Page 3

The none mystery of making a telescope mirror.


( Det er en artikkel på norsk som beskriver hvordan man kan slipe og polere et 150 mm speil i heftet OPTIKK FOR AMATØRER som man kan kjøpe fra Norsk Astronomisk Selskap. Heftet inneholder også nyttige tips bl.a. hvordan man lager holdere for speil, sekundærspeil, okularer og en alternativ måte å teste optikk på. )

What you need for making a telescope mirror are two round Pyrex glass disks which you can buy from an optical company. ( 200 - 250 mm in diameter with a thickness of at least 30 mm is ideal to start with. Also because you get a telescope with big enough mirror to see the wonders of the sky very well. )
( Check out Willmann Bell,Inc. which advertise in Sky & Telescope. ) You need two disks because one should serve as a tool for making the mirror. ( From now on I refere to the pyrex disk which shall be the future mirror as the mirror and the other pyrex disk as the tool. The tool can be common glas. )
An f/8 mirror is recommended because it is not difficult to make. You also need grinding powder ( Carborundum ) with different grain sizes : C-60/80 ( to start with ), C-120 to continue with, C-180, C-220,C-400 and at last the fine grained C-600. You will need approximately 500 cubic cm of each grade. You need a round solid table [ A barrel. ( A corner of a solid table works well too. )] which I refer to as the barrel. You mount the tool on the barrel with 3 wood blocks. Pour approximately 1 - 2 cubic cm C-60 on the tool and mix it with a little water.[ I made a 230 mm ( 9 inch ) mirror without much difficulties and I will try to share the experience with you.]
Put the mirror on the tool and start grinding by moving the mirror in circles and you slowly rotate the mirror too. The mirror should cower 2/3 of the tool nearest to you. You must also slowly walk around the barrel. The curved arrow on Fig.1 show the correct motion of the mirrors center.




Now you can learn some Norwegian.
( Maybe I should learn more English too. )


TØNNE TIL RUNDT BORD ( OVENFRA ) = Barrel for round table ( From above ).

SPEILET = The Mirror.

VERKTØYET = The Tool.

TREKLOSSE TIL Å HOLDE VERKTØYET PÅ PLASS. = Wood block to keep the Tool in place.


The curved arrow show the motion of the mirror.



When the terrible noise end, it is time to wash the tool and mirror and pour new C-60 and water on.
After 1 - 2 hours you can see the mirror curve down in the center and now it is time to change the way you grind. You now continue grinding the mirror by moving it straight back and forth over the whole tool.( See Fig.2 ) Of course you rotate the mirror slowly in your hand and continue walking around the barrel. Each whole movement should be 1/3 the mirrors diameter ( 3 inch ).


The straight arrow show the new motion
of the mirror ( SPEILET ).



After two to five hours ( Depends on your strength.) you must consider how deep the mirror is. If you think of the mirrors surface as a part of the inside of a ( ball ) sphere with radius R, then the mirrors focus is R/2 and for a 230 mm f/8 that is 1840 mm. We set our goal to make f=1800 mm. ( I missed and ended up with f/7, it is very easy to make the mirror to deep. I kept it f/7 instead. ) The radius ( R ) of the mirror is then 2 * 1800mm = 3600 mm. I now call the radius of the mirror r ( 115 mm ), the depth of the middle of the mirror is r squared divided by 2 * R
[ r * r/ ( 2R ) ] which is 115 * 115 /( 2 * 3600 ) = 1.84 mm. If you now make a small metal piece with a thickness of 1.8 mm, you can use this to see your progress in grinding. You put the metal piece on the middle of the mirror and put a ruler over the mirror and the metal piece. When you have grinded to approximately 1.2 mm you wash the table and the mirror and you start with C-120. There must not bee a single C-60 grain left, that is very important. You must of course not mix the bottles of carborundum. Keep them separate. One grain of C-60 mixed with C-120 or higher will make scratches which are difficult to remove later. When you are at approximately 1.4 mm you can continue with C-180. Sometimes you will experience that the tool and the mirrors are stuck together, then put both in a pot of water and warm it slowly until boiling.

NB ! ! This is important. DO NOT COOL the Pyrex plates or they will BREAK. Do something else and wait until the water and the plates are cold enough to touch. ( It is O.K. to warm the Pyrex, but chilling causes them to brake.)

Now the plates will separate. Wait until the plates have room temperature before continuing grinding.

At approximately 1.5 mm you can continue with C-220.[ At least 1 hour work. It depends on your strength.) ]
At approximately 1.6 mm you can continue with C-400.( At least 1 hour work.)
At approximately 1.7 mm you can continue with C-600.( At least 1 hour work.)

If you should go to deep, you just interchange the tool and the mirror. You now hold the tool in your hand and grind in the same manner.
The important thing is that you grind enough with each grade to remove traces from the previous grade of carborundum.
With the C-600 you aim at the 1.8 mm goal the best you can. To fine tune the f value of your mirror you put water on the mirror and test the f value in a dark room by putting the mirror edge on and using a light source and a sheet of white paper. You move the light source and the paper back and forth at the same distance from the mirror. When you get a sharp image of the light source you are at the 2F point ( Fig.3 ) and the focus of your mirror is F. Now you grind with the C-600 ( Interchange the tool and the mirror if necessary until you get ( F ) f = approximately 1840 mm.


