Return to Home Index
Observatory Design Index

Preliminary discussion of Winch Design (still under construction)

The first part of the discussion is directed to a winch design that uses a standard pulley type winch with a single pulley. A later section discusses the two drum winch for two way operation and the final section describes the rather more complex design required for a continuous two way winch design that would be used for a dome that had to be rotated continuously in either direction. Finally, there is a brief discussion about pulling and pushing both roll off structures and dome type structures and the roller arrangements that affect stability of the movement of these structures. Thanks to Chris Vedeler for the very nice Figures.

Part 1. The two-way winch with one driving drum and limited travel

In Figure 1 the problem with a simple basic winch drive is shown. For a winch to work, there must be tension in the output side of the cable. The winch amplifies this tension, called force 2 (F2), as it rotates by cinching the cable around the winch pulley. The amount of amplification you get depends in a complicated way on the number of turns of the cable around the winch and the coefficient of friction between the cable and the winch pulley.

Generally, with two turns around the winch pulley, you can get an amplification of 100 quite easily. The problem with the simple design is that when you turn the winch CCW, in the case shown, F1 needs to be large to pull the load and this tension in the cable stretches the cable. enough that the cable on the left side becomes slack and the force F2 goes to zero or nearly so. Then the winch action no longer works, the winch slips and the force in the right side gets small enough so it can not pull the load. No amount of tension in the cable will help since the cable will always stretch a bit and the winch will slip. Also, any tension in the left side of the cable adds to the load that the right side has to pull. Believe me this will not work. In order to minimize cable stretch, it is essential to use steel cable. I used a plastic covered steel cable for the dome rotation winch on the 12 foot MAS dome which houses our 16" CAS.

To fix this problem we must add, on each side of the load, a tension loader as shown in Figure 2. The action of this tension loader is to maintain a tension in the side of the cable that tries to go slack. A detail of the tension loader is shown in Figure 3. The device, while simple, has several important properties. It consists of a spring which is normally slightly stretched so as to maintain tension in the cables. But you would find that the spring which is just right to maintain tension in the cable will stretch too much on the high force side and make the whole cable system again go limp and the winch will slip. Note that a tension loader is required on each side of the load and must have the following characteristics..

What you have to do is make the tension loader very non linear. This is done by adding a chain across the tension loader spring which is about 150 to 200% the length of the spring. Then what happens is that the spring on the high force side stretches until the chain takes over and the right hand side gets taught and pulls the load. At the same time, the spring on the left hand side is still stretched enough to keep some tension in the left hand side to make the winch action work. In the opposite direction the other tension loader functions the same way.

Some experimentation and balancing of the spring constants and chain lengths is necessary. But when set up properly, this design works very well. You have to have a spring that is strong enough to maintain the necessary tension and still has enough extension to absorb the stretch in the cable. Such springs are available in hardware stores for a few dollars. You also have to keep the chain short enough so that the spring does not permanently deform.

For the dome drive I used springs with about 60 pounds per inch spring constant and stretched them about 2 inches under no load conditions. I made the chains so that the maximum stretch was about 3 inches. You will have to play with these variables to match your load.

I might point out that another solution for this particular drive is that a chain could be used. A chain, captured in a chain sprocket, will not slip even if the low tension side goes limp. This might be a better solution for the roof drive shown. The only problem is the chain typically is several dollars per foot. You can get such chain and chain gear drives commercially but you will find them very costly.

fig1
fig2
Fig 3

Part 2. The two-way dual drum winch with limited travel

For the case discussed above, where the motion of the moving part of the building, roof or dome, is limited, it is probably easier to go to a two drum winch. This design is shown in Figure 4. Here each drum has a cable that is the full length required to move the movable member of the structure as far as necessary. One drum takes up the cable while the other plays it out and vice versa. This is a very simple design which needs no additional springs and the like. The only problem with it is that it requires a two drum winch. These are less commonly available and may have to be devised by replacing a single drum with a modified drum arrangement. It would also be possible to use two winches in opposition. They could be synchronized by placing the motors in parallel electrically.

Fig 4

Assurance that the moving structure will move smoothly is related to the way the forces are applied to the structure and the details of the design of the rollers and their placement. These issues are discussed in a following section.

Part 3. The two-way winch design with unlimited travel

The most trick of all the possible designs is the two way winch with unlimited travel. This would be the design desirable for a dome which is to be rotated round and round without requiring a stopping point and reversal of the motion.

A design for this type of winch is shown in Figure 5. I have just completed this type of winch for a 12 foot dome with excellent results. Note that in this case, the spring tensioning mechanism must be symmetrically placed on each side of the winch. The design depends on getting enough cinching effect on the winch to supply amplification of force through the winch and also sufficient cinching effect on the dome to pull the dome. One problem with turning a dome in this fashion is that the pulling force necessary to turn the dome is all at one point. This tends to want to pull the dome right off of its rollers. The lateral force must be taken up by well placed lateral rollers. This issue is discussed in a later section of this article.

