Jump to content


- - - - -

Notes on the Construction of an Oiled Contact Doublet

Discuss this article in our forums

Notes on the Construction of an Oiled Contact Doublet

This is a description of my adventures with the fabrication of a 6” f/15 oiled achromat.It is assumed you will know all about grinding, polishing and figuring. It will discuss some problems I ran into and what I did to solve them in the easiest and most practical way, as well as things to watch out for. This is not a step by step description of the making of a refracting objective let alone the complete telescope.

I am assuming that a 6” or 5” will be chosen because anything much larger is going to be awkward to mount in any respectable f/ratio. Anything much smaller is hard to make by hand,at least for me it is. Turned edge is the problem. It has something to do with the size of your hands relative to the diameter of the disc which apparently creates a higher pressure at the edge. When these are made commercially they are done on machines with specially made large polishing blocks perhaps 7 to 20 at a time. A 6” is the ideal size for the amateur. Secondary color is practically non-existent at f/15 or so and tolerable down to f/8 or f/10 depending on your sensitivity to the far reaches of the spectrum. In the contact doublet, coma and spherical aberration are simultaneously correctable if the right glasses are chosen. However, these most likely will be more expensive and harder to get than the standard BK7 & F2 or F4 combinations. With these much more common glasses used in a contact doublet, coma cannot zeroed along with spherical aberration but can be held to ¼ to ½ Conrady's limit for OSC'. Don't worry, he's notoriously stern with his tolerances. Even with a non-adjustable lens cell if the tube is carefully cut square you'll be fine. Coma and spherical can, however, be nearly totally corrected together, with the common glasses, only if the elements are air spaced as in the Fraunhofer type of objective. I would not bother with the various other types of objectives such as the Steinheil or the equii-convex crown varieties especially as a first attempt. Fraunhofers are certainly doable but the mechanics of the cell are more demanding as well as decreased light transmission and the possibility of ghosting from the uncoated interior surfaces which will also have to be made to a higher standard than in the contact doublet.

First of all I cannot recommend too highly that you start with a pre-generated set of blanks. This is especially true for anything faster than f/15. They will come off the generator edged to equal size with the wedge at an insignificant or small enough value that can be easily corrected. Believe me, its well worth the extra cost unless you figure your time is worth about $1 /hr. Don't worry about center thickness as its only a concern with lenses that are outside the realm of the amateur. Just specify them thick enough so that flexure can be eliminated with easily made backers. I've found a ratio of around 1/8 – CT/D for the crown will work well and about 1/12 -CT/D for the flint. This applies to objectives of f/12 – f/15. Faster ones willhave a thicker crown as a matter of course so watch that the edge doesn't get too thin or hard to support without flex. You will want to keep the weight within reason too because your tube will become front heavy and be even more difficult to mount. The main thing to watch for with optical glass of the common variety is strain but this should not be a concern if it comes from a reputable source. If specifying glass, grade B is fine and will save money. Grade A is, with modern glass, what schlieren grade used to be.

The main pieces of equipment you will need is an ACCURATE spherometer and a flat, preferably aluminized. A fringe box is very handy but you could get by with fluorescent. A three footed wedge checker is also nice if you don't trust the generator and essential if you're grinding from flat discs. Also a good ray-trace program is a necessity to, in effect, re-design the lens several times, plugging in the value you end up with on the most sensitive surface to determine what the new values for the other three radii should be to correct color and spherical aberration at the focus.

The 4" spherometer w/Mitutoyo digital gauge, an 8" built later and the completed achromat in its cell.The #'s on the spherometers are the radii of the circle passing through the center of the balls squared for use in the spherometer equation.

The spherometer must be accurate enough to measurematching CC, CX surfaces to be equal within about 1 maybe 2 mm on a radius of 500mm or so.This will take a gauge reading to .001mm for a 4” spherometer.I started out with a ring type that had a mic reading to .0001” or .0025mm and for some reason could not get an equal reading on matching radii. I don't know if it was inaccuracy in measuring the ring diameters, the “feel” when the spindle made contact or invisible crap on the ring when zeroed. I later built a 3 ball type with a digital gage reading to .001mm that works much better and is easier to use. Perhaps a gage reading to .01mm would do if used intelligently. When constructing a ball type make every effort to determine the ball spacing. Measuring, doing the math and then back engineering from accurately known surface radii will get you there. Anyway, the more accurate the better.

A flat is a practical necessity and enables double pass testing, a big help. Make the lens cell now. It is much easier to test the lens against a flat with it in a cell. It can be squared to the mirror and put back in the exact same way each time you test. Believe me, it will save a lot of time. There are other ways of testing but you'll never regret acquiring or making one of these flats as they can be used for testing parabolic mirrors or complete telescopes. In this type of usage the flat does not have to be truly flat but must be smooth. A few rings (4-6) CC or CX is perfectly ok. From an optical point of view it could be considerably more curved and still work fine but with more than a handful of fringes you cannot easily tell how smooth it is. This is true of any pair of matching surfaces. To have no uncertainties as to surface quality two or at most four fringes on radius should be the target if used in air. For contact type of objective I would recommend that the R2 & R3 surfaces be oiled rather than cemented and so surface problems and ghosting will be mostly eliminated but it pays to polish them out to within perhaps 10 or so fringes of each other to check for astigmatism and don't forget that ideally the oil forms a very thin meniscus lens of equal thickness from center to edge that has a refractive index all its own and if it has any power to speak of can alter the delicate path traced by the light of various colors on its way through the lens. So unless allowed for in a raytrace it pays to hold the radii quite close to equal.