The optical system.


DYBDEN = The Depth.

SENTRUM I KULA = The center of the ball ( sphere.)

DEN OPTISKE AKSE = The optical axis.



( The polishing process will hopefully get you to 1800 mm. Here it is "not possible" to interchange the tool and the mirror. ) The grinding process can take place outdoors, but the quiet polishing process should be done indoors in a warm room ( approximately 25 degrees C. )
I polished the mirror indoors on the corner of a solid table.
Polishing the mirror.

 Now you must check your mirrors surface to see that you are really finished with the last degree of carborundum, and that there are no traces of previous degrees. You should also try to avoid to much scratches. ( Some scratches don't degrade the optical performance noticeable. )
Now you need the pitch and rough ( polishing powder ) and if you bought a complete grinding set you already have it. You now make a retaining wall on the limb of the tool by using solid tape. Warm the tool in a pot of water ( 50 degrees C ) and melt the pitch in a box or pot and pour it over the convex surface of the tool.
It is on this pitch pizza ( pitch lap ) you are going to polish the mirror in the same manner you grind the mirror.( You now use the polishing powder and water.)
While the pitch still is warm tear of the tape and pour lots of soap and some water on the mirror and the pitch lap. Put the mirror on the warm pitch lap and move the mirror around until the pitch lap has the same curvature as the mirror. If the pitch lap don't curve in the center you can correct this by putting the tool and the mirror in warm water or the sunshine and move mirror around over the pitch lap. It is important that you use soap to prevent the mirror to stuck in the pitch. ( Also cut away the pitch on the outside of the edge of the tool. )
Now there must be parallel channels in the pitch lap. Put some solid tape on the pitch lap ( parallel 5 mm apart. ) and use a solder which is shown in Fig.A with a V shaped iron piece on the warm end. ( If the pitch is not too cold you can use a knife. )





Make a new set of channels 90 degrees on the former channels. Take care, the center square should be offset from the mirrors center as shown in Fig.B. The polishing process may fail if the center square is not offset.



kanaler = Channels
sentrum = Center ( The black dot. )



Air blobs in the pitch will disappear during the polishing process.
When you have succeeded in making the pitch lap you mount the tool on the barrel ( Clean everything one more time. ). You pour some water and polishing powder on the pitch lap and start polishing in the same manner you grinded ( Back and forth, rotating the mirror slowly while you walk slowly around the barrel.) If you are using the corner of a table ( which I prefer.) you walk as far around as possible so that the mirror has rotated 180 degrees. Than you move your position 180 degrees and continue polishing. The point is that it is the mirror which slowly rotate over the pitch lap, not you. ) Each polishing stroke should be approximately 1/3 the diameter of the mirror. The room temperature should be approximately 25 degrees C which are comfortable and makes the pitch dynamical ( Not too hard ).( That is only 12 degrees from your body temperature which are useful. ) What is not so comfortable is that you must use solid leather gloves when you hold the mirror. That is because the mirror expand too much where you touch it and the mirror is unevenly polished. ( The standard method is to bee in a colder room, to mount a round wood plate on the backside of the mirror and hold your hands on this one. I don't recommend it because the mirror is not so easily handled. Some people will disagree in this so you may try both methods if you want. )
When you start polishing you can feel that the mirror is partly stuck some places, but continue with the polishing. The pitch lap will rapidly adapt to the curvature of the mirror. If it is complete stuck, put both in warm ( 35 - 40 degrees C ) water for a while and try to get them apart.
After one hour polish you can see that the mirror reflect light, but you must polish for another 8 hours.
After approximately 8 hours you should consider how the curvature ( the figure ) of the mirror should be.
The goal is a parabola ( Fig C ).


The sphere to the left can not make a sharp image of a star. The parabola to the right can.



Now you need the Foucault test which you can make without difficulties. It is a simple device but extremely accurate when used properly. ( Fig.4 )


The Foucalt test setup
It works like this :

You put the mirror edge on, on a table ( You can use two tables. ) in a dark room. At a distance R ( 3600 mm ) you have a point ( 0.1 mm ) light source a couple of centimeters to the left of the optical axis.
On the other side of the optical axis you have a mounting for a razor blade very close to the image of the point light source. ( You carefully adjust the edge on mirror. )

The mounting should be such that you can slide the razor blade a couple of centimeters toward the mirror and back.

If you look at the mirror from a point R + 3 cm ( 3630 mm ) from the mirror just outside focus of the light source, and you do this right, you should see the whole mirror enlighten.
Now move the razor blade carefully toward the lamp ( You are "cutting" the light rays. ) and you should see the shadow of the razor blade projected on the mirror. The shadow looks different when you have the razor blade at different distances from the mirror.

