If you have been following my blog, you will be well acquainted with the Olympus OM 500 mm f/8 reflex lens from the 1980s. The reflex or mirror lens has the virtue of being compact and light when compared to conventional refractor type telephoto lenses. I also find them to be generally less sharp, but I can’t accept that as an inherent design fault as I’ve used and made many reflector type telescopes that deliver outstanding resolution. I think its a combination of poor manufacturing quality (possibly including poorly collimated mirrors) and difficulty in attaining perfect focusing (especially in the field).
In case you haven’t heard, on August 21st of this year a total solar eclipse will occur that will be visible from coast to coast in the United States. This will be the first actual full solar eclipse that I will experience and I intend to take some images of it. The full corona of the sun will be visible during about two and half minutes of totality when the Moon blocks out the body of the Sun and daylight becomes as dark as twilight. The ideal focal length to frame the sun’s corona with a m43 sensor is about 600 mm. It’s crucial to pick an observation site that will be cloud free that day so it might necessitate an airline flight meaning that I will have to pack light. I’ll need to carry a tracking mount and tripod so I don’t want to carry my Sigma 500mm f/4 telephoto prime (my small tracking mount also has a low weight capacity). I need to carry a decent mirror lens.
Over the years I’ve collected a sizeable number of mirror lenses (much to my chagrin), but really none that I particularly cherished. I needed to temporarily expand my collection to include some legendary mirror lenses and subject them all to a well designed MTF analysis and determine once and for all if there is indeed a mirror lens that can resolve as well as a modern Zuiko/Panasonic m43 lens.
Once again I’ll be performing MTF analysis using the ISO12233 resolution chart and QuickMTF software. Images will be taken in RAW with an Olympus E-PL5 body, sensor stabilization turned off, ISO 200, on a Manfrotto carbon fiber tripod with a Manfrotto 410 geared head and triggered with a USB remote shutter release. Manual focus will be confirmed by enlarging the live view image to 14x. Since some of the mirror lenses operate natively at f/8 where diffraction does begin to have some negative effects on resolution with the m43 sensor, I will also test those lenses using the Metabones Speed Booster T Ultra to raise the aperture by one full stop. As I intend to use the E-M1(II), the smaller pixel size of the new 20 MP sensor will experience diffraction limited resolution at an even larger aperture than the E-PL5. The lenses will end up being tested at a focal length of between 300-500 mm and about f/5.6 All lenses were purchased used and manufactured between the mid 1970s through to the mid 1980s except for the Spiratone, which was likely manufactured in the mid 1960s. Here’s a list of the usual suspects.
1. Spiratone 500mm f/8 circa 1965 MSRP $200, 4 elements in 3 groups.
Spiratone was a highly successful and innovative NYC home business and during its heyday of the 1950s and 1960s was an early importer of Japanese photographic equipment. The company wanted to be the first to market a Maksutov telephoto lens in the United States and asked the LZOS company to add a knurled focusing ring, change the mount from a Leica thread to the universal T2 thread and allow the 3/8″ tripod hole rotate from horizontal to vertical orientations. To their surprise, the Russians agreed and the original MTO Maksutov 500 mm f/8 telephoto lens became the Made in the USSR Spiratone lens. This is a classical Maksutov with both a spherical primary mirror and a relatively thick spherical meniscus lens to correct off axis spherical aberrations like coma. The secondary mirror is fabricated by silver spotting the inner convex surface. Due to relatively cheap Soviet era labor, the lens is heavy and robust in the best Russian tradition. Other mirror lenses tend to be lighter because of thinner meniscus design but field flatness and aberrations must be addressed by introducing additional refractile optical elements and a secondary mirror with its own curvature. Comes with screw on lens shade which helps improve contrast as mirror lenses can be susceptible to off axis light if deficient in baffling.
2. Sigma 400 mm f/5.6 with 7 elements in 4 groups meaning 4 lens elements in two groups and 2 mirrors and 1 meniscus. Comes with very long screw on lens shade.
3. Sigma 600 mm f/8, 1:3 Macro with 6 elements in 6 groups / 3 lens elements.
4. Olympus OM 500 mm f/8, 5 elements in 2 groups MSRP $400 in 1982. Built in sliding lens shade.
5. Samyang 300 mm f/6.3 ED UMC CS,A 9 elements in 9 groups
I wanted to test a currently manufactured lens. It has one ED lens element and modern ultra multi coatings (UMC) to reduce ghosting and flaring which may improve contrast. MSRP $259 Comes with long snap on lens shade.
6. Vivitar 600mm f/8 Solid Cat, 9 elements in 7 groups MSRP $634 in 1975.
This lens was designed in 1970 by Perkin Elmer who manufactured optics for surveillance cameras and spy planes and later the Hubble Space Telescope. The use of a more refractive solid glass structure instead of air meant an unbelievably compact and rugged lens. In 1975, Vivitar contracted Perkin Elmer to make the lens as a high end item to round out their Series 1 lens lineup but production halted after only 3 months due to a prohibitively high sticker price. Vivitar was in the business of designing innovative lenses typically contracted out to small Japanese firms to compete on both price point and quality against the big brand names. The Solid Cat didn’t really fit in. People reporting unusually poor image quality likely are missing the important 9th optical element, the 33.5 mm thread screw on glass filter (UV) at the very end just underneath the lens mount. The lens comes with an integral sun shade.
7. Questar 700 mm f/8, 4 elements in 4 groups, MSRP $995 in 1978.
Questar has been manufacturing its 3.5″ diameter Maksutov f/12 telescope since 1954 and its longevity is a result of peerless hand made manufacturing quality. In 1979 they introduced a telephoto lens based on this telescope model but with an aspherized hyperbolic meniscus matched to a mangin style primary mirror and two field flattening lens elements in the central baffle tube. Light baffling is so good that no lens shade is needed according to the manufacturer. Production ended in 1986 after only 822 lenses were made. If the Vivitar Solid Cat is a cult favourite, the Questar is likely the ne plus ultra of mirror lenses.
8. Celestron 750 mm f/6, 3 elements in 3 groups MSRP $450 in 1978.
Celestron has been making Schmidt Cassegrain telescopes since 1970 and the name is still prominent in the astronomy field although now being under Chinese ownership. Like the Maksutov, the Schmidt uses spherical mirrors but instead of a meniscus, there is a thin corrector plate with a shallow but complex aspherical curve. This thin corrector makes the telescope reach thermal equilibrium faster than a Maksutov, especially as you scale up to larger professionally sized telescopes. A temperature gradient between the interior and exterior of a mirror lens can affect the image quality severely while it is much less an issue with conventional telephoto lenses. It is also more expensive and more difficult to manufacture the Maksutov meniscus lens as the size of the telescope increases so another reason why large Maksutov’s are uncommon, but they are favored in military and industrial applications where ruggedness is essential. The other major design difference is that in Maksutov type mirror telephoto lenses focusing is achieved by moving the secondary mirror. In the Celestron lens, the primary mirror is moved, sliding along the central baffle tube. This is accomplished with an off axis screw system that can cause the mirror to slightly shift or wobble as it moves along the baffle tube. This is known as mirror flop and can affect image quality if the primary mirror position becomes slightly out of collimation. Unlike other mirror lenses, the Celestron secondary mirror has screws that allow you to adjust mirror collimation.
9. The reference lens is the only modern Olympus telephoto lens I have which is close to the testing focal length. It’s the 43rds Zuiko 50-200mm SWD and with a Zuiko 1.4x tele extender I get an effective focal length of nearly 300 mm at f/5. The SWD version was introduced in 2009 and is still available new for $1200, the EC-14 is $439. My copy was bought used for significantly less!
