420 replies to this topic

[freestar8n]

Here is a pointer and description of the ISO 517 standard method for determining aperture. They also make an instrument designed for the purpose:

Entrance pupil measurement by ISO 517.

There are several advantages to this approach. First, the entrance pupil itself is not an object - it is the image of an object formed by unknown lenses in front of it. This measurement is based on looking into the front of the system and focusing on each limiting aperture inside, while moving along an accurate translation slide. There is no eyepiece or projection involved, and the pupil itself is in perfect focus. The measurement is based on a physical translation and there are no fuzzy edges or distortions involved.

The key question in these measurements is - are you actually measuring the true pupil - and by translating along the diameter, you are able to identify the relevant stop inside that is acting as the aperture stop.

Now - this method is not trivial to set up - and it would have similar challenges if the pupil is far from the front of the telescope - since it would appear small and in the distance. But even a crude version of this can be set up by an atm with not much more difficulty than a well prepared collimated laser system.

Frank

[freestar8n]

Here is a way to handle the most challenging case, which is also not uncommon, of an obstruction near focus. This happens with things like binoviewers or things extended off the back, but it requires access to the stop in question - and there can be no lenses between that stop and focus. You also need to measure its distance from focus, and the effective focal length, accurately.

Frank

Attached Thumbnails

• 

[roscoe]

Nils, Frank, Glenn, et. al.:

A pinhole at the focal plane, can you see the edge of the objective, edge of the primary mirror? This has always seemed to me to be a good way to test whether a telescope is operating at full aperture, at least on axis. It can also be adapted for off-axis determinations..

What do you think?

Jon

This is the simple test I've always used - I have an old Huygens EP with the glass removed that I use in 1.25" focusers.....but then I started reading that a wide-view EP, like a 30 or 40mm or the like, needs an unobstructed light passage much larger than a pinhole to develop full capability.
What thinkest thou, o learned ones??
Russ

[freestar8n]

I think simple tests like that might let you know if something is obstructing, like a baffle, but it doesn't actually tell you the entrance pupil diameter. One headache example is a maksutov where you don't know the size of the primary and you don't know if the front corrector is the aperture stop, or the primary, or a baffle. Even if you could determine the primary was the limit, you would need to include the effect of the corrector in making the entrance pupil smaller. And if the baffle on the secondary is a problem, then both the primary and the corrector are involved in determining the pupil size.

But if the secondary itself is too small - and you can measure it - then it reduces to the case of a known obstruction near the focus, with no lenses on the way to focus. And that can be determined directly.

Frank

[hottr6]

My own scope was operating at 140mm effective aperture and I didn't even know it. It's now at full 150mm aperture.

Hey Norme, how did you do this with your Mak, and how did you identify the restriction and "repair" it? Enquiring Mak-minds need to know!

[freestar8n]

(Incidentally, thanks for the mention of MIL-HDBK-141 the other day. I found a copy online and am going to have a look at the chapter on relay systems next week...)

Thanks - yes, that is now just over 50 years old and it is a nice reference. It has a lot of the fundamentals along with examples and diagrams - plus it has a heavy emphasis on visual work, which applies to this forum. And it's free.

Section 7.6.2 has an analysis and diagram of the Galilean telescope, and it shows the entrance pupil - way behind the observer's eye. Things can get crazy when you stick to the actual definition of entrance pupil.

Frank

[Jon Isaacs]

One headache example is a Maksutov where you don't know the size of the primary and you don't know if the front corrector is the aperture stop, or the primary, or a baffle.

Certainly things are not fun with a design such as the Maksutov-Cassegrain, particularly if it focuses by moving the primary mirror as the effective aperture may vary.

Fortunately, I will never be faced with the dilemma of trying to measure the effective aperture of a Mak so the simple minded approach of observing the objective/mirror from the focal plane provides a practical measurement.

Still, I have to think that Roland's method of placing a ruler across the front of the scope and through a pinhole at the focal plane, reading the ruler should provide information as to the effective aperture.

Jon

[Asbytec]

My own scope was operating at 140mm effective aperture and I didn't even know it. It's now at full 150mm aperture.

Hey Norme, how did you do this with your Mak, and how did you identify the restriction and "repair" it? Enquiring Mak-minds need to know!

To avoid getting off topic, PM sent.

On topic, though, one method I used to determine vignetting was occurring somewhere was intruding an obstruction, like a ruler, across the edge of the objective into a de-focused star image. It said nothing about the location of the aperture stop, but that one existed. It took further investigation along with various laser and flashlight tests and measuring optical components to find it. But, the results were consistent with the flashlight test.

