The "Krupa Collimator" Revisited
Submitted: Friday, 15th December 2006 by David C Nicholls
Collimating a Newtonian is something of a dark art to many amateurs. In fact it's not especially difficult, and worth a few minutes effort before starting an observing session, because it can substantially improve image quality.
There are many web pages dedicated to Newtonian collimation, and numerous written articles. For more information, Google on "Nils Olof Carlin" on the Web - and especially his excellent article in the June 2002 issue of Sky and Telescope.
In brief, the usual way it's done is to set up the secondary mirror geometry using a sight tube, align the secondary using a simple laser collimator, then use one of several methods to align the primary. (The laser collimator itself needs to be collimated as they're almost never lined up properly when purchased).
This article is about a new method of collimating the primary mirror, once the secondary has been set up correctly.
The Krupa Collimator
In Amateur Telescope Making Journal, Issue #4, 1993/94, p.25, Jordan D. Marchè II described a novel collimator device for refractors and Newtonians which he named the "Krupa Collimator" **.
Here I present a variation on the Krupa collimator, which I call the "Focused Krupa Collimator" and which provides a very accurate indicator of primary mirror collimation, once the secondary mirror has been lined up as per normal collimation with a sight tube and simple laser collimator.
The original Krupa unit consisted of a small LED mounted in the center of a transparent disc, mounted in/on a tube that fitted into the focuser. (Figure 1)
The LED provides a point source of light that is sent down to the primary mirror via the secondary, then back to the eyepiece tube again. As the LED is very close to the focal point of the primary, the returning beam is parallel light (unfocused). If the primary mirror has a small central "collimation ring", the shadow of this ring will be embedded in the light returning from the primary mirror.
The eye cannot easily focus the parallel rays of the returning beam, so the original design was not readily usable for collimating a Newtonian.
A little experimenting showed a simple way to turn the original design into a very accurate collimating tool for Newtonians. If you hold a 10x magnifier over the top of the LED, when the sight tube is located in the focuser, a magnified "shadow" of the primary "collimator ring" is immediately visible (lower light levels are desirable).
The focusing lens serves two purposes: it provides a sharp, magnified view of the back of the LED, and it brings to focus the returning light from the primary. Thus the image of the shadow of the primary collimation ring and the LED are both seen sharply.
If in place of the hand-held magnifier you substitute a lens mounted in the tube above the LED and transparent disc, you get a compact unit that is a useful tool for collimation.
The LED, centered in the tube, provides a central reference point. If the collimation is slightly off, the image of the primary collimation ring shadow is not centered with respect to the LED. Provided the secondary mirror has already been correctly collimated, adjusting the primary collimation screws to center the shadow around the LED brings the telescope into excellent collimation.
The view seen through the collimator is shown in Figure 2. An offset ring shadow around the LED silhouette is shown if Figure 2a. Figure 2b shows the ring shadow image correctly centered, indicating accurate collimation.
The "focused Krupa" collimator is in effect a direct view equivalent of Nils Olof Carlin's Barlowed laser. With the Krupa, you view the image of the collimation ring shadow directly, rather than looking at the projected image on the base of the Barlow.
Once the combination of the original design and the focusing lens was found to work, it was a straight forward matter (although lengthy) to build a dedicated combination of the LED unit, the tube and an suitable focusing lens.
It turns out that Surplus Shed has a suitable lens (catalog No. L4661), double convex, coated, with a focal length of 35mm and a diameter of 31.6mm (1.25").
My reason for choosing this lens was empirical: it is approximately a 10x magnifier for close objects (the rear of the LED) and fits neatly inside a 1.25" tube. And most important, it was inexpensive and readily available.
The angular magnification (m) of a lens (focal length f mm) close to the eye and close to the object (in this case the rear of the LED) is given approximately by:
m = 1 + (250/f)
The 250 figure in this formula is an approximation to the near-point distance of the eye (in mm) - the closest object it can comfortably focus.
For the selected lens, the magnification turns out to be 8.14x, close enough to my original 10x magnifier to do the job. In my case the near-point is closer to 300mm, so I see a magnification of around 9.6x
A plan for the components that I used to build the unit is shown in Figure 3.
The dimensions are those for the prototype unit, which I've used on my Lightbridge 12" f/5 Newtonian. You can experiment with the lengths of the tubes to suit other applications (eg use with eyeglasses).
