H 
 
 _ 
 
 U. S. FORE 
 
 -^ 
 
 ATLANTA. GA. 
 MAY 1 2 1937 
 
 U. S. Department of Agriculture, Forest Service 
 
 FOREST PRODUCTS LABORATORY 
 
 In cooperation with the University of Wisconsin 
 MADISON, WISCONSIN 
 
 WOODEN DOORS 
 
 By G. E. HECK 
 Engineer 
 
 ~DS1T0RV_ 
 
 I 
 
 Published in 
 
 THE TIMBERMAN 
 
 April, 1936 
 
WOODEN DOORS 
 
 By 
 
 GEORGE E. HECK, 
 Engineer 
 
 The marked pick-up in residential "building 
 throughout the country has led to inquiries at the 
 Forest Products Laboratory as to relative merits of 
 various types of modern wooden doors. Most of the 
 inquiries seek information on strength, the "basis 
 of this article, which should prove of great 
 interest to door manufacturers and the trade alike. 
 
 In making a comparison of doors, it is necessary to consider the 
 kind of strength properties or combination of properties essential to a 
 satisfactory door. Failures in the glued joints of doors are more common 
 than mechanical failures in the wood. It is therefore apparent that 
 weathering, soaking, or conditioning tests are often of greater value in 
 determining the merits of a door than are mechanical tests. Moreover, the 
 modern door is such a complex structure that the results obtained from 
 mechanical tests on only one or two doors are of doubtful value and may 
 even be quite misleading due to the tremendous influence of such factors 
 as variability of the wood itself, type of glue used, workmanship in 
 gluing, species of wood employed; and in the dowelled doors, relation of 
 diameter of dowels to size of hole, moisture content of dowel, number of 
 dowels, and design of door or dowel. 
 
 Several years ago tests were made at the Forest Products 
 Laboratory to determine the superiority of doors of the through-mortised 
 and tenoned, blind-mortised and tenoned, and dowelled types. A total of 
 12. doors were tested. The rails, stiles, and panels were of solid wood. 
 Dowels were l/2 inch in diameter and the tenons, both blind and through, 
 were 5/S inch thick. These tests indicated that the through-mortised and 
 tenoned type of door possessed the highest structural strength and stiffness, 
 The blind-mortised and tenoned doors showed the lowest maximum strength and 
 resistance to swelling, although they showed a stiffness slightly greater 
 than that of the dowelled doors. However, only one design of each type 
 was tested, no attempt being made to vary structural details. It would, 
 of course, be entirely possible to design a dowelled door or blind-mortised 
 and tenoned door that would be as strong as a through-mortised and tenoned 
 door by varying some of the structural details. 
 
 RllOh 
 
 
Assuming that the various members to "be joined are uniformly 
 at the correct moisture content for efficient gluing, one of the most 
 important factors in the construction of doors is the gluing. If poor, 
 improperly prepared, or insufficient glue is used, or its application is 
 faulty, sagging of the doors and opening of the joints under stress are 
 apt to result. Vftien it is uncertain as to the exposure the doors are apt 
 to encounter, it is safer to use a water-resistant type of glue. The gluing 
 of the dowels is especially important in the dowelled door. Tests indicate 
 that threaded dowels give quite consistent superiority in withdrawal 
 resistance over smooth dowels. The threaded dowels tested were 5/8 inch 
 in diameter and were of maple. They had a single spiral groove about l/32 
 inch wide and about two-thirds as deep at a pitch of U- 1/2 turns per inch. 
 They also had two longitudinal grooves on diametrically opposite sides. The 
 longitudinal grooves were slightly under l/l6 inch wide at the surface of 
 the dowel and were ah out 1/32 inch deep. The smooth dowels were simply 
 smooth wooden pine 5/8 inch in diameter. After the tests, the door stiles 
 were ripped through to enable inspection of the holes and the surface of 
 the dowels. 
 
 Presumably the theory of the threaded dowel is that glue will 
 he trapped in the grooves and will he carried along as the dowel is driven 
 and not wiped away as is likely to he the case with smooth dowels. 
 Examination showed, however, that at no point on any of the dowels was the 
 spiral groove completely filled with glue, and in many instances glue did 
 not reach the bottom of the grooves. The same is true of the longitudinal 
 grooves except that sometimes they were filled with glue for l/2 or 3 A °^ 
 an inch from the surface of the block into which they were driven. Small 
 particles of wood adhered to the threaded dowels to a somewhat greater 
 extent than they did to the smooth dowels. This indicated that the threaded 
 dowels were somewhat more efficient in getting the glue to where it was 
 needed and probably accounted for the higher shearing strength and with- 
 drawal resistance obtained on dowels of this type. 
 
 The amount of wood adhering to the dowel, when pulled from the 
 cross bars, however, was so small that it indicated that really good gluing, 
 such as is readily obtained in joint work with the species used, had not 
 been accomplished. Obviously this was due to the dowel not being completely 
 surrounded v/ith a film of glue when it reached its final position in the 
 hole. It is believed that careful experiments would point out how much 
 stronger and more durable dowelled joints could be made by proper control 
 of other features of the gluing operation. Increase in the efficiency of 
 dowelled joints would seem to be highly desirable. Too often the dowel in 
 many factories has been looked upon merely as a fastening and not as an 
 integral part of the door that requires special attention. 
 