Frank, when you know s and d in the diagram above, and compare it to D/f, how do the numbers translate into a quantitative measure of reduced aperture? I'm happy to report my own numbers s/d > D/f indicating full aperture. But, if the numbers were reversed, how can one accurately determine D effective? (D in this case is full aperture, so solving for D doesn't seem to work and the math is an inequality to boot.)

[GlennLeDrew]

No need to know which element defines the stop or entrance pupil. As long as the emergent cylinder of light *is* a cylinder (parallel light), and the edge is not so blurred as to affect confidence in measurement, the result is correct. The emerging cylinder is the same as the incoming cylinder from a distant point. The light cannot but follow the same path through the system. How much simpler can it get?

I just did a test on my C8, using a single-LED flashlight. The eyepiece is a 25mm focal length, and the light was placed about 30cm from it. Near the corrector, the shadow edge is very sharp, which is to be expected because the aperture (either the primary edge or corrector, I'm not sure at this point) is near the measurement point just beyond the corrector. At 3-4m distance, the shadow edge is blurred by about 1mm, which does not impair the measurement to any real degree.

Furthermore, I changed the focus, which caused the emerging beam to taper, becoming non-parallel. I was able to crudely relocate the infinity focus by adjusting until the circle of light near to and at ~3.5m distance was equal.

The flashlight, while not the ideally small source one would desire, works.

I then used the beam expanded laser. Far superior, being so bright and sharp. Indeed, the roof line on the Amici prism was really obvious as a darker line, and dust bits made for nice diffraction patterns.

And here's the kicker. The prism roof line is only some 50mm from the focus, yet it's rather sharply defined at 3.5m from the scope's front end. If one doubts the sharpness of the laser spot imaged at the focus, this will utterly quell that concern.

Bottom line. A suitably small light source at the infinity focus position is an absolutely reliable means of measuring the real system aperture.

[Asbytec]

Bottom line. A suitably small light source at the infinity focus position is an absolutely reliable means of measuring the real system aperture.

The flashlight test, IME, is completely consistent with running an obstruction across the objective and observing the shadow intrusion in a de-focused star. There is nothing like testing on actual star that is actually a point source at infinity. To the degree one can accurately measure, it is consistent with the diameter of the (pretty sharply defined) projected cylinder.

The LED flashlight was very dim and difficult to measure. It was hard to tell how blurry it was. A cheap laser pointer was much better. It was not a point source, but it was essentially a parallel beam. Maybe the width of that beam played a role in the diameter or fussiness of the projected cylinder.

But, neither test shows the location of the aperture stop, just that one might exist and should be seen as such from object space. Right? So, how does one differentiate between the aperture stop and the entrance pupil? And what difference does such a distinction make?

[jhayes_tucson]

No need to know which element defines the stop or entrance pupil. As long as the emergent cylinder of light *is* a cylinder (parallel light), and the edge is not so blurred as to affect confidence in measurement, the result is correct. The emerging cylinder is the same as the incoming cylinder from a distant point. The light cannot but follow the same path through the system. How much simpler can it get?

I just did a test on my C8, using a single-LED flashlight. The eyepiece is a 25mm focal length, and the light was placed about 30cm from it. Near the corrector, the shadow edge is very sharp, which is to be expected because the aperture (either the primary edge or corrector, I'm not sure at this point) is near the measurement point just beyond the corrector. At 3-4m distance, the shadow edge is blurred by about 1mm, which does not impair the measurement to any real degree.

Furthermore, I changed the focus, which caused the emerging beam to taper, becoming non-parallel. I was able to crudely relocate the infinity focus by adjusting until the circle of light near to and at ~3.5m distance was equal.

The flashlight, while not the ideally small source one would desire, works.

I then used the beam expanded laser. Far superior, being so bright and sharp. Indeed, the roof line on the Amici prism was really obvious as a darker line, and dust bits made for nice diffraction patterns.

And here's the kicker. The prism roof line is only some 50mm from the focus, yet it's rather sharply defined at 3.5m from the scope's front end. If one doubts the sharpness of the laser spot imaged at the focus, this will utterly quell that concern.

Bottom line. A suitably small light source at the infinity focus position is an absolutely reliable means of measuring the real system aperture.

Sheeze, this seems like a pretty simple subject for such a spirited discussion. I agree with Glenn that light is reversible so it doesn’t matter which way you send the light to make the measurement. As long as the system is afocal (i.e. properly focused,) the illuminating beam is collimated (i.e. parallel rays,) and the illuminating beam at least fills the exit pupil, the collimated beam at the output will completely fill the entrance pupil. Clearly, the sharpness of the shadow will limit how accurately the diameter can be measured. Placing a point source at a distance of at least 10 focal lengths (of the eyepiece) behind the last surface will ensure that focus errors remain sufficiently small to get a good measurement. This is an accepted rule of thumb in optics for the closest distance for measuring the focal length of a lens. Further is obviously better, but remember that the angular magnification is small in the reverse direction through a telescope. Not having a perfectly collimated input beam will probably introduce errors smaller than your ability to measure the diameter of the output beam. In short, I agree that Glen’s method will work to measure the entrance pupil diameter to within a couple of millimeters—and maybe better (depending a bit on the type of system being measured and where the stop is located.)

John

[Dave O]

Sheeze, this seems like a pretty simple subject for such a spirited discussion. I agree with Glenn that light is reversible so it doesn’t matter which way you send the light to make the measurement.

I don't think there is much argument over that.

As long as the system is afocal (i.e. properly focused,) the illuminating beam is collimated (i.e. parallel rays,) and the illuminating beam at least fills the exit pupil, the collimated beam at the output will completely fill the entrance pupil.

I think that ensuring these conditions are met with the set-up initially described is what is driving the "spirited discussion."

I think Glenn's method is probably 'good enough' for the purpose of most ATM's. It is relatively simple to set up and if one is careful to ensure that the initial conditions are satisfied, any gross differences between expected and actual effective aperture should be uncovered. If the measured results do not agree with the expected value, then further analysis will be required to determine the source of the difference.

However, I do not think that NASA would accept values determined by this method; I am sure that they would want something a little more 'formal'. (But then again, they bought a defective mirror and put it into orbit before verifying its quality ....)

[Nils Olof Carlin]

Jon,

I think the reason Glenn uses a system the requires an eyepiece is that he is one of those binocular guys... It makes it easier... and since binoculars are often not operating at full aperture, it's an important measurement...

With a telescope, the eyepiece can be removed but it is a different situation because you expect it to operate at full aperture and if it is not, then it can be fixed without too much trouble.

The aperture defining the size of entrance pupil might lie just about anywhere. As Glenn demonstrated in that other thread, even at the eye pupil. In this case, Glenn's test may not be easily applicable (nor your EP-less variety)

Nils Olof

[MKV]

[quote name="Asbytec"][quote]A cheap laser pointer was much better. It was not a point source, but it was essentially a parallel beam. Maybe the width of that beam played a role in the diameter or fussiness of the projected cylinder.[/quote]
A simple laser pointer with a 4-5 mm beam diameter is all you need. You can place it right against the eyepiece. The trick is to have the eyepiece exactly at infinity focal setting.

Attached Thumbnails

• 

[John Carruthers]

how does one determine which aperture is causing a stricture? a baffle? objective aperture? eyepiece field stop? eyepiece eye lens, etc?

[GlennLeDrew]

I just did both flashlight and beam expanded laser tests on my C8 again. But this time I added an extension to the back end so as to locate the focus about 250mm beyond the rear port on the scope. This is guaranteed to make the inner (front) opening of the primary baffle the aperture limiter. In any SCT, for on-axis light this will be the obstructor which can be closest to the focus, and hence for the illuminated circle the cause for the worse blurring of its edge.

With the LED flashlight about 30cm from the 25mm eyepiece, the circle's outer edge is blurred by no more than 2mm just in front of the corrector. At 3.5m it's blurred by about 4mm. (The secondary shadow remains extremely sharp, due to it being so far from the light source.) This is as expected (and was why the laser suggested itself.)

The question becomes, is a 2mm or even 4mm blurring of the shadow a liability? Not really. If you measure from the half-light intensity level within the blur zone, you are referencing the center of the extended, non-point light source, which would be equivalent to using a point light. It's not difficult to reduce a blurred shadow's 4mm 'uncertainty zone' to less than 1mm.


With the laser setup, the truly minute light source resulted in sharp edges in all cases, as expected.

Nils Olof,
You raise an interesting variation; the effect of a too-small iris behind the eyepiece. It's perfectly testable. Just place an aperture over the eyepiece whose hole is of the diameter desired. For the on-axis condition, the hole need not lie in the plane of the exit pupil. To the incoming, parallel light, the system cares not if this opening is near or far; the effect remains the same. Now, to test for the off-axis condition, this hole must lie in the plane of the exit pupil.

As for the eyepiece-less case, I haven't given this matter any thought.

[GlennLeDrew]

Mladen illustrates a point I made in my original post; that a laser whose beam is at least as wide as the exit pupil will fill the entrance pupil. If you have the option, switch to a shorter f.l. eyepiece until the laser light fills the aperture. And as always, ensure the focus is reasonably near to infinity.

[Nils Olof Carlin]

Glenn,

Nils Olof,
You raise an interesting variation; the effect of a too-small iris behind the eyepiece. It's perfectly testable. Just place an aperture over the eyepiece whose hole is of the diameter desired.

This may be trivial - what I meant was that you can't do it with the actual eye pupil - doing it with a substitute aperture is of course possible.

Measuring your own actual dark-adapted pupil (not dark-adapted retina) is worth doing.

There have been questions how to determine the limiting aperture if not the objective itself. Doing the test with a reasonable light source (a laser might be uncomfortably bright), you can look down the objective, moving the eye from center to edge and identifying whatever structure cuts out the light first.

Nils Olof

[GlennLeDrew]

how does one determine which aperture is causing a stricture? a baffle? objective aperture? eyepiece field stop? eyepiece eye lens, etc?

Indeed, the flashlight test only shows that a constriction is occurring, not where. You need to use other means to examine the system if a obstructor is to be identified. A start is afforded by the simple expedient of looking into the front end while the light is on (not recommended with a laser source, unless the output light is suitably dim.) Place your eye where the edge of the illuminated circle lies, and try to identify which stop or element is clipping the light. For longer focal length scopes, you can magnify things by using a close focusing, low power finder or bino (just one side.)

[freestar8n]

Glenn - you are still missing the point. If there is a limiting aperture near the focus, it completely changes the robustness of the measurement because the pupil is then far away. When I did the test with an sct, I placed a small aperture about 40mm from focus. I also used an actual flashlight. In order for the edge to be sharp enough, because the pupil is so far away, the light source needs to be very small - which means the flashlight gets pulled away - and the image gets darker.

This test will often work perfectly well and behave just like the diagram - but a common situation is - you are concerned if something near the focus is limiting the aperture. In that situation it all gets harder.

In short:

1) I don't think it is good for newbiews to be holding a flashlight to their eyepieces except as an entertaining light show. Without care, the number they get could be optimistic or pessimistic about the true aperture.

2) With care, a collimated laser system or "collimated laser test" should be pretty reliable in most cases - but care needs to be taken to make sure the beam is collimated over a very long distance - if the location of the pupil is not known. Even then, it will be challenging to measure restrictions near the focus.

3) Anyone doing this stuff should be aware they are measuring the size of an image that is some distance away, and it is caused by some physical object in the 'scope. Knowing which is the obstruction in question will help improve the measurement approach.

Some examples:

Refractor: As long as the stop is in the objective and not caused by a baffle downstream, a simple flashlight test should work fine. The shadow is projected only a short distance and the errors are proportionally very small.

Maksutov with stop known to be at the corrector: Hold up a ruler and measure the corrector diameter.

Maksutov with stop at mirror: Collimated laser test should give a sharp shadow because the pupil is not far away. Simple flashlight test could be off and fuzzy.

Maksutov with stop at baffle: Careful collimated laser test may be ok - but physical measurements and ray trace would be desirable if the baffle is just cutting into the edge. The measurement becomes much harder because the primary and corrector are both making the pupil image.

SCT with ambiguous stop position - either at corrector or primary: This is difficult to measure accurately and really needs a ray trace. The corrector is not a simple lens and the divergence is partly due to the weakly polynomial surface.

Binoculars with obstruction near focus: Careful collimated laser test will probably be ok as long as it isn't too close to focus - but measurement accuracy beyond 1mm in measuring a shadow is not trivial. Since the test result is often binary - as in, "is the aperture less than stated or not?", the test needs extra care.

Large aperture system with obstruction near focus: This will be very sensitive to the source and collimation of the beam and require care to make a measurement based on shadow. If the obstruction is accessible then use the measurement described above.

Final note - there are many recommended ways to measure lens properties that have been adopted over the years in optical metrology. I had heard of the ISO method a long time ago and it immediately made sense because it is a direct physical measurement that does not rely on auxilliary optics or specially prepared light source. I have never seen anyone - ever - recommend measuring aperture based on a projected shadow. If a new method of measurement is suggested - that's great - but I think it would be helpful to compare it to established methods. In the case of the flashlight test discussion in CN, I think I am the only person putting it in the context of an established method - and its inherent disadvantages. Without that context, and without defining the terms such as entrance pupil in the first place, the whole discussion is somewhat in a vacuum and ignores the efforts of people who have thought about these things for centuries, and come up with good ways to measure them.

