Offner tester for 24 inch completed

自製天文望遠鏡 牛頓望遠鏡 杜蘇式 卡式 設計和磨鏡方法<br>
佛科試鏡法 佛科刀片儀 光柵試鏡法 光柵片
回覆文章
Chanlunlun
夸克星
文章: 3847
註冊時間: 週四 09 10月, 2003 21:06

Offner tester for 24 inch completed

文章 Chanlunlun » 週二 09 3月, 2004 02:07

Dear ATMs,

The Offner tester is completed on Saturday, 6th March. The tester is fully fabricated by solid aluminium and it consists of two lens inside – the compensator lens and the field lens.

Traditional Foucalt method for testing a paraboloid is by judging the radius of curvature of different zones on the mirror by unaided human eye. While such judgement is rather subjective it is nevertheless quite accurate for mirror of small aperture and of sufficiently long focal length. Before the advent of Dobsonian revolution amatueur mirrors tends to be no more than 12 inch aperture with a typical f/ratio no less than f/6. In such circumstances a Foucalt method is usually sufficient for figuring mirror beyond 1/8 wave, provided that the mirror maker has capable technique and sufficient experience.

However when f/ratio approaches f/4 or less or when aperture exceed 12 inches the judgement of the position of zones on mirror becoming more and more difficult and exceedingly subjective. In addition, the tolerance involves, especially with mirror of very big aperture, are so tight that traditional Foucalt test becomes very tough to be handled with. Finally if unfavorable zone appears on mirror surface, which frequently occurs with big mirror, such zone is unable to be visualize by a Foucalt test. Hence it is very difficult, if not impossible, to judge the quality of big mirror simply by Foucalt test alone.

It is by no means a new idea to use a lens to introduce just the right amount of spherical aberration to compensate that of the paraboloid under test. In such way all the light reflected from the paraboloid will focus perfectly at a point at the knife edge of the tester. Thus the testing of the paraboloid becomes very convenient – just like the testing of a sphere at its center of curvature. If the mirror darkens at once upon the knife edge, the figure must be a perfect paraboloid. Such method of test is termed as a null test.

Such idea was first proposed by Dall in 1947 and he suggested of using a plano convex lens for such compensating purpose. It is quite surprised that this null test received so little attention among ATMs. Amateurs toil away reading deep Foucalt shadows, with all the uncertainty that this process involves. When using the Dall null test, a couple of evenings』 work could provide them with the means of producing far superior surfaces.


However Dall null test still has its limitation. For mirror of aperture exceeding 20 inches a single plano convex lens can no longer provide sufficient compensation for aberration of the paraboloid under test. Here Offner came to rescue the situation. Instead of using one, in the Offner null test, two plano convex lens are being used together. The first one is called the compensator and the second one is called the field lens. Theoretically speaking, the compensator provide compensation for the first order spherical aberration while the field lens take care the rest of the high order spherical aberration.

The Offner null test is a very powerful test. It can handle mirror of aperture up to 100 inch and f/ratio as low as f/2.5. Thus it is the standard test employed by observatory, and, by the Hubble Space Telescope, for testing their primary mirror. It is not until every recently that the technique of this powerful method is discovered by amateur. I would rather say there are only a few ATMs in the world up to date know about, and practically has carried out, this test method.

In the Offner null test the dimensions has to be very precisely measured and calculated. The story of Hubble Space Telescope can testify above statement. The field lens of the tester for the primary mirror of HST is placed at a wrong position by about 1 mm, the primary mirror thus produced an error of several wavelengths. Nobody knew about this until the telescope had sent into space. Of course everyone knew about the disaster that followed.

Thus Offner null test cannot be trusted alone, just because you will have all kinds of errors - errors in lens, errors in the distance between tester and mirror, errors in the location of knife edge….all will contribute to the final test quality. Consequently the Offner null test must be supplemented by a star test. Actually the star test will be the final judge, the Offner null test only acts as a guide for the figuring process.

We have finished our Offner tester. With the most careful measurement the two lens separation, vertex to vertex, is 121.06 mm, plus or minus 0.05 mm. The distance between the slit to the compensator lens is variable, depends on where you would like to place the knife edge.

With the advancement of technology, we are now being equipped with techniques and materials that are unthinkable 10 years ago – computer optimazation ray tracing software ( such as OSLO ), high intensity monochromatic LED light source, high quality floating glass with very zero stress, finite element analysis for calculation of mirror support, low priced giant diagonal mirror, super fast polishing powder, ultra wide angle eyepiece even corrected for coma and field curvature of fast Newtonians …………

Without these we cannot even dream of making giant mirror such as what we are doing now.

