This is the first of a 2 part Instructable explaining how I built a wooden telescope from scratch. In this part, I will show you how you can build the key element of a telescope: The primary mirror.

A good mirror will show you details on the moon, planets and deep sky objects whereas a poor quality mirror will only give you blurry views of the skies.

A telescope mirror requires an extreme surface precision. However, you will see that this quality is most of the time best achieved by hand rather than machine polishing. This is part of the reason why some people prefer building their own mirror instead of buying a cheap instrument. The other reason being the knowledge you acquire by slowly creating a high quality optical instrument.

Here's a list of website I used during the construction of the telescope:


Mel Bartels' Amateur Telescope Making

More detailed article about my mirror: http://www.thomasjacquin.com/build-your-own-telesc...

Step 1: Materials

Here's a list of items that you will need to build your mirror:
• A glass blank made from a material with a low expansion coefficient (Pyrex, Borosilicate, Duran 50, Zerodur, etc)
• Silicon carbide in different grit size (example: 60, 80, 120, 220, 320)
• Aluminium Oxide (25, 15, 9 and 5 microns)
• Cerium Oxide
• Pitch
• Sharpening stone
• Waterproof plaster (dental plaster)
• Ceramic tiles
• Epoxy glue

I bought a kit from First Hand Discovery than contained all these items for $155 (8" mirror kit).

Step 2: Prepare the Glass Blank

Glass blanks often come with marks on the surface. The circular mark was left by the bottom of the kiln while the top marks are coming from the temperature difference when the glass was cooled down.

We'll start by bevelling the edge of the glass to limit the risk of chips. A sharpening stone is a good tool. Remember to keep both mirror and stone wet since glass dust is very bad for your lungs.

The bottom of the mirror needs to be as flat as possible before we start working on it. To even out the surface, we are using coarse carborundum (Silicon Carbide #60). Spread the powder (and water) on a flat piece of glass and rub the glass blank onto it. After a few seconds, you will see a gray paste. Wash it of and add more wet grit. Continue until the surface gets free from holes and bumps.

More info here:



Step 3: Build a Plaster Tool

This tool will be used to create a concave surface on the glass blank.

Cover the glass with a plastic film. Make a cardboard cylinder around the glass and pour the plaster inside. Let it dry and remove the cardboard. Carefully separate the glass and file off the burs on the edge.

More info: http://stellafane.org/tm/atm/grind/plaster-disk.ht...

Step 4: Cover the Tool With Ceramic Tiles

We need a strong surface to grind glass. That's why we need to cover the tool with ceramic tiles.

Glue the tiles on the plaster tool using epoxy. I used a quick gel epoxy but I would not recommend it since it was slowly eroding over time.

Note that you should avoid placing a tile or a hole in the center of the tool. Instead, slightly offset a tile to avoid any central defect in your mirror surface.

More info: http://stellafane.org/tm/atm/grind/tile-tool.html

Step 5: Start Grinding

You can now start the real work. Put some wet coarse grit on the tool and start rubbing the glass on it.

The stroke here is the chordal stroke.

After a few strokes, turn the mirror in your hands and walk around your polishing station the other way. This ensures a good randomization of the courses and prevents cyclic errors.

More info: http://stellafane.org/tm/atm/grind/rough.html

Step 6: Measure the Sagitta

Continue grinding until you get the desired curvature. To estimate the curvature, we will use a dial indicator to measure the sagitta.

If you want to build a telescope for planets and moon observation, you will need a longer focal ratio (F/8 or higher). This means a shallow mirror.

On the other hand, if you want to view galaxies and nebulae, you will need a fast aperture (F/4 for example). This means a deeper mirror.

My mirror is F/4.75. The sagitta on my 8" mirror is 0.1".

Sagitta calculator: http://stellafane.org/tm/atm/general/scope-calc.ht...

Step 7: Smooth Out the Surface Using Finer Grit

Once the sagitta is reached, we have to smooth out the surface while keeping the same curvature.

By switching mirror on top / tool on top, we keep the shape of the mirror. At this stage, we can use finer grit. The courses are now reduced to 1/3 center over center. Mark the larger holes with a Sharpie and continue grinding until those are gone. That's a sign that means you can switch for a finer grit.

