Takahashi FSQ 85 ED 光學討論

檢測望遠鏡光學部份,赤道儀機械組件,目鏡配件和拍攝器材 CCD等...
頭像
鄧登凳
夸克星
文章: 9493
註冊時間: 週一 03 8月, 2009 17:15
來自: 3rd planet of solar system

文章 鄧登凳 » 週二 15 6月, 2010 21:09

anguslau 寫:
Wah!! 寫:
Subaru 寫:Basically, diffraction spikes or dark band is more visible when there is a 尖角 as the obstacle, or the diffraction will spread around when the edge is a smooth one.
Agree, the diffraction flare/spikes shape only related to the SHAPE of the edges, not related to the vignetting seriousness.
Agreed! But why I say that is because as vignetting becomes more severe, the olive shape is more elongated and has sharper ends. Hence the diffraction spikes more prominent. Of course, there are a few assumptions here.

Anyway, I think discussions over the last couple of days has been fruitful and I believe I have much better understanding of the issue now. :D
The blocking of available aperture of the lens results in a olive shaped opening. The same causes optical vingetting and the non-circular diffraction spikes. Vingetting and seriousness of the diffraction spike distortion are correlated because they have a common cause.

I find this explanation here easy to follow and you can actually see the olive shape openning in this article:
http://toothwalker.org/optics/vignetting.html

頭像
昇仔
夸克星
文章: 2377
註冊時間: 週二 13 11月, 2007 17:40
來自: 香港

文章 昇仔 » 週三 16 6月, 2010 01:48

Another test which verifies that the aperture of the fourth lens in the FSQ causes the vignetting we see.

The first diagram shows the connection I use without the reducer. I am not sure about the names of the connecting rings shown on the left but the purpose is to verify that none of these rings are causing the vignetting. The method is as follows: With all the rings connected, look through the FSQ from the position of the camera. Move the eyes slightly off axis until a small part of the light cone is obscured. With the eyes stayed in position, all the connection rings (including from behind camera rotator) are slowly removed and any change in the shape of the light cone is noticed. No change is observed and the aperture of the fourth (last) lens of the FSQ is seen to obscure part of the light cone as shown in the two photos on the right. The photos are taken with a DC along the axis and slightly off axis with all the rings removed and illuminated by a headlight.

The second figure shows the same test being done with the telescope pointing at distant street lamps. This time the DC is fixed with a tripod slightly off axis behind the telescope. The photo on the left shows a small part of the light cone being obscured when all the connection rings are in place. The photo on the right shows that the fourth lens obscures part of the light cone, with all the connection rings removed and the interior illuminated by a headlight.

The observation is consistent with the ray diagram shown in the third figure, which shows that the fourth lens of the FSQ limits the light cone that can be observed at the focal plane.

The last two figures show a series of photos taken with setup from out focus (behind the focus) to in focus (in front of the focus). The dark band diffraction pattern is not obvious from this photo but they can still be seen upon careful examination. As we move from out to in focus, the pattern of street lamps shifts slightly towards the centre of the photo (i.e., to the right) as expected. This is obvious from the ray diagram. However, the shape of the lamps is clearly not symmetrical about the focus, the in-focus ones look quite spherical but the out-focus ones olive-shaped.

I suggest FSQ-106 users to see if they can observe the same things.

I think now we have understood the cause of the problem well. :D The only thing I can't explain is that why the in-focus and out-focus pattern are different. They should be symmetry about the focal plane and this property should be independent of the optical design. :twisted: They second thing I don't understand is that why Taka don't make the aperture of the fourth lens larger to reduce the vignetting :evil:
附加檔案
in-out.jpg
in-out.jpg (186.92 KiB) 已瀏覽 19234 次
frame.jpg
frame.jpg (163.39 KiB) 已瀏覽 19234 次
FSQ-106ED-lightpath-s.jpg
FSQ-106ED-lightpath-s.jpg (60.68 KiB) 已瀏覽 19234 次
obscure.jpg
obscure.jpg (71.52 KiB) 已瀏覽 19234 次
scope.jpg
scope.jpg (103.58 KiB) 已瀏覽 19234 次

