A low-tech barndoor drive

(Content last modified: 8 May 2001)

This page is devoted to barndoor drives, particularly with a simple, yet effective single-arm design. I also discuss more complex designs in the context of deciding if you need something complex. Before I start, I'd like to thank the people who've given me feedback on this page and some helpful ideas.

The text is divided up into several sections:

Plans

I've been experimenting with single arm barndoor drives, especially with regards to making them as simple as possible to build with a minimum of tools at a minimum cost, yet of high enough quality to use with a 135-200mm lens. I've tried to mention the higher-tech option where appropriate. I go in to what some people may think is excruciating detail in the construction; this is aimed at someone like me when I constructed my first barndoor, all thumbs! I figure that if you have access to a drill press, lathe, electric jigsaw, etc., you can probably do a good job of designing a more sophisticated barndoor based upon my plans or those from astronomical publications.

I believe that this design can be built in an evening by just about anyone, even if you've never really built anything before. I've put together several of these, and I can now do it in about 2 hours without hurrying. In addition to adults, this could be a good project for an older child interested in astronomy. You will probably want to read through the whole thing first before you run out and buy the materials because there may be a few places where you might want to do something slightly different.

This design is of the isoceles-type. That is, the threaded rod used to increase the separation between the boards acts as the base for an isoceles triangle. This is accomplished by allowing the angle of attack of the rod change and thus keep in contact with a fixed point on the top board. The tangent-type has the threaded rod vertical and the contact point on the upper board changes with time. This latter design loses it's accuracy faster and the mount will be less stable due to not being able to fix the contact point.

Much to the dismay of some people, the measurements are all in English units, because that's the way hardware and lumber is sold here in the United States.

Materials list

Lumber

These can all be cut from a single 3- or 4-foot 1 x 4 (I use white pine)

Hardware

Tools

Other materials

First, cut the five pieces of lumber. The cut ends don't have to be perfectly square, but cut carefully. Use a several-foot long board to align the two 13" boards with a small separation. Place the hinges such that you get as good of alignment as possible relative to the pre-cut edges of the board; a T-square may help here. Pre-drill the screw holes, and attach the hinges to the boards, one near each edge of the board with a gap in the center. Check to see that the boards line up pretty good when you close the hinge (it doesn't have to be perfect). DO NOT USE A DOOR HINGE FOR THIS MOUNT. Sorry to shout, but I've never found a door hinge stable enough for a good barndoor drive. The smaller shutter hinges are strong enough for the weight of a 35mm camera with at least a 135mm lens and are very stable. If you plan to use a medium format camera, you might need sturdier hinges.

Now comes the most crucial step. The threaded rod has to pass through the lower board and contact the upper board at a certain distance from the axis of the hinge. In inches, the formula is:

d = 228.56 / tpi

where tpi is the number of threads per inch of the screw (20 in this case), and the 228.56 is the reciprocal of the tangent of 0.25068 degrees, the angle through which the stars appear to move in one minute. This is what makes one revolution of the screw correspond to one minute of time. In our case, d=11.428, or 11 7/16". This measurement needs to be as accurate as possible, definitely to within 1/16" and preferably to 1/32" (this is easier than it sounds when using a small hinge).

Once you have marked the boards for 11 7/16", drill a small pilot hole in the bottom board. Follow this with a 1/2" hole deep enough to hold a 1/4"-20 nut with some room underneath. Drill a 1/4" hole through the rest of the board for the threaded rod.

As the boards separate, the threaded rod must pivot towards the hinge so that it will still hit the top board at 11 7/16". Thus, the 1/4"-20 nut must be mounted so that it can pivot. One possibility is to drill and tap opposite sides of the nut for machine screws. However, there is a low-tech way that is still effective and only requires drilling on one side of the nut. Drill a 1/16" hole at the center of one side of the nut but not all the way to the threads, then follow with a 1/8" bit. If you are careful, you can firmly hold the nut against a board with a pair of pliers, and drill downward. With an all-purpose drill bit it will take a while to drill the hole so don't force it. Natually, if you have access to a drill press, this procedure will be easier. The hole will probably end up being a little larger than 1/8", and in fact, you want the hole to be just large enough to accommodate the 3/16"-32 machine screw. Apply cyanoacrylate glue to the screw end and threads and stick it into the hole. In the few seconds you have before the glue sets, you can twist the screw which will help to attach the end into the bottom of the hole. Once the glue has set, squirt a little more glue at the top of the hole. I haven't tested this over a very long time period, but I haven't had any problems yet in more than 10 nights of use.

