OcularBiker: Telescope Making

New Houghton-Newtonian Project

In my old web-pages I started to make notes of my newest telescope project but unfortunately I never ended the whole project! :( Somehow time flew so fast and other duties and works always came over it. Well, that's normal life I think. You have to decide which one is the most important at the time. Anyway, here is the story from my old web-pages:

STARTED 20-NOV-2005 / Updated 19-JUL-2007

FOREWORD

My present D5.6" (142 mm), F/4.6 Lurie-Houghton Newtonian telescope has performed quite well, so I decided to start a new LHN project if I get all the needed optical glass material together.
The primary will be 25 mm thick, D8" Borofloat blank which I purchased from Stathis Kafalis, Germany. I'm going to make a D7.4" (187 mm), F/5 telescope this time. The odd diameter of the telescope is due to the design of LHN. Always the free optical diameter of the corrector is a little bit smaller than the diameter of the primary. My primary is standard 8" (205 mm), so the corrector lenses (optical glass, for example Schott's BK7) will be little under 8". It's very easy now to try different Lurie-Houghton telescope setups because of Ivan Krastev's java script based program,Lurie-Houghton Telescope Calculator.
I have tried to figure out an ATMaking sequence of Houghton-Newtonian optics so that it would be the easiest one. There has to be possibility to adjust parameters during the making. There is no sense trying to reach some of the values exactly because many times it's impossible. The setup what I have used is so-called "all spherical" setup. Ie. all five surfaces are spherical and also the corrector's lenses form surface pairs as follows: R1(convex)/R3(concave) and R2(convex)/R4(concave). The fact seems to be that the primary is good to make first, so that you can adjust corrector's lenses RoCs to match the reached RoC of the primary.
Without questions it's obvious that ATMakers like the surfaces pairs R1/R3 and R2/R4. Anyway, I guess the different carbo and al-oxide stage works are clear and rather straight forward things if you have made atleast one paraboloidal mirror. In HN's corrector lens cases I like to clearly divide the grinding work to rough and fine grinding phases. That's because of trepaning, edging and de-wedging works. Once again I have to say that these working methods are all well known and not necessarily the best possible ones but have worked for me once. So, if there is even something useful or worth trying hints and tips, good! I'm glad. :)
I try to list and make the work sequences in step by step order to reach the goal without doing any extra stages:

- Making the Primary
- Grinding the lens pairs to 220 carbo stage
- De-wedging both lenses
- Trepaning the lenses
- Fine grinding lenses
- Polishing concave and convex surfaces
- Ronchi testing concaves and the interference test
- Mechanical aspects of LHN telescope

Making the Primary

On the up-left in the photo the grinding of the mirror blank has already started. I already have a fine grinding tool(covered by aluminum plates) from an old project, so by using existing materials I try to minimize the work and material amount. On the photo I'm using a D110 mm steel plate for rough grinding. It's one of the weights I normally use on the overarm to power up the grinding action. I have plenty of those similar weights, so one can be dedicated to rough grinding use.
Manual grinding, putting pressure on the leading edge on the center area of the mirror, the grinding action is more effective. The tool stay on your side and the inside edge is all the time on the mirror's center area. When the concave shape is starting to form I normally let the machine to do the hard work for me. If the deepening of the mirror proceed slowly, I come back to manual grinding.
Another method is to use "edge to edge" style grinding when the tool goes a little bit over the both edges of the mirror. The tool goes over the center of the mirror. In the photo I'm using an old concrete tool which dia is larger than the tile area. In both cases my machine's table rotates 4..5 times faster that overarm.Typical rotation speed for turn table is about 70 rpm. Naturally all the grinding work can be done manually or by using different type of a machine. I just use what I have.
Making the primary is the most familiar thing for us. We have made plenty of them but still it has to be mensioned that also in HN telescope case the primary finally forms the image. The surface should has the perfect and smooth spherical figure. The primary is maybe the easiest one to produce what we ATMers face. Last time I used only different lines/mm ronchi gratings to get the required sphere. Carefully done ronchi test is sufficient. The primary of HN telescope is also good to design so that effective diameter will be a little bit smaller than the real clear diameter. In my case the standard mirror blank's diameter is 205 mm. I will use 2 mm bevel when the clear dia. is 201 mm. The designed setup requires dia. 198 mm.
It's better to ensure that the light path through the corrector lenses fits totally to the surface of the primary. Even 3..5 mm could be designed to leave on the edge if it happens so that you get problems with TDE. In corrector lens cases you have to "hide" at least 3 mm/edge inside the cell to ensure proper supporting of lenses. I would say that 4...5 mm is a good value. Not a bad idea if your lenses in the end have a narrow TDE. The same situation is with the center holes: Better to make clearly smaller holes to lenses than the designed and needed obstruction of the secondary mirror. Very easily you have to leave a few millimeters wide TDE on the edges of the center holes. Fine grinding went the fastest and easiest ever! Maybe I am finally learning from my mistakes. In fine grinding stage with a Mirror-O-Matic style machine it's very important to spend time to find out exactly "the sweet spot". Ie. the best combination of the stroke length and overhang (stationary method with a little side motion) which produces the spherical form from the center to edge smoothly. I have used a pen test always starting from 220 carbo stage.