頭像
willis
中子星
文章: 1804
註冊時間: 週六 14 2月, 2009 08:07
來自: Hong Kong
聯繫:

文章 willis » 週三 16 6月, 2010 07:08

Thanks 昇仔 for detailed experiment on explaining cause of vignetting. It is clear that the 4th lens element blocks part of the light. From the design, I don’t know whether the 3rd element also blocks part of the light. Anyway, blocking does matter and it causes the vignetting.

The mechanism of dark band is still unclear to me, maybe I am not knowledgeable to get the gist of why blocking part of the light in the middle of the light cone will cause dark band. The cross in Newtonian, is a kind of blocking, will cause bright diffraction spikes and how come the blocking in FSQ will case dark band not bright spikes or bright band?

Just a thought, the phenomenon of asymmetry light spot around the focus may not be understood in term of light cone. If we think in term of light cone, we think in term of geometrical optics. It is symmetrical around the focal point anyway. However, the wave nature of light is omitted. Think about the geometrical optical cannot explain the finite size of light spot. Geometrical optical tells you that a point, no area, can be formed, which is not true.

Maybe an experimental approach can be used. We can try to put an olive shape mask and place it in the middle of the light cone in ordinary design either doublet or triplet deign. In this case, we can put the olive mask in front of the reducer. So the blocking is in the middle of the light cone and see if it produces:

1. Dark band in doublet or triplet design.
2. Asymmetric spot around focus in doublet or triplet design.

** I cannot do it today because all my telescopes are in school lab now.

About昇仔’s question,
昇仔 寫: They second thing I don't understand is that why Taka don't make the aperture of the fourth lens larger to reduce the vignetting :evil:
This is an easy one. As I mentioned in my previous post that large rear lens are needed for less vignetting.
willis 寫: Secondly, Wah’s explanation on large rear lens are needed for less vignetting is correct. If the ray diagram provided by Tak is correct, it explained already. However, if a smaller rear lens is used, some light will be blacked. But the point is that the so-called image circle provided by a telescope maker depends on the tolerance. If the tolerance is big, the image circle is large. So without the tolereance, the size of the image circle quoted by the telescope maker may not reliably reflect the vignetting performance. As far as I know top class reducer/flatteners are very big in size. They need 3” or even 3.5” focuser to hold them.
Big size reducer/flatteners provide vignetting free images and at the same time, they are very expensive. Only really top class telescope makers provide this option for their perfectionist customers. See the Astrophysics flatteners below:
http://www.astro-physics.com/index.htm? ... /photo_acc
The so-called 2.7” flattener costs US$525 to US$700. From the description, it stated that it supports Pentax 6x7 Camera. However, there is another so-called 4” flattener and it costs US$2100!!! [吐血1] It also supports Pentax 6x7 Camera. BUT in the description it stated clearly that
“This optical marvel was originally designed to take full advantage of the medium format Pentax 6x7 camera, and will provide pinpoint stars all the way to the edge of its field with no vignetting!” [愛的轟炸]
In addition the telescope must be upgraded to a 4” focuser. [冷汗直流]
So the reason why Tak doesn’t use a bigger 4th lens is simple, the cost. But the trade-off is 60% luminosity of the image circle. Just like the cheaper Astrophysics 2.7" flattener which is so-called Pentax 6x7 supporting. As far as I know for truly support Pentax 6x7 reducers, 88mm image circle, all are above HK$10,000! [吐血1] There is no free lunch...