Go down to Silicon Carbide #320. Once you've reached this step, you should start to see some reflection when viewed from the side.

More info: http://stellafane.org/tm/atm/grind/fine.html

Step 8: Make a Pitch Lap

Now is the time to get a shiny surface. We need another tool with a smoother surface. You can make one with plaster or thick plywood. We will cover it with a soft material: pitch.

Pitch comes from conifers. It is sticky and hard to clean. I recommend that you use it in a safe place and wear old clothes.

Make another cylinder around the new tool. Melt a good amount of pitch and pour it on the tool. Let it cool down and remove the cardboard siding before it is too hard. Using your hands, you can start shaping the surface to make it a bit convex. Making channels will also help you when pressing the glass on it.

More info: http://stellafane.org/tm/atm/polish/pour-lap.html

Step 9: Polishing

Put some wet cerium powder on the pitch lap and press the mirror against it. This will make the cerium penetrate the surface of the pitch. Use some soap the enhance lubrification if you need to.

The courses are still 1/3 diameter center over center. When the aspect of the surface has changed on the whole surface of the mirror you can switch for a finer powder.

Remember to keep groves on the pitch for a good abrasive circulation.

More info: http://stellafane.org/tm/atm/polish/polish.html

Step 10: Build a Foucault Tester

A Foucault Tester is an instrument designed to analyze the surface of parabolic mirrors. It has a light source that shines on the mirror. When the light comes back, it focuses in a different area depending if it came from the edge or the center of the mirror.

The tester uses this principle to let you visually see errors in the range of 1 millionth of an inch. Adding a Ronchi Screen to the tester will save some time as you will get an idea of the surface without making any measurements.

To make your life easier, make a mirror stand. A screw in the back lets you adjust the tilt.

More info: http://stellafane.org/tm/atm/test/tester-main.html

Mirror Stand: http://stellafane.org/tm/atm/test/tester-stand.htm...

Step 11: Make a Paraboloid

The polishing stage left us with a fully polished mirror with a nice spherical surface. However, a sphere won't work for astronomical purposes. We need a paraboloid.

The difference between a sphere and a paraboloid is small (order of 1 micron). To achieve this paraboloid, we will use the Foucault tester. Since we know what the reflection should look like, we will make special courses with cerium oxide until the reflection on our mirror matches the theoretical one.

The general course is a "W" with several push/pull moves. The amplitude should be 4/5 diameter laterally and longitudinally.

There is also a full list of different courses to correct specific surface errors.

More info: http://www.bbastrodesigns.com/JoyOfMirrorMaking/Pa...

Step 12: Control the Surface With the Foucault

This is what the reflection looks like in a foucault tester equipped with a Ronchi grid.

Depending if the grid cuts the light before the radius of curvature or after, you can interpret the lines and deduce the shape of the surface.

More info: http://stellafane.org/tm/atm/test/ronchigrams.html

Step 13: Finished Mirror

Once the surface reaches certain criteria, you can consider your mirror finished.

Lord Rayleigh proved that a mirror having a wavefront error less than 1/4 of green light wavelength (520 nm, also referred to as Lambda) is indistinguishable from a perfect mirror. As a mirror amplifies errors by a factor 2, you will need a mirror smoothness of 1/8 of green light wavelength minimum which means a peak to valley distance inferior to 65nm.

On the surface error analysis above, you will see that this distance is 28 nm, which equals to 1/20 Lambda on the glass and 1/10 Lambda on the wavefront.

The Strehl ratio is also an important rating since it will be used to determine the overall quality of the final instrument.

The Couder mask is used to take measurements with the foucault tester. The column "Average knife reading" on the analysis software corresponds to the reading on the dial indicator of the foucault. The "Zones" are the cut-out parts on the mask with 1 being the center and 7 the outermost one.

More info to build your own mask: http://stellafane.org/tm/atm/test/mask.html

Step 14: Aluminization

Now that the mirror is finished, we need to send it for aluminization. Currently our mirror only reflects 4% of the light. An aluminium deposit upon its surface will increase this percentage to more than 90%.

An optional additional SiO2 coating can protect the metal from any source of oxidization.

Most companies will add a true center mark upon request. This helps during collimation and does not impact the quality of your mirror since the center does not participate in the image you see in the eyepiece.