The flint element on its wood backer having its CX surface figured.

Tooling is somewhat a mixture of taste and experience. Trying as always to do the least amount of work I took the inner R2 & R3 through fine grinding each on the other. The outer surfaces R1 & R4 were ground with cast plaster and tiles. Finish up with 5u grit and the polish will go quicker but watch out for seizing with glass on glass surfaces. Even if you are working by hand you must use a backer to give an even pressure on the rather thin glass blanks.The backers for holding the elements were fashioned from high quality plywood disks of the same diameter as the blanks.Waterproof these with polyurethane or some such and seal the edge with tape.You will be using these MOT and MOB so they will get wet. In order that the curved back surfaces were supported evenly I got some Plumbers putty and fashioned it into long spagetti like strands about 1/4” dia.These were laid in a tight spiral on the plywood disk.Cut a circular sheet of plastic the size of the disk about .010 thick and then press the blank down with the plastic on the putty so as to spread the putty out to cover the whole disk.The whole form a sandwich with the blank, plastic sheet, putty and plywood.Don't let the putty come into contact with the glass especially the flint because there is something in the putty that can stain the glass.This doesn't matter much with a contact doublet but should certainly be avoided -- I learned this the hard way. Tape the blank to the disk with plastic electrician's tape but be careful not to stretch the tape too much as you pull it around the blank and disk.If you do, the compressive force from the tape is larger than one would think and it can warp the surface as its being polished and figured.And yes, I've learned this the hard way too.

I dispensed with test plates for the surfaces because I reasoned that the inner surfaces would take care of themselves and in any case R3 could be null tested at center of curvature and R2 can be matched to that. The rear R4 would be the one to figure to correct the spherical aberration remaining from inaccurate radii. For the remaining R1, I would trust to the fates, given my peerless ability to make smooth spheres. Was I wrong! At least on my first objective. I ended up with a turned down edge on R1 I could not correct working on R4. Thankfully it was only about 2mm wide and the lens still performed fairly well with only a slight halo around out of focus bright stars. I was more careful on the second identical lens being made at the same time which turned out to be as perfect as I could see on a star test. I would have to say that it is very probably worth the extra effort to make a test plate for the front surface. If you do, remember that it would only have to be 2/3 the diameter of the lens to be perfectly adequate. Just use the center to check the edge zone.

I won't describe building the mechanical parts of the telescope as that is more than adequately described elsewhere. After all this time and work it was ready to be tried out on the sky. When it was finally out on its mount Jupiter was high in the southern sky and would be my first target. What I saw was appalling. I thought I was seeing things or it may have been the something in the homemade wine I was drinking that made me hallucinate. Jupiter's disc and moons were fairly sharp but the disc of the planet was totally featureless and without any color.After the wine wore off and thinking about it during a sleepless night I was convinced that my ring spherometer may be the problem. Sure enough after finishing my ball spherometer and calibrating it I remeasured the surfaces again. After a ray-trace I could see that the red and blue ends of the spectrum were not intersecting at the 70-80% zone, in fact they were not intersecting at all. The ray-trace also showed that the only thing to do was to regrind the R4 surface to a shorter radii, about 100mm shorter out of a radius of around 12,000mm.It only amounted to, if I remember correctly, a difference in sag of .002mm over the span of the spherometer. This also had the effect of decreasing the focal length by around 20mm which could be easily measured under the ronchi test to check if what I had done was done correctly. If it was this would ensure that red and blue intersected at the proper place.

Finally, I got the lens back in the telescope where, mirable ductu, it looked near perfect. There were, I found, colored bands on Jupiter - who knew!

If you decide to build one of these you'll never regret the experience even if it's not successful. But if it is you will have a regard for it that is of a different character than that for a mirror you have made. Why this is so I'm not sure. Perhaps it has something to do with the elegant and delicate calculations that went into it. Or maybe its each of the four surfaces you have made splitting the light into its constituent colors, directing them to the next surface, complimenting each other and bringing them all to the same tiny spot at the focal plane inside the Airy disc. They can be a work of art in a way that no mirror can be.

Jan Bentz

  • dan_h, Snickersnee and happylimpet like this


Nov 18 2015 04:33 AM

I've never made an optical surface in my life (and probably never will) but this is was a fascinating read. Cheers!

Nov 18 2015 04:36 AM

I will echo happylimpet's sentiments.


I have two left thumbs, so I will stick to using instruments only, while also enjoying a nice read into making one's own optics!

Nov 22 2015 08:07 PM

Lotsa hard work - congrats on your success!

Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics