the failures resulted from a rapid degradation of the sealant, presumably from ultraviolet radiation.

The results for some 67 sets of units received from November 1960 to July 1963, analyzed on the basis of the standards set by CMHC in 1964 (-40 °F (-40 °C) initial dewpoint and 0°F (18°C) after weather cycle) and on the basis of the present CGSB specification (-60 °F (-51 °C) initial and 0°F (-18 °C) after weather cycle), are as follows:

1964

115 V LINE

The results indicated that many manufacturers could produce units having no deposit or only a faint deposit. Furthermore, there was no evidence of field problems on brands having only faint deposits. A viewing arrangement was therefore developed in which a faint deposit is not apparent but a medium deposit is readily visible.

CGSB Specification 12-GP-8 is now being applied widely in the specification of sealed double-glazing for federal government buildings. The test apparatus has been reproduced by the testing laboratories of the Department of Public Works and results of tests in accordance with the standard are being used by an Inter-Departmental Qualification Board to develop a list of qualified brands. The results of laboratory as well as outdoor exposure tests indicate a steady and marked improvement in the quality of units produced since the program was started.

Interim results for 33 sets of units received before 1961 are given in Reference [1]. Only five of these sets would have passed the 1964 CMHC requirements and three sets the CGSB requirements. Six units from 29 of the sets were exposed outside and dewpoint temperatures measured periodically. After one year all units had failed on seven sets; after two years all had failed on 14 sets; after three years all had failed on 21 sets; and after six years all had failed on 22 sets. After seven years only one set was free of failures. Stains from materials in the sealing system appeared on at least three sets. At least two of

After one and a half years' outdoor exposure on two units each of 39 sets, one unit in each of three sets has failed; and three of the sets show signs of staining.

These figures include the results of tests carried out for manufacturers for purposes of product development and qualification by Central Mortgage and Housing Corporation. Some of the sealing systems were never marketed or were marketed for only a brief period. A substantial improvement in quality of specimens submitted since the program began is, nevertheless, apparent. Approximately 9 percent of the units received up to November 1960 would have passed the current CGSB requirements; 17 percent received from November 1960 to July 1963; 20 percent received

from August 1963 to July 1965; and 44 percent received from July 1965 to the present. Essentially, all of the units currently marketed incorporate a design that has met the test requirements of the Canadian Government Specifications Board standard. Although there has been no formal survey of field performance, the incidence of seal failure reported to the Division has greatly decreased. It seems, therefore, reasonable to assume that the average quality of units has greatly improved since the beginning of the research

program.

4. Conclusion

The procedure for evaluating sealed doubleglazing now in wide use in Canada appears to provide a reasonably good basis for judging the quality of assembly and the relative ability of the

various sealing systems to withstand mechanical stresses in service. It is mainly deficient in not identifying the effects of aging on the required physical properties, and some further consideration of this is desirable.

The severity of the acceptance requirements set by CMHC were gradually increased during the period of development of procedures, so that there was continuing pressure on the industry for improvement of the product. Competent manufacturers have responded and there has been a major increase in the average quality of units since the program began, to the benefit of both customer and producer. The CGSB standard now provides a good technical basis for specifying sealed double-glazing and for further development and improvement of both the methods of test and the product.

H. E. Robinson: Has your group been able to get field information or experience as to the durability or failure of units in service in Canadian buildings?

Mr. McKinley: Grant, would you like to respond?

A. G. Wilson: I think the answer is, as far as we're concerned, that we haven't had any planned program of getting information from the field. Most of our contact with conditions in the field has come directly from the manufacturers. We have enjoyed a very good relationship with all the manufacturers, have gotten to know them very well, and we've received from them some sort of feeling for the situation in the field, but we have had no planned program of our own.

The feedback mechanism of Central Mortgage and Housing Corporation as far as I know is again not a planned one, and I think the answer is that there is not good information on the statistics of failure in the field. We have been brought into a number of situations, mainly in the earlier days, when there were certain specific types of units that were involved in widespread failures. We have this kind of evidence, but we don't have any statistics on what is happening in a general way in respect to units as they are presently being manufactured.

Mr. McKinley: Mr. Rosen.

Mr. Rosen: Mr. Moderator, here's a question to you as a representative of one of the major American glass companies. In manufacturing insulating glass units, have you taken advantage of the Canadian program to utilize the Canadian standards of testing for units?

Mr. McKinley: Yes, like most manufacturers, we have submitted units for Canadian approval tests. We find them effective.

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

Mr. R. O'Shaughnessy: Mr. Wilson, in the Canadian programs do all manufacturers submit units to you or only those who are selling units in programs in which government monies are involved? Is there a mandatory program for this before sales can be made to the public?

Mr. Wilson: The incentive for manufacturers to submit the units to us in the first instance, aside from their interest in getting assistance with their own developments, was to get acceptance or listing by the Central Mortgage and Housing Corporation. And there is a very strong incentive because a very large percentage of the housing units which are constructed in Canada are constructed under the National Housing Act. This means that their mortgages are guaranteed by the federal government, although the mortgage money usually comes from private sources. So there's a strong incentive for any major producer to have his product listed by the Central Mortgage and

Housing Corporation. This means that a very large percentage of the manufacturers submitted units for this reason.

Participant: One of the tests you describedthe fogging test-appears quite severe for some types of units. How can excessively severe, or unrealistic, tests be guarded against?

Mr. Wilson: This is one of those things we are very much concerned about. We are always aware of the fact that some condition imposed in an accelerated test may be completely ridiculous and that units that can withstand it may be practically impossible to make. We try to avoid ruling good products off the market because of some unrealistic test condition. To get back to the fogging situation, there has been a considerable amount of evidence of fogging in the field, and I think some manufacturers almost went broke replacing units that were showing stains. To avoid getting back into this situation the CGSB committee decided that they should have some test to insure that this didn't happen. As I said, one of the things we are most concerned about in a test like this is to avoid unnecessary tests, or tests so severe that they rule out all the good products on the market. Mr. McKinley: Thank you.

