Tools and materials :

• Aluminum disk, 10.25" diameter, 3/8" thick, commercial grade
• Aluminum bar, 1/2" thick, 1" deep, 6" long, commercial grade
• 3 chrome ball bearings, 5/32" diameter
• 3 10-32 set screws, 3/4" long, cupped end
• 1 1" long 10-32 capped set screw
• 3 brass knurled nuts to fit set screws
• Mitutoyo digital dial indicator, 1/2" travel, 1 micron resolution. #543-252 Model ID-C112E
• 1 drill bit for metal, 3/8" diameter
• 1 drill bit for metal, ???" diameter
• Basic tap and die set. (Costs about $40 at Home Depot)
• Metal protractor. (Probably about $10 at Home Depot).
• Compass (for drawing circles)
• 1 tube of Crazy Glue (Dave Rowe suggests JB Weld)

The aluminum disk cost $50 and was purchased at a metal fabricator. Thealuminum bar was scrap that I was given for free.

The dial indicator was ordered from Enco and cost a bit more than $200.The ball bearings, set screws, knurled nuts were found at Home Depot, Sears Hardware, or Tru Value Hardware and each cost about $0.40.

Steps

I basically followed the directions given by Dave Rowe on his web page.See

Frankly, this is all you'll ever need to know about the subject.

The first step was to draw concentric circles for each of the diameters I envisioned.The protractor was used to identify the 120 degree points and these were marked off on eachcircle with a pencil.

Next,

• What I Did : Since the shaft of the dial indicator is 3/8" in diameter, a 3/8" hole was drilled in the center of the disk using a drill bit designed for metal. Initially, a hand drill was used to drill this hole. I later borrowed a friends drill press for the 3/8" hole in the scrap aluminum bar upon which the dial indicator rests. The hole forthe indicator's set screw was also drilled at this point. A couple ofdrops of Crazy Glue were put on the base plate and the aluminum bar was aligned with thebase plate. This needed to be carefully done, as the shaft of the indicator was slightlylonger than the height of the bar and as a result protruded into the base plate slightly.
• What I Should Have Done : Drill and tap a hole for the indicator set screw into the barstock. Glue the bar stock onto the base plate, then use a drill press to drill one 3/8"hole through both. It ended up not making any difference, but I could tell there was aslight difference in the angle each hole was drilled at. I also noticed that one side ofthe bar stock seemed to have a slightly larger hole side than the other. I attributed thisto a "wobble" in the drill bit, but it had not effect on alignment.

For what it is worth, if I had had a drill guide for a handheld drill, I don't believe a drill press would have been necessary.

Still using the drill press, holes were drilled for the "legs." Threads were then tapped in these holes. As aluminum is somewhat soft, this was not hard to do at all.

A handheld Dremel was then used to debur the edges of all the holes drilled.

Ball bearings were then "Crazy Glued" to the cupped ends of the set screws andthese were threaded through the appropriate holes. The knurled nuts were then threadedonto the set screws from above the plate, firmly fixing their positions.These adjustable legs can be repostioned in other holes to allow for otherdiameters to be tested. Their depth is also adjustable to allow fordifferent sagittas (both concave and convex) to be measured. The drawbackis that the distances between the legs needs to be re-measured each time theyare moved. I thought this was a relatively small price to pay,considering the cost of getting different aluminum plates made.

The dial indicator is held firm by the longer set screw. Tightening withone's fingers seems good enough. The block it rests upon is also "CrazyGlued" to the base plate.

I have not yet tested this unit, but expect that it will work well. If I were to do itagain, I would probably use something like the "JB Weld" mentioned above. It only costsabout $4.00 in a local Sears Hardware. If the Crazy Glue fails for some reason then thatwill be what I turn to.One other thing I've noticed is a slight wobble to each of the bearings as their set screwsare turned. I am not sure if this is due to the ends of the set screws being at slightangles, or if the balls were not fully seated. In any event, it will reinforce the needfor accurate measurements between the legs once a diameter is chosen and the legs aretightened down with the brass knobs.