Once you have the pivotable nut assembly finished, lay the nut into position and mark on the wood where the screw will lie. You then need to create a groove in the wood for the screw. This groove should just barely be deep enough for the screw, and it's very important that you make the groove such that the head of the screw protrudes into the wood. This is what will prevent the assembly from moving linearly. If you have a small chisel, you can use it to carve out the groove, but you can actually make due with a screwdriver and a small piece of sandpaper to smooth things out if you are using a soft wood like pine.

Once the assembly is in place, secure it with a 4" mending plate. The plate should be wide enough to almost completely cover the screw, and once the plate is secured the assembly should only rotate with no other significant motion possible. Also, make sure the screw rotates freely, otherwise you risk popping the screw out of the hole drilled in the nut. Keep in mind that the nut doesn't have to pivot very far for the exposure times that are appropriate for this device.

Here is a close-up view of the pivotable nut assembly:

[11K JPEG]

Now we need to improve the contact point on the upper board. Use a large drill bit to drill out a "bowl" into a mending plate and when you attach this plate to the upper board, make sure this bowl is centered exactly 11 7/16" from the hinge. Round the end of the threaded rod with a file, making sure you smooth out at least the first thread. The rod should make a smooth contact when it is rotated, and should not slip back and forth. A couple of rubber bands (not too tight, though) around the opening of the barndoor will help insure a solid fit.

Here is a view of the "inside" of the barndoor, unfolded; note that the contact point on the upper board (on the left) is a 2" mending plate in this model, and not a 4" plate as described:

[11K JPEG]

We need a drum on the threaded rod so that we have a place to mark times as well as to facilitate rotating the screw by hand. Probably the lowest-tech route here is to use a 2 1/2" hole saw on your power drill. This is the one "unusual" thing on the parts list that you won't be using for much else (they're designed primarily for installing door knobs) and as such you might want to make a special effort to borrow this from someone. A larger disk is better, but you would probably need a drill press or lathe as I couldn't find a hole saw larger than 2 1/2" for a small hand drill. Carefully cut the disk, widen the center hole to 1/4", and then make the following stack on the threaded rod:

1/4"-20 nut
3/8" washer
Drum
3/8" washer
1/4"-20 nut

Tighten the nuts down as firmly as you can (this is where two crescent wrenches or two pairs of pliers help out). The washers should dig into the wood a bit, because the last thing you want is to turn the drum but not have the screw turn!

To use the mount, you need to turn the drum and screw by a discrete amount every several seconds. It's impossible to maintain a steady rotation by hand, and you don't need to do so. Some good rules of thumb are to turn the screw by 1/30th of a rotation once every 1/30th of a minute (2 seconds) for a 135mm lens, 1/15th of a rotation once every 1/15th of a minute (4 seconds) for a 50mm lens, and 1/10th of a rotation once every 1/10th of a minute (6 seconds) for a 28mm lens. The calculations page gives more information on the time tolerances for various lenses. The main idea is that you only want to adjust the screw often enough to keep the image sharp on the film. The more often you touch the mount, the more likely you are to bump it or otherwise cause vibration. With the wide-angle lenses, eventually you will get good enough that you can sneak a quick peek at the sky to look for meteors in between screw adjustments. These numbers are rather conservative, so if you accidentally miss an interval, immediately rotate it up to where it should be, and your photo should still be OK.

To make the time scale on the drum, you need to measure the exact circumference of the drum. You can do this by wrapping a piece of paper around the drum. I recommend marking the paper every 2 seconds, to match the intervals given in the previous paragraph. Thus, mark the paper every 1/30th of the total length and label each mark or every other mark. Tape the paper in place on the roll. You need to read the time scale relative to some fixed point, and a place to put a watch. You can kill two birds with one stone by using a curtain hanger. You can also come up with something else here, but the important thing is that the watch be attached to the drive near the drum so you can see both at the same time.

