Mr. McKinley: Thank you. I think, we can all join in genuine admiration of Mr. Gjelsvik's ability to speak to us in our own language. We have about 15 minutes before lunch is due and I am sure that there are many questions you would like to present to Mr. Gjelsvik. Yes sir?

Participant: I would like to know if the report that you referred to several times, your No. 44 report, is available in English?

Mr. Gjelsvik: We have some of our reports available in English and in English only. The numbers I have referred to are Reports Nos. 33, 44, and 48 and Reprint No. 145, and all of them are available in English.

Mr. McKinley: Very good. I am sure that Mr. Gjelsvik will be glad to give you detailed references later on. Yes sir?

Participant: I would like to know what types of units were tested in your programs, and I would also like to know what drying agents were used.

T. Gjelsvik: The units I have referred to and that we have tested during the period of years have not only been manufactured in Norway. In fact, only a few of them were manufactured there. Most of them came from abroad, from Denmark, Sweden, Germany, France, or England and we also had one set from the United States and one from Canada. The manufacturing methods are completely different. We have, I would say, all types of units. We have had the all glass units made in the United States, we have had units with the glass to metal seal made in different countries, and we have had a fairly large number of units with the edge seal made with completely different types of sealing materials and different types of space and edge construction. The desiccants have partly been silica gel and partly activated alumina.

Mr. McKinley: Yes sir, Mr. O'Shaughnessy.

R. O'Shaughnessy: You have evidently been taking the opposite method, not method but procedure, in using large units versus the small units we use. Have you in your tests correlated some difference in the test results utilizing the same test with the different size units?

ex

T. Gjelsvik: Our test method is very strong, at least partly, on simulating gusty wind pressures, and applying variable wind pressures on tremely small units would be really nonsense. So we have generally tested the fairly large units. We have not only tested the size mentioned but also some smaller sizes, but most of the units have been the large size. We have field experience with different sizes of units and, in general, we have found a reasonably good correlation. But we also found that the very small units are weaker and much more likely to fail than the larger sizes in practice.

Mr. McKinley: Yes?

R. Fentress: When you have your wind gust test as part of your test, how long a duration of the different gusts do you put on?

T. Gjelsvik: Well, as I said before, we are using five wind gusts a minute. The units are subjected to the pressure pulsations throughout the 24 hours of the day, except for the few hours when units. are being cooled to 14 °F on one side.

R. Fentress: The peak wind load is sustained how long?

T. Gjelsvik: The top of the peak lasts only for a part of a second. The shape of the wind gust is approximately sinusoidal.

Mr. McKinley: Mr. Robinson?

Mr. Robinson: Mr. Gjelsvik, in view of the extensive field investigations that you have made, I would like to know if you found any correlation or difference in the failure or breakage rate of windows that face north and have little sun upon them as compared to those that face the sun?

T. Gjelsvik: We tried to find if there was any real difference, but we were not able to prove any difference.

Mr. McKinley: Yes sir, Mr. Beatty?

J. A. Beatty: In view of the requirement of the Canadian specification for 32 oz. glass, do you specify one particular thickness of glass to make your evaluation of units, and can you make the same evaluation on other thicknesses of glass?

T. Gjelsvik: Well, to be able to get results which you can compare directly, you should always use the same thickness of glass. If you use a different thickness you can never be sure how you should interpret the final results. When we tested a very large size of units, we always specified the glass thickness of 4 mm which corresponds roughly to your double strength. When we wrote the Scandinavian specifications, we reduced the size of unit and we reached a size where most manufacturers would supply the units with glass thickness of only 3 mm, but some of these manufacturers do not make units with 3 mm, they make it with at least 4 mm, so in fact we have run into trouble due to this. And that's one of the reasons we want to increase the size of units again. We want to get up into sizes where all manufacturers will supply the units with the same glass thickness and that will be 4 mm.

Mr. McKinley: Do we have perhaps one more question? Yes sir?

Mr. Robinson: Would you agree, sir, that you can affect your test results by making the glass thinner rather than thicker on the sizes of units that you are talking about? By this I mean that you would reduce the pressure difference from inside to outside because of the greater deflection of thinner glass near the central area of the unit.

POLYSULPHIDE SEALANT

FIGURE 1. Polysulfide unit.

Variation of aging and exposure ranges were adopted from 2 days to 30 days. Sets of these glass/aluminum samples are exposed as follows: Room temperature Oven aging at 70 °C. Water immersion at 50 °C. UV/Water-Ambient Linseed oil

Generally the samples are tested for peel adhesion after exposure at 2-day, 6-day and 30-day intervals. For the purpose of long term experimentation the samples are tested additionally at 30-day intervals up to one year.

Over a six month period several hundred peel adhesion coupons were tested. Values averaged 12 to 15 pounds per inch width after six months exposure to the above mentioned aging conditions for the conventional manganese dioxide-cured polysulfide systems. For a system which is highly resistant to various glazing compound vehicles (generally vegetable oil based), the values were in the magnitude of 30 to 34 pounds per inch width.

In addition to the long-term study of adhesion, moisture vapor transmission (MVT) data were obtained using ASTM E 96 test method. MVT rates range from 0.354 g/m2/24 hr to 0.533 g/m2/24 hr. The average specimen thickness used was 35 mils to correlate to the normal thickness of sealant between the spacer and the glass.

FIGURE 2. Test panels-glass/aluminum samples.

Based on these preliminary data of peel adhesion values and of MVT rates, sealed insulating glass units were then made and subjected to tests for seal integrity, initial dewpoint, accelerated weathering (dewpoint rise) fogging for both architectural and refrigeration applications and resistance to glazing compounds.

It should be noted that the data being presented are for commercially built sealed insulating glass units, not laboratory samples. Therefore, the type of workmanship generally used was indicative of what can actually be obtained in field units and makes the results more realistic. Our study involved several hundred sealed units of all descriptions.

3. Type of Study

The initial seal test was adopted to determine the seal integrity or seal leakage prior to subjecting the units to long-term accelerated interior weathering. The units, after being subjected to vacuum (3 in of mercury) for 2.5 hr, must show no signs of seal leakage and must not deviate from the zero deflection reading by more than 15 percent. This test has also proved to be a valuable research tool in determining glass deflection, effects on various thicknesses of glass, and the capabilities of sealants to withstand strain and stresses.

This change of 3 in Hg represents an altitude of 3,000 feet, so you can readily see the severity of this initial test.

Figure 3 illustrates the test chamber used for checking the seal integrity. In this phase of our test program we evaluated 450 organically sealed insulating glass units. The failure rate was approximately 10 percent; these failures were attributed generally to poor workmanship. There was no significance as to the type of cured polysulfide (PbO2 or MnO2).