Frank

[freestar8n]

Still, I have to think that Roland's method of placing a ruler across the front of the scope and through a pinhole at the focal plane, reading the ruler should provide information as to the effective aperture.

This is a situation where there is no free lunch. As soon as you say the pinhole is at the focus, you are making the measurement dependent on the careful setting of another variable - and if the obstruction is close to focus, the sensitivity to error - and the difficulty of seeing the pinhole - increases. The ISO measurement doesn't even know the focal plane exists. If you know the offending physical aperture, you could put the optical system through a band saw and chop off the whole back end and throw away the extra optics - and the entrance pupil measurement would be unaffected.

Frank

[GlennLeDrew]

When restricting to the on-axis condition, where the entering and emerging light is parallel to the optical axis, from the standpoint of pupils this test could in a fashion be called 'dimensionless' (I can't think of a better term right now.) The identical light path is followed no matter where the obstructor or puoil lies.

Whether the obstructor is a 50mm stop immediately behind a 100mm objective, or a 25mm stop midway between objective and focus, or a 12.5mm stop 3/4 of the way from objective to focus, the effective aperture is always the same one-half of nominal aperture. The light does not care where the obstructor/pupil-defining stop lies.

[freestar8n]

Any measurement process like this has associated error. To know the error in the measurement, you need to model it and propagate the errors based on that model. In many situations, errors will increase dramatically under certain conditions - and then become strongly dependent on other factors that are otherwise negligible.

Your original note specified accuracy and criteria. I don't know how you would state it all now, but it had to do with light source 10 fl's away for a measurement accuracy of 1mm. Can you do the error propogation in a model based manner to guarantee this holds for an arbitrary optical system with unknown stop? What are the criteria to guarantee 1mm accuracy for this measurement?

There is yet another non-trivial matter I forgot to mention. Lasers are regarded as these great things that shoot out perfectly straight and collimated beams - but the reality is quite different. Although use of a laser will help by creating an intense and small light source for the measurement - it involves propogating a clipped Gaussian beam through a multi-component optical system - and then making measurements of its width to make sure it is "collimated." This is yet another thing that looks great in a diagram - but in reality departs from ideal - and cuts into the ability to claim 1mm accuracy in a general sense.

A good measurement needs to be robust against contributing errors - and all this light source preparation and the additional eyepiece and backend optical elements add complexity and errors to the measurement - that are avoided by the ISO method. That's why it is used and cited.

Frank

[Nils Olof Carlin]

Glenn - you are still missing the point.

Unfortunately, so am I.
The ISO517 might be a good way for industry and for optical systems in general, but I suppose the CN members would be interested (if indeed at all) in testing telescopes and binoculars of varying designs, mainly for identifying unnecessary aperture-limiting obstructions. Maybe an aperture stop near focus (like a too long and narrow focuser drawtube, a badly undersize secondary, or what else?) will not be accurately measured using a plain flashlight in Glenn's test - but who would rasonably care about the exact size of the vignetted entrance pupil as seen on axis? If you find such an aperture stop, you would identify and remedy it, and go on observing.
I think that it may be of more practical interest to determine the size of the unobstructed (if any) field of view. As well as its position relative to the center of the EP focal plane, and its shape (circular, elliptic?). I have outlined a rough method of finding it somewhere on CN...

Nils Olof

[GlennLeDrew]

Frank,
We have the benefit of direct feedback, which allows to reduce uncertainty to comfortably acceptable levels. Calculation of layout error is obviated by *two* measurements.

Is the emerging beam diameter the same near the objective and at some distance (one telescope focal length or more) from the near measurement?If yes, the light is collimated. That's all we need to know in order to be confident that the projected light is near enough to the correct placement at the focus.

Refocusing as necessary to achieve parallelism at the output automatically corrects any error introduced by an incorrect focus distance of the image of the light produced by the eyepiece.


The other source of uncertainty comes from a blurred edge on the projected cylinder of light. The laser absolutely eliminates this. For a flashlight, even a few millimetres wide blur can be largely compensated for by measuring at the estimated half-light point in the blur zone.

These two aspects, parallelism of output beam and sharpness of shadow, provide all the confidence we need. Where are the pupils and stops? Irrelevant, because we are working on axis. Is the entrant light sufficiently parallel? The shadow edge sharpness answers that. There is no other aspect in which uncertainty rears its head.


It remains for you to show how a *sharp-edged* and *collimated* cylinder of emerging light can possibly be doubted as being indicative of the effective aperture. Bearing in mind that the practical astronomer does not need to *quantify* error to the 0.1% level.