Also polishing was this time quicker because I had also saved the polishing tool from an old project. Here is a photo of the tool after cleaning the channels. Watery or milky cerium oxide mixture works very well and fast. This time it was suprising to see almost totally polished surface after 3 hours. Ok, I continued 2 more hours to be sure and now I have polished app. 6 hours totally because have to continue from the undercorrected sphere to closer of the target. By adjusting once again the stroke almost the perfect sphere is obtainable. If the sphere is the target I have left the surface a little bit undercorrected and by using 50% tool manually figured the surface. 50% tool combined with the W-stroke (TOT) is very powerful combination for fast mirrors. The tool is rotated only after 20...30 pcs. back and forth W-strokes! The rotation is app. 1/8 of the full turn. The W-stroke is only like a single W-letter. Mirror on the table has to rotate or you have to walk around the table. That way you also can make really smooth strokes without any jerkiness. The edges of 50% tool just touch the edges of the mirror in every point of the W-stroke. The edges of the tool can go slightly over the mirror edges. Important is that the motion of the tool you generate is smooth, light weighted and uninterrupted. Also the cerium mixture should be even thinner than during polishing! Milled rouge or cerium oxide would be excellent for the figuring work. Of coarse, the tool has to be well pressed and furnished with micro-facets. I have tested am. method for a 6", F/6.4 mirror suggesfully. It seems to be so that 50% tool works better for the mirrors F/4..F/6. If the mirror is slower, the normal 75...80% machine polishing or full size tool will do the figuring job very well.
I have read that sub-diameter laps would cause easily problems to the surface. Ok, my experience is still quite limited but at least I have to say that I have opposite results! Every time I have used 50% tool with watery thin cerium mixture, the surface of the mirror has seemed to get smoother surface compared to one polished by a machine. Even if I have always made the machine polishing with quite light weights. Ie. app. 3..5 kg (6,6...11 lbs) tool weight.

To make a list of the important things of manual sub-diameter figuring:

- always TOT, very well pressed tool, it's must!
- clean and really open channels between pitch squares, micro-facets usefull
- pitch layer thickness app. 8 mm, Gugolz 64 normally
- cerium mixture like coloured water, better to use milled rouge or cerium-oxide!
- very light weights, only a part of your own hands' weight
- concentrate to produce smooth motions as you can, no jerkiness allowed!
- do not turn the tool on your hands after every stroke...
- ...it's enough to turn the tool after 20...30 strokes...
- ...then stop the work and turn the tool in the middle of the mirror, not on the edge area!
- the W-stroke as a W-letter, kissing the mirror edges or slightly over!
- 10...15 minutes work before testing
- if the center gets more correction than the edge or other way round....
- ...change the W-stroke's sides "\" and "/" or center "/\" widness...
- ...just a few minutes figuring, test to see what happens
- patience, patience, patience!

