Opticallimits

Full Version: next PZ lens test report: Sigma AF 150mm f/2.8 EX HSM DG APO macro OS (EOS)
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4 5 6 7 8
[quote name='miro' timestamp='1312964621' post='10567']

The problem with encoder is not resolution. We have novadays rotary encoders with 37 bits/revolution. However the accuracy is still counting in percents % /e.g. 0.01...1%typical value /. When you look at high end encoders from Heindenhain you will come back to Ziess. They use the Zeiss high quality glass for their encoders.

However I woud say that there are not extreme expansive – e.g price of 2000..3000Euro for such encoder or target is still acceptrable???



What will be more exaensive is

1. Keep clean target

2. ALignment of Target

3. Alignment of camera

4. Precisision foucus

5. Environmental disturrbanceses. E.g trilling

======================

I have few ideas. But all of them are very expensive.

If you have some ideas how to solve all 5 issues above please let me/us know.

I'm interested of such setup too

[/quote]



I think these issues are solvable. An absolutely clean target is not necessary - only the edge to be measured must be clean.

Alignment is primarily a focus issue and you can refocus for the spot to be measured. That would mostly rely on CD-AF which is usually extremely accurate at least at max. aperture.

It's still a quite hypothetical discussion I think. We just can't afford to put vast resources on this specific aspect. There are further characteristic which are generally more important than macro performance (other than for macro photographers of course).
[quote name='Christos' timestamp='1312965130' post='10568']

Klaus other than the fact that the EF 100 L IS is 100mm which one is the better lens?

[/quote]



They are darn close. The Sigma has virtually no bokeh fringing whereas the Canon provides a better bokeh. The IS is superior to the OS in macro scenarios. The Sigma has a bigger working distance.
[quote name='Klaus' timestamp='1312970601' post='10572']

They are darn close. The Sigma has virtually no bokeh fringing whereas the Canon provides a better bokeh. The IS is superior to the OS in macro scenarios. The Sigma has a bigger working distance.

[/quote]

The Canon also is weather sealed.



Both are nice, as is the Tarmon 180mm f3.5 (which lacks both the HSM/USM and thew IS/OS).

Guest

[quote name='miro' timestamp='1312966079' post='10569']

Ooops,

The current Imatest target is reflective while encoders base is transluctive.

Target has dark and white areas.The targhet is with high contrast. While most optical encoders are transluctive – metal grid over glas or trans[perant plastic.

Back illuminating is not an option. Since another side effects will occures – scattering, refraction etc. The encoder manufactures do a lot of trics to acheve resonacble sharp images. remember the encoder head is looking for 0 and 1 while we must look for nuaces of light

SinCos encoder is exection of 0s rule, but high end manufacures guaranties 1% THD by perfectly aligned encoder

[/quote]



miro, I am not sure but I think you made the wrong assumptions.

The optical encoder I used needed a reflective surface. This was achieved the traditional way by sputtering the wafer with chrome after the lithography process. So this requirement of Imatest can be served.

Concerning the sensor signal, I might misunderstand you. The sensor does not look for 0 and 1, thats the microcontroler input. The sensor measures level of light sensivity, which are transformed into digital outputs by the quadrature encoder via A and B channels. I think this is what you meant when talking about the SinCos encoder function and zero points. But then, what influence does this fact have on what we need? Is it sharpness of the pattern you are concerned about? The lithography mask we used was created without optical processes involved. The optical errors that can lead to low contrast edges on the final pattern on the wafer during lithography, yes. But this is a problem in general and the magnitude of the error is related to the structure size we need.



Klaus suggestion of 20k lines per inch calculates back to ~1.27µm distance from line to line, making it ~0,63µm line width. I have to confess, this is probably out of reach for the capabilities I could provide. But in these dimensions, structure contrast should still not be a problem.



Christian
[quote name='Yakim' timestamp='1312961304' post='10566']

Yes, but the price and weight increase and WD decrease also puts it further away from the excellent (and cheap) Tamron 180/3.5. Not an easy decision.



If most of my macro subjects were insects I'd pick the Tamron, as WD would be more important than OS or max aperture. If not, the 100/2.8 IS L looks to be a better all-around lens because it's much smaller and lighter.



As a consumer I applaud Sigma for thinking out of the box and giving us unique lens options but despite its unique capabilities, I think that the target market of the 150/2.8 OS is rather small.

[/quote]

While I agree that if pure WD is the motivation, any 180mm would be superior to a 150mm or shorter, I think the bigger question is to stabilise or not. There you only get two choices (in Canonland) the 100L and now the Sigma 150 OS. While the 100L was somewhat interesting to me, I never bit as from memory it doesn't easily take teleconverters so would significantly limit its potential. The Sigma 150mm OS is a more natural choice there. I also think a 150mm with optional 1.4x TC is the more versatile choice over a straight 180, and now you can get the 150 with OS it is pure win.



The Sigma 150mm non-OS was and still is hugely popular, and I do not doubt the OS version would be different there. I am wondering if I should pay the extra to upgrade from the non-OS...
[quote name='TheChris' timestamp='1312975929' post='10574']

miro, I am not sure but I think you made the wrong assumptions.

The optical encoder I used needed a reflective surface. This was achieved the traditional way by sputtering the wafer with chrome after the lithography process. So this requirement of Imatest can be served.

Concerning the sensor signal, I might misunderstand you. The sensor does not look for 0 and 1, thats the microcontroler input. The sensor measures level of light sensivity, which are transformed into digital outputs by the quadrature encoder via A and B channels. I think this is what you meant when talking about the SinCos encoder function and zero points. But then, what influence does this fact have on what we need? Is it sharpness of the pattern you are concerned about? The lithography mask we used was created without optical processes involved. The optical errors that can lead to low contrast edges on the final pattern on the wafer during lithography, yes. But this is a problem in general and the magnitude of the error is related to the structure size we need.



Klaus suggestion of 20k lines per inch calculates back to ~1.27µm distance from line to line, making it ~0,63µm line width. I have to confess, this is probably out of reach for the capabilities I could provide. But in these dimensions, structure contrast should still not be a problem.



Christian

[/quote]



Well, technically we would only need a "perfect black/white" transition (actually a very dark gray, white). A normal printer can provide an exceedingly sharp edge as long as the test quadrant it is not tilted (which would introduce the typical edge "stepping" issues). We just tilt the print in order to have the suggested 5 degree tilt of the edge which must be measured.

So we don't need something like fine line-pairs. I reckon this is much easier to achieve but I'm clueless regarding the details of the lithography process.

I may be interesting to do some proof-of-concept testing based on suitable scrap material. As mentioned any significant investments are totally out-of-scope anyway. This is a small scale operation here - basically a side job of three enthusiasts rather than a really serious business.



cheers



Klaus
[quote name='Brightcolours' timestamp='1312974816' post='10573']

The Canon also is weather sealed.

[/quote]



Sigma claims sealing for the 150 OS, too.



Regarding the discussion around 150/180 mm: the 150 non-OS felt more like a shorter lens (physically, which means: in the bag). That's the biggest drawback of 180mm lenses IMO, they are significantly larger than the Sigma 150. Or to put it differently: that's what probably made the 150 so popular, it's quite short for such a long lens.



-- Markus

Guest

[quote name='mst' timestamp='1312990811' post='10580']

Sigma claims sealing for the 150 OS, too.



Regarding the discussion around 150/180 mm: the 150 non-OS felt more like a shorter lens (physically, which means: in the bag). That's the biggest drawback of 180mm lenses IMO, they are significantly larger than the Sigma 150. Or to put it differently: that's what probably made the 150 so popular, it's quite short for such a long lens.



-- Markus

[/quote]





....It's quite short for such a long lens.



Is there Irish blood somewhere in you, Marcus?



Sounds like the punch line to a Paddy & Michael joke



Beautiful
I wish I could claim it was intended <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/wink.gif' class='bbc_emoticon' alt='Wink' />



-- Markus
[quote name='mst' timestamp='1312990811' post='10580']

Sigma claims sealing for the 150 OS, too.



Regarding the discussion around 150/180 mm: the 150 non-OS felt more like a shorter lens (physically, which means: in the bag). That's the biggest drawback of 180mm lenses IMO, they are significantly larger than the Sigma 150. Or to put it differently: that's what probably made the 150 so popular, it's quite short for such a long lens.



-- Markus

[/quote]





Ok, 5* in this case in the mechanical section.



Pages: 1 2 3 4 5 6 7 8