Here is a close-up view of the threaded rod, drum, and curtain hanger; note that the drum in this photo is about 3" across, not 2.4"; for sentimental reasons I use the 3" drum from my original barndoor drive from 1988. Note the markings every 2 seconds, and the numbers given every 4 seconds. Also note that the numbers increase from right to left as you operate the screw.

[13K JPEG]

Next, we will take care of the mounting of the swivel head camera mount. This should be mounted near the center of the top board. Center one of the 3/4" x 1" x 1" blocks near this center point and screw it to the top board of the barndoor. Then screw the other 1/2" x 1" x 1" block on top of this one. The idea here is that for a 2 1/2" screw, you need a total wood thickness of about 2". If you use a 2" screw you can eliminate the 1/2" thick piece, but the extra 1/2" will make it easier to see through the camera's viewfinder when the camera is trained on certain parts of the sky. Then, drill a 3/8" hole through the center of the two blocks, and all the way through the top board of the barndoor. Insert the 3/8"-16 cap screw through this hole, using a washer above and below. Verify that the swivel mount securly fits. If your swivel mount is some other size than I have specified, you will have to adjust this step accordingly. I would advise that you take the mount with you when you buy the cap screw so that you can verify the actual size and pitch.

The platform must be attached to a tripod. I'm assuming a tripod head with a 1/4"-20 screw. For this, locate the center of the bottom board and drill a 1/2" hole exactly deep enough to hold two 1/4"-20 nuts stacked together. Use cyanoacrylate glue to stack the nuts. This can be a little tricky because you can't let any glue get on the threads, and the nuts have to be accurately stacked or else the screw on the tripod head won't go far enough in. Insert the nuts into the 1/2" hole and use two 4" mending plates, overlapping the nuts just enough so that the nuts are held firmly, but the plates don't prevent a screw from being inserted.

You can also greatly shorten this step by buying a piece of hardware that you can pound or screw into the bottom of the board with a female 1/4"-20 thread.

Here is a close-up view of the bottom of the mount:

[13K JPEG]

Polar alignment is critical for excellent astrophotos. I'll give two possibilities for construction techniques that allow you to get a finderscope aligned with the axis of the hinges. The next section will detail a technique for the actual polar alignment.

Measure from the hinge axis to draw a line on top of the top board parallel to the hinges as accurately as possible. A square can make this easier. The exact distance you choose isn't very important, but all of the line must be the same distance from the hinge axis. Very carefully align the pre-cut edge of the 3/4" x 3 1/2" x 1-2" board to the line, making sure that you are accurate to better than 1/32" of an inch from one side to the other (this is easier than it sounds), and screw it down. The edge is then used to align the small finderscope, riflescope or a narrow tube. If you are exactly 1/32" in error relative to the hinge, you will be off by 0.5 degrees, which is a good minimum accuracy to shoot for. In my design, the edge of the board provides a resting point for a removable small finderscope. You may want to permanently mount the finder on the platform, but this is not necessary.

Here is a top view showing the finder support:

[10K JPEG]

To get even more accuracy, you may want to allow the finder support to rotate. Then you need to pick a distant object and sight it through the finder (you can use a bright star or a planet, but it's probably easier to do this technique on a daytime object). Watch the object through the finder as you manually unfold the barndoor. If the object remains in the center then you have really good alignment. If not, rotate the finder board a very small amount, adjust the tripod mount to compensate and try again. Continue iterating and when you are done, the finder should be very accurately aligned with the hinge axis. Once you reach this point, then you can add another screw to the mount so it can't rotate anymore.

And finally, a view looking south of the whole mount, similar to the view you will have while using it:

[11K JPEG]

Polar alignment and accuracy

Assuming you have measured the 11 7/16" distance correctly, 0.5 degree accuracy in polar alignment will allow for about a 16 minute exposure with a 50mm lens and about a 6 minute exposure with a 135mm lens, and this is a pretty good idea of the minimum limits of this setup.