I like to get closer and closer slowly, in the same time the surface seems to get even smoother. The most impotant thing is to have a clean edge without any signs of TDE after the polishing is over. Many times machines produce very nice, almost spherical figure if done correctly. Anyway, I just use the same figuring method for a Lurie-Houghton primary than to a normal paraboloidal mirror of a Newton telescope. As I already told I try to leave the surface figure a little bit undercorrected sphere before figuring. In the HN primary case a good sperical figure is quite easy to achive.When you are happy with the straight ronchi lines (2..3 visible), it's time to change the ronchi grating. Finally 5..6 lines/mm ronchi would be quite good to have. Of coarse 10 ln/mm would be nice! Also it's good to remember the old, good knife edge test.

Today (02-Feb-2006) I continued polishing my primary. Now the spent polishing time is app. 7 hours totally. The figure is now very close to the sphere. Have to continue still for awhile to get rid of the hint of TDE. I tried to take a couple of ronchi photos but had to stop because of our "Terminator Dog" liked to play with my camera... :) Nowadays (13-Sep-2008) our dog has reached adult age and is everyone's friend.

Today (05-Mar-2006) I had a few hours to continue the polishing of the primary. I had to make a new pitch lap earlier because the old one's thickness was too thin (under 5 mm). If the pitch layer is too thin you may see very odd difficulties, such as the TDE won't disapear even if you know you have a right stroke! If pitch squares are too thin they won't be as flexible as they would be to work properly. Of coarse softer pitch might cure the problem.Anyway, now the satus of the primary (9h) is that TDE is gone but there still is a app. 10 mm wide area on the edge which has a longer radius than the rest of the mirror. It's not bad, start to be clearly seeable when 4..5 ronchi lines (3 ln/mm) are visible. I will let the primary cool down and will test again. Otherwise the figure is very good. Ronchi lines are straight even 1..2 lines visible. I won't start even talk about manual figuring until the ronchi lines will end to the bevel as straight as they do now on the other areas of the mirror.
Once again, it has been clear that a good spherical figure won't be achived if the tool (75...80% TOT) is rotating on the primary different speeds in different points of the stroke. Sometimes it's difficult to get the tool rotating in the same speed. If the tool's micro-facets are too small the tool starts slowing down on the edge areas (normally). I have added scratches on the surface of the tool under running warm water with a sharp knife. Immediately after that the tool is very jerky and you may get a feeling that the whole machine is breaking down. Normally after a few minutes the jerkiness is gone and tool is rotating very nicely.
I would say that the same rate rotating of the tool and "free floating tool" with light weights are the key factors to produce the nice spherical figure. "Free floating tool" means that the over arm structure has to be made so that it won't produce any side forces. Ie. a small play in the vertical hinge of the over arm is good! I do not know for sure but that kind of experiences I have...
Today (25-Mar-2006) finally came the test in which the ronchi lines looked straight all the way to the edge. All in all, it just was a matter of the smooth and equal rate rotating tool. I polished about 30 minutes and the final slight TDE disapeared. On the left is the latest ronchi photo after the polishing is done. Here is the second photo and third one. The lines are also quite straight with 4 and 5 ln/mm. Have to check now careful the need of figuring level. I would say that the figure is just a little bit undercorrected sphere if there is only 2 lines visible. It's just difficult to take a good photo of that situation with my camera.

