Comparison of tests on experimental 15-inch metal spars and 11-foot chord metal wing ribs. Parts 1 and 2 (Airplane Section report)

              ~
Auburn University Libraries
1111111 1111111111 lllll lllll lllll 111111111111111111111111111111 \I\\\ \II\\ II\\ 1111
File D 52.33/264
(AVIATION>
PUBLISHED BY THE CHIEF OF AIR SERVICE, WASHINGTON, D. C.
Vol. VI March 1, 1926 No. 556
COMPARISON OF TESTS
ON EXPERIMENTAL 15-INCH METAL SPARS AND
11-FOOT CHORD METAL WING RIBS
PARTS 1 AND 2
( Airplane Section Report )
Prepared by J. S. Newell
Engineering Division, Air Service
McCook Field, Dayton, Ohio
August 3, 1925
WASHINGTON
GOVERNMENT PRINTING OFFICE
1926
Ralph Brown Draughon
LIBRARY
MAY 2 9 2013
Non·Depoitory
Auburn University
PART I-Spars:
~; ~.""=!'\ .~ ... ·~~1;~
.. .J .... ·' j
r
INDEX
Results and conclusions ___________________ ___ ____ _______ ________ ________ ____________ _____ _
Method of testing~ _____________________ ____ ______ __________ ___________________ ______ ___ _ Mild carbon-steel tube spar, welded joints __ ____ ______ __ ________ ______ '._ ____________________ _
Chrome molybdenum steel tube spar, welded joints __ ____________________________ ___________ _
Nickel-steel tube spar, welded joints, wire diagonals ____________ ________ ____ ___ __ ______ ____ __ _
Combination steel and duralumin spar_ ___ ____ ___ ______ __ __ __________ ______________________ _
Duralumin tube spar, welded joints ____ ____________________ _________ __ _________________ ___ _
Built-up duralumin spar, lightened web t ype ___ ____ _____ ___ ___ ____ _____ ____ ______ __________ _
Built-up duralumin spar, plate girder type ________ ___ ____ __________________ ________ ____ ____ _
Wood box spar ____ ___ _________ ______ ______ _____ _____ __ __ __ ________ __ _____________ ______ _
Rating of the different types __ ___ ____ ____________________________________________ ____ ____ _
' PART II-Ribs:
Results and conclusions ____________________________ ________ __ ________ _____ _____________ __ _
Method of testing ___ __ ___ _____________ ____ -'- _____________________ ______ _____ ______ ____ __ _
Steel tube ribs, welded j•oints_ - -- - -- ------- ________ ____ ______ ______ ____ _________ _____ ___ __ Duralumin tube ribs, riveted joints __ __ ___ ___ _____ __________________ ____________ ____ __ _____ _
Duralumin tube ribs, welded joints __ ________ __ _________ __ _______ _________ ____ ____________ _ _
Duralumin ribs, hollow rectangular members _ _____ ___ ___ __ ___ __ ______ _______ _____ ____ ______ _
Duralumin ribs, flanged chords, tubular webs _______ ___ _______ ______ _____________ ____________ _
Duralumin ribs, stamped chord and web members _________ _______________ _______ ____________ _
Duralumin ribs, stamped truss type chords ______________ _______________ ____________ _______ _ Rating of the different types ____ ______________ __ _______ ____ ___ __ ____________________ _____ _ _
Page
2-3·
3
4-5
6
6-7
7-8
8
8-9
9- 10
10-11
11-13
14
14
14-15
15
16
16'
16-17
17'
17- 18
18-19
CERTIFICATE: By direction of the Secretary of War the matter contained herein is published as admin­istrative
information and is required for the proper transaction of the public business.
(n)
..,
COMPARISON OF TESTS ON EXPERIMENT AL 15-INCH
MET AL WING SPARS AND 11-FOOT CHORD MET AL
WING RIBS
PURPOSE OF TEST
Early in 1923 contracts were let for the design and
construction of five types of experimental metal wing
beams and seven types of wing ribs. The spars and
ribs were each designed for a definitely specified load
system which was arranged to develop stresses of the
magnitude of those in a 16,000-pound airplane, so that
the resulting structures, when tested, would furnish
data from which a satisfactory type of all-metal wing
could be developed for use on large airpla nes. Each
type of spar and rib was tested under ide ntically the
same conditions as the other types, so that the test
data would be truly comparable and afford a positive
means of determining the merits of each type.
The purpose of this report is to scrutinize the data
collected during the tests for the purpose of rating the
various spars and ribs as to their desirability for
use in airplane wings, and if possible to develop means
for analyzing and designing the various types of spars.
In this report certain conclusions have been drawn as
to the merits of the spars and ribs and methods are
suggested for taking care of some ' of the more difficult
points in the design of the metal spars, but it must be
borne in mind that the data on which the conclusions
and the methods of analysis are based are very meager,
so that both may be revised if furt her tests are made
and more data accumulated. The report is given in
two parts, the first of which deals with the wing spars,
the second with the wing ribs. Each part is complete
in itself as to conclusions, methods of testing and
rating, and comments on the behavior of the spars or
ribs during test.
. strength
The spars are rated on a basis of the . ht and
we1g
strength .
weight X deflection rat10s and are compared to a 15-
inch spruce box spar which was designed and built by
the Engineering Division to be used as a criterion for the
metal types. The ribs are compared on the basis of
t h e swtree1.n gg htth rat1. 0 w1" th some cons1' d erat1. 0n f or the 1. r
production and maintenance features.
(1)
Part 1.-COMPARISON OF 15-INCH EXPERIMENT AL MET AL SPARS
RESULTS AND CONCLUSIONS diagram, provided reasonable allowance is made for
play in the type of connection used. Because of the
These tests emphasized several points of importance paucity of data on which this conclusion is based and
relating to the design and construction of metal spars, the wide differences between types of connections this
the most important being the marked tendency of such problem can not be considered as solved, however, and
spars to fail by buckling or crinkling of the compression designers should not put too great faith in the secondary
members. The majority of the spars tested failed stresses computed in this way, although it is recom­either
by the buckling of one of the compression diag- mended for use until more test data are available to
onals in the web system or by the buckling of the vindicate or disprove it.
compression chord itself. That buckling failures A comparison of computed and observed deflection
occurred in the chords is particularly worthy of note, curves on three spars, two metal and one wood, indi­since
the distance between bracing used in the tests cated that for beams having solid webs the deflections
was redu'ced to 45 inches · from the original design could be quite closely predicted by use of 'the precise
specification value, which called for 90-inch spacing. equations for members under combined loading. But
No crinkling failures occurred, but that may be ex- for a beam having a lightened web system these
plained by the fact that none of the spars were made of equations did not give satisfactory results, the tendency
extremely thin material, such as is often proposed for of such a beam being to deflect to a much greater
metal wing spars. · Had such material been used it is extent than was indicated by the equations. It is
highly probable that parts . of the test spars would concluded, therefore, that the precise equations are
have failed by local crinkling, as in a number of cases satisfactory for use in designing beams having solid
local failure of fittings did occur. Designers of metal webs, but that these equations should not be used
spars should keep these two types of elastic fai lure, for when the web is lightened by cut-outs. Such members
·neither of which have practical methods of analysis appear to behave more as trusses, and should b-e
been developed, in mind and carefully design all analyzed accordingly.
compression members so as to reduce their tendency These tests indicated that welded duralumin was
to buckle or crinkle to a minimum before using them not sufficiently reliable for use in wing spars, a weld in
in a spar. To do this the designer must rely mainly the one spar of this type tested having failed under
on his experience and judgment. less than three-quarters of the design load, at which
Results obtained from a study of the failure of one t ime a majority of the remaining welds showed cracks.
of the test spars indicate that the buckling of the Another type of spar tested proved that it was
chord members of a tubular truss type spar can be possible to weld heat-treated nickel-steel tubes, but
prevented if the tubes are made large enough to carry tests on minor ·parts cut from the spars on which this
the average axial load between lateral supports as a was done showed that such a procedure was not
column load, assuming C= 1.0 and L equal the length desirable, since it reduced the· strength of the material
between supports. This method is probably some- to about that of the annealed steel. This reduction of
what conservative, but because of the lack of data the strength of the material at the welds obviously
on this subject no better one can be suggested at the nullifies the effects of the heat treatment on members
present time. designed to take tensile loads, and it would be danger-
It was found from data obtained during these tests ous on short · columns, in which the fiber stress was
that the welded tubular trussed spars could be analyzed liable to exceed the reduced yield point and cause the
and designed as pin-jointed trusses, and the indications tube to crinkle a·t the joint. Since little is gained by
are that they will deflect very nearly as pin-jointed heat treating long columns, whose strength depends on
structures, having no play in the pinholes. ' The the modulus of elasticity alone, a property which is not
latter conclusion is reached from a comparison of affected by the heat treatment, it seems futile to at­observed
and computed deflections for only two types tempt to weld highly heat-treated steel in any case.
of spars, neither of which deflected very much, and it is T his does not mean that it is undesirable to weld high­probably
not safe to extend it blindly to all types of strength alloy steels. A study of the weights of the
trussed spars. From a study made on a third spar, welded 1020 steel and the welded chrome molybdenum
however, in which bolted fittings were used, it was steel spars will show that considerable gain can be made
found that good agreement could be obtained between by using a material with a high yield point and tensile
observed deflections and those computed by a Williot- strength, so long as these properties are not obtained
Mohr diagram if sufficient allowance was made to by a process which is counteracted and nullified by the
provide for the pley in the pinholes of the fittings. It heat of welding. Tests on minor parts of the chrome
therefore appears that the magnitude of the secondary molybdenum steel tube spar showed that efficiencies
stresses in trussed spars can be determined with suffi- around 80 per cent .could be obtained with welded
cient precision for use in design by obtaining the joints, where about 90 per cent would be expected if
deflection of the spar under load from a Williot-Mohr mild carbon steel were used. While welding has a
(2)
r
'3
deleterious effect on chrome molybdenum steel tube, '
it is nowhere near so great as on nickel steel, where the
ultimate tensile strength was very greatly r educed.
