CLEO III Beam Pipe
Beam Pipe Picture

The Technical Specification for the CLEO III beam pipe. It was written by Jeff Cherwinka and David Cinabro. This document provided the basis on which vendors bid to manufacture the beam pipe.

Drawings of the CLEO III Beam Pipe.

  1. Full Assembly: 6058-25. Updated to revision B to reflect changes listed below.
  2. Beryllium Sub-Assembly: 6058-26. Updated to revision C to reflect the changes listed below.
  3. Copper Sub-Assembly: 6058-27. Updated to revision B to reflect the changes listed below.
  4. Copper Brazement: 6058-28. Updated to revision B to reflect changes listed below.
  5. Inner Central Skin: 6058-29
  6. Outer Central Skin: 6058-30. Updated to revision C to reflect minor drawing changes and the change to a rounded from a square stop on the ends.
  7. Central Manifold Ends: 6058-31. Updated to revision C to be more clear about what the part should look like before the final machining by the Beryllium manufactuer and to add a rounded corner to join to the stop of the outer central skin.
  8. End Collar: 6058-32.
  9. Updated to revision B to reflect the as built collar to match to the inner central skin.
  10. Inner Central Skin/End Collar Sub-Assembly: 6058-40

Related Drawings.

  1. Outer Magic Flange Jacket: 6058-33
  2. Inner Magic Flange Section: 6058-34.
  3. Updated to revision B to reflect tighter inner radius.
  4. SS/CU Transition (Test Version): 6058-35
  5. E-Beam Backup Disk: 6058-38. Updated to revision B to reflect added hole.

Vacuum Test of the Mechanical Strength of the Inner Central Skin

Here at Wayne State we made in aluminum a model of the Inner Central Skin. The idea is that aluminum has about a fourth of the strength of beryllium and if the aluminum model of the tube does not collapse when there is vacuum inside it is very unlikely that the real beryllium tube will collapse. We did this with an older version of the drawing which did not include the stops at the ends. The procedure was to start with 2 inch diameter aluminum bar stock which had a central hole drilled out at a local job shop. Center caps were made to fit inside the tube and this was placed on the lathe. This was then machined down to have walls of nominal thickness of 0.040 inches. All the wall thickness measures were done with Hitoshi Yamamoto's tube thickness measurement device. Shim stock was then introduced between the tube and caps to adjust for errors in either the centers of the caps or in the lathe. The tube was then brought down to have a wall thickness of 0.030 inches. Measurements of the tube found all the walls within in the (0.029 + 0.002 - 0.000) inch specification except at one spot near the center where it was 0.0285 inches and a diameter within 0.001 inches of 1.710 inches. A Delrin resin mandrel was made and inserted in the tube to support it during the next operation. We originally tried a mandrel of Woods Metal, but this did not completely fill the tube leaving some gaps and was so soft that it did not add much to the strength to the strength of the tube. Also a new set of center caps was made.

The tube was then moved to a 3 axis CNC mill. The tube was mounted on a rotating index head on the bed of the mill to give us 4 axis functionality. A program was written to carve the 0.016 inch deep "pockets" in tube. This was tested on some scrap and then run on the real tube with the depth of the pocket set to zero. The pockets were carved in the order as shown in this little diagram. Pocket Carving Order Diagram Our first try at this we simply went in order clockwise around the tube. After the fourth pocket the tube lifted up off of the Woods Metal mandrel and during the carving of the fifth pocket the tool broke through the tube. Three things caused this to happen. 1) The tube's inner diameter was not constant. This led us to go out to a job shop to drill the center hole. The pocket that failed probably was 0.006 inches thick from measurements we did of the nearby region after the disaster. 2) The Woods Metal mandrel was soft and had gaps and thus did not add much to the strength of the thin tube. We moved to the Delrin mandrel. 3) Pockets carved in order caused the tube to be forced out on the edges of the region of carved pocket and forced in at the center of the region of carved pockets and the center of the uncarved region. We went to the back-to-back order for carving pockets which minimized this effect.

After carving the pockets and giving the whole thing a good polishing we measured the thickness of the pocket walls. The table below shows the results. The two ends of the tube were distinguished by where the number labels were painted. The numbers went in order clockwise around the numbered end of the tube.

Thickness of Pocket Walls (in 0.001 in)

Pocket Number Number End Middle Unnumbered End
1 12.5 11.5 14.0
2 12.0 14.5 13.5
3 13.0 10.0 13.5
4 13.5 15.5 15.0
5 13.5 15.0 16.0
6 13.5 14.0 12.5
7 13.5 13.5 13.5
8 13.5 12.5 14.0
This is not so bad. The specification from the drawing is 13 + 2 - 0. Pocket 3 is the only one that was bad for much of its length, but it is on the thin side and thus does not spoil a mechanical strength test. We also checked the ribs and found that they were all within the nominal thickness of 0.029 + 0.002 - 0.000 inches. The pockets had the nominal width (actually a little wider than the specification) and the 4 of them had the nominal length. The other 4 were 0.040 inches longer than nominal. This was half the width of the carving tool, but how the CNC mill "forgot" about the tool width for some of the pockets and not the others is a mystery. Finally the tube on its mandrel was put back on the lathe and cut to length with a parting tool.

We then made end caps with grooves for o-rings to seal the ends. One of the caps led to a T-coupling with one port for a vacuum gauge and the other for a vacuum pump line. These and the completed tube are shown in these pictures. (Click on them for big versions.)
Tube + End caps View 1 Tube + End caps View 2

These where then assembled, liberally coated with vacuum grease and hooked up to a vacuum pump. Here are three views of the tube under vacuum. You can see the numbers that we painted on the pockets in the picture on the right.
Under Vacuum View 1 Under Vacuum View 2 Under Vacuum View 3
Unfortunately you can not read the vacuum gauge on these pictures, but it got to 15 miliTorr and was kept there over night. Note that the tube did not collapse demonstrating that we have at least a safety margin of four (aluminum has a modulus of 10x10**6 psi and beryllium has a modulus of 44x10**6 psi) for the real beryllium tube. Jeff Cherwinka has done some ANSYS modeling of this tube and based on that work thinks the safety margin may be more like seven. We may use some of this margin up by running a coolant at pressure in the gap between the two tubes, although in CLEO II we suck the coolant through and have less than atmospheric pressure on the outside of the inner tube.

Last we have the proud parents. James Roy Barlow (aka Roy) did all the machining, Dave Cinabro oversaw the work, and advice was offered by Jeff Cherwinka, Anton Sternad, and various others. Roy and Setup Dave and Setup