頭像
Wah!!
夸克星
文章: 13283
註冊時間: 週一 15 3月, 2004 21:28
聯繫:

文章 Wah!! » 週三 16 6月, 2010 07:59

willis 寫: The mechanism of dark band is still unclear to me, maybe I am not knowledgeable to get the gist of why blocking part of the light in the middle of the light cone will cause dark band. The cross in Newtonian, is a kind of blocking, will cause bright diffraction spikes and how come the blocking in FSQ will case dark band not bright spikes or bright band?
Have you heard about "curved spider" for Newtonian before?
A curved spider doesn't create spikes but flares around stars.
附加檔案
Spider.jpg
Spider.jpg (302.41 KiB) 已瀏覽 19157 次

頭像
Skyobs
白矮星
文章: 724
註冊時間: 週日 16 9月, 2007 22:37
來自: HKSAR

文章 Skyobs » 週三 16 6月, 2010 10:06

By the way, how does FFT pattern (is that a plot of magnitude, phase, or combined?) relate to diffraction pattern? :roll:
最後由 Skyobs 於 週三 16 6月, 2010 10:41 編輯,總共編輯了 1 次。

頭像
willis
中子星
文章: 1804
註冊時間: 週六 14 2月, 2009 08:07
來自: Hong Kong
聯繫:

文章 willis » 週三 16 6月, 2010 10:06

Wah!! 寫: Have you heard about "curved spider" for Newtonian before?
A curved spider doesn't create spikes but flares around stars.
Thanks Wah! I don't know this. Now I have one more info. to go. Can u generate the FFT of the attached graph? Many thanks!
附加檔案
cir_L.gif
cir_L.gif (3.49 KiB) 已瀏覽 19116 次
asym_cir.gif
asym_cir.gif (4.13 KiB) 已瀏覽 19116 次

頭像
Wah!!
夸克星
文章: 13283
註冊時間: 週一 15 3月, 2004 21:28
聯繫:

文章 Wah!! » 週三 16 6月, 2010 10:39

:D
附加檔案
asym_cir_144[1].jpg
asym_cir_144[1].jpg (58.86 KiB) 已瀏覽 19095 次
cir_l_904[1].jpg
cir_l_904[1].jpg (68.71 KiB) 已瀏覽 19095 次

anguslau
中子星
文章: 1188
註冊時間: 週五 15 4月, 2005 20:18

文章 anguslau » 週三 16 6月, 2010 11:51

昇仔 寫:I suggest FSQ-106 users to see if they can observe the same things.

I think now we have understood the cause of the problem well. :D The only thing I can't explain is that why the in-focus and out-focus pattern are different. They should be symmetry about the focal plane and this property should be independent of the optical design. :twisted: They second thing I don't understand is that why Taka don't make the aperture of the fourth lens larger to reduce the vignetting :evil:
I admire your diligence and thoroughness in performing these experiments! :wink: I haven't got the drive to go through all these myself. :P

To me, it doesn't really matter whether the 3rd or 4th element is the bottleneck. So long as I am pretty sure that I am using the largest opennings throughout the connector train, I know I have done all I could and unlikely to have introduced further obstructioins. I am happy. :P

If you look at the mechanical diagrams for the two scopes, I think you can appreciate the difficulty in increasing the size of the 3rd and 4th elements. Spaces are very tight already. Any significant increase in the 4th element may have tremendous implications on the entire mechanical design. Engineering is always a tradeoff between many factors - usually cost and a list of performance goals. Frankly, I think Tak has already done an excellent job on the FSQs. :wink:

I am curious to see why the in/out focus aperture shapes can be different too. Theoretically they should be the same unless we have overlooked something. I might do some experiments when I have the opportunity. I am also interested to see the shape of the partially obstructed aperture on the FSQ106.

