<

[2/4] 488 mm F4.85 Dobson

Making the Primary Mirror

After ordering the half a metre mirror blank there was no stepping back. The blank arrived from Hanko to Helsinki 8.5.2006 and cost me around 400€. I was excited and started to work immediately. The first task was to bevel the sharp edges to prevent chipping. This blank was a beast! It took me six hours to bevel all edges since the total length was more than three meters! The diameter of the blank wasn’t exactly 500mm but 490mm. Therefore after bevelling the final diameter became 488mm.

Before I started working on the concave mirror surface I ground the back side of the blank flat. This both prevents astigmatism to form and also looks nice. The surface of the glass was very uneven and there were some heavily rising edges. First I ground these bulges off using both a whetstone and #60 carborundum. Then I used a 36cm tile tool to smooth the surface down to #220 carbo. First I managed to make the surface slightly concave but then I applied more pressure on edges while doing long W strokes. This returned the surface back to flat. I measured the flatness roughly by putting a straight aluminium profile over the mirror blank. When the back side of the blank was finished I covered it for some protection during rough grinding. Smoothing the back side took around 20h.

Then I moved on working on the actual mirror side of the blank. Again I used first the whetstone to grind off the biggest bums. The mirror side of the blank was already slightly concave having around 1mm deviation in the centre compared to edges. My most important tool in rough grinding was a heavy 17cm iron disk. Bottom of the disk I glued some 14mm nuts with epoxy.

My strategy was to grind first with the 17cm iron tool, then 20cm glass blank and finally with the 36cm tile tool. The idea was to repeat this sequence and end up with a smooth spherical surface. However, I noticed that the shape of my tile tool doesn’t follow the curved mirror blank surface very fast and attends to turn the surface back to flat. Therefore I decided to leave the tile tool for awhile and continue my glass digging with two remaining tools. After using the small 17cm tool for awhile the surface was no longer sphere. I returned the surface back to sphere by working only with the 20cm and 36cm tools. After 44h of work I had around 6mm deep hole middle of my glass. Now I was more or less done with rough grinding.

"Rough grinding is a caveman job, so do it like a caveman. Eat well, sleep well and work like hell". - John Dobson

I had some bad luck with my plaster tools. I made a new tool for fine grinding and bought one litre of epoxy. This epoxy was not easily running resin but it turned out to be some kind of thick paste. I still managed to finish my new tile tool and I left the room for washing my hands. Then I heard a BIG BANG and found not one but two pieces of my tool on a shop floor, great! Now I was out of tiles and out of epoxy on which I spent 50€. For my new tile tool I bought the right epoxy resin and also decided to try not ceramic but tiles made of glass. The glass tiles wore out fast but they lasted till the end of fine grinding.

After rough grinding I went on for fine grinding. It was pleasant and smooth to work with the new tile tool and had no problems. Every time I moved on to finer grit size I washed the tool carefully with a dish brush. The finest grit I used was 15 micron aluminium oxide. I followed the grinding progress by illuminating the back side of my blank with a flash light and then observing the surface roughness with an inverted eyepiece. The fine grinding for a certain grit size was finished when there were no traces visible from the previous stage. After the last grinding stage the focal length was equivalent to 2367mm. My goal was 2500mm but I decided to leave it as it was. There is no harm when the tube is going to be a bit shorter. I finished fine grinding my mirror 3.8.2006 and it took also 44h just like the rough grinding.

Polishing the Primary Mirror

For polishing the fine ground surface I needed to cast a new pitch tool. I no longer used special hard dentist plaster that turned out to be a nightmare. Adhesion to the plaster was very poor even for epoxies and therefore I decided to use Ardurapid floor screed instead. I made the tool quite thick so its own weight applies most of the pressure in polishing process. The diameter of the pitch tool was 40cm but despite the large size I had no major problems casting the pitch. It seemed like I finally got it! I used medium-hard pitch for polishing.

Before I started the first polishing session for that specific day I always made at least an hour long cold press. As a weight for the press I used a bucket filled with macadam. I washed the mirror using a water hose so I didn’t have to risk moving it around. The pitch tool I washed in a sink. The goal was to have a spherical surface after the mirror is polished. In order to achieve that the strokes during polishing must be varied. The most I used the common W-strokes but also straight I-strokes cross the centre and straight off-center strokes were used. I also applied some V-strokes and large elliptical ones. I always did at least one full round using a certain stroke pattern and didn’t change it during the round. I worked for 20min and then I made a 20min long cold press. Usually I was able to polish around 60min per day (effective polishing time). Polishing a mirror this size by hand turned out to be good workout.

To prevent astigmatism I used certain procedures. In my mind I divided the work stand in 12 different points (like 12h). Every time I started a new 20min polishing session I turned my mirror to the next adjacent point clockwise. During cold presses I positioned the pitch tool always to the same direction respect to the work stand. Therefore also the position of the tool changed in controlled manner respect to the mirror. I was walking around the work stand anticlockwise during polishing while the pitch tool was rotating slowly clockwise.

I made good progress on polishing from beginning and even the edges seemed to polish very well. I had the feeling that finally after three finished mirrors I had learned the secrets of polishing. It is important to apply enough pressure on the pitch tool. I understood now that on previous mirrors I didn’t use enough pressure and that’s why it took so long time to polish them.

It is important that the pitch tool moves across the mirror smoothly without any problems. It should also make a low “ssssss” sound and you should feel that there is great deal of friction between the pitch and the glass. If this was not the case then more pressing or pressure was required. I used cerium oxide for polishing.

I examined the progress of polishing using the same method as in fine grinding. In the end I was able to see nothing but some free particles on the surface. I also compared a reflection of a laser pointer between my polished surface and other polished surfaces. It seemed like the reflection from my surface was less than others. The main mirror was now polished! The effective polishing time was only 22h which was half of the time I needed for my 30cm mirror! However, the total time needed for polishing process with tool makings etc. was around 70h. Polishing went very well and beyond all my expectations.

Polishing Progress

00h 40min Gloss everywhere, 90mm from the edge starts to dim.
01h 25min Starts to dim 60mm from the edge. Gloss increased but not fully polished anywhere.
02h 10min Starts to dim 40-50mm from the edge. Gloss increased. I made my strokes longer.
03h 20min Gloss increased and now it's difficult to see the excact border where the dim region starts.
03h 40min Gloss on the edge increased and I'm able to see clearly through edge regions.
04h 40min More glaze but still some dim areas especially close to edge.
05h 40min It is difficult to see any residual roughness in the centre region.
08h 00min Extremely difficult to see any scattering in the centre.
11h 00min More gloss on the edge regions. Not a big difference between the centre and edge.
12h 20min Starts to be difficult to see roughness on the edge regions.
13h 20min Little pits can be still seen across the mirror by using an eyepiece.
15h 00min Density of the pits decreased.
20h 20min It starts to be difficult to see the pits with an eyepiece.
21h 20min The mirror starts to be polished. The reflection of the laser beam less than reflection from other polished surfaces.
22h 00min The mirror is polished! With the eyepiece only small free particles can be seen. Very low reflection of the laser beam.

Figuring and Measurements

Before I could start figuring the mirror I needed to finish the frame of my main mirror cell and the mirror stand for Foucault measurements. For the stand I used 50x50mm plank that was painted black. The mirror cell is then hanged on the two hooks on the stand. For safety I attached two temporary aluminium profiles to the main mirror cell so that the mirror does not touch any surface when the cell is lifted to vertical position. I recognised that the seat belt I used to support the mirror stretched slightly when the mirror was hanged to the stand. Therefore I had to wait for the measurements until next day when the belt was stabilized. This is also the reason I decided to use the aluminium belt in the final version of my mirror cell.

Now when the mirror was polished I needed to figure the spherical surface into paraboloid. For figuring I casted some smaller pitch tools with soft pitch. The two main tools I worked with where 200 and 150mm in diameter. I also casted 80 and 100mm tools just in case but luckily I didn’t have to use them.

This is where I realised that despite many instructions in literature there is no point to make any pre-grooves into pitch when pressing the pitch and the mirror together for the very first time. Every time I made pre-grooves I ended up with some air pockets. The main reason is that the edges of the grooves rise up slightly and have the first contact with a mirror. When I casted my tools without any pre-grooves I managed to get very smooth surfaces without any air pockets or other defects. I always kept the tools covered when not used.






Notes of Measurements (Finnish)

In order to save time I first made Ronchi measurements only. The shape of the mirror was pretty close to sphere and I was able to continue straight to parabolising the mirror. The Ronchi lines started to bend quite nicely from beginning. When the Ronchi lines were curved enough I changed to Foucault measurements. For Foucault measurements I made two masks. One conventional with four rectangular holes and one Couder type mask with seven holes. Practically I used only the Couder mask since it was easier to use and covered larger area of the mirror.

After every 15min figuring session I measured the mirror. There was one point where it was difficult to get the error under 120nm. Then I changed my measuring policy and determined the difference only between the inner and outermost holes until the mirror was no longer clearly under corrected. This made my measurements much faster. First I covered all the holes in the mask that I didn’t measure at the time but later I realised it was not necessary. The open holes in the mask didn’t much disturb my measurements and I didn’t have to touch the mirror or walk between my Foucault tester and the mirror.



I noted down all my strokes and session times in order to learn the effect of different strokes. I think this is vital for learning. Many times when I wanted to correct a certain error I went through my notes to pick up the right stroke patterns.

At the end of figuring my results from Foucault test start to vary because of unstable environment. After measuring the mirror ten times I got the average value of 45nm for the final error in the mirror. My goal was 1/10 wavelength (55nm) so the mirror seemed to turn out even better I expected. I also tested whether there is any astigmatism by observing the Foucault light source with an eyepiece. The resulting diffraction pattern was round and I didn’t see any ellipticity caused by astigmatism.

One way to understand how precise the surface is, is to imagine that the mirror is as large as the Baltic Sea. In that case the highest wave in the sea would be only around 12 cm high (P-V value used).

P-V WaveFront Error This value tells you what is the difference between the highest peak and the lowest valley in your mirror. In my case it’s 0,15 wavelengths. However, this doesn’t tell you how many of these peaks and valleys you have and is therefore not a very good measure to describe the overall performance of a mirror.
Strehl Ratio This value defines the ratio of intensities between the resulting airy disk and a perfect airy disk without any aberrations. In my mirror it’s 89% and commonly 80% is considered as a limit for diffraction limited optics.
Surface RMS Error Statistically averaged error (RMS) at the mirror surface in nanometres. The limit for diffraction limited optics is around 19nm. In my mirror it’s around 15nm. According to these Foucault results my mirror seems to be pretty good. Interferometric measurements would give more accurate understanding of the overall performance of the mirror.

I sent my figured mirror to Orion Optics in UK. They aluminized the mirror with their special enhanced multi layered Hilux coating. According to Orion Optics the reflectance of their coating is 97% over a wide range of visible spectrum. The corresponding common value for a conventional coating is 86%. I also ordered my 88mm secondary mirror from there with the same coating. So if the numbers they represent are true then my 488mm telescope collects as much light as 530mm telescope with conventional coating. Below is a photo of the coated mirror when it arrived back from UK. I put a round sticker middle of the mirror to make collimation easier.

Previous Page

1 2 3 4

Next Page