There is more than one technique for alignment. The one I will describe depends upon having the finder position fixed and sighting on the North Celestial Pole (NCP). First, do not align with Polaris for polar alignment, as this will be about 3/4 of a degree off. The rule of thumb is that for every 10 arcminutes of polar alignment error, you can expect the tracking to be off by as much as 2.5 arcsec per minute of exposure. If you can get 15 arcminute polar alignment, which is feasible by using a direct sighting method through a finder, the natural tracking errors that occur in a single-arm drive will allow exposures of 20-25 minutes with a 50mm lens and about 10 minutes with a 135mm lens. Remember that 15 arcminute alignment does not mean 15 arcminutes relative to what you see through the finder, but 15 arcminutes relative to where the hinge pivot points, which will be two slightly different things even in the best case.

OK, so how do you get 15 arcminute alignment? First, the finder alignment must be as good as possible. Once you've set up the equipment in the field, then "think lambda". Lambda Ursa Minoris, that is. This +6.4 magnitude star lies a little more than a degree from Polaris. More importantly, the midpoint of a line between these two stars lies just 20 arcminutes from the NCP. If you imagine a perpendicular bisector of this line, then you travel along this line for about 1/4 of the distance between Lambda and Polaris in a direction that takes you to the concave side of the handle of the Little Dipper. Here's an ASCII diagram (I'll have a graphic for this sometime):

                                                  *
                                                 Zeta
              

  
                                   Epsilon
                                      *
        
                      Delta   
          NCP+   .      *
  Polaris*     Lambda

As long as you have cross-hairs on your finder, you should be able to get the finder centered within 10 arcminutes of the NCP, plus the error in the alignment of the finder with respect to the hinge. The only down side of this technique is that you need enough of a finder to be able to see Lambda. I can usually see this star naked-eye at my site so my finder with a 10mm clear aperture is fine. If your skies aren't so good, you may prefer a 30mm aperture.

If you have a spotting scope, or a short focal length refractor, you can mount the telescope to the platform and you might be able to modify the usual "drift alignment" procedure (do a Google search to get instructions) to the barndoor mount. This may be non-trivial, though.

One important aspect of barndoor drives is that unless you can get polar alignment better than about 15 arcminutes, a double-arm drive gives no significant advantage over a well constructed single-arm drive. With 15 arcminutes of polar alignment error, the absolute limit to exposure times with a 50mm lens on a target near the celestial equator is about 32 minutes. Double-arm drives can be built with sub-arcsecond natural errors during this time, but the natural errors in a perfectly-built single-arm drive for a 32 minute exposure will be only about 20 arcseconds, negligible compared to the 100+ arcsecond error due to polar alignment.

The following is a photo of the M44 area (near +20 declination) taken with my prototype for this design. This is a 6-minute exposure with a 135mm f/2.8 lens stopped down to f/4.

[PHOTO]

The tracking is essentially "perfect", despite a light breeze. I believe the design is capable of longer exposures at low declinations, but I haven't been able to rigorously test this drive on a calm night. For something like the Cygnus Milky Way near +45 declination, you should be able to go 1.4 times as long. However, as I keep emphasizing, you really need to nail your polar alignment. I found that during Hale-Bopp, 4 minute exposures with my 135mm lens gave consistent dead-on tracking unless it was windy. However, Hale-Bopp was bright enough that I really didn't need to go longer, thus the lack of testing of longer exposures.

In reality, errors in construction may place a more stringent limit on the total exposure time. This is where a motorized drive can help because the motor speed can be "tuned" to correct for this type of error. However, as with a double-arm drive, unless you simply want the construction challenge and the freedom from manually turning the screw, a motorized drive won't help unless you can get really good polar alignment.

I just want to add here that contrary to what some people say, it's not particularly difficult to use a manual drive and get really good results. Yes, manual tracking can be tedious on long exposures and while you are tracking you can't be observing the sky. This is just like manually-guided prime focus astrophotography. Part of the satisfaction comes from the fact that you did all the work yourself. Having said that, the next section gives some objective reasons for and against using more advanced designs including motorized designs.

Pros and cons to more complex drives

Here is my laundry list of pros and cons for various complex designs (that don't require a telescope) relative to a hand-powered single-arm drive.

Motorized single-arm drives

Pros

Cons

Hand-powered double-arm drives

Pros

Cons

Motorized double-arm drives

Pros

Cons

Motorized equatorial platform

Pros

Cons


Contact information

Back to A Wide-Field Astrophotography Page

File last modified: 08 January 2005

Valid HTML 4.01! Valid CSS!