  Apparent convergent error < 2.6°

  Apparent divergent error   < 1.3°

  Apparent convergent error < 0.66°

 

These are among the more slack requirements I have seen. Note that he has been using these standards "when working with cheaper imports". 

 

 

Best regards! 

Erlend


Edited by ronnekex, 14 December 2018 - 07:25 PM.


#20 MartinPond

MartinPond

    Fly Me to the Moon

  • *****
  • Posts: 6,530
  • Joined: 16 Sep 2014

Posted 14 December 2018 - 11:21 PM

Right....so I think we are aligned amongst ourselves now.

Seems the USN std. is way tighter than anyone could notice,

But the stds suggested by the tables are near the trouble zone.

 

The lack of power or field consideration for the USN probably

has to do with how almost all USN binoculars were 7x50 or 6x30...

Rolling on the waves at sea makes 10x pointless.

Do to fixed machining tolerances and big dimensions, 7x50s

are almost always super-aligned. When they are cross-eyed,

either fixing a cross-threaded front or cleaning the prism shelves 

almost always fixes things. Yet most oldies still have eccentric rings,

in a pinch.



#21 vanderpol4

vanderpol4

    Lift Off

  • -----
  • Posts: 4
  • Joined: 03 Nov 2018

Posted 15 December 2018 - 09:43 AM

There is a nice short online primer on afocal  optics:

 

http://www.photonics...book/v2ch02.pdf

 

It seems from their research Donald Harry Jacobs published his notes on binoculars and prism misalignment here within this text's 487 pages :

-----------------
Fundamentals of Optical Engineering
jacobs, donald
ISBN 10: 0070321507 / ISBN 13: 9780070321502

Published by McGraw-Hill Book Company, i 1st ed.
Description New York : McGraw-Hill, 1943
x, 487 p. : ill. ; 24 cm.
Notes
An expansion of the author's notes for a course in applied optics given under the Engineering, Science, and Management War Training program at George Washington University.

Includes bibliographical references and indexes.

-------------------------

 

They note:

"Jacobs is one of the few authors to discuss this problem ( reference above).

 

Jacobs divides the axes misalignment into three categories : 

(1) misalignments requiring a divergence D of the eye axes to fuse the images ,

(2) misalignments requiring a convergence C 

(3) misalignments requiring a vertical displacement V .

The tolerance on allowable misalignment in minutes of arc is given by Jacobs as
Divergence:   7 . 5 / ( M - 1) ;

Convergence:   22 . 5 / ( M - 1) ;

Vertical:    8 . 0 / ( M - 1) (25)

Note that the tolerance on C , which corresponds to convergence to focus on nearby
objects , is substantially larger than the tolerances on D and V"

 

No e-version of Jacobs' book seems available, he seems to be the scientist who studied this for the war dept, the machine that is an Army on war footing probably spit out the other number set based on his work.

 

I found a copy on abbes books and am ordering it, should be here in a week or two, let you know what I find, it is almost 500 pages, might take a little digging.

 

I thought the photonics handbook primer comments on visual accommodation as a function of distance were were insightful as well.

Clear Skies

Lou

Prism alignment tolerance
Prism alignment tolerance


#22 vanderpol4

vanderpol4

    Lift Off

  • -----
  • Posts: 4
  • Joined: 03 Nov 2018

Posted 15 December 2018 - 12:52 PM

So my opinion,  some things as always, especially in physiological studies, are subject to interpretation.

 

Claim seems to be it is close to diffraction limited vertical. Perhaps that can be true, experienced observers seem to match that criteria on, for example, 20x 100mm night time observation.

I think there is some modern work that can show why that is (just) possible.

 

A)

The human eye under low light with 5mm iris has a calculated edge diffraction resolution limit of 1.2 10^-4 radians or .4 minutes of arc

Maybe surprisingly your brain and rudimentary optic system can actually get to around 4 10^-4 or 1.5 minutes of arc angle ( 1.5 * 1/60 of a degree). 

http://hyperphysics....opt/Raylei.html

http://www.kylesconv...rcminutes#.0004

 

B)

Austin Roorda, Ph.D.,University of Houston College of Optometry, put this deck together online and gave permission for use. It has some great points.


http://voi.opt.uh.ed...orda_Optics.ppt

 

First lets see what 1-15 minutes of arc angle looks like projected on to the retina:

Sampling by Foveal Cones

20.20
20.5
 
They also make the point that this photoreceptor array is "sampling" spatially across your retina at a rate of 120 cones per degree or you can detect a two bright points separated by one arc minute, again at the diffraction limit of your perfect eye, or 4 or five times better then your actual eye. 
 
Said another way Ed is using 4 or five photoreceptor cones in each eye for each tiny star he sees with both eyes. Thats not a lot of info, your brain is working pretty hard to reconstruct that image. Perhaps given Jacobs empirical study, Your brain is possibly using vertical as reference, this could explain the observation fatigue.
 
Daytime use has lot other phenomena including your brain actually "sees" better in brighter light, and also it its interesting to note most people eyes are have the visual axis about 4 degrees inward, 360 minutes of arc, or in other words looking straight forward mechanically, your eyes converge on a spot about arms length away. This probably explains why divergence is awful and convergence is much more tolerated.
 
and of course none of this matters if your not using a close to diffraction limited optical train and device performance....
 
I imagine proving diffraction limited vertical collimation is also a bit tricky.

  • MeridianStarGazer likes this

#23 MartinPond

MartinPond

    Fly Me to the Moon

  • *****
  • Posts: 6,530
  • Joined: 16 Sep 2014

Posted 15 December 2018 - 01:16 PM

Haha......right!

Alignment machines might see/feel the tightest limit.....but people? Not likely.

 

It is true that stars give the cortex less to work with than a grove of trees, of course.

Still, cognitively aligning views with 10-200 stars is not too tough..  



#24 STEEL

STEEL

    Messenger

  • -----
  • Posts: 484
  • Joined: 24 Jan 2007
  • Loc: Italy (Genova)

Posted 21 December 2018 - 07:06 PM

 

 

Rich V. sited Claudio:

"Vertical step: 15 arc minutes  

Divergence: 20 arc minutes
Convergence: 60 arc minutes"

I found even more discussions by Claudio on various standards here: https://www.cloudyni...ested/?p=323377

 

 

 

I tested my binoculars APM 82 mm 45° ED at 147x with objects distant 9.3 miles (15 km). I noted the apparent alignment error which is 35' horizontal convergence and vertically 14'.

Moving the image into the 3.2 mm 60° eyepieces (Tecnosky planetary ED) dist. 9.3 miles:

Horizontal convergence 39.4 inches (1m).

Vertical 15.7 inches (40 cm).

What satisfaction this APM binoculars!!



#25 NDfarmer

NDfarmer

    Viking 1

  • -----
  • Posts: 502
  • Joined: 02 Apr 2009

Posted 31 December 2018 - 06:12 PM

Allignment or collimation are common in binoculars even new ones from
the most well known mfrs.

It was posted above about a collimation test for vertical and horizontal
alignment. I have done that with a flat rest tripod mount, the easiest to
use, think Bogpod or Berlebach.

I have many binoculars and I live in a rural site, and I like a target of
1/2 mile to over a mile away. Road signs, power poles are very good targets
and are easy to use, I have done 10 binoculars in a short time.
This is all from my living room.

I have some 15-16-20X power binoculars, from several sources, the Nikon Monarch
5 models are a great value in this area.

The Celestron Skymaster, should include with instructions on how to turn
the screws.!!!! They are easily found on the internet and needed.

Jerry


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics





Also tagged with one or more of these keywords: collimation, equipment



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics