Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
Reading another thread got me thinking about the claim that the QV series objective is just a re-labelled "normal" mitty. So I did some tests.
Here are setup info:
Mitutoyo MPlan 5x 0.14NA and Mitutoyo QV 2.5x 0.14NA
Canon 100mm macro lens is used as tube lens, giving magnification of 2.5x for both objectives
Sony A7M3 in crop mode because the QV and the 5X on 100mm do not cover full frame.
Both stacks were done 212 images, step size of 40um
My test result: I think the 'normal' 5x APO is better than the QV in terms of telecentricity even though QV series are suppose to be telecentric. I know I can use Zerene to check telecentricity, but I think the difference between the two is very small, so I used another way -- stack them twice, the "normal" loading order and intentionally reversed order. Then compare the two stacked images for both objective.
"normal" 5X mitty QV 2.5x mitty As you can see in circled areas, the images stacked using QV objective have larger gaps, implying there are more differences between stacked in loading order and reverse order. The difference shown here is largely due to changes in images size (could be other factors, such as alignment due to difference in scene). By default, Zerene is aligning images against the first image (hence the importance of stacking order), whether separately or not, bigger gap in comparison means larger difference in size seen by the camera. The more telecentric an objective is, the LESS difference.
Finally, here is a video, (on my personal account, please DO not subscribe :-)) https://youtu.be/onCNdsqtDSw
Here are setup info:
Mitutoyo MPlan 5x 0.14NA and Mitutoyo QV 2.5x 0.14NA
Canon 100mm macro lens is used as tube lens, giving magnification of 2.5x for both objectives
Sony A7M3 in crop mode because the QV and the 5X on 100mm do not cover full frame.
Both stacks were done 212 images, step size of 40um
My test result: I think the 'normal' 5x APO is better than the QV in terms of telecentricity even though QV series are suppose to be telecentric. I know I can use Zerene to check telecentricity, but I think the difference between the two is very small, so I used another way -- stack them twice, the "normal" loading order and intentionally reversed order. Then compare the two stacked images for both objective.
"normal" 5X mitty QV 2.5x mitty As you can see in circled areas, the images stacked using QV objective have larger gaps, implying there are more differences between stacked in loading order and reverse order. The difference shown here is largely due to changes in images size (could be other factors, such as alignment due to difference in scene). By default, Zerene is aligning images against the first image (hence the importance of stacking order), whether separately or not, bigger gap in comparison means larger difference in size seen by the camera. The more telecentric an objective is, the LESS difference.
Finally, here is a video, (on my personal account, please DO not subscribe :-)) https://youtu.be/onCNdsqtDSw
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
Just in case, here are stacked (loading order) images for both objectives.
Mitty 5x QV 2.5x
Mitty 5x QV 2.5x
-
- Posts: 1617
- Joined: Sat Apr 14, 2018 10:26 am
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
The telecentricity difference is interesting, though mounted on a QV system the situation may be different (possibly? Does the tube lens or spacing play into this at all?). They have tiny camera sensors so even with a 100mm lens only the very center would be used, not that it seems to vary much from center to edge here.
I'll admit I'm most curious about how the color correction and sharpness compare (and to a lesser extent planarity). For being on a 100mm lens on aps-c, effectively a field number of ~56mm, the results of both look quite nice and very similar across both. I'll admit I'd love to see a couple wafer shots with 100% crops, but I didn't actually do any myself when i had a chance so I certainly can't complain if you leave it at this.
I'll admit I'm most curious about how the color correction and sharpness compare (and to a lesser extent planarity). For being on a 100mm lens on aps-c, effectively a field number of ~56mm, the results of both look quite nice and very similar across both. I'll admit I'd love to see a couple wafer shots with 100% crops, but I didn't actually do any myself when i had a chance so I certainly can't complain if you leave it at this.