Grinding the lenses to carbo 220 stage

I got the lens blanks 20 mm thick, 195 mm by diameter. All surfaces were grinded to flat and the blanks were edged. The material is crown glass marked to be K9 which is chinese equivalent to BK7. Normally lenses are about 1:10 thick optical glass disks. Lurie-Houghton setups show that it would be better if the lenses are thinner but the influence is very small. So, it can be say that normally you start grinding 1:10 glass blanks. I have started grinding R4/R2 just by putting one of the disks on the turntable and manually grind them together. After reaching the correct radius with 150 carbo I have continued with 220. If you change the disks places on the turntable the radius is quite easy to keep very close to the target.
After R4/R2 are done with 220 carbo I have made concrete (or plaster) support tools for those grinded surfaces. For now on it's better to support the lenses evenly on their surface areas, so that you do not cause flexing of the lenses during grinding R3/R1. There is rather much work with those tools but absolutely they are worth of making. After R3/R1 are also grinded with 220 carbo you have to make those tools 6 pcs. more. 2 pcs. for supporting the surfaces R3/R1 and 4 pcs. for polishing each of the four surfaces. One possibility would be to make only 2 pcs. wooden support tools and use some kind of a filling material between a lens and tool. If you have a grinding machine it would be nice to use 3 pcs. plastic holders instead of taping for the lenses in order to quickly change lenses place concave-on top / convex-on top. In a lathe it would be nice to make separate individual plywood tools.In my case once again I felt frustrating so slowly the work of R3/R1 proceeded. I just had to let my grinding machine to do the hard work! Here is a photo of my setup while grinding R3 on R1. R3 on R1 by using a stroke that the upper disk's minimum overhang was 1/4 of the dia, as shown in the photo. The max. overhang naturally was 1/2 of the dia., ie. the upper disk's center was on the lower's edge. That way I got about 0,05 mm (reading of D150 mm spherometer) glass away with 2 or 3 wets. The weight was app. 5,5 kg, the upper disk included in. When grinding glass on glass a good happit is to change the blanks from up to down peridiocally to ensure even surface quality and maintain the radius. If you have a spherometer you do not to have a reference flat to calibrate it. In LH telescope case you already have a good reference, the primary. Normally the primary is good to make first and you can measure it's RoC quite accurate. If you are careful you can set the spherometer to show exactly the correct reading on the primary. Or you can adjust it to show zero and then take the relative readings from the concave surfaces R3 and R4. If the radiuses are ok enough and the surfaces smooth it's time to start measure wedge of the lenses.

Fine grinding lenses

If you have a Mirror-O-Matic type machine, ie. the turntable is rotating faster than the overarm I have found the following sequence for lenses to be best for me. First both pairs will be grinded manually on each other to achive the radiuses. The grid size about 150 carbo. Then I have manually continued with 220 carbo, in the same time reducing the wedge of the lenses. It's good to have two flat supports so that you can quickly change disks places. The lenses can be roughly fit to the flat support by using, for example, thin rubber rings. In any case no hard grinding pressure is allowed. After reaching the wedge on acceptable level (app. 0.1 mm or less) and the surface is smooth I have moved on using my machine and individual grinding tools.
This time I tried first time "aluminum plates glued with pitch tool". ALP tool seems to work very fast and really efficently! Important is to grind bevels on each side of the each plate. Otherwise in the finest grades you can have nasty skraches. Also no heavy weights needed and allowed. I used about 4 kg weights including in the lens and support tool. A disadvantage is that aluminum + pitch cause dark grey waste which coloures your fingers.Before beginning machine grinding I zeroed the spherometer on the grinded surface, so that I was able to detect even slight radius changings. If the tool is in contact well and you have the correct stroke/overhang the surfaces will not change. Mine was not at the beginninig so the radius of R4 came about 70 mm longer. By adjusting the stroke smaller I get the radius back (deeper) to the spherometer zero.Anyway, I personally think that absolutelly those pitch tools are worth of trying in lens making. I would choose aluminum plates but tiles work too. Aluminum flat bars are easy to buy from local stores and they are not very expensive. This time I used round al-plates.
If liked the surface pairs can be manually grinded for awhile to ensure they absolutelly have the same radius after each grid sizes. If you do not have machine manual method, of coarse, is the best way. By grinding manually the pairs you do not need to make extra grinding tools and you can be sure pairs match together for sure.

De-wedging both lenses

In my case I ended up to have about 0.3 mm wedge in the both lenses after 220 carbo stage. I would have de-wedged earlier but I wanted to do rough grindings first. As Texereau illustrates in his book de-wedging can be done either on the top or bottom of two disks. I even this time tried to do the both disks de-wedging in the same time. Ie. I didn't rotate neither of the disks. Just kept the thick sides towards me and moved the upper disk back and forth on the edge area of the lower disk. Maybe it's better to de-wedge one disk at once. In the beginning of each wet I did de-wedging for awhile an then continued normally until it was time to make a new wet. This way the wedge disapears quite slowly. 10..12 carbo 220 wets were needed until the front lens has now wedge 0.08 mm and the rear 0.06 mm. Without notice it has to be continued normal grinding actions about 15...30 minutes depending on the worker.
I have read that wedge should be about 0.1...0.125 mm after carbo 220 before changing to a finer stage. In finer stages the wedge reading should be split in half until after the finest grade the wedge is 0,025 mm or less.

