View Full Version : Lecture on Petzval Lenses, Tucson, Oct. 5

Mark Sawyer
2-Oct-2017, 19:54
For anyone in Tucson, AZ who's interested in historic optics, Dr. Jose Sasian of the University of Arizona's College of Optical Sciences will be presenting a lecture on "Joseph Petzval's Lens Design Approach" this Thursday, Oct. 5, 2017, at the monthly meeting of the Western Photographic Historical Society. The meeting is at 7:00 pm at the Ward Six City Council Meeting Rooms, 3202 East First Street, Tucson, AZ.

Hope to see a few of you there!

Nodda Duma
3-Oct-2017, 19:50
After many years of working as a self-taught lens designer, I went through U of A's Optical Sciences graduate program to get some formal education. Dr. Sasian was my professor for a couple of lens design courses. Very good instructor, and I was always fond of the historical perspective he provided to the design forms. He taught Petzval's design as part of his course, amongst many other well-known forms.

If you are in the area, I highly recommend attending his talk. I would love to watch a live feed if it was available.

Steven Tribe
4-Oct-2017, 02:15
If I could attend, I would ask the following questions. I have never found a satisfactory answer to them!

1. What did the redesign of the Petzval/Voigtlander Petzval - in order to bring together the visual and chemical focal lengths - consist of?

2. The Dallmeyer revision of the rear elements in order to introduce softness into the Petzval was considered as bogus by many of his competitors. They claimed that increasing the distance between the front and rear elements of the standard design acheived the same sort of "flou". Has anyone done any testing of the differences?

Mark Sawyer
4-Oct-2017, 10:31
I'll pass on your questions, Steven, but wasn't it Lerebours who did the redesign to synchronize chemical and visual focus around 1850?

Dallmeyer did several modifications to Petzval's design starting in 1860 or earlier. The 1867 design was the one that introduced softness, but it was done to spread the depth of field, not create a "flou" effect. I've noticed on other lenses that have adjustable softness that increased spherical aberration does increase depth of field, "sort of"...

Introducing enough spherical aberration to spread the Depth of field creates a very soft effect overall, so while there starts to be a sharp core behind the out-of-focus area, it's still dominated by softness, and the main focal plane has a noticeable soft overlaying image too. The overall effect is more soft than spread-depth-of-field, and if you close down the aperture to gain conventional depth of field, the spherical aberration goes away and you're back where you started.

In my limited playing with My Dallmeyer 3b and Vitaxes (same diffusion design), my observation is that the limited travel of the rear elements doesn't provide much softness, just a touch, like a still-restricted Velostigmat Series II with the diffusion ring still factory-limited, or the Cooke Portrait Lenses. Enough to soften skin texture a bit, but not a Pictorial effect, or enough to noticeably effect depth of field. For what it's worth...

4-Oct-2017, 15:46
History records Lerebours as optimizing the color focus of the Petzval design.

See you at the lecture Mark. It will be nice to see what he's been up to since our dinner that time.

Nodda Duma
24-Oct-2017, 12:28
How did the lecture turn out?

Mark Sawyer
24-Oct-2017, 14:07
The lecture was well-attended and was quite interesting, at least to the optics aficionados. In response to Steven's question about using spherical aberration to spread depth of field, Prof. Sasian cited a passage by Rudolf Kingslake that spherical aberration "created the appearance" of more depth of field. He also showed images of some of Petzval's original calculations, and cleared up an old story that Petzval used the services of some forty Austrian artillery men to help with the design mathematics. Petzval did all the original design work alone, while the artillery men did later calculations on variations of Petzval's design, (presumably various focal lengths and apertures?).

Steven Tribe
24-Oct-2017, 16:38
It is nice that some of the old repeated stories have a grain of truth in them. I was checking various addresses of the early Vienna Petzval maker Waibl on a old map of Vienna from the 1850's and found the location of the address on my lens was next door to the Imperial Artillery College. Thanks for asking.

27-Oct-2017, 08:18
I liked it. I asked a question but didn't explain it well, so didn't really get an answer. It was, how precisely matched are the different rear elements to each other, and the rear group to the front? I ask because I wonder if each piece of glass is hand ground and optimized for it's refractive index, relative to the other. In cases where some is missing an element or the front group, would the image quality suffer if they replaced it with one from another lens?

He said the design was very forgiving. But I don't think he understood I wanted quantifiable data.

Nodda Duma
27-Oct-2017, 15:43
You're basically asking about the fabrication and assembly tolerances that can be accommodated by the design. Tolerances include lens decenter (element offset from the ideal optical axis), wedge, the tolerance on the surface curvature, lens spacings, amongst others.

Do a Google search for "optimax tolerance chart" (but without the quotes). The chart you want to look for has three sets of tolerances: commercial, precision, and (state of the art or limit, I don't recall which).

His "forgiving design" response to me says commercial-grade tolerances or looser...perhaps even as loose as the catalog optics you can buy from, say, Edmunds or Thorlabs (you can see these values in the online catalogs).

As for optimized for refractive index:

The indices of the achromat elements are calculated to balance color. However, the design is not adjusted for melt index data...that doesn't fly for mass production. In practice, the assembly tolerances (like decenter and lens spacing) are adjusted to accommodate the expected variation in index. For the classic Petzval, you don't even need to know what the index tolerance is...focus adjustment and - if needed - spacing between objective and field groups accommodates even the variations of the glass indices of the 1800s.