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
Hi Peter,
Best, ADi
I think so too and use the normal 5x with the tube lens 100mmQV series objective is just a re-labelled "normal" mitty
Best, ADi
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
Good for you. I have ALWAYS believed that the QV is better in terms of telecentricity until I did this test, mostly because QV is suppose to be telecentric and it is marketed as so by Mitutoyo. Of course, there are probably many reasons for this, one example is the environment they are used, I THINK (not a fact) QV is used for industrial automation, measurements and thus the different requirements.
Lesson learned for me is that it is always a good thing to actually get some hands-on experience.
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
Just to clarify it a bit: I am trying to "dis-prove" (sorry, lack of English words here) telecentricity of both objectives and by stacking images in both orders (normal loading and reversed), compare the results. The more difference, the less telecentric they are and the results are visual as I am more or less a visual person. Zerene has a "hidden" feature that allows you to quantify it, but since the two are so close, those numbers will be close, too, less dramatic.Scarodactyl wrote: ↑Thu Oct 14, 2021 11:23 pmThe telecentricity difference is interesting, though mounted on a QV system the situation may be different (possibly? Does the tube lens or spacing play into this at all?). They have tiny camera sensors so even with a 100mm lens only the very center would be used, not that it seems to vary much from center to edge here.
I'll admit I'm most curious about how the color correction and sharpness compare (and to a lesser extent planarity). For being on a 100mm lens on aps-c, effectively a field number of ~56mm, the results of both look quite nice and very similar across both. I'll admit I'd love to see a couple wafer shots with 100% crops, but I didn't actually do any myself when i had a chance so I certainly can't complain if you leave it at this.
I really do not think you can do this test, telecentricity, with wafer as you need to do a deep stack. The tests I did had about 8mm in depth, this is needed as both are "near telecentric", without dramatic depth, it is hard to see the difference.
Lastly, I am not sure if there are difference between center and edge because the subject is mostly in the center and edges are blank space. Anyways, the tests are not about sharpness, nor any other factors except telecentricity.
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
2nd thought, wafer might be an excellent subject to test: shoot it at an angle to create depth . . . and by having an angle, lines, patterns on the wafer can be made closer (to camera). Unfortunately, I do not have a wafer and also too lazy
-
- Posts: 1617
- Joined: Sat Apr 14, 2018 10:26 am
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
I understand the test was for telecentricity, and I think it's an interesting and puzzling result. I don't know enough about the logistics of telecentricity to know how much the tube lens affects it, though the spacing to the tube lens must somehow(?). Leica also has a measuring system with an added spacer on an otherwise normal stereo objective for telecentricity so it's not just the QVs.
And much as I'd be interested in a more conventional image quality comparison a couple 100% crops of these tests would probably give a good idea on that front too.
And much as I'd be interested in a more conventional image quality comparison a couple 100% crops of these tests would probably give a good idea on that front too.
- rjlittlefield
- Site Admin
- Posts: 23561
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
Telecentricity is determined by the location of the limiting aperture, as seen from the front of the lens.
In most of the systems we work with, the limiting aperture will be within the objective, and in that case telecentricity depends only on the objective, not at all on the tube lens.
In the case of Leica's spacer, the limiting aperture must be located above the objective. This would be common for stereo microscopes, which change their working NA depending on zoom settings in the head.
In some cases there can be different zones, for example the optics may be telecentric in some small area around image center, but then become non-telecentric farther out. See viewtopic.php?t=39724 for discussion.
See for example the latter section of the first post at viewtopic.php?t=39677 .
--Rik
In most of the systems we work with, the limiting aperture will be within the objective, and in that case telecentricity depends only on the objective, not at all on the tube lens.
In the case of Leica's spacer, the limiting aperture must be located above the objective. This would be common for stereo microscopes, which change their working NA depending on zoom settings in the head.
In some cases there can be different zones, for example the optics may be telecentric in some small area around image center, but then become non-telecentric farther out. See viewtopic.php?t=39724 for discussion.