 It seems very doubtful whether joints with as little glue as 
 found in the foregoing tests would remain effective under the conditions of 
 vibration and changes of moisture content to which doors are subjected in 
 
 RllOU -2- 
 
service. It is understood that some door manufacturers apply glue to the 
 walls of dowel holes "by means of spindles operated "by automatic machinery. 
 The efficiency of the dowel joints produced in this manner is not known. 
 
 Shrinking and swelling with its attendant sticking is another 
 trouble encountered in doors of all kinds. This is, of course, caused by 
 moisture changes in the door itself and may he greatly retarded "by proper 
 seasoning and painting. Warp, which includes how, cup, twist, and crook, 
 may come from unequal moisture content values in the various component 
 pieces at the time of gluing, or from purely mechanical defects. Lack of 
 refinement in the manufacture of the door, such as poor matching of veneer, 
 is a common cause of warp,. A cross hand of a slab door, laid with its 
 grain at a pronounced angle with that of its mate, is likely, for instance, 
 to pull a slab door all out of shape. It is therefore well to store a new 
 door for several weeks hefore hanging under conditions as nearly as possible 
 like those under which it is to he used in order to allow it to reach a 
 moisture content approximating that of actual use. After all fitting is 
 done, it should be finished all over and especially on the top and "bottom 
 edges. Thorough finishing will retard the excessive moisture changes which 
 may cause swelling or warping of the door. 
 
 The moisture content correct for the service to which a door 
 is to he put of course varies with conditions, which for example are not 
 the same in a bungalow on the Gulf of Mexico as they are in a steam-heated 
 apartmint in Minnesota. Where the exact conditions are known they should 
 govern; seasonal changes naturally require the use of average figures. 
 
 A study has been made at the Forest Products Laboratory to 
 determine the changes that occur in the dimension of doors made from lumber 
 of different moisture content values. The results are by no means intended 
 to indicate proper or improper moisture values, but rather are intended 
 merely to show what happens in one representative type of door when, for 
 example, sample doors are made up from lumber at 9j 12, and l6 percent 
 moisture content values, respectively, and are hung in rooms in which the 
 atmospheric conditions simulate those inside a home in the northern states. 
 
 Six three-panel doors were used in this study. The wood was 
 air dried ponderosa pine of quality suitable for a No, 2 grade of door. 
 The stiles and rails were held together with oak dowels at each joint, two 
 in the top rail, three in the middle, and four in the bottom, which is a 
 total of nine to a stile. 
 
 The finished doors were hung on hinges in the usual manner, 
 three of them in a room where the relative humidity and the temperature 
 were held constant at the summer values of oO percent and 80° F. 
 respectively, and three in a room where the winter conditions of ~$0 percent 
 and 80° F. were maintained. Measurements and weighings were then made 
 periodically to determine twisting, cupping, and changes in sizes and 
 weight. 
 
 rhoU -3- 
 
The doors made up from lumber of 9, 12, and l6 percent moisture 
 content, developed openings at the muntin and shrank noticeably in the 
 panels when subjected to conditions similar to those in a residence in a 
 northern state in winter. The opening in the lo percent door was very 
 noticeable, that in the 12 percent door less so, and that in the 9 percent 
 door still less. Changes in dimension of the doors, called working, upon 
 transfer from conditions approximating summer in the northern states to 
 those approximating winter in the same region were practically the same as 
 the converse changes. Changes in dimension were greater than twisting and 
 cupping changes. Only a snail amount of twisting took place even in the 
 upper part of the doors, where the low stiffness permitted the maximum 
 amount. TOie cupping tool- \ilace nearest the unhir^d side of the d-x.r, The 
 change in width of the rtij-s was not so great at the ends as at the mid- 
 points, because the dowels restrained shrinkage at the ends. 
 
 The kind of wood generally used for core stock in veneered panel 
 doors is one of light weight and low shrinkage. The species most commonly 
 used are white pine and ponderosa pine. Little information is available on 
 the best thickness of veneer to use for stiles and rails. The experience 
 at the Forest Products Laboratory has been that less face checking results 
 when the veneer is thin, than when it is thick. Veneer as thick as l/k inch 
 is very likely to check on drying. The thicker the veneer the greater the 
 difficulty from shrinking and swelling due to moisture changes, particularly 
 in cases where no c^oss banding is used. Tests fail to show any appreciable 
 difference between sliced sawed, and rotary-cut veneers when mc.de into 
 plywood. There is a slight advantage in the plywood raa.de of sawed veneer 
 in that the plies are all practically quarter-sawed material and lend 
 themselves to a minimum of twisting and warping with shrinkage. Sawed veneer 
 also has less tendency to check and split under severe exposure conditions. 
 Tor the same reason, vertical grained veneer can probably be dried with 
 less trouble from splitting and curling. 
 
 With proper gluing little difficulty should be experienced with 
 either sliced, rota: r y-cut, or sawed veneer. They can all be glued 
 successfully. It would be difficult to state which would give the best 
 panel in three or five-ply construction. 
 
 RllOU -k- 
 
UNIVERSITY OF FLORIDA 
 
 llll III llll llll llllll llll II! 
 
 3 1262 08928 6354