Of the t ypes of spar which showed up well the results
of this study point to the combination steel and duralu­min
spar as being best from the strength to weight ,
standpoint. If stiffness is also considered, as it must
be because of the effects a very flexible spar has on the
aerodynamic properties of a wing, the welded chrome
molybdenum steel tube spar is the best of the metal
spars, but the spruce box beam which was built for
comparative purposes appears to be even better. The
stiffness and weight of this wooden spar, which carried
the same load as the metal ones, indicate that even for
spars as deep as 15 inches a wooden box section is
practically as stiff as the best of the metal spars, while
it is lighter than most of them.
Where a depth as great as 15 inches is not available,
as, for instance, in a thin-winged bo!Ilber , the duralu­min
plate girder type of spar is preeminent. These
tests showed that such a spar having a depth equal
to about half of that used on the trussed types can
be built to carry the same stresses and be about as
light as the trussed tubular types. Because of the
increased weight entailed in stiffening the web of a
deep plate girder this type is not so economical where
large depths are available, but it is the only practicable
one of the types tested for use in a heavily loaded spar
for a thin-winged airplane.
It is concluded , then, that of the seven types of
metal spars tested, three are particularly good. The
best from t he standpoint of ease of analysis, design,
and construction, and from a consideration of the
strength-weight-stiffness ratio, is the chrome molyb­denum
steel tube spar. The next best appears to be
the combination steel and duralumin spar having a
steel web system and duralumin chords. The deflec­tions
of this type of spar are relatively large, and might
affect the aerodynamic properties of a wing in which
it was used. Moreover, they are not readily deter­minate,
so that a certain amount of weight would have
to be added arbitrarily to provide for the secondary
stresses unless a series of spars were built and tested
to develop one for the case at .hand. Where depth is
not available the duralumin plate girder type is the
best of the types tested. It can be closely analyzed,
and it is a relatively easy spar to which to attach ribs
or fittings. This type should prove satisfactory in
shallow wings, where depth is not available for a
tubular trussed spar, but where depth is available one
of the trussed types will proba bly prove lighter and
stiffer, hence more satisfactory.
METHOD OF CONDUCTING TESTS
The spars were placed in the test jig, shown dia­grammatically
in Figure 1, which was constructed espe-
Adm. No. Serial No.
cia lly for these tests. Figures 2 and 16 show spars in
the t est jig and give a clear idea of its construction.
Each spar was 20 fee~ jong, .having a 15-foot span and
a 5-foot cantilever overhang. Loads were applied
at the third points of the 15-foot span and at the
cantilever tip, as shown in the figure. The load on
the tip platform was 50 per cent greater than those on
each platform in the span, the design loads being 3,000
pounds at the tip and 2,000 pounds at each third point
of the span. The steel straps which supported the
spars at the 15-foot point were sloped in 'Such a way
as to develop an axial load in the span of 20,000 pounds
when the full design loads were put on the plat forms.
This combination of loads gave a maximum bending
moment of 180,000 inch-pounds, and axial load of
20,000 pounds and a maximum shear of 3,000 pounds,
all of which are comparable to the loads on the upper
wing spars of a conventional biplane weighing about
16,000 pounds.
The spars were supported against lateral failure
every 45 inches, and deflections were obtained at
intervals of from 20 to 30 inches after each increment
of load was added.. In some cases gauge points were
established at several places along the spar and extensb'­meter
readings were taken to determine the actual
stresses at these points. These stresses \vere ·desired
for purposes of comparison with those computed by
analytical methods to determine whether or not t he
analytical methods used for design were sufficiently
accurate to be depended upon.
Two spa.rs of each type were t ested in all but two
cases. The first spar was designed to fail slightly
below the required load, so that t he second articles
could be strengthened to hold the load. The first
article of the welded duralumin spars demonstrated
that this type of conetruction was impracticable, so
no second article was tested. The first article of the
welded chrome molybdenum steel tube spars failed
under 105 per cent of the required load, so no second
article was constructed. In all other cases two spars
of each type were tested.
DISCUSSION OF THE INDIVIDUAL TESTS AND
THEIR RESULTS
Each type of spar will be discussed separately as to
its behavior during t est and as to the more prominent
structural merits or demerits emphasized by the tests.
Investigations of the deflections, extensometer read­ings,
and such other data as were obtained during the
tests will be ma.de to determine which methods of
analysis agree best with the test results.
The data on deflections, loads carried, and failures
of the spars were obtained from the static test reports
of the individual spars. The following is a list of these
reports, with their McCook Field serial numbers :
Title
575 . _ __ _ 2274 . _ _ _ _ ___ Static test of Huff-Daland test spars (botb articles).
689 . _________ - -- --------- - - -- - Static test of Engineering Division experimental steel spar.
r,02 and 667 .. 2321 and 2434 ___ Static test of Boeing experimental wing beam.
644 ____ ______ 2401. _____ ____ __ Static test or Kerber Boulton experimental wing beam.
633 ____ ___ ___ 2384 _____ ____ ___ Static test or Douglas experimental wing be(>m.
573 . . -- --- --- 2215 ______ _____ _ Static test or L. W. F. test spars (botb articles) .
622 and 704. _ 2370 and 2501. .. Static test or aeromarine experimental duralumin wing beam.
601 and 656 .. 2322 and 2430 .• . Static test of Engin~ering Division experimental wood box wing beam.
A comparison and rating of all of the spars tested
will conclude this section of this report.
Type 1.-WELDED LOW-CARBON STEEL TUBU­LAR
SPARS
The Huff-Daland Co., of Ogdensburg, N. Y., sub­mitted
two welded low-carbon steel tube spars for test,
the first article weighing 75.9 pounds and the second
articl~ 74 pounds. The type of construction is indi­cated
in Figure 2, which shows the spar in the test jig.
Figures 3 and 4 show the arrangements and sizes of the
members used in t he first and second articles. The
first spar failed at 75 per cent of the design load, the
failure being in the cantilever tip near the point of sup­port.
This portion of the spar was redesigned, the
structure at both points of support was changed some­what,
and, though the second article weighed about 2
pounds less than the first one, it carried 110 per cent
of the design load and failed under 115 per cent.
The first spar failed due to the heavy compression
in the lower flange about 6 inches outboard of the
outer point of support causing one of the t ubes to
buckle. This failure took place with 75 per cent of
the design load on the platforms. A new cantilever
tip section was then welded on and a second test made
to bring out further weaknesses in the spar . In the
second test failure occurred under 105 per cent of the
required load, the bearing of the pin at the outer sup­port
point being so great that the pin tore out of the
fitting assembly. It was also noted during this test
that the fishtail splices used in the chord members were
so made that the smaller tube or the swaged portion of
4
the larger one was permitted to extend a short distance
into the bay where the large tube was required. Local
failures were noted at these points during the test,
and it is recommended that whenever tubes are to
be spliced in this type of spar care be taken to extend
the larger-diameter tube a short distance into the bay
where the small tube is suffi cient. Slight changes were
made in the sizes of tubes and the fitting details used
in t he second article.
This type of welded tube spar showed up very well
in the tests so far as any structural t roubles were con­cerned.
Except for the local failures where splices
were made at panel points no difficulties were encoun­tered,
and these were not encountered in the second
article, as this detail was handled correctly.
During the test of the first a rticle readings were
taken on several of the chord members and diagonals
with a Berry extensometer. The stresses indicated by
the elongations or contractions shown by t he extenso­meter
were checked against t hose computed for the
various members by the ordinary methods of analysis.
During the test on the second ar t icle extensometer
readings were made on practically all members, the
majority being made on an 8-inch gauge length. On
the members that were too short for an 8-inch gauge
or were obstructed by the jig bracing a 2-inch gauge
length was used . The results obtained on this short
gauge were of no value, however, as the stresses and
consequent elongations were insufficient to give a good
series of readings. The following tables give the
stresses in the members as indicated by the extenso­meter
and as determined by the stress analysis.
