It is not unfrequently stipulated in a specification for roof work that one or more bays of roofing erected complete, as far as the ironwork is concerned, but without the roof covering, shall be tested in the contractor's yard prior to taking down and erection on the site, with such dead loads as shall represent those conditions of wind pressure or weight of snow which the roof principal has been designed to meet.

Fig. 341. Scale 1 inch = 1 foot.

Fig. 342. Scale 1 inch = 1 foot.

The loading is frequently adjusted so as to represent either an equal or unequal distribution of wind pressure, and is usually made up by a sufficient number of such sections of steel or iron bars as may be handily available for the purpose, laid across from principal to principal; or, it may be, in the case of lofty roof principals of large span, by pig-iron suspended by straps from those points of the main rafter where the load arising from the purlins would occur in practice.

Where, as is usually the case, the actual roof covering is not present, then of course the weight per foot super of the covering must be allowed for, and added to the load representing wind pressure or snow.

It is customary, in the course of such a testing of the strength and stiffness of a roof principal, to ascertain the actual amount of deformation or deflection which the framed structure undergoes, and it may not unfrequently be the duty of the designing draughtsman to supervise the operation of testing, and to adopt such methods of practical measurement as will show such deformation or deflection. Such opportunities of witnessing the behaviour of any framed structure under conditions of actual loading should always be welcomed and made the most of by the student, who will find in them at all times instruction and interest, while possibly some disturbing thoughts on the accuracy of the results of his stress diagram may arise, if, for example, he finds occasionally certain tension members manifesting a distinct inclination to behave as struts.

Fig. 343. Scale 1½ inch = 1 foot.

The conditions of practical testing of a roof principal or girder in the contractor's yard will, as a rule, be found to include, first, a more or less loosely adjusted platform of balks or sleepers upon which the steel work is erected, and forming the supports or abutments; and, secondly, a foundation stratum of the loose ashes so frequently forming the surface of an erecting yard, and which is to a large extent compressible and yielding.

Fig. 344. Scale ¾ inch = 1 foot.

These two conditions taken together imply an amount of settlement of the supports which may possibly exceed in value the amount of deflection or deformation proper to the structure itself, and which is the result of the bending moments due to the loading. It follows therefore that vertical settlement of the temporary supports due to compression of loosely packed timber or soft soil, or both together, must not be mixed up with the measurements of the actual deflection of the structure itself due to span and loading, otherwise the results will be of little value.

There are various methods by which this little difficulty may be overcome, and the actual deformation of the structure ascertained. One of those very commonly adopted is to stretch a string or wire from end to end of the principal or girder, to serve as a datum from which careful measurements may be taken to the point or points at which the deformation of the structure is to be ascertained. If pegs (not liable to disturbance) are firmly driven into the ground on the line of the stretched string or wire, and the line of string marked thereon before and after loading, the total settlement can be ascertained.

Long strings or wires are, however, apt to be troublesome in wet and windy weather; and in certain classes of structures, such, for example, as the bow-string or sickle type of roof principal shown in Fig. 345, a better result may be obtained by the use of the ordinary spirit-level and staff, readings being taken at all convenient points, including, of course, the levels of the abutments or supports, before and after loading. A bench mark not liable to disturbance should be selected for reference and check. Where it is desired to ascertain the deflections of the upper member of the roof truss at points inaccessible to the staff, suspended wooden rods, slung from the points referred to, may be used, the intersection of the cross wires of the level being marked upon the rods before loading.

The whole series of levels being reduced to a common datum, and the settlement of the entire structure (as apart from the deflections due to bending moments) being ascertained, the actual deformations can be calculated, as though the structure had been placed upon an unyielding foundation.

Roof principals of considerable span show a tendency to stretch, due to elongation of the main tie, and this small increase of span should be watched and measured.

If the centre lines of main tie and principal rafter do not coincide at the shoe, but at some point outside it, certain other indications of strain will possibly become apparent, of which the student should take note.

The amount of actual load to be piled upon the rafters of a roof principal, of any cousiderable span, to represent an assumed wind pressure of, say, 40 to 50 pounds per square foot of surface, is very considerable, and in cases where only two principals, or one bay of roofing is erected, possibly without bracing, care should always be taken to avoid any risk of failure by lateral flexure of the principals, a type of collapse to which principals are particularly exposed, and which has led to sometimes disastrous failures in course of erection.

This can be obviated by sufficient lateral strutting, but such support should be so arranged as not to carry any component of vertical test loading to the relief of the principal, but only to act as a safeguard against lateral flexure.