Hey everyone,
Today I want to take a closer look at mirror lenses for cameras. So, last year I’ve spent quite a lot of
time on designing and making a compact monolithic telescope. This item was made of a single piece of glass
and it has a long focal length compared to its size. In this video, I will discuss a few conventional
mirror lenses, which have quite similar designs and also have a relatively long focal length
for their size because of a folded optical path. These lenses can be found online quite abundantly
and be bought for relatively little money. And, I was curious whether they would be suitable
to serve as the basis for a compact travel telescope. So in this video, I will take a detailed look
at the optical performance of this type of lens. And in an upcoming video I will show you how
they are constructed and if they can maybe be tweaked to perform a little better. In this video I’ll have a look at three
lenses: the 800 mm focal length Vivitar Series 1 lens which is of fairly recent production
date, a 600mm Sigma mirror telephoto lens, probably from the 1980’s and a 500mm Soligor
lens, which is probably also from the 1980s. I bought these second-hand with prices varying
between 50 and a 100 Euros. The first lens I got in was the Vivitar 800mm
focal distance lens. It has an aperture of approx. 100mm, making it an F/8. It also has a large central obstruction that
is about 47mm in diameter. The lens looked brand new and unused and had
clean and very clear optical coatings. It came in a box, with a pouch and even life-long
warranty. No serial number though, but I was pretty
excited to take a peek through this “high definition” lens. So, I mounted it on a Canon 90D, using a T-mount
adapter and just pointed it out of the window. Now, my initial excitement disappeared in
just a few seconds. Because it was immediately clear that this
lens, despite it being in excellent cosmetic condition, wasn’t sharp at all. Images remained fuzzy, no matter how hard
I tried to focus. I must admit that, never before had I see
a commercial lens perform so poorly. Just to give you a frame of reference: when
you compare the sharpness of this lens to a conventional 400mm telephoto lens, you can
clearly see a huge difference: even though the conventional lens has half the focal distance,
it outperforms the 800mm Vivitar lens on sharpness by far. But in fact, you don’t even need a high-power
telelens to do this: even a simple 135mm focal distance lens, which I bought in a thrift
shop for about 10 euros, reproduced more detail when you zoom digitally into the images. Now, my initial reaction was that the mirror
lens was probably a bad apple in a batch because I remembered the Vivitar brand and especially
the “Series 1” indication to be generally of good quality. So, I did a search online about the experiences
of others. It turned out that this lack of sharpness
is actually very common with this particular lens, leading to quite a few disappointed
buyers. Okay, so maybe I should have done a little
research before I actually bought this lens. In fact, it seems to be a far departure from
the original Vivitar Series 1 mirror lens, which was of good quality. The original lens was actually called a solid
catadioptric. That name is actually kind of misleading,
because it suggests that the optics are made of a single solid piece of glass which is
not the case. It actually contains 7 individual optical
elements. Anyway, it turns out that the company behind
the Vivitar products is not the same as the one back then. The Vivitar brand was registered by Ponder
and Best Inc. in the early 1960s and in general, this company rebranded Japanese and German
products under the Vivitar name. After the founders of the company died, the
brand knew various other owners and in 2006, Vivitar was bought by Syntax-Brillian, a consumer
electronics manufacturer. Unfortunately thought, they went bankrupt
in 2008, at that point a company called Sakar International acquired the brand name and
intellectual property and continued to sell products under the Vivitar name. But these products were now made in China
or Korea. And apparently, they are on a different quality
level than the Vivitar products previously manufactured in Japan and Germany. Now, I guess there will always be quality
differences between companies and products. But if there is one thing I hate, it is when
companies put products on the market whose quality is so bad that they are basically
unusable. Because you cannot take a single decent picture
with lens, so the best thing you can do is turn it into a nice plant pot or maybe a funny
hat. Anyway, this lens was kind of a disappointment. But luckily a few days later, a box arrived
with a different lens: the 600mm Sigma mirror telephoto. Contrary to the previous lens, this one is
quite vintage. It was first introduced in 1979 and I estimate
