Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.apo.nmsu.edu/Telescopes/SDSS/eng.papers/19930517_ferrule_wear/19930517_ferrule_wear.html
Дата изменения: Tue May 19 04:55:56 1998
Дата индексирования: Sun Apr 10 08:44:28 2016
Кодировка:
Wear of a Stainless Steel Ferrule

Wear of a Stainless Steel Ferrule

Sloan Digital Sky Survey Telescope Technical Note 19930517

Russell Owen

Introduction

Fibers are retained in the plug plates using high precision ferrules. Each ferrule will be used to plug approximately 300 holes. We must be sure the ferrules will not wear excessively during use; 1 mil reduction in diameter of a ferrule is enough to reduce light throughput by approximately 1%.

Ferrules are typically made of either stainless steel or ceramic. Ceramic ferrules are very high precision and hard enough that wear is not a concern. However, they are more expensive than stainless steel and may crack, resulting in a defect that could be very difficult to find yet cause erratic loss of throughput. Also, ceramic ferrules may be difficult to find with the required fiber hole size. Stainless steel ferrules offer lower precision (which will lose us approximately 1/2 to 1% light compared to ceramic), but they are relatively inexpensive, readily available with the hole diameter we require, and immune to cracking. The big question about stainless steel ferrules is whether they are hard enough to stand up to repeated plugging; the tests reported on below indicate that the answer is "yes".

Measurements

One ferrule made of 303 stainless steel was tested. It was a high precision Ofti SMA 906 ferrule with a few low-precision modifications, exactly the kind proposed for the project. The diameter of the tip of the ferrule was measured using a micrometer. Then the ferrule was plugged into holes in aluminum test plates (the plates used were from the 3/93 drill tests). The diameter of the ferrule was re-measured after each plate of 50 holes was plugged. I applied some torque to the ferrule during each insertion and removal to simulate the torque provided by the stiff jacketing material that will be attached to the ferrule (this torque is necessary to hold the ferrule in the hole). I applied the torque approximately uniformly to all sides of the ferrule (on the average over all holes). In the real world the torque may be concentrated on one side of the ferrule if the jacketing is significantly stiff in torsion. If so, the results reported here are optimistic.

Results

It is interesting to note that the ferrule was inserted into each hole with no difficulty. The nominal clearance of the ferrule was 2 mils. This value, derived from the 1/92 drilling tests, was chosen to be as small as possible and still have clearance for 99.9% of the holes. However, the holes in these plates were drilled to much tighter tolerances than those of the 1/92 drilling tests, so it is not surprising that there was ample, and likely excessive, clearance on all holes.

The ferrule and holes were inspected under a 35 power stereo microscope. The spiral grooves left by the tools were clearly visible, but both the ferrule and all inspected holes looked excellentЁvery round and clean. After the plugging operation the ferrule was visibly worn; but the tooling marks were still quite visible under the wear marks. Hence the wear was not very deep. The ferrule also had many new fine axial scratches, but so fine and so widely distributed as to have no possible effect. The holes (each of which had been plugged twice by the ferrule) had one or two narrow roughly axial bands of wear, but they covered less than 1/10 of the area of the hole and appeared to be so shallow as to be merely cosmetic.

Figure 1 shows the change in diameter of the ferrule as a function of the number of holes plugged. The results agree with the visual inspection: the ferrule exhibits some wear, but not enough to cause concern. Each ferrule will be used to plug approximately 300 holes, at which point the ferrule's diameter will be reduced by approximately 0.06 mils. This translates to a tilt of 0.5 mrad, which will have no measurable effect on throughput. Extrapolating the data, we would have to plug well over 5000 holes before reaching 1 mil of wear.

These tests show that stainless steel ferrules should serve us quite well, and that using aluminum for plug plates is also an excellent choice.

Figure 1: Change in Ferrule Diameter vs. Number of Holes Plugged