Trepanning the lenses

Last time I trepanned my existing telescope's lenses at the beginning of carbo 220 stage. I figured out that it would be a good idea wait until the major of grinding work is done. Especially that is valid (I think) in bi-concave lens case. The thinest point of the lens is in the middle. By trepanning the lens after grinding both sagittas I eliminated the need of making rather huge bevels to the lens hole's edge and the core itself. It's just not fun to make and estimate the needed bevels' size if you have trepanned the lens blank before any grindings! After 220 carbo app. 1..1.5 mm bevel is enough. If a bevel is grinded off totally during grinding rather big pieces could break from the glass edge.
On the contrary, the bi-convex front lens's the thinest point is naturally on the edge, leaving the thickness of the center almost untouched. Almost the original lens blank thickness can be found in the center after both sides grindings. In this case the bevels will "stay alive" to the end. Also if you think about the trepanning situation in reality. You are normally using a drill press and some kind of a drilling table. Both sides flat lens blank is much better to adjust on the table than convex one! Especially if you want to drill from one side 3..4 mm and turn over the lens and drill from that side finally through the glass.
Ok, speaking of trepanning methods there is at least two main styles. Either you are going to use a traditional bicuit-cutter and loose abrasive or you use a diamond cutter. I haven't tried the last option but have heard positive things of using it. Instead of that I have always used a biscuit-cutter and carbo 150 quite suggesfully. Last time I tried that "both sides drilling" method and it worked very well. Only you have to remember that it's recommended to use a really good drilling table to ensure that the lens will be centered accurately after turning around.

Polishing concave and convex surfaces

I have always made polishing tools more or less traditionally. Ie. I have poured the melted pitch on a surface of a concrate or plaster tool and after a short cooling time pressed the pitch on the glass surface. The process has been succesful only partly. Some times there has been left a large area on the center which is far from in good contact to a mirror or lens.
So, I read from the Net a few articles concerning of mould usage to make a pitch lap. I had bought a few years ago a car mat but at that time my experience of the mat was not very good. Ok, now I thought it would be so nice to manage to use a mould and get the pitch layer ready to rather good contact and best of all, get the pitch squares ready made in the same time. At first I tried to use wood oil as a separating element between mould and pitch. It worked very well and the mould came off nicely but the oil melts pitch surface so no good. Then I tried to use liquid Sun Light soap and that has worked quite well for me. Now I have made 3 pcs. pitch tools with that method succesfully. Important is to try cut the car mat as round and accurate as possible. I have used a Borofloat (flat surface) blank as a moulding base. On the surface I has put the mould on and around the mould/glass blank a plastic straip. If the mould would be tight enough to the plastic straip the moulded pitch layer would be straight away perfect. Mine is not as can be seen from the photo so I have to clean the edges which is not a big deal. Anyway, after a short cooling time I slide off the mould with pitch from the Borofloat blank. I put the mould/pitch cake on the lens surface and press gently the pre-heated concrete tool in to still soft pitch surface. While gently pressing the tool stick permanently to the pitch and the squares copy the right curvature of the lens. Ok, the mould is in between and normal hot pressing in needed before starting polishing but normally only short pressing is needed. Here is a photo of the polishing tool of R2 surface after using it 4 hours with light weights. I did not need to re-channel it.

Now (19-July-2007) I have polished R3, R1 and R2. There is nothing different to polish a lens than a mirror. Only it's better to support the lens carefully in order to not introduce any flexing of the lens. Actually I have always used light weights on the machine to eliminate any flexing possibilities. But it's better to be sure, so much work has to be done before you can test the optics.As it can be seen from the photo of R1 that there is no center hole. I did not have a suitable biscuit cutter at the time of grinding so I decided not to trepan the lenses at all. I try to make a short secondary support as possible. Ok, I will loose one of the benefits of corrector lenses but on the other hand making the lenses has been easier and faster.

Here is the latest OSLO-file to download. The setup is still preliminary. Here is Ukyo Chen's proposed set-up. Thanks Ukyo!
....................................................
Anyway, I will try to add here the different stages of making my new LHN optics.