TABLE 1.- Results of extensometer measurements on first article of Huff-Daland experimental spars
Extensometer readings Extrapo- iated Load from
Point for 85 Observed Tube Observed stress
per cent s tress area load analysis'
25 per 75 per load 1
cent load cent load
--- - --- - -
Ac _a_n__d_ B__._ -_-_-_-_-_-_-__-_-_-_-_-_-_- 43. 5 59. 0 +26.5 -18, 500 0. 4878 -9,050 -9,800 43. 0 38. 0 -8.5 +6, 000 .1079 +650 +1,080
D---------- ----- ------- 46. 0 54. 0 + 14.0 - 9, 800 .1079 -1, 060 -1,080
E ______________________ 44. 0 37. 0 - 12.0 +8, 400 .1079 - 905 +1,080 F and o _____ ______ ____ 51.0 66. 5 +26.0 -18,000 . 3313 - 5, 960 - 6, 540
Hand J_ ___ ____________ 51.0 68. 5 +30.0 -21, 000 . 4878 -IO, 250 -11 , l()()
Kand L ___ __________ __ 63.0 77. 0 +24.0 - li,000 .3313 -5, 640 -6, MO M __ __ ______ ____________ 56.0 50.0 -10. 0 +7,000 . 1079 +755 -1, 080
N-- --- -- --- ----- ------- 63.0 72. 0 +15.o -10, 500 . 1079 -I, 130 +1 , 080
0 and p __ ____ ____ ____ _ 55. 0 72. 5 +30.0 - 21,000 . 4878 -10, 250 - 11, IOO
R and$ ___ ________ _____ 50. 0 67. 0 +29. 0 -20, 300 . 3804 - 7, 740 - 9, 150 T __________ ________ ____ 40. 0 49. 0 + 15.0 -10, 500 . 1079 -1, 130 -1 , 080
U and v _ ----------- --- 58. 0 77. 0 +32.0 - 22, 400 . 3804 -8, 530 -8, 500
wand x __ ____________ 99. 5 59. 0 -67.0 +47,000 . 1849 +8, 690 +8,430
zy ____________________________________________ 129. 5 112. 5 -29.0 +20.300 . 1272 +2, 580 +3, 270 147. r, 170. 0 +38. 0 -26, f>OO .1272 - 3, 390 - 3, 270
a and fJ ______ _________ __ 50. 0 74. 5 +41.5 - 29,000 . 2041 -5, 925 - 6, 320
1 Obtained by taking!~ of the difference between readings at 75 per cent and 25 per cent of the design load.
' This load is obtained by taking 85 per cent of the design load computed in the stress analysis.
One point difference in extensometer reading equals 0.0002 inch per 8-inch gauge length. This is equivalent to a change or stress of 700 pounds
per square inch when E= 28,000,000 pounds per square inch.
As a result of these tests and the investigation of the by t he shear on t he spar, a re sufficiently accurate for
stresses in the members as given in T ables 1 and 2 it use in design although they probably are not in abso­has
been concluded that the ordinary methods of lute accord with t he actual stresses developed in the
computing the loads in members of a truss of this type, members. The agreement between t he observed and
i. e., t he stresses in th e chord members being determined computed stresses in the chords was reasonably close,
by the bending moment and the stresses in t he diagonals but considerable discrepancy existed between the
5
TABLE 2.-Results of extensometer measurements on second article of H-uff-Daland experimental spars
-
E xtensomet er
reading
Member Differ-
No load \·Designed
enee
i---- --
I and 2 ____ ____ __ ______ __ Results are useless
3 and 4 __ _ --- --- - --- -- -- - 34
I
454
I
420
5 and 6 __ - - --- -- - - ----- - - 63 616 5.53
7 and 9 and 8JO . _. _-_-_- _--_-_-_-_-__--_-__-_- -_-_ 92 545 453 II and 12 ___ __ __________ _ 8 440 432 13and14 ___ __ ___ ______ __ Resul ts are useless 15 and 16 ___ _____ ____ __ __ 760 920 740 593 17 and18 ... ____ ______ ___ 610 1,280 19 and 20 ___ __ ___ ________ 600 I, 152
2213 aanndd2 224 -_-_-__--__- _-_- -_-_-__-_-_--_-_ 989 l , 180 25 and 26 ___ _______ ___ ___ 947 1, 301 690 731
2279 aanndd 3208 __._ _- -_-_-_-_-_-_-__-_-_-_- -_-_ 568 471 31 and32 ____ ___________ __ 820 1,03() 33 and 34 ___ ___ ________ __ 530 710 660 1,230
35 and 36_ .. -- --- - - - -- - - - 875 1,430
37 and 38 ... - -- -- - - - - - - - - 900 1, 290
3419 aanndd4 420 ____ _- -_-__-_-_- -_-_-_-_-__-_- -_ 880 1, 340 950 1,360
43 and 44 __ -- - ---- - -- - --- 880 880
4475 aanndd 4468 ____ ._ _-_-_- -_-_-_-_-_-_-__-_- -_ 965 l, 110 800 510
5419 aanndd 5520 ____ _. _-_-_- -_-_-__- -__-_-_-_- -_ 785 1, 210 53 and 54 ____ ____________ 815 1,315 930 1,340
5575 aanndd 5568 _____. -__-_-_- -_-_-__-_--__- -_-_ 655 950 735 380
59 and 60. -- --- -- - - - - - - - - 880 1,370
61 and62 . - - -- -- -- - - - ---- 1, 070 430
63 and 64 _. - - --- - - --- - --- I, 045 530
65 and 66_ - . -- -- - - -- - - - - - 940 1, 265
6679 aanndd 6780 ____ _. -_-__-_- -__- -_-__- -_-_-_-_ 1,085 475 71 and 72 ___ __________ ___ I , 030 1, 630 73 and 74- __ __ ___ ____ ___ _ 810 410 940 530
Gauge length =8 inches.
One division on dial-
50
,
1
000
inch.
For E=28,000,000one divisionon dial=~8xC:::: =70 pounds.
Increasing readings on dial denote compression.
160
-247
670
552
191
354
41
-97
210
180
570
555
390
460
410
0
145
-290
425
500
410
295
-355
490
-640
-515
325
-610
600
-400
-410
two values of the stresses in the diagonals. This is
probably explained by the fact that some bending was
induced in the diagonals due to their rigid connection
to the chords, and also to the fact that both chords of
the spar were in compression inboard of the outer sup­port,
which would induce additional stresses in the
diagonals.
Owing to the fact that the deflections were small, it
has been impossible to obtain satisfactory data for a
study of the effects of secondary bending on this t ype
of spar. It was expected that information on this
point would be obtained during this series of tests
and it is grea tly regretted that satisfactory data were
unobtainable. The effect of the large load on the
cantilever was to cut down the deflection of the bay,
and this effect was so pronounced as to eliminate
practically all of the deflection and hence practically
all of the secondary stresses due to the deflection
moments. The large tip load caiJSed seve re local
distortions in the members of the first article near
the outer point of support, with the result that the
curve of observed deflections was irregular even under
a load equal to only half the required. Owing to the
I
Observed spqosuturanersdess i inpnec hr Tube size Taruebae I Olbotsauedbr viene d Colomtaudpb ueint e d
- ---·-
Plate 2XYs o. 2500 --- --------- --- --- --- -- - -29,400 1'4XO. 065 .3440 -10, 100 -11,540
-38, 700 1'4X . 065 . 3440 -13, 300 -13, 080
-31, 700 l V.X . 065 . 2930 -9, 300 - 9, 230
-30, 200 l ~ X . 065 . 2930 -8, 850 . -7, 690
'4X . 049 . 1079 1----- ------- + 1, 260
-II, 200 '4X . 049 .1079 -1,200 -1, 260
+17,300 '4X . 049 .1079 +1, 870 + 1. 260
-46, 900 1'4X . 065 I . 3440 -16, 100 -14,620
-38, 600 1'4X . 065 .3440 -13,300 -13, 080
-13, 400 l ~ X . 065 . 2930 - 3, 950 -6, 150
-24, 800 l V.X . 065 . 2930 -10, 200 -6, 150
-2, 880 '4X . 049 . 1079 -310 +1,260
+6, 800 '4X . 049 . 1079 + 730 +1. 260
-14, 700 '4X . 049 . 1079 -1, 590 -1,260
-12, 600 '4X . 049 .1079 -1,360 +1,260
-40, 000 2 x .065 .3950 -15,800 -II, 540
-38, 900 2 x . 065 . 3950 -15,400 -10,000
-27, 300 1'4X . 065 . 3440 -9,400 -9, 230
-32, 200 1'4X . 065 . 3440 -11, 100 -12, 460
-28, 700 '4X . 049 .1079 -3, 100 -1,260
----- ------- '4X . 049 .1079 ------------ +1,260
-10, 150 '4X .049 . 1079 -1,100 -1, 260
+20, 300 %X .054 . 1395 +2,830 +3,810
-29,800 1'4X . 065 .3440 -10, 250 -5, 100
-35, 000 2 x .065
I
.3950 -13 800 -17, 380
-28, 700 2 x .065 .3950 · -11:300
11
-20, 700 Plate YsXl . 1250 2,600 22,480
+24, 900 %X .054 . 1395 +3,475 +3, 810
-34,300 %X . 054 .1395 -4, 785 - 3,810
+44, 800 l ~X. 065 . 2420 +10. 800 +9,980
+36,000 l~X . 065 . 2420 +8, 110 +4, 840
-22, 700 l ~ X .065 . 2675 -6, 075 I -7, 500
+42, 700 %X . 054 . 1395 + 5,960
I
+3,810
-42,000 YsX .054 . 1395 -5, 860 -3, 810
+28,000 YsX . 054 . 1395 +3,910 +3, 810
+28, 700 YsX . 054 .1395 +4, 000 +3, 810
small deflections developed in this spar a slight irregu­larity
of this sort was sufficient to ruin the precision
of the curve of observed deflections, and for this
reason no definite information was obtained from the
first test . After the cantilever tip had been rebuilt a
second test was run, in which these distortions were
not so great and a fairly good curve of deflections in the
span was obtained. Deflections were computed by
the Williot-Mohr diagram for this spar under 50 per
cent of the design load. It will be seen from Figure 6
that the computed deflection curve agrees quite well
with the observed values obtained during this second
test, although a comparison with the irregular points
obtained during the first test shows a very poor
agreement. It is thought that for welded spars of
this type the deflections obtained by a Williot diagram,
which assumes all joints to be pinned but to have no
play in the pinholes, will give reasonable results when
used for computing secondary stresses.