By the way, I have examined my images with 106+reducer again very carefully. Because its amount of vignetting is similar to FSQ85 (without reducer) and may therefore have observable diffraction dark bands as well. In deed, I can now see the dark bands upon careful examination. Though they are quite faint and not very prominent. Perhaps its partially obstructed aperture has less sharp corners compared to the FSQ85. Would be interesting to compare.
附加檔案
FSQ-106EDtube.jpg
FSQ106
FSQ-106EDtube.jpg (117.63 KiB) 已瀏覽 19065 次
FSQ-85EDtube.jpg
FSQ85
FSQ-85EDtube.jpg (112.36 KiB) 已瀏覽 19065 次

頭像
Skyobs
白矮星
文章: 724
註冊時間: 週日 16 9月, 2007 22:37
來自: HKSAR

文章 Skyobs » 週四 17 6月, 2010 02:05

Skyobs 寫:By the way, how does FFT pattern (is that a plot of magnitude, phase, or combined?) relate to diffraction pattern? :roll:
With some search on the web, now I understand why use FFT to show diffraction pattern. (It's because far field diffraction pattern has the same form as Fourier transfomer of the aperture. Here's a lecture on the topic http://academicearth.org/lectures/appli ... iffraction )

However, FFT diagram with fringes as shown has some drawbacks:
1) It only represents a single wavelength.
Star light is broadband. Fringes and dips (of multiple wavelengths) beyond the first few would mix together and become a smooth flare rather than distinct fringes.
2) It contains artifact caused by discrete rectangular processing (during transform and display, and maybe after JPEG compression).
You can expect diffraction from circular aperture is the same at any angle. But the diagrams show particular "spikes" in 0,45,90,...deg.

So I would suggest apply maximum filter of width approximately 2-3 fringes on the FFT result to smooth out those fringes. This could hopefully get a better visualization of the diffraction pattern.

頭像
MANDII
夸克星
文章: 2951
註冊時間: 週二 28 7月, 2009 15:13
來自: HK

文章 MANDII » 週四 17 6月, 2010 11:11

在整个讨论过程中,PAUL NG 首次明确地,鲜明地指出问题核心是镜子设计上( Petzval类(有问题) 和非Petzval类(没问题))造成之别。
虽只列举出一个例子,但从其后的众人深层的光路理论上的探讨,的确也是能从理论上支持了这点见解。

不过,话说回来,我深信那还是不能以“是一个个别问题或者属不于属是一个问题来”来看待FSQ85 ,假如那是一个个别FSQ85的问题,
那么我想世上也不会有那么多Petzal 类的镜子的诞生了。

頭像
昇仔
夸克星
文章: 2377
註冊時間: 週二 13 11月, 2007 17:40
來自: 香港

文章 昇仔 » 週四 17 6月, 2010 11:42

willis 寫: Think about the geometrical optical cannot explain the finite size of light spot. Geometrical optical tells you that a point, no area, can be formed, which is not true.
The shape of diffraction pattern is a matter of wave optics, but the shape of the light spot in and out focus is a matter of geometrical optics. It is the latter one that we don't understand.
willis 寫:Maybe an experimental approach can be used. We can try to put an olive shape mask
Yes we may try to produce the diffraction pattern artificially
willis 寫:
Big size reducer/flatteners provide vignetting free images and at the same time, they are very expensive. ... The so-called 2.7” flattener costs US$525 to US$700. From the description, it stated that it supports Pentax 6x7 Camera. However, there is another so-called 4” flattener and it costs US$2100!!! [吐血1] It also supports Pentax 6x7 Camera. BUT in the description it stated clearly that
“This optical marvel was originally designed to take full advantage of the medium format Pentax 6x7 camera, and will provide pinpoint stars all the way to the edge of its field with no vignetting!” [愛的轟炸]
In addition the telescope must be upgraded to a 4” focuser. [冷汗直流]
So the reason why Tak doesn’t use a bigger 4th lens is simple, the cost. But the trade-off is 60% luminosity of the image circle. Just like the cheaper Astrophysics 2.7" flattener which is so-called Pentax 6x7 supporting. As far as I know for truly support Pentax 6x7 reducers, 88mm image circle, all are above HK$10,000! [吐血1] There is no free lunch...
You have already scared me with the expensive cost! [熱] No money for these expensive options!!!!