As I live in Canada, I looked for a canadian company and got a great service with Normand Fullum Telescopes

More info: http://stellafane.org/tm/atm/coatings/atm_coatings...

<p>Truly amazing,Thanks for sharing!!!</p>
<p>You might want to mention somewhere that it is simply the stroke that makes the concave surface. It took me a little while to figure that out. I went back to see if you had curved the grinding tool somehow. Rubbing two flat surfaces together and getting a curve doesn't make a lot of sense. </p>
<p>Just a note, spinning epoxy on a platter produces a parabolic shape.</p>
You're right, any liquid in rotation will produce a parabolic shape. Some people use this property to make liquid mirrors. A container is carved to a circular shape and a small quantity of liquid metal (gallium alloy usually) is poured inside. Once you start spinning the container, the metal spreads onto the surface creating a parabola. However, you need to isolate that system from any kind of vibrations and you can only look at the zenith.<br> <br> More info here: <a href="https://en.wikipedia.org/wiki/Liquid_mirror_telescope" rel="nofollow">Liquid mirror telescope</a>
<p>Awesome work!</p><p>Unreal...</p><p>Thank you!</p>
<p>Though I love astronomy, my big passion is Special Effects. One of the effects I am looking in to requires a large, concave mirror. These can be expensive, so the knowledge I glean from this Instructable will come in very handy. Thanks!!</p>
<p>Maybe this technique using a piece of mylar distorted by vacuum would work for you also if your &quot;concave&quot; isn't too specific:</p><p>https://www.youtube.com/watch?v=_8sd9UgjXLE</p>
If you don't need an extreme precision on the surface mirror, this is indeed a a quick and cheap way to get a large concave mirror. You might have to pump some air out once in a while as I don't think it would hold its vacuum over a long period of time.
<p>I wonder if anybody ever tried to build a home made cnc (+rotary) grinding / polishing machine for parabolic mirrors. Using computerised machines (arduino or rasperry for example) could reduce the errors and production time. just an idea maybe someone can pick it up :)</p>
<p>My late uncle was a genius and built a mirror grinder to make a telescope from scratch in the late 60s when our space program got underway. I was too young at the time and once I got old enough to get interested, he wasn't around long enough to teach his methods to me. He had his mirror coated by Edmund Optics and they were so impressed with his 6&quot; mirror, they asked him where he got it. They could not believe he made it in his garage using a flywheel from a tractor engine as his turntable. It took him several weeks to grind it. He used two blanks; one became the tool for the other. I still have some 12&quot; blanks he started but never finished. Check out these photos of his hand-built telescope. He claimed to have been able to clearly see the Apollo 10 spacecraft orbiting the moon. The saddest thing about this story is that he didn't take the time to fiberglass the tube and it eventually deteriorated from weathering in his driveway.</p>
<p>CNC machining is already the standard way of grinding and polishing telescope mirrors at an industrial level. This method usually produces consistent and good results. Some amateurs actually use grinding machines to cut most of the grinding and polishing time. </p><p>Figuring is actually the most sensitive part. Machines can achieve a good result for longer focal ratio (f/6, f/8, etc) but hand figuring is probably still the best option, especially for faster mirrors.</p>
<p>Well done.</p>
<p>Well done.</p>
Wow - absolutely incredible! This is one of the best instructibles I've seen. It looks like a very enjoyable project. It must be so satisfying to use something you've made with your own hands in such a painstaking way. This is true craftsmanship. If I get the time to devote to it, I'll definitely be doing this. Thanks so much for sharing.
<p>It looks very interesting. How many hours did it take from start to finish to complete the process?</p>
<p>I won't lie, this takes a long time. I did not work on it every day but here's a rough schedule. I started in May and got it polished by the end June. July and August where spend on the figuring part (going from a sphere to a parabole).</p><p>If you work full time on it, the rough carving, smooth grinding and polishing can all be done within a month. The time consuming part comes from the figuring process. I probably hit all the possible surface errors from turned down edge to central hole and oblong spheroid. 5 minutes of the wrong stroke can add 2 weeks of work.</p><p>Between each figuring session, you need to clean, dry the mirror, let it cool down and take measurements with the Foucault. This multiplies the effective working time by 2 or 3.</p>