I disagree. The best indication of telecentricity comes from comparing images where the same features are nearly in focus at all times. With an arbitrary subject and a deep stack, the alignment process is very likely to be misled by changes in the appearance of the subject as it goes in and out of focus. Even optics that are accurately telecentric can appear to be not so telecentric, if evaluated on images that are too much OOF.I really do not think you can do this test, telecentricity, with wafer as you need to do a deep stack.
See for example the latter section of the first post at viewtopic.php?t=39677 .
--Rik
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
OK, I still do not get what you intend to say .Scarodactyl wrote: ↑Fri Oct 15, 2021 8:12 pmI understand the test was for telecentricity, and I think it's an interesting and puzzling result. I don't know enough about the logistics of telecentricity to know how much the tube lens affects it, though the spacing to the tube lens must somehow(?). Leica also has a measuring system with an added spacer on an otherwise normal stereo objective for telecentricity so it's not just the QVs.
And much as I'd be interested in a more conventional image quality comparison a couple 100% crops of these tests would probably give a good idea on that front too.
The word "spacer" usually means (to me) an empty space between two objects. If that is what you meant and suggesting the QV series is just a normal 5x mitty + some empty space, then I do not think that is the case at all. Light rays coming out of the mitty 5x (an infinite one) are the so called parallel rays, you can add as much empty space, or in your words, spacer, as you want without changing anything (maybe losing some intensity) as long as you have a tube lens to re-focus these parallel rays onto sensor plane. So there must be something optical added in the extra segment at the end of QV.
Adding some optics at the end of another lens makes it a complete different optical system, even if the front part is a normal 5x mitty.
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
@Scarodactyl
for your spacer idea, you can do this experiment: get an infinite objective, add 5cm extra space between it and your tube lens and see what happens
for your spacer idea, you can do this experiment: get an infinite objective, add 5cm extra space between it and your tube lens and see what happens
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
Totally agree.rjlittlefield wrote: ↑Fri Oct 15, 2021 8:39 pmTelecentricity is determined by the location of the limiting aperture, as seen from the front of the lens.
In most of the systems we work with, the limiting aperture will be within the objective, and in that case telecentricity depends only on the objective, not at all on the tube lens.
In the case of Leica's spacer, the limiting aperture must be located above the objective. This would be common for stereo microscopes, which change their working NA depending on zoom settings in the head.
In some cases there can be different zones, for example the optics may be telecentric in some small area around image center, but then become non-telecentric farther out. See viewtopic.php?t=39724 for discussion.
I disagree. The best indication of telecentricity comes from comparing images where the same features are nearly in focus at all times. With an arbitrary subject and a deep stack, the alignment process is very likely to be misled by changes in the appearance of the subject as it goes in and out of focus. Even optics that are accurately telecentric can appear to be not so telecentric, if evaluated on images that are too much OOF.I really do not think you can do this test, telecentricity, with wafer as you need to do a deep stack.
See for example the latter section of the first post at viewtopic.php?t=39677 .
--Rik
Then again, relative comparison between two objectives, a deep stack can make it easier to tell which one is less telecentric (my dis-prove clarification) between the two. Unless there are situations this is not the case.
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
OK, this is something new for me (to learn), thanks.In some cases there can be different zones, for example the optics may be telecentric in some small area around image center, but then become non-telecentric farther out. See viewtopic.php?t=39724 for discussion.
- rjlittlefield
- Site Admin
- Posts: 23561
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
Sure, if you want to assume that both stacks are affected equally by possibly serious errors caused by too much OOF.
Personally I do not like to make such assumptions, when accurate measurements are a simple alternative.
--Rik
Re: Mitutoyo MPlan 5x vs Mitutoyo QV 2.5x Comparision
Isn't telecentricity errors suppose to have central symmetry? Clearly, misalignment on the both images doesn't - it appears moved to the upper right corner for the entire field. So that's not scale change. Probably the difference caused by QV objective being longer (due to the spacer), so when tilted on some angle, the same as PlanApo, this results in greater shift.