Figure 7 gives a comparison of the stiffness of the
first and second articles under 50 per cent of the.design
·load. These curves are of little value in this case
because of the irregularity of the observed points.
Type 2.-WELDED CHROME MOLYBDENUM
STEEL TUBE SPAR
In order to obtain a test on the practicability and
reliabilHy of a welded chrome molybdenum steel
structure, the Engineering Division designed and
constructed a spar of a Warren truss type somewhat
similar to that submitted by Huff-Daland. A different
type of fitting was used, the spar was subdivided into
longer panels, and chrome molybdenum steel tube was
used instead of 1020 carbon steel, with a resultant
weight saving of some 15 pounds. Since the first
article held 100 per cent of the design load and failed
under 105 per cent, only one of these spars was built.
Its weight was 60 pounds, which was somewhat heavier
than it should have been, due to the fact 'that tubes
of the required gauge were not available, so that all
tubes were heavier than required by the stress analysis.
This was particularly true of the smaller web members,
where both gauge and diameter were increased in
order to u·se stock sizes and avoid delay. It is probable
that the weight of this spar could have been reduced 4
or 5 pounds without affecting the strength.
6
Except for the fact that the fitting at the inner
point of support buckled badly under 85 per cent of
the load, although it did not fail even with the 105
per cent load finally put on the platforms, no structural
defects were noted during the test. Figure 8 shows
the spar ·after it had failed by buckling in the middle
of the bay, and it gives a good idea of the type of
construction used. The box fittings were made of
steel and were welded to the chord members. Figure 9
shows the general arrangement of the tubes and their
sizes. No extensometer measurements were made on
this spar but fairly good curves of deflection were
obtained. Figure 10 shows the curve of deflection
obtained under 50 per cent of the required load and
Figure 11 gives a comparison between the observed
deflection and those obtained by the use of a Williot­Mohr
diagram. Although deflection readings were
not obtained under 50 per cent of the design load, they
were recorded at 35 and 60 per cent loads, from which
data the curve shown- in Figure 11 was obtained by
interpolation. It is fairly smooth and will be seen
to be in quite close agreement with the deflections
obtained by use of the Williot diagram. Due to the
small deflections involved the percentage error in these
curves is fairly high, but the actural errors involved
are so small that reasonable results could be obtained
if these deflections were used for computing secondary
stresses.
An investigation of the cause of the chords buckling
indicates that chord members designed to take com­pression
should be strong enough to carry as a column
load the average axial load to which they will be sub­jected,
the length of the column being equal to the
distance between lateral supports and C being equal
to unity. In the case of this spar the upper chord
member at the point where failure occurred was a
1.Y:;- 0.058 tube, and as such should have failed under
an axial load of 11,100 pounds when C= 1, 12,500
pounds when C=, 2, assuming L to be 45 inches, the
distance between supports. ·The average load between
supports was 11,700 pounds and the weighted load,
giving double weight to the load which extends over
two-thirds the span, was 11,320 pounds. For the
lowi:ir chord member a 1%-0.065 tube 45 inches long
should carry 8,200 pounds with C= 1, 12,000 pounds
with C = 2. The average load between lateral supports
was 11,660 pounds and the weighted load 10,870
pounds. From these data it appears that fairly de­pendable
results would be obtained when the smaller
load extended over a majority of the bay by using
the average load between supports and C = 2, though
a more conservative result would be obtained with
c~~ 1 and a weighted load. Until more data are
obtained on this subject it is recommended that the
latter practice be followed as being a little more con­servative.
Tests on minor parts showed the following properties
for the material in this spar:
Proportion limit: 48,800 pounds per square inch.
Ultimate tensile stress: 105,100 pounds per square
inch.
Modulus of elasticity: 23,500,000 pounds per square
inch.
Average stress in two welded joints at failure: 82,700
and 86, 700 pounds per square inch.
Average tensile strength of tubes: 106,000 pounds
per square inch.
Efficiency of joints: 78 and 82 per cent.
The modulus of elasticity of these tubes was low,
much lower than it should have been, but the other
properties of the steel are about a·verage. The ef­ficiencies
are quite good for welded joints and indicate
that the strength of a chrome molybdenum tube is not
materially affected by welding. The efficiencies are
somewhat less than those to be found in welds in mild
carbon steel, but are much be;.ter than those for
welded heat-traated nickel steel, as will be seen from
the results of similar tests on type 3 spars.
Type 3.-WELDED NICKEL-STEEL TUBE SPAR
Figure 12 shows the electrically welded steel tube
spar submitted by the Boeing Airplane Co. This
photograph was taken after the second article had
failed under 120 per cent of the required load, and it
shows the type of construction very well. The chords
and vertical web members were nickel steel, heat­treated
and then electrically welded, while the diagonal
members were all swaged nickel-steel tie rods. Tests
made on minor parts cut from the spar showed the
following properties:
Carbon content: 0.13 to 0.32 per cent.
Nickel content: 3.42 to 3.52 per cent.
Elongation: 20 to 22 per cent.
Ultimate tension: 116,000 to 136,000 pounds per
square inch.
Ultimate compression: 110,000 to 136,000 pounds
per square inch.
The strength of the welded joints ran from 62,000
to 86,000 pounds per square inch in tension.
The first article weighed 54 pounds and failed under
75 per cent of the design load, the failure (see fig. 13)
being in the first bay outboard of the outer point of
support, where the breaking of a diagonal wire re­sulted
in the complete rupture of the cantilever tip.
Figure 14 shows the dimensions of the first article and
the sizes of members. It will be noted that this spar
was tapered somewhat, whereas the second article,
shown in Figure 12, has parallel chords. This type
of spar was originally designed by the Boeing Co.
for use in a tapered wing, so the first of the test spars
was tapered. The taper was omitted from the second
article, to make it comparable to the other test spars
that had parallel chords. This second article weighed
62.5 pounds and failed under 120 per cent of the
design load, the failure being caused by the buckling
of the chords near the inner point of support.
No extenrnrnete1 readings were taken on this spar
and no studies were made to check the observed de­flections
with those obtained by any analytical method.
Curves of comparative deflections of the first and
second articles under 50 per cent load are given h
Figure 15.
From the tests made on minor parts of this spar it
appears to be undesirable to weld highly heat-treated
steel tubes. Nothing is to be gained by heat treating
tubes which are to be used as Euler columns, since
the treatment has no effect on the modulus of elasticity.
For short columns or for tension members where the
working stress is liable to exceed the low yield point of
the material after welding nothing is to be gained,
since this y ield point is practically the same as that for
the annealed steel.
The fact that the yield point on annealed alloy steel
tubes is higher than that for welded mild carbon steel
indicates that some gain can be had by using a high
yield point unheat-treated alloy steel in a welded joint,
where such a joint can be made satisfactorily, but the
above tests indicate that no further gain can be de­pended
upon by heat treati11g the material before
welding.
Type 4.- COMBINATION STEEL AND DURALU­MIN
SPAR
The most satisfactory type of spar tested, when
rated entirely on the ratio of load carried to weight of
spar, was one having a built-up web system of mild
carbon steel tubes which were welded to steel saddle­shaped
plates to which duralumin tube chord members
were bolted. Figure 16 shows the spar in the test
jig and indicates the type of construction quite well.
The first article of this type weighed 61.5 pounds and
held 135 per cent of the required load when the fitting
at the outer point of support failed. The second
article weighed 50 pounds and carried the required
load, failure occurring at 105 per cent of the load.
Lighter saddles and lighter tubes were used in the
second article, which failed by buckling of the com­pression
chord near the outer point of support. Figure
17 shows how the saddles were beginning to fail at the
time the chord buckled, indicating that the spar was
working about at its limit.
Dural tubes
Extensometer readings were made on the first article
during test and compared with the computed stresses
in several of the members. The results are shown in
Figure 18, the upper values being those obtained from
the extehsometer readings, the lower from the compu­tations.
A study of these results shows that while the agree­ment
between computed and observed values is not
exact the quantities are in all cases of approximately
the same magnitude, indicating that the methods of
analysis are in sufficiently close accord with actual
conditions to be satisfactory for use in design, which
checks the conclusion reached from a study of the
results obtained on the welded 1020 steel tube spar
type 1.