Best regards

Chan Yuk Lun
8-3-2004
附加檔案
Offner_tester.jpg
The finished Offner tester. It looks just like a torch, just that it sents out a divergent beam, with just the right amount of spherical aberration to compensate a 24 inch f/3.5 paraboloid.
Offner_tester.jpg (19.22 KiB) 已瀏覽 10522 次
Offner_components.jpg
The components of the Offner null tester. The white lens is the compensator, taken out from a 26 mm Orthoscopic eyepiece. The blue lens is the field lens, from a toy binocular. The left top corner is the lens barrel, next to it is the slit housing. In the
Offner_components.jpg (31.13 KiB) 已瀏覽 10522 次

頭像
alyf
主序星
文章: 174
註冊時間: 週日 29 6月, 2003 21:58

文章 alyf » 週二 09 3月, 2004 12:51

製作非常認真!!!!!!
陳sir可否提供光學圖讓我們更清晰的理解null test的概念呢?
many thanks~~

頭像
Wah!
夸克星
文章: 9073
註冊時間: 週六 28 6月, 2003 01:19
聯繫:

文章 Wah! » 週二 09 3月, 2004 12:52

有無結構圖睇下呀?
同埋想知field lens有乜用... :roll:

苏鲁锭
主序星
文章: 280
註冊時間: 週六 19 8月, 2006 21:34

文章 苏鲁锭 » 週一 28 6月, 2010 13:55

折射补偿检测,可以将剩余球差控制到很小。
但是折射透镜的曲率半径,厚度,间距,玻璃的折射率。。都要测量到很高的精度。
特别是凸面的曲率半径的测量和面型的精确检测不容易。
反射补偿检测,虽然对大口径小焦比的非球面补偿后任然有一定的剩余球差。补偿球面也比补偿透镜大很多。
但是补偿球面是凹面,很容易精确检测。曲率半径要求也不严格,只要有严格的面型就行。
所以业余爱好者最好做反射补偿检测。这样容易一些。

Chanlunlun
夸克星
文章: 3847
註冊時間: 週四 09 10月, 2003 21:06

文章 Chanlunlun » 週二 29 6月, 2010 08:32

苏鲁锭 寫:折射补偿检测,可以将剩余球差控制到很小。
但是折射透镜的曲率半径,厚度,间距,玻璃的折射率。。都要测量到很高的精度。
特别是凸面的曲率半径的测量和面型的精确检测不容易。
反射补偿检测,虽然对大口径小焦比的非球面补偿后任然有一定的剩余球差。补偿球面也比补偿透镜大很多。
但是补偿球面是凹面,很容易精确检测。曲率半径要求也不严格,只要有严格的面型就行。
所以业余爱好者最好做反射补偿检测。这样容易一些。
Dear Mr. So,

Sorry again I have to answer in English because I am being an idiot in typing Chinese by computer keyboard.

You are absolutely correct. the parameters of the lens are not readily available and thus Offner test can be ideal only in theory. However field test indicated that even our primitive Offner tester could produce usable big ( 24 inches ) and fast (f/3.5) paraboloid mirror.

If one can spend money to purchase first class off-the -shelf lens from USA ( e.g. from Edmund ) he is able to make a precise enough Offner tester to do the job. But the cost of constructing such precison tester is very high, not quite the cup of tea for average ATMers.

The development of Computer Aided Foucault Test (CAFT ) has greatly enhanced the accuracy of this traditional method, especially in the area of the big and fast paraboloid mirror. Recently we re-tested our 24 inmch mirror using the CAFT and just discovered that central region ( about 10 inches diameter ) has a slight bulge which was not indicated by the former Offner test. I figured out this bulge away using the CAFT and star tested the mirror again - it seemed that the star images are a little bit sharper than before.

At least this indicated that the CAFT is usable. All we need is a computer and a very precise simple ( yes, very precise, but simple, they don't contradict ) Foucault tester.

Best regards
Chan Yuk Lun

頭像
funoooo
夸克星
文章: 7831
註冊時間: 週五 30 6月, 2006 16:34
聯繫:

文章 funoooo » 週二 29 6月, 2010 09:09

真係好 prof. 期待鏡神影相鏡出世。

苏鲁锭
主序星
文章: 280
註冊時間: 週六 19 8月, 2006 21:34

文章 苏鲁锭 » 週四 01 7月, 2010 09:39

用软件翻译了一下。看了一个大概意思。
谢谢!!
附加檔案
7.1.1.jpg
7.1.1.jpg (149.88 KiB) 已瀏覽 7391 次

苏鲁锭
主序星
文章: 280
註冊時間: 週六 19 8月, 2006 21:34

文章 苏鲁锭 » 週四 01 7月, 2010 09:54

我初步计算了一下。用口径450的球面就可以检测600/F3的双曲面。而且剩余误差很小。可以达到成像点1/22(相当于镜面1/40)波长(550nm)。
当然口径450mm的球面很大了。但是相对于口径100mm左右的二片式的折射镜成本不是很高。
如果为了节约成本,也可以用小一点的球面。只是剩余误差会稍大一点。
附加檔案
7.1.3.jpg
7.1.3.jpg (69.31 KiB) 已瀏覽 7386 次
7.1.2.jpg
7.1.2.jpg (26.2 KiB) 已瀏覽 7386 次

苏鲁锭
主序星
文章: 280
註冊時間: 週六 19 8月, 2006 21:34

文章 苏鲁锭 » 週四 01 7月, 2010 10:09

在这个检测中我们只有测准二个镜子之间的距离就行。
一般可以测量到1mm的精度,就是误差3mm。最后引入的误差也只有1/20波长左右。
足够使用的了。
附加檔案
7.1.4.jpg
7.1.4.jpg (62.27 KiB) 已瀏覽 7381 次

bbsdma
原恆星
文章: 60
註冊時間: 週一 19 1月, 2009 19:56

文章 bbsdma » 週四 01 7月, 2010 13:50

軟件翻译对CAFT部分介绍不是很清楚,手工翻译如下:

陳先生致王先生:
在此抱歉我隻能以英文回復,因我不善于通過鍵盤輸入中文。

你是正確的。透鏡的參數不能精確獲得,于是Offner測試未能達致理論計算的理想程度。但是,据我們的實踐,我們簡陋的Offner測試儀仍然能(協助我們)做出可用的24英吋f/3.5拋物面主鏡。

如果能直接購買現成的頭等質量的美制(如從Edmund公司)透鏡,那么就能製作出足夠精確的Offner測試儀。唯製作這樣精確的測試儀成本高昂,非普通業餘製鏡者所能及。

---下面陳先生介紹了一種新的測試手段CAFT,譯註
計算機輔助佛科測試法(CAFT)的進展,使得歷史悠久的佛科測試法的精度極大提高,尤其是對大口徑快速拋物面鏡。 最近我們用CAFT法重新檢測了我們的24英吋主鏡, 髮現了中央10吋部分有輕微的鼓起--而這各現象在之前的Offner測試中並沒有被髮現。我通過CAFT法修掉了這部分鼓起并重新進行星點檢測,似乎星點的銳度有些微的提昇。

這說明了CAFT法的有傚性。 我盟隻需要一臺電腦和一臺非常之精密而簡單的佛科儀。

此致。
最後由 bbsdma 於 週四 01 7月, 2010 13:56 編輯,總共編輯了 1 次。

bbsdma
原恆星
文章: 60
註冊時間: 週一 19 1月, 2009 19:56

文章 bbsdma » 週四 01 7月, 2010 13:55

另請教陳Sir,
搜索了一下,網上對CAFT的介紹十分有限。陳Sir能否縯示其原理和實現,以及它所能到達的精確度的分析?

bbsdma
原恆星
文章: 60
註冊時間: 週一 19 1月, 2009 19:56

文章 bbsdma » 週四 01 7月, 2010 14:00

請問王師傅: 球面反射補償方法,兩反射鏡如何實現對中?
主鏡中心有孔,如何準確測量其頂點處曲率半徑?是不是直接測量主鏡邊緣到曲率中心的距離代替?

苏鲁锭
主序星
文章: 280
註冊時間: 週六 19 8月, 2006 21:34

文章 苏鲁锭 » 週四 01 7月, 2010 22:12

bbsdma 寫:請問王師傅: 球面反射補償方法,兩反射鏡如何實現對中?
主鏡中心有孔,如何準確測量其頂點處曲率半徑?是不是直接測量主鏡邊緣到曲率中心的距離代替?
非球面的曲率半径可以测量的很准确。
中心有孔的非球面。可以用细丝代替刀口测量。测量接近于孔边缘的曲率半径。
二个反射镜的位置,稍难一点。
主要有二种误差:一个是上下错位,一个是倾斜。
上下错位没有影响,球面口径大一点,这种误差可以不考虑。
倾斜误差,可以从刀口阴影图的不对称中看出来。所以能够调整到正确位置。
附加檔案
无彗差.jpg
无彗差.jpg (23.82 KiB) 已瀏覽 7299 次
倾斜.JPG
倾斜.JPG (21.2 KiB) 已瀏覽 7299 次
错位.JPG
错位.JPG (20.12 KiB) 已瀏覽 7299 次
彗差.jpg
彗差.jpg (36.7 KiB) 已瀏覽 7298 次

回覆文章

回到「自製天文望遠鏡 Amateur Telescope Making 」

誰在線上

正在瀏覽這個版面的使用者:沒有註冊會員 和 8 位訪客