40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
  OMC elog  Not logged in ELOG logo
Message ID: 116     Entry time: Thu Apr 18 11:43:59 2013
Author: Koji 
Type: General 
Category: General 
Subject: Mounting Glass Bracket Failure 

[Jeff, Koji]

- While we were working on the optics alignment, one of the mounting brackets made of glass god tore apart into two when a holding screw was removed.
The glass component had a crack at the very middle of the part.

- We borrowed a setup for photoelastisity measurement from Garilynn. This is a set of polarizer configured to have cross polarization. If there is no photoerastisity
the image is colored in blue (somehow). When the polarization is rotated, the color is changed in red, yellow, or white.

- The cross polarizer was tested with a polycarbonate face shield for the UV protection. It seems doing its job.

- We took a set of photos to see any residual stress in a block. The entire inside of the channel is frosted glass so the technique didn't yield much.
In one orientation we did see stress near the ends but the orientation didn't allow us to see exactly where.

- We had 30 brackets and one OMC requires ten of them. This means that there was no spare and now we don't have enough.
  So we decided to spend more as test pieces.

- We tested three scenarios this afternoon. In all three cases both screws were snugged (estimate 0.5 in*lb) before torquing by a torque wrench with a dial meter.
  The divisions on the dial of the wrench are 1 in*lb. We were not so confident in the exact measurement but we felt good about the repeatability of the values.

  1. Duplicated the original mounting with the chamfers of the PEEK bar facing into the channel. Cracked as the torque wrench read 1 in*lb.
      Crack initiation at the first screw, starting along the longitudinal centerline.

  2. Turned the nut bar over so the flat side faced into the channel. Successfully torqued both screws to 1 in*lb and removed them.

  3. With a razor blade, made fairly large reliefs (countersinks) around the holes in the PEEK. Successfully torqued both screws first to 1 in*lb then 1.5 in*lb.
      The block did crack (again at the screw along the centerline) when the torque was ~1.9 in*lb.

It occurs to us that we need micro-compliance AND structural rigidity to distribute the load. The PEEK bars are small and particularly thin where the #4-40 helicoil holes are.
The load is probably concentrated way too much at the holes because it is too weak. Perhaps a good solution, among others, would be to use an aluminum nut plate
with a thin (.02") kapton or viton layer to give the micro-compliance. Additionally, a kapton layer could be used between the block and the aluminum shim,
though this one is probably to be avoided so as to ensure rigidity of the bolted assembly to the bench. Lastly, the nut bar should be shaped such that the area
around the holes and the end of the channel (pretty much the same area) are less stressed than the center portion. 

After the discussion with Peter and Dennis, we decided to reinforce the bonded glass piece with invar shims.
Each shim will be threaded such that we don't need to stress the glass piece any more. EP30-2 will be used as the glue.

ELOG V3.1.3-