A study of the deflections of this type of spar, how­ever,
shows that some difficulty may be encountered in
determining the deflections for use in computing the
secondary stresses. While they can be determined by
the use of a Williot-Mohr diagram, as is shown by the
curves in Figure 19, the effect 'of play in the pinholes
must be provided for in drawing the diagram, and
unless the correct play is assumed the results will be
of little value. Figure 19 shows three curves of com­puted
deflections in comparison with the observed
values and indicates that a small variation in the
amount of play assumed in the pinholes has a compara­tively
large effect on the deflection of the spar. In
this case reasonably close agreement was obtained
when a play of 0.01 inch was assumed at each end of
each diagonal web member, but there is no reason ·to
believe that 0.01 inch would give satisfactory results
in every case. It is, therefore, difficult to determine
the secondary stresses in a spar of this type. Due to
the heavy cantilever load, which reduced the deflections
i11 the span in this case, these secondary stresses were
small and had little or no effect, but for the usual
conditions existing in a wing the deflections would be
relatively great a11d would have to be provided for by
some assumption as to the play in the pinholes. The
result of such an assumption is to add material more or
less arbitrarily to provide for the secondary stresses
with a consequent increase in the weight of the spar.
Because of the difficulty of analyzing a spar of this
type, and because of the probable effect its large
deflections will have on the aerodynamic qualities of
a wing, it can not be rated as highly as some of the
types which lend themselves to more exact analysis
and deflect less, although it should be emphasized that,
a very light spar can be developed in this type of con­struction
by more or less cut-and-try methods of build­ing
and testing until the optimum sizes are obtained.
Figure 20, which gives the curves of deflection for the
first and second articles, indicates how greatly the stiff­ness
is reduced by the lighter members used in the
second article.
Tests on minor parts gave the following properties of
the materials:
Steel tubes
Proportional limit_ ______ ____ _______ 21,000 pounds per square inch ______ _ ,_
Ultimate tensile strength ___ __ __ ____ 57,700 pounds per square inch ____ _____ 62,800 pounds per square inch.
Modulus ofelasticitY-- - --- --- - ----- 12,825,000 pounds per square inch ____ Load to break welded joint_ ___ _____ --------- -- - -- ----- -- -- - ------------- - -- 54,250 and 58,600 pounds per square inch.
Efficiency of weld ____ ______ __ ____ __ --- --- -- ---- --- ------ -- - ---- --------- --- 86 and 93 per cent.
84520-26--2
One joint failed in the saddle just at the edge of the
weld, the other in the tube a .short distance from the
weld.
Type 5.-WELDED DURALUMIN TUBE SPAR
Figure 21 shows the welded duralumin tube spar sub­mitted
by the Douglas company. The spar weighed
56 pounds, hut it did not stand up satisfactorily during
test. It held 50 per cent of the design load and failed
as the jacks were being let down under the next incre­ment
of load, 25 per cent of the design value. One of
the welded joints pulled away in tension, as is shown in
the photograph, Figure 22, and upon examining the
remaining joints carefully it was found that about half
of them showed cracks in the welds similar to those· in
Figure 23. Figure 24 gives the construction details and
sizes of members used in this spar. ,
8
Because of the pronounced failures in the joints of
the first article under a load between half and three­quarters
of that for which the spar was designed, it was
decided that welded duralumin tube spars were not
practicable and no second article of this type was built.
The failures in the welds are explained by the fact that
the welded joints were stiff and had a very considerable
influence on the web members, causing them to bend
and maintain a copstant angle with the deflected
chords. The resistance of the web members to this
bending tendency was sufficient to overstress the welds,
which were relatively weaker than the surrounding ma­terial,
and caused them to crack. Due to the low modu­lus
of rupture of duralumin, which results in larger de­flections
and distortions than are encountered in steel
structures of the same type, this effect is quite pro­nounced,
and it is not thought good practice to weld
duralumin tubes as was done on this spar. Figure 25
shows the curve of deflection of this spar under 50 per
cent of the design load. A comparison of the deflec­tions
in the span of this spar with those of the welded
steel spars will show the difference in stiffness of steel
and duralumin.
No extensometer measurements were made on this
spar, nor were any studies made to compare observed
and computed deflections.
Type 6.-BUILT-UP DURALUMIN SPAR,
LIGHTENED WEB TYPE
Figure 26 shows a type of built-up duralumin spar
having lightened web plate and C-shaped flange
members which were formed from flat sheet stock.
All joints were riveted. Figure 27 shows the construc­tion
somewhat more clearly and gives the dimensions
of the various details.
These spars were designed and built by the L. W. F.
company. The first article weighed 747':! pounds and
failed under 95 per cent of the design load. The
failure occurred about 20 inches inboard of the outer
point of suppor..t, the lower flange buckling and shearing
the cover-plate rivets, as shown in Figure 28. A
heavier cover plate was used at this point on the
second article, which weighed 74~ pounds and which
successfully carried the design load. Due to the large
deflection of the cantilever tip, no further increments
of load were applied to the outer platform but the loads
in the span were increased. The spar failed, as shown
in Figure 29, when a 10 per cent overload was applied
to these platforms. It will be noted that the failure
occurred at practically the same point as in the first
article, but it was due to the buckling of one of the
verticals rather than a flange member. Figure 30
shows the failure in more detail and also shows how the
compres.sion diagonals of the web system buckled
under load. The same buckling action, though not
so pronounced, was apparent under 50 per cent of the
design load, and the question arose as to what would
happen to a wing spar of this type under vibration or
changing loads. Since the diagonals buckled under
half the ultimate load, or at a load which would repre­sent
the working load on an airplane spar, considerable
doubt was expressed as to the reliability of a structure
of this sort due to the possible fatigue failure of the
thin sheet metal in the diagonals, which were alternately
buckled and straightened out with every fluctuation
in load. While no d~finite statements can be made as
to the effect of this action, since 110 service tests on this
type of spar have been made, jt is a point of considerable
importance in t he design of spars of this type and its
possible effects should be borne in mind.
Another questior which arose in connection with
this spar was whether it would behave as a beam or,
due to its lightened web structure, as a truss. Ex­tensometer
readings were made at a number of points,
as shown in Figures 31 and 32, which give the plotted
results from these readings. An excellent set of lines
was obtained, and it was hoped to determine from them
whether the stresses in the chord members could be
better approximated by designing the spar as a beam
or as a truss. Table 3 gives a resume of the r esults
obtained and shows that it makes little difference
which system is used for actual design, since both ·give
approximately the same results and neither agrees
very well with the stresses indicated by the extenso­meter.
In computing the stresses as a beam the
moment of inertia of the web system was' neglected,
the I of the span being taken as that due to the chord
members only. In computing the stresses as a truss
the panel lengths were taken as 12 inches each and the
depth between centers of gravity of the chords as
10% inches. Both of these figures are approximate,
but will serve for this purpose. All stresses were
obtained at 50 per cent of the design load.
9
TABLE 3
Comparison of stresses in L . W. F . spar Computed stresses against observed
Member Extensometer reading
I j_ At Ol oad cAetn 5t 0l opaedr Deinffceer -
------ -- -
U1 and U2 ____________ 540 800 260 u,and u, _________ ___ ."80 890 310
Uo and 10-- - -------- - 545 938 393
U11 and U12 --- --- - ---- 515 798 283
Utiand U1s -------- -- - 610 868 258
U23 and U:?• ----------- 320 310 -10 u,, and u,, ___________ 860 520 -340
L1 and L, ______ ____ __ 365 588 223
L3 and L~ -- - ------- - -- 270 425 155
Lu and L12-- -- --- --- - 290 423 133
L,. and L,, __ --------·1 :i20 538 218
LJl and J..,3~ - - - ------ · - 300 605 305
It will be seen from a study of the results given in
Table 3 that neither system gives stresses equal to
those accorded by the extensometer, though both do
give stresses of about the same magnitude.
A study was made on the second article to ascertain
how well the observed deflections checked the com­puted,
and it was found that the deflections computed
in the ordinary way for a beam, using the moment
of inertia of the chords only, would not agree with
the test results, as will be seen from Figure 33. The
computed deflecti ons were obtained by the use of the
precise equations for beams under combined loads
the moment of inertia being taken as that of the chords
only, flanges and cover plates, the webs being neg­lected.
Even so, the computed deflections are only
about a third of the observed and show that a spar
having a lightened web system such as this does not
deflect as a beam having a solid web, but, clue to the
higher stresses existing in the vertical and diagonal
inembers, deflects to a much greater extent. Owing
to the com'plicated web system used, no effort has
been made to check the observed deflections assuming
the spar to act as a trus:;, but it is thought that much
closer agreement with the actual deflections would be
obtained by considering a spar of this t ype as a truss
rather than a beam.
Type 7.-BUILT-UP DURALUMIN PLATE GIRDER
Figure 35 shows the type of built-up cluralurnin
plate girder spar submitted by the Aeromarine Plane
& Motor Co. The weight of t he first article was 60Y2
pounds and it held 70 per cent of the design load,
failing under 75 per cent. The failure was caused
by lateral buckling of the compression flange. The
second article was strengthened by the extension of
the cover plates, and it failed under 110 per cent
of the design load after having successfully carried
105 per cent. The second art icle weighed 65 pounds.