<p>I've been hand lapping a 4X4X1 hardened steel blank. Because of cooling mine is concave on both sides and the surface is already a very rough parabola - which I'm trying hard to eliminate. I desire ultimately a flat surface true to about .001 in. For machining I don't expect to need greater accuracy as most frequently accuracy into .01 of an inch are required for my tasks. But we're discussing machines, not optics. It is indeed challenging work. And at my chronological advancement, proving very painful. :(</p>
Thomas,<br>Compliments for the detailed instructions with photos and illustrations which will help for beginners.<br>I built an 8 inch telescope about 6 years. The grinding of the mirror was done at home. It took more than a month to finish the mirror grinding and getting it silvered.<br>The mount was also built and enjoyed viewing the craters of the moon and especially Saturn rings.<br>Used the links you have provided.<br>A. S, Bhasker Raj<br>Bangalore<br>India
<p>Thank you Bhasker,</p><p>I'm glad that successfully built your own mirror. It always seems like a huge amount of work at first but the view of Saturn rings through the eyepiece is worth all of it.</p>
Thanks Thomas<br>
<p>I'm blown away by the amount of work you guys are willing to put into such a project. </p><p>Just out of curiosity, is there some way to do this using a turntable or lathe, instead of by hand over months? </p>
<p>Yes, there are some machine that will do it for you. They are quite useful when grinding low aperture mirrors with a large diameter since the volume of glass to remove can be quite large. This kind of machine is not hard to build actually. You can google &quot;Telescope mirror grinding machine&quot; and you will find a few models and even plans.</p><p>This is also a good page explaining how much work is involved when making your own mirror: https://stellafane.org/tm/atm/general/myths.html</p>
Interesting! <br>Thanks for sharing that. <br>I'm truly amazed at how much work you enthusiasts are willing to put into making a telescope mirror like this.
Why wouldn't one put time into making an ultra high quality instrument. If you want something done right you need to do it yourself.
<p>Of course! If... one happens to be an astronomy buff. It's entirely reasonable. </p>
Fair enough.
<p>Yes, but how can I use it to build a death ray?</p>
<p>Place a powerful source of heat/light just further out than the focal point. This heat/light will be focused at a point far away (how far away is determined by how far you hold the source from the mirror).</p><p>My friend had one of those &quot;million candlepower&quot; lights, which he used for diving- you could feel the heat from several metres away if it was pointed at your face. I don't know how much heat you'd need to get a &quot;death ray&quot; this size to actually set something on fire remotely- why not give it a go? :)</p>
I think it's awesome that you actually had an answer for that.
<p>Please keep this thread going! Inherited Sci American's 3 vol &quot;Amateur Telescope Maker&quot; and have been sorta kinda pondering this for years. Eager to see your mount solution</p>
<p>&quot;I bought a kit from First Hand Discovery that contained all these items for $155 (8&quot; mirror kit).&quot;</p><p>While I love making things myself, including the learning experience involved, I find that sort of expense hard to justify when I can get an already silvered 8-inch primary *and* secondary mirrors for under $150 on eBay.</p>
<p>Unfortunately, an &quot;already silvered 8-inch primary &amp; elliptical&quot; at that price will very likely be machine polished and will probably not even reach the Rayleigh criteria of Lambda/4 minimum on the waivefront.</p><p>To take an example, the first 8 inch mirror I found on ebay claims to have a L/16 RMS surface quality. This means only L/8 on the waivefront. Also, this is the Root Mean Square value of all readings, which is always better than the Peak-to-valley reading. On my mirror I end up with L/96 RMS while my PTV value is only L/10. That's almost a 10x factor. (PTV gives the shape, RMS the smoothness)</p><p>If we apply this reasoning (this factor varies from a mirror to another) to the mirror on ebay, we end up with a value somewhere between L/1 and L/2 which is really far from the Rayleigh criteria.</p><p>This mirror combined with a cheap elliptical will probably have a lot of diffraction and will give poor views of the stars in most circumstances.</p>
<p>Consider me educated. Thank you. :)</p>