The point is that you should be able to move the razor blade back and forth around focus ( R ) and at every position you should be able to carefully move the razor blade toward the optical axis and cut the light rays while you are looking at the enlighten mirror. If you get this right you have a Foucault test.
If you move the razor blade approximately 20 mm inside R ( That is at a distance 3580 mm from the mirror. ) and cut the light rays from right to left, you should see the shadow coming from right to left on the mirror.
If you move the razor blade approximately 20 mm outside R ( That is at a distance 3620 mm from the mirror. ) and again cut the light rays from right to left, you now should see the shadow coming from left to right on the mirror.

Now it is going to be very interesting. Try to put the razor blade at distance R and cut the light rays. Now you will see some strange shadows. Try to leave the razor blade at the point where you see the strange shadows. ( Not when the mirror is black or not when the mirror is fully enlighten. ) If the figure ( the curve of the mirror ) was part of a sphere ( Fig. C above ) you should not see any shadows. The mirror should be half dark all over so these shadows tells you something about the figure of the mirror. ( You can of course build a much better Foucault test than the one described here where you can use fine screws to move the razor blade back and forth, and toward the optical axis. )

The interpretation of the shadows.


When you observe the shadows, think of the sphere ( sfære ) as a completely flat surface. If you see this flat surface from the side it should be a straight line. Any other figure should deviate from this line. Fig.5 show a parabola ( parabel ) relative to this line with the corresponding shadows just beneath. From left to right you can see that the parabola relative to the sphere first rise to the 70% zone then descent to the 0 % zone ( center of the mirror ), then rise to the 70 % zone and drop of to the 100 % zone ( edge ). ( This is how the finished mirror should look like. ) You also see that where the figure rise the mirror is enlighten where it drops the mirror is dark.




Fig.7 show how you can think the shadows come to existence with an Imagined Light source to the left which enlight the up hills and the down hills are in the shadow.

Fig.8 show a figure which is part of an elipsoide. It is quit opposite to the parabola. If you should see this figure than polish approximately 5 minutes with long strokes ( 2/3 the diameter of the mirror. ), than the figure moves toward the sphere ( 0 % corrected ) and if you continue for another 2 - 3 minutes it moves toward a parabola ( 100 % corrected ). If you continue with long strokes it will end up as a hyperbola ( over corrected ).

Try to get it spherical first.

If you are so lucky to get a spherical mirror take a break and price the lord ( or whoever you prefer.) From the sphere you should try to get to the parabola ( working "downhill" ) by polishing with long strokes ( 2/3 ) for 2 minutes. [ The goal is to get it between 60% and 95% corrected ( under corrected by 1/10 wavelength ), because when you use the telescope in the evening the mirror gets colder and colder which in turn make the mirror very close to 100 % corrected ].

After polishing for 2 minutes test the mirror again and read the rest of the article.

After reading the rest of the article and testing the mirror you are going to finish the production by polishing by overhang ( Fig.13 ). 1/6 of the mirrors diameter should be off the tool all the time.
When you are going for the parable, then you can avoid the problematic turned edge by polishing off center for 2 minutes. ( Never polish more than 2 minutes without testing and you polish as before with 1/3 strokes ).
After now 4 minutes polishing you test the mirror again and you should see the "parabolic" shadows. If the temperature in the testing room is different from the working room, leave the mirror in the testing setup for 10 - 20 minutes.

Now take a look at the parabolic mirror in Fig.9 and notice that the parabola under the Foucault test has a shorter focus in the center than the edge because the parabola have the strongest curvature in the center. The difference should be ( r * r / R ) and for this mirror that is 3.7 mm . 60% to 95% correction make the range from 2.2 mm to 3.5 mm which is your goal. ( Aim at 3 mm )
Now move the razor blade closer to the mirror ( approximately 1 mm ) and cut the rays. If the center zone ( 0% to 70 % ) looks spherical ( The mirror darken evenly in the center. Don't mind the outer zone. ) than mark the position of the razor blade ( Point 1 in figure left ). Now move the razor blade from the mirror and cut the rays. When the outer zone ( 70% to 100% ) darken evenly you mark the position of the razor blade ( Point 2 ).
If you should get the distance between the marked points close to 3 mm don't ever polish again because further polishing can introduce a turned down edge and other undesired figures.

Put the razor blade again at the position R and look at the shadows. They should look soft and similar to Fig.6.

You don't need to read further unless you have a different situation than the one described above.

Often the mirror has a depression in the center ( Fig.10 ) or a turned down edge or both. To remove a depression in the center, use short stroke ( 1/6 ). The same is for a turned down edge ( Fig.11 ) ( You may now produce a hill in the center ).
You can remove a hill in the center by polishing with overhang ( Fig.13 ). That is you polish off center of the tool now with 1/3 of the mirror outside the tool.
Sometimes it is difficult to detect a turned edge, than you can move the razor blade 4 - 8 mm closer to the mirror. Move the razor blade to the left and stop before the shadow enter the center. You should get a shadow which look like one of the shadows in Fig.12.
a. Show a perfect mirror.
b. Show a mirror with a turned down edge.
c. Show a mirror with a turned up edge.

( The oblate is absolute a wrong figure for a mirror. )



Home.