Failure occurred at practically the same place in both
of these spars, it being a lateral buckliLg failure in
each case, as shown in Figure 36 for the first article
and Figure 37 for the second.
Figure 38 shows the general arrangement and details
of construction of this spar. The bulb angles used
Computed stress due
Total ob- Stress due Stress due
to bending
served to axial to bend-stress
load ing
As abeam As a truss
-6, 500 -5, 625 -875 -l, 100 -1, 270
-7, 750 -5, 625 -2, 125 -1,835 -1,900
- 9, 800 - 5, 625 -4, 175 - 3, 300 - 3, 175
-7, 050 -5, 625 - 1, 425 -2, 575 -2, 530
-6. 450 -5, 625 -825 - 370 -635
+ 250 -5, 200 -5, 450 +4, 975 +5. aoo
+8,500 0 +8,500 +6, 950 +5, 300
-5, 550 -5, 625 +75 +1, 100 +635
- 3,875 -5, 625 + 1, 750 +t,835 +!. 270
-3,315 -5, 625 +2. 310 +2,575 +1,000
-5, 450 -5, 625 + 1n +1, 100 +635
- 7, 625 0 - 7, 625 -4, 975 - 5, 250
for flanges were extruded duralumin sections, the web
plate was sheet dural, and the web stiffener angles
were formed from flat sheets. This spar was especially
interesting structurally, as the proportions between the
more important dimensions were determined upon by
an extension of the methods used for steel plate
girders.
This spar was only 6Y2 inches deep instead of 15,
as were most of the other spars t ested, but it showed
up remarkably well and was not unduly heavy. The
depth used was in the opinion of the designer the
optimum, as any greater depth would have increased
the volume of the webs and stiffeners required and
rn increased the total weight. The flange angles were
about the minimum size that could be used without
reducing the lateral rigidity to point where failure
would occur by buckling under less than the design
load, so that little could 11ave been gained by lighten­ing
them. They were, moreover, about the minimum
thickness that could be extruded. This spar should
be admirable for a heavily loaded thin-wing airplane
where the truss types, either steel or dural, would be
impracticable and very heavy.
At the time the first article was tested a paper by
Mr. S. Timoshenko was presented in the 1924 Trans­actions
of the American Society of Civil Engineers on
"Beams without lateral support." A method was
given there for determining the stress causing a beam
without lateral support to fail by buckling, bttt no
loading condition similar to that used in these tests
was provided for and the method was so condensed
as to be difficult to extend to the test conditions.
Mr. J. Prescott discusses the stability of thin plates
in his book on "Applied Elasticity" and derives ex­pressions
for the critical loads on beams under certain
conditions, none of which are similar to those used
in the test. It is probable tha t either Timoshenko's
or Prescott's methods could be extended to provide
for the loading conditions encountered in an airplane
spar, but this has not been done as yet. There is,
therefore, no method which can be offered here whereby
the critical loads on a girder or other beam may be
determined. But as practically all metal beams fail
by lateral buckling of the compression chord, the
designer must give this subject his consideration and
10
make such prov1s10n as his judgment and experience
dictate to prevent this buckling action in spars of
this type.
Tests were made on minor parts of the first article
and the following properties were obtained for the
duralumin:
Modulus of elasticity: 10,400,000 pounds per square
inch.
Tension yield point: 29,500 pounds per square inch.
Ultimate tensile strength: 57,000 pounds per square
inch.
Elongation in 2 inches = 27.75 per cent; in 8 inches=
18.9 per cent.
Ultimate shear strength : 42,000 pounds per square
inch.
The specimens used in the above tests were cut
from the extruded flange angles. The deflections of
the first article were computed by the precise method
for obtaining the stresses in and deflections of beams
subjected to combined axial and lateral loads to de­termine
whether or not this method would give as
satisfactory results when applied to a metal spar as
it does with wooden ones. Figure 39 shows the
curves of deflection for this spar as computed for 50
per cent of the design load and as observed for this
load during the test. It will be noted that the agree­ment
between the observed and computed curves is
excellent throughout the entire span, indicating that
the precise equations will work satisfactorily for
designing spars of this type.
Figure 40 gives the curves of deflection observed on
the first and second articles under 50 per cent of the
design load. It will be seen that, due to the shallow­ness
of this spar and the low modulus of elasticity of
duralumin, the deflections are large as compared with
some of the other types of spars tested. The result
of this is to increase the secondary bending stresses
in the span and so tend to make it heavier than would
. be necessary were ·it stiffer. Since this type can not
be made deeper economically it does not seem so
desirable for use in a thick wing, but it most certainly
is for a thin-wing airplane.
A great advantage of this type of plate girder spar
is that it is not made up of a multiplicity of small
elements, the failure of any one of which will result in
the failure of the entire structure, but it is composed
of three units, two chords and the web, and the failure
of tire spar entails the failure of one of these main
structural elements, each of which is capable of quite
exact analysis and none of which is liable to be stressed
or cracked by forming or injured by the process of
fabrication. For these reasons it should be possible
to develop a spar of uniform strength and depend­ability,
since no reliance is placed on variable welds
or light riveted joints in thin material.
Type 8.-WOODEN BOX SPAR
In order to have a satisfactory criterion for judging
the merits of the various types of metal spars which
have been described above a spruce box spar was
designed, built, and tested by the Engineering Division.
It will be noted from Figure 41, w'hich shows the de­tails
of construction of the second article, that no
effort was made to remove the last possible ounce of
weight from this spar by tapering the flanges or varying
the depth at any point. Figure 42 is a photograph of
the first article, which weighe~ 66 pounds and failed
at the points of support under 85 per cent of the design
load. Figure 43 shows the failure at the inner support
clearly. While the failure appears to be due to shear
through the filler block, which was of spruce and had its
grain parallel to the chords of the beam, it actually was
a combined shear and bending failure. Due to the
depth of the filler block, about 12 inches, it bent quite
noticeably under 50 per cent of the design load, and, as
the tensile strength of wood at right angles to the grain
is very small, the fibers separated. Cracks then
started on the tension side of the block and progressed
·toward the bolt holes, thus weakening the block and
permitting it to fail in shear.
The second article was designed with plywood filler
blocks to obviate such failures, and it is interesting to
note how much the blocks were reduced in weight "yet at
the same time increased in strength. The diaphragms
were lightened somewhat, having been solid in the first
article, but the flanges and webs were kept the same in
the second article, which weighed.6072 pounds and failed
just as the jacks were released from the platforms under
the design load, the failure occurring in the compression
chord about 9 inches inboard of the outer support.
In developing the filler blocks for the second article a
block was made of quarter inch. spruce laminations the
grains of which were alternately vertical and horizontal,
and this was glued to short sections of chord to simulate
the conditions in the beam. The detail was then tested
under a load parallel to the chords and showed a yield
at about 30,000 pounds, failure at 34,500 pounds. The
block in the first spar, which was not plywood, yielded
at about 10,000 pounds end load during the test and
failed under 17 ,000 pounds. Since the required
axial load was only 20,000 pounds the design of the test
block was revised, reducing it in size. A second end
detail with the smaller block was built and tested. It
yielded at 20,000 pounds and failed at 28,600 pounds,
so it was regarded as a satisfactory size for use in the
second article. This small block weighed about 80
per cent as much as the original solid one, yet it carried
about twice as much load. The use of plywood blocks
therefore appears desirable when the axial loads in a
deep spar are heavy, subjecting the block to bending as
well as shear. Tests of minor parts obtained from the
second article showed the following properties for tlie
spruce in the flanges:
Lower
chord
Upper I As used in
chord I design
Moisture content _____________ ___ _________ ________________ _______ ____ ____ __ _______ ___ · ___ ________ ___ _______ _ 11. 18
0. 379
9, 310
11.42
0.380
8, 925
10. 0
Specific gravity ___ ------------------------- - -------- ____ ____ __ __ __ ___ ___ _________ ______ ___ ___ ___ ____ _______ _
Modulus of rupture, pounds per square inch . ... -- ------- - - · ---- - --- ---- -- -- ---- -·- -- - --- ---- -- - ---- - ---- --­Modulus
of elasticity, pounds per square inch·--- --- ---- -- --- ---- --- -- ---- --- ----- ---- --- - --- --- -- - - --- - ---­E;
lastic limit in bending, pounds per square inch.------- -- --- --- --- --- -- -- -- - ---- ---- -- ---- ---- - -- -- --- -----
Maximum compression parallel to grain ______ ___ ___ ----- - -------- - ---------- - -- ------- _______ : ____ ________ _ 1, 778, ()()()
5, 910
5,340
1, 782, ()()()
5, 625
5,230
10, 300
1,600, ()()()
5,500
11
The material was, therefore, about average and con­formed
quite well to the standard values used in design.
In order to have the failure of the first article occur
slightly below the design load the size of chords to be
used was determined for the critical section and the
thickness of each was then reduced one-sixteenth of an
inch. The same-sized chords were used in the second
spar, since no failures were developed in the first under
85 per cent of the design load. Since the critical point
was at the outer point of support these undersized
chord members were reinforced somewhat by the fill er
block, and the failure when it occurred was just inboard
of the end of this block.