<p>Making your own telescope mirror by hand grinding/polishing is an interesting activity--I did it as a 12 year old at the Adler Planetarium in Chicago, and it lead me into a 40 year career as an optical engineer, but frankly, these days I really can't justify it on a cost/quality basis. An 8&quot; mirror, aluminized and figured to 1/16 wave rms, can be had from one supplier for $220. The kit mentioned costs $155 and getting it aluminized at the suggested vendor another $136, so the DIY cost just isn't justified, unless you have a specific f/number or focal length requirement. On the other hand, if you want to learn how to be an optician, by all means, go for DIY, except, of course that modern opticians depend extensively upon machine-driven tools, even CNC. Walking around the barrel for endless hours is so 20th century. </p>
<p>Don't see myself undertaking this project, but I found it absolutely fascinating from a craftsmanship, problem solving and engineering point of view. Please update us with the rest of the telescope as I'd love to see more of this.</p>
<p>An Instructable about the rest of the telescope is currently in a draft state. I'll publish it soon.</p>
<p>Thomas,</p><p>Great instructable. Have you ever seen this site? You can do it fully automated.</p><p><a href="http://www.mirror-o-matic.com/" rel="nofollow">http://www.mirror-o-matic.com/</a><br><br>Maybe worth an instructable too. </p><p>I made myself a Dobson Newton Reflector from a PVC pipe. I bought the mirror, But maybe I will grind my self a mirror using your Instructable.</p><p>Cheers.</p>
<p>Thanks for the link. I might eventually build one if I decide to build more scopes (I'm sure I will).</p>
<p>Fascinating 'ible....you must get allot of joy looking at the heavens in something you built.</p>
<p>Great instructable. Good detail, references &amp; photos. Very clearly written. Well done!</p>
<p>FYI, if you happen to be in the San Francisco Bay area, there is a Telescope Makers Workshop at Chabot Space &amp; Science Center (http://www.chabotspace.org/telescope-makers-worksh...</p><p>They meet every Friday evening. Many people have successfully made there own mirrors and telescopes there over the years. There are folks on hand to give guidance, and advice. And its fun.</p>
<p>Thanks I will keep that in mind if I ever fly to San Francisco.</p>
<p>Maybe I am being dense: You made a tool from plaster with two flat sides. You glued flat tiles to one of the flat surfaces. How does this tool yield a concave surface when used to grind the glass? Doesn't the tool+tiles need to be concave?</p><p>Thanks!</p>
<p>Dean is right, the chordal stroke consists of back and forth moves with the mirror on top. The center of the glass rides the edge of the tool. Gravity + pressure from your hands apply a downwards force at that exact area. If you keep doing that while rotating around the tool, you will eventually grind the outer circle of the tool faster making is convex. </p><p>In the meantime, the center of the glass is the place always being eroded by being rubbed on the edge of the tool. It eventually gets concave.</p>
<p>It's mathematical! If you grind a lot in some places (middle) and a <br>little in others (edges) then more gets stripped from the middle, less <br>from the edges. The key is to OFFSET the grinder and the mirror: the center of the grinder does not stay over the center of the mirror, it gets moved! If you rotate the mirror and keep repeating the <br>process... CONCAVE!</p>
<p>OKAY, I've toyed with this for years and I have to say, you have done it perfectly. Where were you in 1999? NOW, here is the next step, and YOU are the one to take the challenge: put the whole thing on a rotating table so you spin your own parabola. Simple enough in principle, I've just never had the time. But if you can figure out the technique, scaling up to 48 inches should be easy. Then it is all polishing with a robo-jig running a polisher moved by old printer steppers and the same rotating platter spinning the mirror while the polisher moves back and forth, all coordinated by raspberry pi. <br>BRILLIANT Instructable. Thank you.</p>
<p>A friend of mine just gave me a six inch blank and tool. Now I HAVE to do this! Thanks for the fine instructable.</p>
That's quite an accomplishment. I'm not sure I have the time to build this right now, but I'm definitely interested.
<p>Thank you.</p>
I have wanted to do this for years now. I have owned the glasses since <br>about 2007, but have always been afraid to start. I liked the way you grinder against tools as opposed to another glass blank. I really look forward to see the rest of this build!!!!

About This Instructable




Bio: Most of the things I build usually relate to either astronomy, physics or woodworking in general.
More by ThomasJ1:Planets Photography Guide Newspaper Stand See Through Star Map 
Add instructable to: