Window Material Testing

Standard models used in predicting vacuum window performance do not predict the actual performance of Kevlar laminate fabrics accurately [2,4]. Therefore, extensive tests of the Mylar/Kevlar composite windows have been done and more are planned. A summary of pressure tests on windows composed of the custom laminated fabric currently used in construction of all Hall C vacuum windows is given in Tables 2.6 and 2.7, below. Both vacuum and hydrostatic test tanks are available for the HMS circular windows and for the SOS rectangular window. Figure 2.21 shows a picture of the SOS vacuum test tank. The large round HMS vacuum test tank uses the inch vacuum extension piece shown earlier in Figure 2.19 (but removed from the spectrometer) with a inch thick aluminum blanking flange. For hydrostatic testing, windows (both round and rectangular) are mounted directly on blanking flanges with appropriate water fittings installed.

Table 2.6: Tests on the Hall C Vacuum Windows (1 of 2)

Table 2.7: Tests on the Hall C Vacuum Windows (2 of 2)

The tests show that the large HMS exit window does not begin to leak until the pressure on it reaches over atmospheres. The small HMS window did not begin to leak until over atmospheres of pressure was applied. This is consistent with the load being transferred to the outer circumference, the circumference of the small circle being four times smaller than that of the large circle. The large rectangular SOS windows began to leak at around 10 atmospheres. It may be possible to construct thinner SOS vacuum windows for improved spectrometer resolution if required.

In addition to determining the absolute failure pressure of the windows, a major goal of the testing effort was to observe the failure modes of the window material and flange design. The large windows failed by ripping around the opening perimeter. The flange ring was then rounded more and a better result was obtained. It may be possible to further round the flange if future tests indicate this to be necessary or advantageous. The small windows failed by ripping out at the bolt holes.

Another question the tests in Table 2.6 were designed to answer was if a biting aluminum clamping O-ring was necessary or advantageous. It was found that windows with aluminum clamping rings failed at lower pressures than windows without aluminum rings. The aluminum tended to tear the Mylar, allowing leaking and uneven stress. The only window which delaminated had water between the Mylar and Kevlar which leaked in through a tear in the Mylar caused by the aluminum ring.

In conjunction with outlined full-size window testing, small samples were subjected to stress and tensile analysis using strain gauge instrumentation. Through this work, it was found that the composite window material maintained the Young's Modules of Kevlar, and had the same ultinate failure load as Kevlar pieces of the same size. Further, it was determined that the failure tears were along a 45 angle to the woven fibers, consistent with observations of the rips in the destructively-tested full-size windows.

In addition to the above testing efforts, two large round HMS exit windows composed of the custom laminate material were ``knife" tested. The windows were installed on the vacuum test tank and placed under vacuum. They were then cut by a blade at the end of a long pole (so as to protect the personnel testing the window from the results of a catastrophic failure). The window did not rip any further than the cut and the material did not delaminate. No evidence for catastrophic failure was observed in either of these tests.

Although more long term reliability testing is necessary (see below), results have so far been favorable. Windows previously used on the HMS spectrometer are routinely hydro tested to failure after biannual removeable and found to hold up as well as freshly made windows (see, for example Table 2.6). Windows which have undergone multiple cycles fail at nearly the same pressure as uncycled windows (see, for example, Table 2.7). A small round window did not fail after being left pressurized for a week at psi. (this test is not entered in the tables). Entrance and exit windows composed of the Mylar/Kevlar laminate material here described have held vacuum on the Hall C spectrometer effectively and safely for over three years.