The chords were, therefore, just about the right
size to carry the maximum loads, but they could have
been reduced in weight by tapering. The plywood
webs would have been reduced in thickness and the
whole. cantilever portion would have been lightened
by being tapered, so that this spar could probably
have had from 3 to 5 more pounds taken out of it
and still passed the test.
Considering this fact, it is interesting to compare
this wooden spar with the rnetar ones both as to
strength and weight and as to deflection. Figure 44
gives curves of deflection of the first and second articles
under 50 per cent of the design load, and a comparison
of these curves with those for the metal spars will
show that the wooden spar is stiffer than any of the
others of equal weight, except the chrome molybdenum
spar, which it just about equals. As for strength, it
shows up very well in comparison with the metal
spars, much better, in fact, than was anticipated when
it was built. The fact that this wooden Rpar compares
so favorably with t he metal ones throws considerable
doubt on the claims of the all-metal spar enthusiasts,
who state that either steel or duralumin structures
will be inherently much lighter and stronger than wood.
It is to be noted that all of the spars tested were of
fairly heavy gauge material and that no crinkling
failures were noted in any of the members during test.
The use of such material permitted the designers of
these spars to analyze their structures and use tubes
or other sections to carry the computed stresses with
reasonable assurance that there would be no local
failures due to crinkling. It may well be, therefore,
that spars considerably lighter than this wooden one
or than any of the metal types tested could have been
obtained if each member had been made of light-gauge
material locally reinforced against crinkling failures.
Such members can not be analy:r,ed satisfactorily but
must be developed by cut-and-try experiments, which
are extremely costly and which can not be guaranteed
to produce a light, stiff spar in every case. The
question therefore remains as to whether a lighter
spar can be built in metal than in wood when rigidity
of construction is included, as it must be, in the r e­quirements
for a satisfactory airplane wing spar.
Figure 45 shows the curve of deflections of the second
article as computed by the precise formulas for beams
in combined bending and compression in comparison
with the observed deflections. The agreement between
the curves is quite good, especially in the middle of
the bay, and it indicates that these formulas will give
satisfactory results for computing the stresses in spars
of this type. Due to the small deflections in tqis case,
the error indicated by the discrepancies between these
curves is relatively large but should not affect the
stresses appreciably.
RATING OF THE TYPES TESTED
Figures 46 and 47 show curves of deflection for the
first and second articles of each type of spar under
one-half the design load. Figure 48 shows the same
data for the second articles under 95 per cent of that
load, the highest load at which data .on all types
were obtainable.
It is fair to assume that the first articles were purely
experimental from ·the point of view of the designers
and were built to fail at a load below the required,
so that the second articles could be revised and built
to sustain the design load. As such they are hardly
comparable and so will be neglected, the ratings of the
various types being made on the second articles alone.
In rating these spars three things will be considered,
the strength, the weight, and the deflection of the
spar. The first comparison will be made on the basis
of the strength to weight ratio. since weight is extremely
important in the design of airplane spars. A second
comparison will also be made on the basis of strength
to deflection, stiffness being an essential in a successful
spar to maintain the form and aerodynamic character­istics
of the wings. The final rating will be made by
combining these two ra tios and comparing the spars
on the basis of strength to weight times deflection.
The following table gives the values of these various
ratios :
TABLE 4.-Comparison of experimental spars
50 per cent design load 95 per cent design load Maximum load without failure
------ 1-------~--~-~~-
8 ·~ I .;; i·§ I ;; .§
""' ~ 0 ""' 0. 0 s.. 0. c::
Type or construction ::; ~ 0 -5 ~ ~ 8, -E 5 ~ ~ ~ .a -5 ~
c I ~ -0 ~ , ~ ~~ ~ :::.. I ~ 1= ~~ ~ I ::. c::I IC ~~
z I o .c .... ..c - bOI.. :-:.. ~ o - ... ... ... bO... ... 1... o ... ... ..c: ... bO... ... -
.... .c · ..... g -5~ ex 0 ° I.cl .!? -5 .S? l:X c::i o .::: 1 .... 2 5 .g f::X
:_, 'O ~ ~ -~ g ~g V.1 1~ 'O £ \ ~~ ~ gg (/) \~ 'O ~ gfjl•~ ~ §~ {/) ~ t ! .3 ~ I ~ I~ ~ I JE:~ ~ .3 ; ~ ~ ~ }E ~ I~ .3 ~ ~ ~ ~ ~ ~ ;
i -~-:-l-~:-~-1c~-~0-s-~-~e-11-0~-. 1 -3-.-50-0 -74. 001 -:;:;~ o 05 70,000 945 6, 650 74. 00
1
00
1
o. 09 74, 000
1
1, 0001
1
8, 050 74. 00 ·~ o. 21 ~8,400 ~
3
H~~~.~~~a~:.':i st~~~kei - 3, 500 60. 00 58. 3 . 07 50, 000 834 6, 650
1
60. 00
1
101 . 15 44, 4001 740 7, 000 60. 00 117 . 21 33, 300 555
steeL .. --- - - --·-- -· 3, 500 62.501' 56. 0 . 35 10,000 160 6,6.50 62.50 106 .65 10,200 164 7,700 62.50 123 . 76 10,100 162
4 Steel apddurai__ ___ __ 3,500 50.00 67.31 .28 12,500 250 6,650 50.00 133 .54 12,300 246 6,650 50. 00 133 .54 12,300 246
5 Welded d uraL ____ __ _ 3, 500 No second article of this type built
6 Built·upduraL _·- ··- 3,500 74.701 46.81 .241 14,6001 1951 6,6501 74.751 891 .561 11 , 9001 1591 7,350, 74.751
7 Dural girder_·-· - - - - -- 3, 500 65. 00 53. 9 . 35 10, 000 154 6, 650 65. 00 102 _ 65 10, 200 157 7, 350: 65. 00
8 1 Spruce box. ------ - ·-- _ 3,5~50 57.9 .07 50,0001 826 6, 650
1
60.50 110 . 13 51,100 846 6,6501 60.50
9s: . 15 I 9, 800!
1131 ' . 73·, 10. 100
110, . 13 51, 1001
131
155
846
1 This deflection was obtained by taking 105 per cent of that for 7,000 pounds total load, the maximum at which deflections could be taken due
to distortion of this spar.
12
For a satisfactory comparison of the spars on the
basis of strength to weight it is necessary to consider
the maximum load condition on the spar, since the
weight is a maximum at all times. From Table 4 we
find the following ratios of strength to weight for the
maximum load on the spar.
Spar Ratio
type
R
. ..r Strengthb' . ·
ating o; spars on -W~eig h asis
Type of construction
Ratio
to box
spar
------- !------~----------!--
4 133 Steel and duraL .. ---- ---- ---- -- ---- -- -- ------ -- - 1. 21
3 123 Heat-treated nickel steeL-- -- -- ---- -------- ------ l.12
2 117 Welded chrome molybdenum steel.____________ _ l.06
7 113 Dural plate girder------------ ------------ - -- -- -- J. 03
8 110 Spruceboxspar. ___________ _________ __ __ _______ _ 1.00
1 · 109 Welded 1020 steeL_ ___ __ __ __ __ __ ___ ______ ____ ___ . 99
6 · 98 Built-up duraL. - -- ----- -- -------- ------ -- -- -- -- . 89
The combination steel and duralumin spar stands
out quite a . ways in advance of the other types, with
the heat-treated nickel steel spar a fairly close second.
Any one of the next four types could fill third place,
there being very little to choose between them on the
strength-weight ratio basis.
If we consider stiffness instead of weight as our cri­terion,
the rating of the various types is as follows:
Rating of spars on Streng~h basis
Deflection
Spar Strength
type Defiection Type of construction
Ratio
to box
spar
8 51, 100 Spruce box spar ... ---- -- ------ -- ---- -- ---- - I. 00
1 38,400 Welded 1020stee!____ __ __ ______ ____ _______ __ . 75
2 33, 300 Welded chrome molybdenum steel.. __ _____ . 65
4 12,300 Steel and duraL---- ---- ------ ------ -- - - - - -- . 24
3 10, 100 Heat-treated nickel steeL ---------- -- ------ _ 20
7 IO, 100 Dural plate girder __ ____ __ __ __ ______ __ __ __ __ . 20
6 9,800 Built-up duraL ______ ____ ________ ___ __ _____ . 19
It is interesting to note that the wood box spar is
the stiffest of all of the types. This is probably ac­counted
for by the fact that the web of this spar was
solid and not subjected to high stresses. A study of
the above table indicates that the stiffness of the web
system is a factor of the utmost importance in deter­mining
the stiffness of a spar. The larger and more
lightly loaded tubes in the web of the welded 1020. steel
spar result in a stiffer structure than the welded chrome
molybdenum spar. The large difference between this
latter t ype and the combination steel and duralumin
spar; both of which had web members of approximately
the same size, is due partially to the lower modulus of
elasticity of the duralumin chord members, but chiefly
to the play in the pinholes at the points of attaclJing
the chords to the webs. This play in the joints also
exists in the heat-treated nickel steel spar, and, when
combined with th e elongation of the highly stressed
wires in the web system and the equally highly stressed
alloy steel chords, gives a spar that is more flexible
than the combination steel and duralumin type. The
two duralumin spars proved to be the most flexible of
those tested, as was expected, but it is interesting to
note the shallow duralumin plate girder with its solid
web was stiffer than the built-up duralumin spar, which
was approximately twice as deep but which had a
lightened web system.