this lens is between 30 and 40 years old. Cosmetically, the Sigma lens looked okay on
the outside, but as for the optics themselves, they were not in too good a state. Inside fungus and various damages to the coating
were clearly visible, indicating that, at some point in the past, condensation had taken
place inside the lens. Also, the coating looked yellowish, probably
due to oxidation or even nicotine. So, after these initial discoveries, I was
afraid that I was in for another disappointment. But what do you know: this lens turned out
to have much better image quality, despite all the fungus and bad coating. Here you see a comparison where I zoomed in
on details in the images of the Sigma and Vivitar lens. Quite a difference, right? So obviously, the flaws in the reflective
coatings of this lens do not have too much negative influence on the optical quality. I must admit though that in general, I did
not find the images of mirror lenses particularly pleasing: especially the circular shape of
the bokeh. And if you search online, this is actually
one of the most frequently mentioned drawbacks of catadioptric camera lenses: the unnatural
looking out-of-focus effect due to the central obstruction in the lens. For proximity photography this effect will
always be an issue. However, if you were to use the lens for astronomy,
this would not really be problem, because basically everything is at infinity. So when evening came, I pointed this lens
up to the sky. And with a full moon and a lot of light pollution,
the only other things I could see were basically the planets: Jupiter and Saturn. Thing is that the 600mm focal distance is
actually kind of limited if you want to see planetary details. This is because the projected size of the
planets on the camera chip of the Canon 90D are too small to show a lot of detail. Still, they are somewhat recognizable. So I guess this lens is nice if you stick
to an occasional look at the moon or maybe at some deep sky objects However, I would
not recommend this type of lens for viewing small objects like planets, mainly because
of the limited magnification. As a bonus, I bought this little Soligor mirror
lens, with a focal length of only 500mm, so the shortest focal length of all the 3 mirror
lenses. It came in the original 1980s style box, with
a lens hood. Cosmetically it looked fine, also the coatings
on the inside. But as with the Vivitar lens, it displayed
a significant fuzziness although not as bad. While testing, I noticed it had these asymmetric
out-of-focus ring patterns, even in the center of the image field. This indicated that something inside was not
positioned correctly with respect to the optical axis. So, will I put this lens aside for now, but
in the next video, I’m going to address this particular problem. So to summarize the previous in one sentence:
I bought three lenses named Blurry, Moldy, and Squinty. After I had checked out the first Vivitar
lens, I considered whether it would be possible to improve the performance by making a small
change to one of the optical surfaces. Since the lens was a total failure anyway,
I thought it was worth a try. But before I could do this, I first wanted
see what the problem with the optical performance was. Now, there are many different ways that you
can measure optical performance. For example, you can take a high-resolution
test chart and look at how the patterns are reproduced in the image by the lens and then
put this as a result in a graph. And this gives you a measure of optical performance. But by itself, this test does not give you
any information about which aberration in the lens is responsible for image errors and
how to correct it. So in order to measure this particular aspect,
I decided to evaluate the lens using interferometry in an auto-collimation configuration. And because you might not be familiar with
this particular method, let me briefly discuss how this works, using a few schematics. For interferometry, we generally use a well
collimated parallel beam of coherent light. And in auto-collimation this is also the case. But here we use a collimated beam of light
which is generated by the optic under investigation itself. Assume you have a nice point-like source of
light on the imaging side of the lens and you have a very good optic, then basically,
this light will exit the optic in a well collimated beam. If we now place a mirror in front of the lens
that directs the light back in the reverse direction, this light will be imaged back
into a very small spot. However, if the optic is not perfect then,
first it will not leave the lens perfectly collimated. But second, on the way back it will encounter
the same aberrations again and this will result in a much larger focal spot. And because the light passes the device 2
times, this method is 2 times more sensitive to optical imperfections. Now of course this is not all, because we
