I own two stereo microscopes. One is a secondhand Meiji EMZ-1, a sturdy Greenough achromatic design with 1-3X zoom and a well regarded workhorse seen in many labs. I have the binocular version and although do possess the eyepiece adaptor to allow photo work to be carried out, this is rather clunky in use. For dedicated photo work and to extend the limited zoom range, this was supplemented a few years ago by a Leica S8 1-8X, a model with fully apochromatic optics. As stated in their brochures, Leica adopt Greenough optics in this design rather than a common main objective to optimise the optical performance.
Left: Meiji achromatic EMZ1 1-3X, right Leica S8 Apo 1-8X. Both are continuous zoom Greenough designs.
Microscopy hobbyists by nature are curious souls and although I was pleased with the performance of both stereos in their respective roles I was interested in the comparative performance of the optics—how does a modern apo' compare with an older modest achro'? The Leica is stated to have a maximum NA of 0.10 at 8X with its native optics without auxiliary lens. I haven't found to date published NAs for the Meiji range, but accurate measurements of the acceptance angle to front element allows a good value of NA 0.075 to be derived at its max mag of 3X, a typical value for this type of stereo when the data is available.
It is not possible from these single NA values to comment on how the NA varies across their zoom range. In practice when studying more demanding subjects, although subtle, I had the impression that the Meiji had the edge in definition up to its maximum mag of 3X; the Leica having the edge on contrast and colour. To see if formal NA measurements supported this observation I sought out methods of measuring the NAs accurately.
A quick but useful single point assessment of a stereo is to study a typical micrometer slide calibrated to 0.01 mm at 1X optical zoom (with typical 10X eyepiece). If a stereo can resolve this, it has an NA of at least 0.0275 (using R=λ/2NA). The Meiji EMZ-1 did clearly resolve this type of slide at 1X, whereas the Leica S8 Apo did not, which needed a zoom of ca. 1.25X to resolve. Comparative images are shown below. As this test is near the limit of visual acuity, it is a good test of the stereo optics both visually and with a camera.
Above: A 0.01 mm micrometer slide, unmodified except resized crop from master using 1X optical setting on both stereos. Left - the Leica S8 and right, the Meiji EMZ-1.
|
The Meiji crisply resolves it at 1X, the Leica does not. The Leica did begin to resolve at 1.25X as shown right. Note the absence of chromatic abberation in the Leica images, as expected from the apochromatic optics cf the achromatic Meiji. Projection distance adjusted so that the micrometer occupied the same number of sensor pixels on both stereos. The same epi LED ringlight used for both to avoid any differences in transmitted illumination. Sony NEX 5N body in Silent Shutter' mode (which I've confirmed to be essentially vibrationless) and with 10 sec shutter release delay.
|
|
Using the Web to research ways of measuring the NAs accurately across the full zoom range, I came across 'PierreH''s valuable article Formules mathématiques d'optique appliquées aux (stéréo) microscopes where some methods are summarised. One of the two links was to Gérard Weiss' article Ouverture numérique d'une loupe binoculaire : comment l'établir? describing how to measure the NA of microscope optics, including stereos, by measuring the exit pupil diameter at the eyepiece (both are hosted on the excellent Le Naturaliste website; Google translates them well for non-French speakers like myself). Gérard Weiss describes and illustrates the theory and practice of the technique very well. The simple equation linking exit pupil to NA is derived in his article.
Measuring the exit pupil diameter at eyepiece at different zoom positions using a loupe with built-in 0.1 mm calibrated scale was straightforward (varying from 0.65 to 1.3 mm for the Leica and 1.2 to 1.9 mm for the Meiji). The loupe as recommended was supported on card collar at correct distance but did have to be careful about potential parallax error with the loupe I used. The NA results for the Leica S8 and Meiji EMZ1 derived from these measurements are presented below.