頭像
昇仔
夸克星
文章: 2377
註冊時間: 週二 13 11月, 2007 17:40
來自: 香港

文章 昇仔 » 週四 17 6月, 2010 11:49

anguslau 寫: I admire your diligence and thoroughness in performing these experiments! :wink: I haven't got the drive to go through all these myself. :P

To me, it doesn't really matter whether the 3rd or 4th element is the bottleneck. So long as I am pretty sure that I am using the largest opennings throughout the connector train, I know I have done all I could and unlikely to have introduced further obstructioins. I am happy. :P .
I am only doing things in a way that an "old" physics student may do, driven by curiosity :wink: . And I am not quite sure that whether the connector rings has obscured the image, as Mr. Tang mentioned. I feel happy when I discovered that the rings are not the cause. :P
anguslau 寫: By the way, I have examined my images with 106+reducer again very carefully. Because its amount of vignetting is similar to FSQ85 (without reducer) and may therefore have observable diffraction dark bands as well. In deed, I can now see the dark bands upon careful examination. Though they are quite faint and not very prominent. Perhaps its partially obstructed aperture has less sharp corners compared to the FSQ85. Would be interesting to compare.
Interesting, it sounds like 106 is a bit better than 85 in terms of all these effects, would like to know why if possible. :P

頭像
Wah!!
夸克星
文章: 13283
註冊時間: 週一 15 3月, 2004 21:28
聯繫:

文章 Wah!! » 週四 17 6月, 2010 14:18

昇仔 寫: Interesting, it sounds like 106 is a bit better than 85 in terms of all these effects, would like to know why if possible. :P
If we scale down 106 to fit 85, the 3rd glass of 106 is larger than 85, isn't it? :roll: :roll: :roll:
When we use the same camera, the field is narrower on 106.
Can these be the reasons why 106 is a bit better than 85?

頭像
昇仔
夸克星
文章: 2377
註冊時間: 週二 13 11月, 2007 17:40
來自: 香港

文章 昇仔 » 週四 17 6月, 2010 14:39

Wah!! 寫:
昇仔 寫: Interesting, it sounds like 106 is a bit better than 85 in terms of all these effects, would like to know why if possible. :P
If we scale down 106 to fit 85, the 3rd glass of 106 is larger than 85, isn't it? :roll: :roll: :roll:
When we use the same camera, the field is narrower on 106.
Can these be the reasons why 106 is a bit better than 85?
The CMOS size remains unchanged for the same camera, but the focal length of 85 is shorter, so the off axis angle will be larger as we go from the centre to the edge of the CMOS, suffering from more geometical obsucre. Sounds reasonable. :roll:

頭像
Subaru
夸克星
文章: 6693
註冊時間: 週三 02 7月, 2003 19:27

文章 Subaru » 週四 17 6月, 2010 14:59

昇仔 寫:I think now we have understood the cause of the problem well. :D The only thing I can't explain is that why the in-focus and out-focus pattern are different. They should be symmetry about the focal plane and this property should be independent of the optical design. :twisted: They second thing I don't understand is that why Taka don't make the aperture of the fourth lens larger to reduce the vignetting :evil:
It's due to the blocking of the entrance pupil's outer light beam by the back element (or baffle in other refractor) is less when the focal plane being observed is inside focus, and more severe on outside focus.

This is very common when observing this effect in Newtonian with the secondary. It's not uncommon the ATMer put an insufficient size secondary, when he just check the coverage of the seconday inside focus. Actually, checking the coverage at outside focus (at least some mm outside focus) is a more secure way to ensure the secondary can cover the whole primary. That's why ATMer is trying to reduce the backfocus of the Newtonian as much as possible to have a small size secondary and on the other hand able to cover the required field illumination.

回覆文章

回到「望遠鏡 廠制儀器 測試報告專欄 Telescope Testing Report Center」

誰在線上

正在瀏覽這個版面的使用者:沒有註冊會員 和 6 位訪客