Note: The red figures in Figure 3 above are for a second version I made - they give a slightly better view [for someone not using glasses]. The double convex lens I used separates the spacers (#2 and #3) by about 4mm, where the spacers have a wall thickness of 2.5mm. If you use thinner or thicker tubing for the spacers, you'll need to adjust the length of those two spacers slightly, otherwise the back of the LED isn't quite focused properly. Remember, this is experimental!
The lower three tubes (#2,3,4), the lens and the disc in the above diagram slide snugly up into the topmost tube (#1), with the top of tube #2 flush with the top of tube #1 and around 27mm of the lowermost tube (#4) protruding from the bottom.
The top of the plastic disc should line up with the 1mm hole in tube #1 to allow the LED wires to exit. Likewise the notch in tube #3 should line up with the same hole.
The double convex lens is inserted with the flatter side down.
In use, the 27mm (35mm) exposed section of tube #4 is inserted into the focuser.
To build the prototype I used scrap parts (PVC piping, cardboard tube from an old fax paper roll), a 15 cent 3mm clear red LED, and surplus 1.5mm thick clear plastic sheet. A drill press and/or lathe is very useful, but a hand drill on a bench mount is all that's needed to make the tubes and clear plastic disc (and is what I used).
I also cut off the rounded nose of the LED to show the small square light emitter unit. This makes the source more accurately a point source, but probably isn't strictly necessary, if you use a small enough LED. Polishing the cut face of the LED is necessary, cerium oxide or "Brasso" on a soft cloth work well.
In building the unit, the most tricky part is soldering the wiring. I used "wire wrap" wire because it is very thin. I trimmed all but a short stub of the LED "legs" off and soldered the wires to these stubs, then connected the other end of the wires to a small dual-AA switched battery pack, via a 100 ohm resistor. The battery pack I used was a dual-AA battery box with switch from Dick Smith Electronics (part number S6155)
When the soldering is finished, blacken the back and sides of the LED to cut stray red light (black marker pen and model paint work well). It's not essential to get the rear blackout perfect, but it is aesthetically more pleasing.
Using a red LED has the advantage of not upsetting dark adaptation, when used at the telescope while observing. I also tried a white light LED, but this conferred no visual advantage, and did upset dark adaptation.
LEDs and Image Brightness
The brigthness of the image depends on the type of LED used and the current through it. I used a 3mm 2800 mcd clear red LED, Jaycar cat. # ZD 0104, with a 150 ohm resistor in series with 3 A76/LR44 alkaline button cells. Prior to that I used 2 1.5 volt AA cels through a 100 ohm resistor. A refinement might be to add a small 1K linear potentiometer or trimpot in series with the fixed resistor. I find the 150 ohm 4.5 volt combination is bright enough to collimate at normal daytime inside light levels, and is still OK at night.
Some Photos of the Prototype and Components
The 2xAA battery pack is a bit cumbersome, due mainly to the size of the batteries, but it's hard (impossible?) to find a commercial battery holder for 2 or 3 A76/LR44/AG13 alkaline button cells, so I built one from a spent lip balm tube(!) and a small spring. The tube holds the button cells quite snugly. The balm advance/retract control provides a convenient on/off switch, too. Cheap and effective. See photos below.
How well does the Krupa work?
Very well indeed. It's easy to use. It may not have the slick anodised finish of a commercial product, but it leaves you in no doubt about how well lined up your Newtonian optics are. For a total cost of around $12 including batteries.
It's not about to replace sophisticated instruments like the Catseye autocollimator, but once you have lined up the secondary precisely using an autocollimator, the Focused Krupa is actually easier to use than anything else I've tried, to give a final adjustment to the primary mirror collimation.
The designs and ideas here are free for anyone to use. If reference is made elsewhere to this information, due credit should be given to the author.
If you wish to manufacture the unit commercially (or a development of it), you are free to do so, but I'd appreciate a sample!
** Republished by Willmann-Bell Inc. in "The Best of Amateur Telescope Making Journal" Volume 1.
An earlier and slightly edited version of this page has been published in the December 2006 CAS journal "Southern Cross".
Article by David C Nicholls (dcnicholls) revised 5th January 2007. Discuss this article at the IceInSpace Forum.