The NA of 0.07 at 3X of the Meiji agreed well with my geometrical measurement of the front optic (0.075). Shortly after this study an email colleague informed me that Leica can send very detailed optical characteristics of specified stereo models in their range if asked. Their data for the S8 is also presented in the graph (with the kind permission of Leica). My own measurements derived from the exit pupil data closely matched Leica's stated values; an excellent endorsement of the effectiveness of the method described by Gérard Weiss. This in turn gave me confidence in my data derived by this method for the Meiji.
The graph shows a number of interesting aspects of the NA / resolution of continuous zoom stereos:
NAs decrease with zoom to give a value at the lowest mag which is much less than the single NA value at its highest mag typically quoted by makers. (Very few makers specify in their brochures how the NA / line pair mm varies with zoom.) At 1X the Meiji's value of ca. 0.04 is comparable to the lowest power 1X planachro' objective for a compound microscope, e.g. the Zeiss 1X NA 0.04, 160 mm tube length.
(Note that the continuous optical zoom of a typical stereo seems to be distinct from an Optovar zoom on a compound microscope, i.e. where the images from an objective of fixed NA is inspected at increasing mags, albeit at the possibility of encroaching 'empty mag'. Comments welcomed.)
The Leica, although a more modern design and the better corrected, has a markedly lower NA at its lowest power of 1X and these lower values than for the Meiji are retained in the shared zoom range of 1-3X. These values therefore support my impression remarked on earlier that the Meiji in use has the edge in resolution.
The results also support the inability of the Leica to resolve the 0.01 mm micrometer slide at 1X as its NA of 0.026 (Leica data) is below 0.0275, i.e. that required to resolve it.
Do these results suggest that NA at the lower mags of a continuous zoom stereo is sacrificed when it has a wider zoom range, e.g. in this case the 1-8X of the Leica S8 APO cf the 1-3X zoom of the Meiji EMZ-1? The NA's across the zoom range stated for the two Leica flagship monocular macroscopes do support this, i.e. Z6 APO max mag 3.6X, NA at 1X 0.0358; Z16 APO, max mag 9.2X, NA at 1X 0.027.
If this is more widely applicable across makers' continuous zoom models, it may be worth bearing in mind if a user particularly requires optimal resolution for their work at the lower mags e.g. if visually studying or photographing detail of larger subjects or to maintain depth of field, in which case there's a case for seeking out a quality lower zoom range stereo; with a modest zoom range achromatic possibly outperforming a wider zoom range apochromatic. (The humble achromatic Meiji EMZ-1 with NA 0.03 at 1X would just out resolve the Leica Z16 APO flagship at 1X!). I'm not clear if the observations on NA also holds for fixed mag zoom Galilean optics—does a maker have more control of the NA at each mag with this design? I'd be interested in hearing from readers with models of various ages and designs as to whether the optics pass what seems to be a demanding 10X micrometer slide test at 1X zoom setting, which may give a single point hint of the scope's optical quality. Or even better, the full exit pupil tests.
Gérard Weiss shares his results for the AO-570 stereo in his article mentioned earlier, and reports an impressive NA of 0.036 at 1X. PierreH also shares his results using three methods for the Wild M8 with 1X planapo objective on the Le Naturaliste site's Forum.
Comments to the author David Walker are welcomed.
Footnote: I'd be interested to learn if the classic apertometer approach as described by Dushan Grujich in his excellent article last month for measuring optics of compound microscopes is also applicable to stereos. My own quick trials were unsuccessful but admittedly didn't pursue the method on a stereo when learned of the exit pupil method.
Acknowledgements:
Thank you to 'PierreH' and Gérard Weiss for their valuable articles hosted by the Le Naturaliste website which alerted me to the valuable method of deriving the NA from exit pupil measurements.
Thank you to Leica Microsystems (UK) for kind permission to share their numerical aperture data in a non-commercial educational context.
Thank you to a fellow hobbyist for letting me know that Leica can supply detailed optical data for specified models in their range on request.
Related articles:
Micscape Review April 2007: Three stereo microscope models / designs compared (Meiji SKC-BT, Meiji EMZ-1 (Greenoughs) and Leica MS5 (CMO)).