Even after the machine shop flattened the most strongly warped segment, the rotation ring is still not flat. The ring continues to deform to a conical surface, though not as severely as before. Half the ring slopes upwards toward the center, while the opposite half slopes downwards. The distance from the highest point to the lowest point on the ring is five millimeters.
Why is this a problem?
The cogwheel of the dome drive motor bears against the cog rim on the rotation ring. Irregularities in the rotation ring causes the cog rim to rise four to five millimeters above the cogwheel for each quarter rotation of the dome. This may cause the cogwheel to slip on the cog rim, thereby stalling dome rotation and stripping the cog teeth.
How flat is the rotation ring?
To quantify the ring's flatness, height measurements were recorded with a caliper at two handles, - one at the location of the dome drive motor and one on the handle adjacent to the motor. Measurements were taken at thirty-two points evenly distributed along the circumference of the ring.
The first photo below shows the location of these two handles (drive motor removed).
The second figure below shows the results. The amplitude of variance is greatest (5 mm) at the handle to the right of the drive motor. One would expect the curves to be identical. This discrepancy may indicate that the ring is climbing the side roller. This can probably be alleviated by fastening the cog rim to the rotation ring at more points along its edge. However, this discrepancy is only minor. The overall trend clearly shows that the ring is warped most severely in segment 1. The neighboring segment 2 was originally very warped, but this segment was rolled flat in a machine press, as documented in the previous post.
Note that the peaks in the curves are consistently highest near the joints between segments. The four dovetailed joints are the most flexible areas on the ring, which may account for the higher amplitudes.
Segment 2 was delivered to the machine shop this morning for rolling flat in a machine press.
The machinist flattened the 10 mm thick sheet metal by applying six thousand kilograms of pressure across rollers 15 cm apart. Several passes and a lot of eyeballing made the quarter segment of the rotation ring flat.
The machinist reported that warping is caused by internal strain in the metal. The strain causes warping when the metal plate is cut into smaller pieces.
This is the ring segment that has delayed the dome assembly. It is bent 55 mm out of plane, which corresponds to over five degrees at both ends. As the dome was test-assembled in the factory, this 10 mm steel plate must have been bent during shipment to Norway, probably during on or off-loading.
This segment was bolted securely to the floor of a four-meter long pallet which also contained half of the fiberglass dome components. The three remaining segments were strapped to this floor-bolted segment. Perhaps strain on the pallet during lifting before or after transport bent the ring without bending the other steel segments strapped on top of it.
On a flat surface, this steel plate (which forms a quarter circle) assumes the shape of a cone. That is why the segment was warped in a conic form after mounting on the lower rotation ring.
A large machining firm will attempt to flatten the segment this week.
After a long hiatus (day job duties, family activities and holiday travel), the lower rotation ring is now in place.
Thirty-four 16 mm bolts cast in concrete hold the ring in place. The ring consists of four equal segments of steel plate. The dome is secured against wind by eight C-shaped assemblies of thick steel. These contain vertical and horizontal rollers. The horizontal rollers prevent lateral movement of the rotation ring.
Before bolting, the diameter of the ring was adjusted to 350 cm, with a margin of error of only one millimeter in all directions. This high accuracy is ensures that the upper ring will rotate freely, without being squeezed by the horizontal side rollers.
The next step is to mount the upper rotation ring.
After several hours of angle grinder work, the fine adjustments of the observatory crown are finally complete. The surface is very level with very few irregularities. Two hundred point measurements with a laser indicate a surface accuracy of 0.5 mm.
The next operation is to bolt the lower rotation ring to the crown. The ring is made of four segments. These must be adjusted such that the ring is circular to millimeter accuracy.
The lower half of the telescope pier’s foundation is visible through the entrance. This is cast directly on the basement rock, approximately forty centimeters below the surface of the gravel. The upper, pyramid-shaped part of the pier’s foundation will be cast on top of this after the observatory dome is in place.
Click the image to enlarge.
The laser shows a level surface on the crown. Some small adjustments with a diamond wheel and angle grinder were needed to remove some slight banking in a few places. These irregularities were caused by the self-levelling screed that cured far too quickly on a warm summer day in July. If I were to make the crown again, I would have made the top edge of the crown's form level, and used this as the reference surface for the trowel. Having done that would have made the screed unnecessary.
Click image to enlarge.
To partially conceal the observatory, and also to prevent the occasional escape of cows and sheep, I am building a wooden fence along the border to the pasture. Because of the fence, there has been a hiatus in observatory construction. Galvanized steel rods welded to flat iron serve as anchors for the wood columns. This is then planked vertically.
I underestimated the minimum amount of fiberglass-reinforced, self-levelling screed necessary to fill the irregularities in the concrete crown. I used 50 kg of screed, which produces about 24 liters of product. This was far too little, - half as much as I should have used. In order for screed to flow to a level surface around the ring, the fluid mass must be at least 2 cm deep, preferably 3 cm. I will try again, this time using 100 kg of screed (48 liters). Most of the bolts are long enough to allow for this extra thickness. In those few places where the bolts are not be long enough, I will drill new holes for expansion bolts.
The crown will be about 9 or 10 cm thick. The lower 8 centimeters is steel-reinforced concrete. This is capped by a 1 cm layer of self-levelling poured screed. This is a non-viscous, fiber-reinforced cement that flows to a level surface.
It is a simple task to make the formwork for the crown. Sub-roofing panels (commonly used underneath ceramic roofing tiles) are flexible, thin fiberboard that are waterproof on one side. These are easily fastened to the expanded clay aggregate blocks with wood screws drilled through wooden stiffening ribs (see photos below).
The telescope will look like this
This is Freiwald observatory in Upper Austria. The instrument is a 60 centimeter Cassegrain which is practically identical to the instrument in Norway, both the truss tube (background) and the mount (foreground).
I 2010, I visited the production hall of Astrosysteme Austria (ASA) in Freistadt, Austria. I was also present during the installation and testing of the fabulous ASA DDM 160 direct-drive mount.
I happily received my first telescope when I was six years old. It was a very small newtonian with an eight centimeter primary mirror. As the years passed, my telescopes became larger and better quality. When I was nineteen, I worked part-time at the SFA university observatory in Texas. The main instrument was a 46 centimeter (18 inch) Cassegrain telescope. NASA used this instrument in the early 1960's for infrared mapping of dust on the moon. The result of these studies were used to determine suitable landing sites for the Apollo landers.
For many years, I planned a telecope that was larger than the 46 centimeter university instrument. These plans will soon be realized. It will be a 62 centimeter diameter cassegrain. That is 6006 times as large as the surface area of my first telescope :-D