Here’s a group photo of all the mirror lenses used in this study. I have neglected to include the well regarded Zeiss Mirotar or the Sony/Minolta lens because I don’t know of anyone I can borrow them from and I’ve spent enough already. (Note: I forgot to include the diminutive Samyang lens in the shot, it’s that small!!!)
And here are the MTF results. From left to right in the top row: Spiratone, Sigma 400, Sigma 600. Middle row: Olympus, Samyang, Vivitar. Bottom row: Questar, Celestron, 43rds Zuiko. MTF10 curves are in red, MTF30 are in green.
(you can view fullsized chart at https://dl.dropboxusercontent.com/u/4852049/MegaMTFmontage.jpg)
(You can view 50% sized chart at https://dl.dropboxusercontent.com/u/4852049/ResolutionChartMontage.jpg)
I’m happy to discover that the Olympus mirror lens is as sharp in the center as the 50-200mm SWD and is in fact better in the corners with a very even flat field. And edge contrast of the mirror lens is very similar to the refractor type telephoto lens. The Olympus lens is also ergonomically the superior lens. The focusing ring is correctly dampened to provide just the right amount of resistance and the focus zone is relatively wide. Lenses like the Spirotone were too stiff, and the Sigmas far too loose and their focus zones razor thin. Just breathing on them could cause them to surge past the focus point. The Samyang performed well and was the only lens that truly fits the m43 dimensions. The Vivitar was a disappointment and its poor performance giving washed out images as the low MTF10 curve (red) illustrates. This could be a one off variance and sometimes it pays to spend extra and obtain that pristine example which has been used by a knowledgeable enthusiast and properly stored (I have since learned that the Vivitar lens was infected with fungus hence the reason for its terrible performance). I had high hopes for the Questar but its performance was only average. The Celestron performed just like a telescope would, it had a sharp center but the periphery was terrible since there are no corrective lens optics to flatten the field.
Finally, I managed to track down archival test results from the magazine Modern Photography. The magazine was bought by the publishers of Popular Photography in 1989 and then promptly shut down but for many decades prior it was the go to authority for its comprehensive and standardized bench testing of camera lenses.
My results don’t align with the magazine’s, some of this could be attributed to the wide variation of mass produced products and the possibility of testing a dud. But the Questars are renowned for their uniform quality and hardly made on a large scale, and the Olympus mirror lens were also made in very small numbers. All the lenses appear to be physically in very good condition despite their age.
In my study, it’s clear that the Olympus reflex lens is the one I’ll be taking with me to the solar eclipse.
Since we’ve crossed over so much to my other hobby (astronomy), I always wanted to settle the long running rivalry between who builds better telephoto prime lenses, telescope makers or camera makers. I favour telescope makers because the high end ones are essentially handmade professional instruments that will last several lifetimes. The kind of things you hand down to your grandchildren. Refractor telescopes tend also to be optically much simpler so fewer things to misposition. I imagine that they will have devastatingly sharper centers and weaker peripheral performance since that’s what telescope observers want.
So in the left corner is my highly coveted Astro-Physics Traveler, a 620 mm f/ 5.9 refractor with a triplet objective lens with one super ED element. The glass is hand figured at the end to 1/10th wave accuracy and the three elements stacked with such precision as to take upwards of 5 hours. The optical tube assembly, lens cell and focuser are all made in house and precision milled. About 600 were manufactured during the 1990s and like all collectables they continue to appreciate in value. In the right corner is my Sigma 500mm f/4.5 (shot at f/5.6) prime consisting of 11 elements arranged in 8 groups with two ED glass elements in the front lens element.
The results are a little shocking. Both are remarkably similar having high edge contrast and sharp centers and equivalent peripheral performance. Of course MTF curves are only one measure of an optical system but it is impressive that a mass produced telephoto prime lens is seemingly as good as a hand built telescope made at an annual rate of under a hundred. The Sigma is about 2 decades younger and likely benefited from improved technology but still managed to cost less brand new. Now 7 years onward, the Sigma is worth half of its original price while the nearly 30 year old AP Traveler is likely worth double its original price.