need a quantitative way to measure the nature of the aberrations and we can do this by using
interferometry. In this case, we take a coherent light source
like a laser diode, we focus the beam tightly into a very small spot and from this create
a divergent beam of light. We then use a beam splitter to split this
beam into 2 parts: one part is used to create a reference wavefront and is directed into
a spherical mirror, which basically focusses the light back into a very small point on
this side. And after the light has passed the focal point,
this wavefront is nicely spherical. The other fraction of the light goes into
the optical system under test, which is in auto-collimation and also returns the light
to that same point. But this wavefront now contains information
about the aberrations of the lens. And so, by evaluating the interference pattern
between the 2 different wavefronts, we can calculate the errors present in the lens. And in this way, we can measure the wavefront
error and identify the problems in the optic. And of course, this setup is not only useful
for mirror lenses but can in theory be used for any confocal lens. Now, this setup does not really look very
complicated, right? But in practice, it is not that easy to accurately
perform this type of experiment, for various reasons: you need quite a lot of high-quality
tools, like a fairly good beam splitter, a high-grade confocal lens, a good spherical
reference mirror and optical flat of sufficient size. And on top of that you need tools to accurately
place and orient the different components correctly to achieve good alignment between
interferometer, lens and optical flat. So let me show you the actual setup that I
built and explain a few things. Here is the coherent light source, a red laser
diode, which is on an aluminum platform together with a gradient index lens, the beam splitter
and the spherical reference mirror. And, this creates a spherical wave front in
the direction of the camera. By the way, to validate the quality of this
part of the setup, I referenced it to several very high-quality spherical Perkin-Elmer wafer
stepper mirrors. For those of you interested, here you see
the calibration measurement, showing that the interferometer part has sufficient accuracy
to do measurements with. Next, we want to use this setup to measure
a lens in autocollimation and it is mounted in this contraption, which allows for the
manipulation of angle and position over a total of 6 degrees of freedom. This lens manipulator is pretty vintage. I saved it from the trash bin a few years
ago and have used it on quite a few occasions because it is a very versatile instrument. On the other side of the lens under test,
I placed a large optically flat mirror, which is a Zygo zerodur dichroic mirror also bought
on Ebay. Although the reflectivity at the laser wavelength
is only 50% it is a very good large diameter optical flat. On this side of the setup, we can now record
the interference pattern with a camera. Based on the recorded images we can create
a wavefront map using the open-source software DFTfringe. I’ve discussed this process in earlier videos
in some detail and I will put a few links in the description in case you want to download
the software or want to know how to use it. To check if the setup is working, I first
tried a few other lenses like the conventional 400mm Panagor lens, previously used in the
visual comparison. Using the interference patterns recorded with
this lens, we can calculate the wavefront error. It turns out that the measurement mainly shows
a spherical aberration, which you can identify by the bump in the center. We also observe this peculiar type of astigmatism,
which shows up as variations in the wavefront error as a function of the angle with respect
to the center. This type of aberration is called trefoil-astigmatism. Now, trefoil is frequently caused by mechanical
stress, for example by fixation screws, which in a lot of optics are placed under 120 degrees
angle with respect to each other. When too tightly fastened, they can result
in the deformation of an optical component, which can then cause this type of wavefront
error. So, the trefoil wavefront error indicates
that there might be a stress related issue somewhere in this lens with a mechanical cause. So, given the presence of these 2 main aberrations,
how do they influence optical performance? Well, as said before there are many different
ways to define this property. but for this video I just wanted to use something
simple that can be represented by a single number. I chose the Strehl ratio, because to touches
upon the essence of what a good optic should actually be doing. Okay, what is the Strehl ratio? Well, say we have this perfect lens which,
for this example, is focusing light from a point source at infinity into the smallest
spot possible. As you can see here, the light is not focused