Mr. Solvason: I might comment further that practically all the manufacturers in Canada can produce units that meet the requirements of CGSB, so while this testing as I said involves a fair number of development tests, eventually practically every manufacturer in Canada becomes able to produce sets that will meet the requirements. If something is sent in that doesn't quite meet them, the manufacturer goes back to try and find out what's wrong with his operation and straightens that out, until eventually he can get a product out that will meet the requirements. Mr. McKinley: Yes?

Mr. Robinson: Are the units that are placed out on the board for field exposure samples those that have previously passed the laboratory weathering test?

Mr. Solvason: The later sets that have gone on the field are samples of those which have passed the laboratory tests so that of the samples that are out on the board, there are none from sets that failed in the weathering test. The selection put outside was a fairly good selection, much better than the overall selection as reported in the weathering test.

Mr. McKinley: Yes?

Participant: Does CMHC require periodic resubmission of units for testing in order to keep their acceptance list current?

Mr. Wilson: The Central Mortgage and Housing Corporation, and I shouldn't really be speaking for them, because we're not representatives for them although we are fairly aware of their requirements, does not have any fixed period for

resubmission of products for their acceptance list. In the case of sealed glazing units, there was a requirement from CMHC for resubmission much more often than would normally be the case, because as Mr. Solvason said, the information on unit performance, the tests themselves, and particularly the requirements that were being used by CMHC were gradually evolving. As these changed, and as information came in, and as they began to know more about what was reasonable and what was possible, they would ask for resubmissions in the history that Mr. Solvason has given you. The most recent has been in connection with the CGSB specification which came out in 1966. Subsequent to that, all of the units in Canada wanting CMHC acceptance have been reexamined.

Participant: How does CMHC police its acceptance list?

Mr. Wilson: I'll begin trying to respond on behalf of CMHC. The Corporation operates on an honor system, believe it or not. They have no policing systems. They have certain requirements for listing and one is that the listing is only good so long as the manufacturer is manufacturing the product on which the original acceptance was based. That's understood by the manufacturer when he gets his product listed. If he chooses to ignore that, I suppose it's a matter for his conscience to begin with. Of course, there is always the concern that he might be found out, in which case I suspect, it would be rather difficult for him. to get on the list again. CMHC doesn't, as I say, run any sort of continuing certification program. When a manufacturer is listed on the CMHC list,

all this indicates is that the manufacturer has shown he could meet whatever happened to be the CMHC requirements for listing at that time.

Mr. McKinley: I think we might temporarily postpone further discussion of this paper until perhaps around the luncheon table. The Canadian fabricators have said on the one hand, that they do endorse the views that have been expressed and yet in this insurance idea they are going beyond the present situation. That has a great deal of significance for us here. I would like to thank Mr. Solvason and Mr. Wilson again for their contribution.

The gentleman who in many respects has gone to the greatest trouble to join us and share with us his experience is our next speaker. His subject is "Norwegian Experience with Accelerated Test Methods for Sealed Glazing Units and Their Correlation with Field Experience." It's a pleasure to introduce Mr. Tore Gjelsvik, Senior Research Officer, The Laboratory, Norwegian Building Research Institute, who has flown all the way from Trondheim and arrived promptly on schedule. Mr. Gjelsvik, we are very much pleased to welcome you, sir.

Mr. T. Gjelsvik: Mr. Chairman, ladies and gentlemen. I am very glad to be here to talk a little to you about our accelerated test methods and also something of our experience from field studies. But before I start I have to say a few words about our Norwegian Building Research Institute, which is somewhat different from your Building Research Institute here in the United States. We are doing research into building programs and we are testing materials and constructions.

1. Introduction

The work on the subject of sealed glazing units at the Norwegian Building Research Institute started back in 1958, independent of similar work in other countries. The first part of the project was sponsored by a Norwegian company, and led to the construction of an apparatus for accelerated aging. At that time, the accelerated aging tests constituted the whole test program.

Systematic field studies were introduced in 1959, to check the results of the accelerated tests and to gain more general experience. The results of the field studies and the information available from other sources have resulted in successive modifications of the accelerated aging tests. The test program has been changed, and the apparatus itself improved several times. The basic apparatus has, however, been the same all the time.

2. Stresses on the Edge Seal

The actual strains on the edge seal of sealed glazing units were thoroughly examined before the apparatus for accelerated aging was designed. The following types of strains were considered as actual:

Details shall not be given here, reference is made to earlier publications [1], [2].1

Of the types of stresses mentioned above, transportation and installation strains must be considered as more or less arbitrary. Transportation strains can easily be reduced by suitable measures, and with the present installation recommendations [3], the assemblage strains can be virtually eliminated. The real climatic strains must be said to be variations in the atmospheric pressure, changing temperatures, wind and sunlight. Water and vibrations can certainly be of importance in special cases, but whether they shall be included in normal test methods or not, is an open question.

In general, there seems to be agreement between scientists in the different parts of the world about the types of strains acting on sealed glazing units. The importance of the different types of strains, however, is judged to be somewhat different by different scientists. This is unpleasant, perhaps, but not really surprising. Some of the strains on sealed glazing units are fairly well known, while for others, the available information is rather limited. The different judgment is then only a natural result of the differences in the basic material. The situation is now considerably better than in 1958, but still an accelerated test program has to a high degree to be based on common sense.

1 Figures in brackets indicate the literature references at the end of this paper.