How do you
turn a chunk of glass into a telescope mirror, with a surface
accurate to a couple millionths of a millimeter ?
And why do you wish to do so?
Printer friendly version of this tutorial (*.pdf)
|
There are many reasons
for grinding your own mirror. Either you want a unusual
size, or focal length, or to satisfy your curiosity, or
to have the "pride of achievement" by fabricating
something as precise as a telescope mirror. You could
also live in a part of the world where telescope optics
are not commonly available, and this might be the only
way to have a telescope. Does it save money? Until you have purchased the blank, abrasives, polish,
and pitch, your expenses are very near to a ready made
mirror. The finished mirror needs to be aluminized,
there are also shipping costs for this, and of course,
your labor.
|
Lets learn and practice first
| A good
starter size is the 150 mm (6") mirror. A 200mm (8") will
also work. Most ATM books and websites recommend a Pyrex® mirror blank, but in my opinion, annealed plate glass is better for this first, "learn and practice" mirror. It is cheaper, it is softer and grinds faster, needing less abrasives and with these small sizes, the low expansion glass like Pyrex has no practical (visible at the eyepiece) advantage. |
 |
You might purchase a mirror grinding
kit, and have everything you need included. If such a
kit is unavailable, you need to do some scrounging. You
can order a 19 mm thick disc at your local glass shop.
Such glass is usually used for tabletops and window
displays. A ship porthole will also do very nicely. You can use sandblasting abrasives, make your own polishing pitch. Or even silver the finished mirror if aluminizing is unavailable or too expensive. |
|
typical grinding kit |
|
Now you need to decide
about the focal length of your future mirror. If you are making this mirror to practice your skills for a future, large mirror you want to make, a fast, F4 - F5 mirror is probably your best choice. The future, large mirror will also be something like F4.5 so this is a good way to practice parabolising.
|
First phase : rough grinding (or hogging out)
| In this stage of grinding we want
to dig out the flat glass blank, in other words, create a
saggita. Saggita is a section of a sphere, its radius of curvature (ROC) is TWICE as long as the focal length of the mirror. Example : a mirror with a 2000 mm ROC will have 1000 mm FL. |
"Depth table"
|
|
150 mm
mirror |
200 mm
mirror |
|
FL= 1200
mm |
1.17 mm |
2.07 mm |
|
FL= 1000
mm |
1.4 mm |
2.5 mm |
|
FL= 910
mm |
1.55 mm |
2.75 mm |
|
FL= 750
mm |
1.87 mm |
3.33 mm |
|
As you can see from the
table, the "faster" the telescope (shorter focal ratio)
a deeper curve is needed. This of course means that more glass need to be removed, which takes more time and effort. And, parabolising such a fast telescope becomes more difficult. The 200 mm , 750 mm focal lenght mirror shown in this table is a bit of a extreme for visual use,... but it might be very good for astrophotography. |
| To start
hogging out, we need the glass disc, rough grit (#80
typically) and a tool. But, before any grinding , the glass disc must be beveled. |
 |
The bevel is needed to prevent chipping of the disc edge while grinding. Use a tool sharpening stone, wet it with water, and grind at 45 degrees, around the disc, (red arrow) . A 2 mm bevel is enough |
Making a tile tool
| "In the old
days" it was usual that a mirror kit is delivered with 2
glass blanks. One as the mirror, and the other as the tool. If you have such a kit, save the second glass, and make a tile tool. Tile tools grind faster, and you can make 2 mirrors instead of one. Or you can share the costs of a mirror kit with a friend and make a mirror each. |
 |
A tile tool is a disc poured from cement, with a layer of hard, porcelain tiles epoxied on it. Make it the same size as the mirror in diameter, 50 mm thickness is fine. Also, it is important to impregnate it. "Painting" the whole tool with epoxy or boat varnish will be just fine. On all the ATM webpage's and books you will see neat, equally sized, equally spaced tiles. The key to successful mirror grinding is randomness in stroke numbers and tool and disc rotation. Adding more randomness with a tile tool made from tiles that have been smashed with a hammer cant do any harm :) |
|
It is important that the mirror is suspended. A piece of old carpet, a old towel, or a stack of old, wet newspapers will be just fine. This will prevent astigmatism due to uneven support of the glass disc...
|
|
It is the same thing as the chordal stroke, only now you don't grind with the mirror center over the tool edge, go over the center. Again, rotate randomly, both tool and mirror. As soon as the curve has reached the edge, and your saggita is equal or slightly deeper (depending on method you use later) you can go on with the fine grinding. |
There are 2 ways to
proceed. You can slowly reduce your overhang, and
convert to the 1/3 center over center stroke, then
alternate the "tool on top" (TOT) and "mirror on top"
(MOT) to maintain the saggita depth. When MOT , the
curve becomes deeper, with TOT it becomes
shallower.
|
|
But first,
there is a major cleaning operation ahead :) Close your
#80 grit container and put it away. In short, you need
to clean and dispose of every LAST grain of #80 grit
from your mirror, your tool, your rinsing bucket, your
mirror support, and yourself.
|
|
Fine grinding
|
After we have hogged out
the saggita, we use smaller and smaller abrasives to
remove all the remaining pits from rough grinding and
prepare the glass disc for polishing and figuring. A
typical grit sequence can look like this : #120, #220, #400, #800, #1000, and one finer grade, if available. Grind with 1/3 center over center strokes. Measure the saggita, but don't be worried if the mirror is not exactly a F8 or a F6. To perform well, the precision of the optical surface is important, and the mirror being a F6.2 or 7,9 is completely insignificant. You will make your own custom fit tube for this telescope anyway :) As we have thoroughly cleaned all the tools and working space, we can start with 120 grit. After a wet or two, this is a good time to test the sphere on the mirror. The test is very simple |
|
 |
Rinse and
wipe the mirror dry. Draw a grid on the mirror surface
with a permanent marker (<--picture) Add the usual half teaspoon of grit, sprinkle with water and grind for a couple of minutes, rotating tool and mirror often. Rinse the mirror and inspect the grid lines. If the grid is worn out evenly, you are doing fine. If you grinded with TOT, the edges will be more worn than the center, and vice versa if MOT. This is okay. Any other irregularities indicate that there is something wrong either with your strokes, your tool or mirror support. Also, check your bevel! Fix if necessary. Take a magnifier, or a 25 mm eyepiece (use reversed) to inspect the surface. When all the pits look uniform, and there are no remaining pits from previous grit sizes, you can switch to finer grit. Check the edge, and the center, these are the places most likely to "hide" pits. Take your time, and remove all the remaining pits. |
 |
Mirror surface after #220 grit |
|
There is not really much to say about fine grinding. You will notice that the grits last longer and longer, as smaller in size. Grinding noise becomes more and more silent. Perform the grid test once per grit size. Clean thoroughly after going to finer grit. Past#400, the focal length will not change (significantly) anymore, and the glass becomes very smooth, even shiny if observed at a low angle. It starts to look like a telescope mirror :) |
|
The #400 grit is the finest size you will add with the teaspoon. Finer than that, they need to be premixed with water. Take a plastic bottle (from drinks etc) 0.5 liters, add a few spoons of grit in the bottle and add 6-8 times the volume of water. Shake well before pouring the grit on the mirror. It settles back in just a few minutes. Also, you can add a drop of glycerin into the mixture, this will prevent tool and mirror getting stuck, something that happens quite often at very fine grit sizes.... |
 |
© 1999-2005 Berislav Bracun