to a single point but is a spot of certain size with some rings around it. And this is because of the wave character
of light. This smallest spot possible is also called
the Airy disk. If you make a cross section of the spot and
plot intensity vs. position we get this curve. Let’s say that the spot is centrosymmetric
and that this curve represents the light distribution imaged by the lens. Now say we have a less than perfect lens that
gives us a somewhat different intensity pattern. This means that some of the light ends up
in the wrong places. And of course, if you measure the intensity
curve, this results in us a different intensity distribution. Now say we have normalized these curves such
that the area under both curves is the same. For this situation, the Strehl ratio simply
represents the fraction of the light that ends up where it should have, if the lens
were perfect. So basically, if we take the curve of the
perfect optic, overlap it with the imperfect one, the Strehl ratio is defined by the fraction
of the area that is located under both curves. So, what this means is that if the Strehl
ratio of an optic is 0 then light ends up everywhere but where you actually want it
to end up. And if the Strehl ratio is 1 then the optic
is perfect because both curves are identical. And so, every value between 1 and 0 specifies
a certain degree of optical imperfection. For the 400mm telephoto lens test, our software
calculated the Strehl ratio of to be around 0.38. And even though this indicates that it is
not perfect, it is not as bad as you might think. Anyway, now that we know the setup is working,
let’s measure the wavefront error and Strehl ratio of the Vivitar lens. And here is the result of that measurement. It shows us that is indeed this is one really
bad lens. For starters, the Strehl ratio is calculated
to be 0, which means that virtually all of the light ends up in the image plane where
it was not supposed to go. I know this sounds funny because you still
see an image when you look through it, but it explains the extreme fuzziness of the lens. The light distribution would actually look
something like this, which is a far departure of the ideal airy disk. And the main cause is a large spherical aberration,
which is in the order of 1.5 lambda RMS and about 4 lambda Peak to Valley. It’s also got quite a lot of astigmatism
in the X and Y direction. And this suggests deformation or misalignment
of one of the optical components inside. And the value of the wavefront error could
certainly explain why we get such unsharp images with this lens. Now, let’s quickly compare this to the 600mm
sigma lens, which demonstrated much better optical performance. Here is the result of the Vivitar lens for
reference, and here is the identical wave-front error measurement on the Sigma lens. I deliberately put these two wavefront plots
on the same Z-scale so they are easier to compare. What we see is that the Strehl ratio of the
Sigma lens is only around 0.12, which is not all that great. But it is of course a huge improvement over
the 0 value of the Vivitar lens. Also, the RMS-wavefront error is 0.23 lambda,
7 times smaller than the value of the Vivitar lens. And if you look at these two measurements
of the wavefront error, the huge difference in the image quality between the 2 lenses
is not really surprising. Now, before you say that the Sigma lens apparently
is not a very good lens either, please keep in mind that the designs of lenses for photography
always contain a lot of compromises. Contrary to the monochromatic laser light
that is used in this experiment, a photography lens has to deal with multiple wavelengths. We want the different colors originating from
a white light object to end up together in one spot and of course we want this property
in the complete image field. Also, we want as little image deformation
as possible. On top of that, we want the lens to produce
acceptable images over its complete focusing range. And in some cases, we require planar projection
properties of a lens. So, what I mean to say is that image sharpness
is not just about measuring a Strehl ratio on the optical axis using monochromatic light. It is determined by many other parameters. And the measurement method presented here
is just a way to determine which aberrations are present, given a specific optical configuration
and wavelength. Okay, so now that we identified the main problem
with the Vivitar lens, it might be possible to find ways to improve the optical performance
a bit. But this is actually quite an elaborate process. So, in a follow-up video, we are first going
to take a look what’s inside mirror lenses and discover how they work. And I’ll try to improve the sharpness of
the Vivitar lens a bit, by modifying one of the 2 reflective surfaces inside. In addition, I will show you how I gave Mr.
Squinty the Soligor mirror lens his razor-sharp eyesight back. So, till next time!