Fig 5

This arrangement, which seems complex at first glance, is quite simple to understand. The design is and must be perfectly symmetrical if it is to run in either direction without limit on the motion. In practice, the cable is run twice around the dome so as to get enough cinching effect. The cable is run two to three times around the winch pulley also to get enough cinching effect and thus enough amplification of the in-going tension as compared to the out-going tension.

Since the spring/chain tensioners cannot be in line with the cable, they are but effectively in line using two pulleys. As shown, the upper pulleys transfer the winch mechanisms forces at right angles to the dome. It takes about 200 pounds force to move the dome in the particular example I have designed. The winch supplies an amplification of about 10 times. This means the springs have to have a spring constant something in the order of 50 pounds per inch. I found some replacement springs for a hobby horse at the local hardware store that were just right when I put two in parallel. The chain allows for about 3 inches of extension of the springs. Some playing around with the spring strength and extension limit was required to get this just right for the application. I broke two springs before I got it right. Note that the force in the springs is twice that in the cable.

The whole mechanism works very nicely and is sort of interesting to watch as the dome is juggled back and forth. I will put up a photograph of this mechanism as soon as I can get one ready. (Figure 6 photo not yet ready)

Part 4. The motion of pulled and pushed structures

The motion of pushed and pulled structures which are on rollers is very interesting. When I build my first roll off structure, it was the one shown on my home page, I was concerned about the forces required to move it and how well it would stay centered on the rails. The building rolls at its base and is quite heavy. To keep it moving straight, I decided to use V groove wheels and have them ride on a single slender rail. Four 8 inch wheels, which are steel sheaves, are used on each side. They are mounted on hardened steel drill rod and lubricated with heavy pump grease. The rail is an L iron with mounting brackets welded to the sides. The design is self clearing for any ice that might build up in winter. It takes nearly 100 lb force to get the building rolling from a dead stop. It takes only about 50 pounds force to keep it rolling once it has started to move. This is due to the static stiction that the roller system has.

It was my original intention to install a motorized mechanism to move the building. This has not been done since the building is so easy to move by hand. In fact the design is so stable that one can easily push on one side of the building only and it does not skew enough to cause any binding. This is largely because the building is very well built. It was done by a professional contractor. Others, especially with larger structures, like large roll off roofs may not be so fortunate. If the structure tends to flex and especially to skew, it will likely be necessary to apply the force required to roll the roof in a nice symmetrical fashion. All of these issues depend greatly on the resistance of the roller system and the tendency of the structure to skew. In the worst case unacceptable binding will take place when force is applied to one side of the roof. In this case, the lateral integrity of the roof should be checked to see if cross bracing or diagonal bracing can be used to stabilize the roof to keep it from skewing.

In some cases the application of lateral rollers will help keep the roof symmetrical and relieve binding of the vertical rollers or wheels. If that does not work or is not possible the point of application of the force will have to be changed to be more symmetrical. In the worst case, with a rather flimsy roof, it may be necessary to apply the force at two points, one at each side of the roof and in line with the roller mechanism. Each case is unique. With some roofs and shutters, a rather complex system of pulleys and cables is necessary to insure smooth movement of the parts. For example, both Ash and Home domes have elegantly designed mechanisms.

This brings us to the mechanisms for rotating domes and how domes must be stabilized with rollers to insure that they move smoothly. Domes are almost always on rollers. tiny 3 inch rollers are fine for domes up to 10 feet but may be much larger. The dome at Yerkes, 90 feet in diameter, runs on flanged railroad car wheels. Rather appropriate since Yerkes was a railroad transportation mogul.

In the case of non flanged rollers, it is clear that the dome would skitter right off of the rollers if it were not restrained by lateral rollers. The forces used to rotate the dome are applied in various ways, but all tend to push the dome off the rollers. In the case described above, the force is applied to one edge of the dome and would pull the dome right off but for the lateral rollers that keep it centered. Thus, the placement and quality of the lateral rollers should not be under estimated. In the case of the 12 foot dome I have worked on, the use of 16 vertical and 16 lateral rollers very greatly improved the smoothness of the motion over the previous setup which used 8 of each. The method of moving this dome causes very asymmetrical force on the dome. Still even very sophisticated domes, like the Ash dome use only one motor and apply force to the circumference of the dome in one place. In that case the motor is on the inside and uses a perforated track and cogged wheel to move the dome. In the case of the Home Dome two motors are used at opposite sides of the dome. This applies a much more symmetrical force to the dome and moves it more easily. However, the Home Dome drive is slightly less positive since the drives are friction drives instead of a positively engaging cog wheel drive. In larger domes, there are generally several/many motors that work around the entire rim of the dome to move it smoothly. The issues are reliability and cost as with most designs.

Return to Beginning

Go to Home Index for Doc G's Info Site