Assuming weight and deflection as factors of equal
importance in rating a spar, we find that the order
of merit of the different types changes but little.
. . Strength
Rating of bpars on WeightXdejiection basis
Type of construction
Spar Strength
type WeightXdeftection
I Ratio
to
box
spar
8 846. __ -- -- - -- ---- ___ Spruce box spar_ _______ _ . __ ___ __ __ I. 00
2 555 ___ _____ _________ WeldedchromemolybdenumsteeL 0.66
1 518 ____ ______ _______ Welded 1020steeL__ ______ __ ____ __ 0.61
4 246 __ ____ ___ __ __ __ __ SteelandduraL_ .__ __ __ __ __ ____ __ 0.29
3 162 ___ __ ____ __ __ ____ Heat-treated nickel steel. ______ ___ 0.19
· 7 155 __ ______ ____ _____ Dural plate girder___ ____ __ ____ __ __ 0. 18
6 131. ______ ____ __ ____ Built-up duraL __ __ ____ __ __ __ _____ 0.15
I
Omitting the wooden spar from consideration, we
find that the chrome molybdenum steel tube spar is
·lhe best of the metal types when weight and deflection
are given equal weight. The welded 1020 steel spar
is then a very close second, and the combination steel
and dural spar a rather poor third. It is, however,
hardly jus.tifiable to consider deflection of the same
importance as weight in an airplane spar, although it
is an item of great import since it affects the aero­dynamic
properties of the airplane in flight.
It is unfortunate that the welded 1020 steel spar
was so far over in both strength and weight as compared
with these other two, because it is as difficult to say
just how much its weight could have been reduced
and still had it carry the required load as it is to say
how much its stiffness would have been affected by
removing this material.
It seems fair to say that its weight could not be
reduced by 14 pounds to put it on a par with the
chrome molybdenum type without reducing its strength
below the required and greatly increasing its deflection
And there is practically no chance of its weight being
reduced to that of the steel and duralumin spar and
still having it carry the load. For this reason the.
welded 1020 steel spar is not rated as highly as either
the chrome molybdenum or steel and duralumi11
types for use in an actual airplane where weight is o
greater moment than deflection.
As for the merits of these latter types, it will be
noted from the data in Table 4 that with a reductio1
of about 16 per cent in weight the steel and duralumir
spar deflects about four times as much under working
load, or under 95 per cent of the design load, as the
chrome molybdenum type, and it would, therefore
not be so desirable for use in a wing structure. It is
<lifficult to judge between t hese types, as it is bard
t.o decide on the relative importance of weight and
stiffness, but it is believed that the chrome molybdenurr
type, which may be analyzed more accurately thar
the steel and duralumin type, which is not susceptible
to corrosion due to the electrolytic action between
two different metals, and which is considerably stiffer
than the latter type with its bolted 'oints will be more
13
satisfactory for use in an airplane structure and it is,
therefore, given the highest rating of the metal types,
with the combination steel and duralumin spar a
very close second.
Of the other types little need be said, but it is worth
while to note that the shallow plate girder type with­stood
the required load without undue deflection and
without being unduly heavy . For a shaUow wing
carrying loads of the magnitude of those for which
these spars were designed this is the only one of the
t ypes tested that could be used profitably, since the
weight of all the others would increase rapidly as the
depth was reduced.
Of the seven types of metal construction tested, then,
three are worthy of especial note. The ·chrome
molybdenum steel tube spar is probably the best,
with the combiJ1ation steel and duralumin type second,
where a depth of approximately 15 inches is available.
Where such great depths are not available, but where
heavy loads must be provided for, the plate girder
t ype of spar seems to be the best type available, as
it is readily analyzed and it is relatively light and
stiff:
ADDENDUM I
The t ype 4 spar was straightened out after the test
of the second article and the duralumin lower chord
was replaced by carbon-st eel tubes of the same size
to determine the effect of the differential expansion of
the two metals when the spar was subj ected to changes
of temperature. The spar was supported at the same
points as during the static tests and deflectometers
were arranged to measure the change in deflection at
the third points of the long span with changes of
temperature.
Observations a re stiJI being made, but the indications
of a few preliminary results are that while a change of
temperature does make the spar deflect a measurable
amount it is hardly sufficient to cause serious stresses
in a spar of this size. The deflection at the third
points over a range of 30° change in temperature
appears to be between 0.00125 and 0.00150 inch per
degree Fahrenheit. For a change of 150°, which is
about the maximum liable to occur in an altitude
flight, the deflection due to temperature would be
about 0.2 inch, which, with the maximum axial load
of 20,000 pounds, would give a secondary moment of
4,000-inch pounds, a negligible amount on a 15-inch
section when compared with the primary moment of
120,000-inch pounds in the bay, but an appreciable
amount compared with the secondary moments caused
by the deflection of the spar.
It is concluded, then, that the effect of a change in
temperature on spars built up of two materials is not
liable to be serious, although it does cause a measurable
deflection and its possible effects should be taken into
consideration when computing the secondary stresses
for designing such spars.
ADDENDUM II
The method of rating the spars used on the foregoing
pages is admittedly not a satisfactory one, and further
consideration of the problem has led to the following
method, which, while not perfect, is felt to offer a
better comparison of the merits of the various types.
The method is to give the ratio on the strength to
weight basis (see p. 12) three times the weight of
that on the basis of strength to deflection, the two
ratings so obtained being added and divided by four
to get the final rating. This gives the following
ratings:
Spar I type Type of construction
iY Spruce box spar ___ ______ _________ ____ _ _
4 Steel and durnL ------ - -- ----------- -- - -
2 Welded cbrome molybdenum steel- __ _ _
l Welded HY.!O steeL --- - --- - ----------- --
3 Heat-treated nickel steel. __ __ __________ _
i Dural plate gird er_ ____ __ ___ ___________ _
6 Built-up dumL. ------ --- --- - -------- -_
4.00
3. 87
3 . ~1
3. 72
3.56
3. 29
2.86
I. 00
. 9i
. 96
. 93
. 89
. 82
. 71
----------------
The above list, in which the metal types are all
compared to the standard wooden spar, appears to.
give a much more satisfactory idea of the merits of
the various types, the range of the ratios being more
reasonable. While this new method of rating gives
slightly different positions for one or two types of spar,
the differences are so slight that it is not thought
necessary to revise the general conclusions given on
page 12.
Part II.-COMPARISON OF 11-FOOT METAL RIBS
RESULTS AND CONCLUSIONS
It is impossible to say that as a result of t his series
of tests on 11-foot metal ribs any one type ·of con­struction
stands out noticeably ahead of the others.
So far as picking an optimum type of rib for the load
conditions specified is concerned these tests are of
little value, since every t ype but one carried con­siderably
more than ·the required load and was, there­fore,
heavier than need be. But as a means for
determining the difficulties to be met in designing and
building ribs of the types tested these tests were very
satisfactory.
They prove that on a rib of this size it is difficult to
reduce the amount of material sufficiently to get a
light rib without getting one that fs extremely flexible
laterally . In fact, most of these ribs carried from 50
to 150 per cent overload because they were made of
members of the minimum practicable size.
The tests also prove that welded duralumin joints
are too unreliable to be used even in rib construction,
since their efficiency varies from 20 to 80 per cent of
the tube strength.
They also show that it is difficult to split and flatten
duralumin tubes for riveted joints without anneaJin·g
them, because of the tendency for cracks to form and
creep up the tube while it is being worked.
These t ests show that satisfactory ribs can be
stamped out of duralumin sheet if some method is
used to give them lateral rigidity. If flanging is to
·be used the flanges must extend over the length of
all members, especially the gussets at points of con­nection
of chord and web members, or buckling of the
flat sheet will occur. The use of trussed chord mem­bers
having no web bracing connecting the upper and
lower chords, as was done in the type 7 ribs, is shown
to be undesirable for ribs of this size, as it resuits in
the use of such small members that the resulting rib
has very little lateral stiffness.
The welded steel tube ribs t ested demonstrated
t hat even with the smallest practicable sizes for the
members the resulting rib would be much t oo Rtrong,
and al so too heavy to give satisfaction in an airplane
of this size.
METHOD OF CONDUCTING TESTS
Six ribs of each t ype were purchased, and in most
and its center of pressure at 41. 7 per cent. The design
load in high incidence was 750 pounds, in medium 650,
each of these loads being distributed through a lever
system which applied 32 equal loads to the ribs, about
70 per cent being applied to the upper chord.
All of the ribs were designed for a Gottingen 387
wing section having a chord of 132 inches, the spars
being assumed at 20 and 85 inches from the leading
edge. The test data are, therefore, strictly comparable,
since all ribs were designed for and tested under
identically the same conditions.
Deflections were obtained on the upper cap strips
at a point midway between the spars.
DISCUSSION OF THE INDIVIDUAL TESTS AND
THEIR RESULTS
Each type of rib will be discussed separately as to
its behavior during test and as to its more prominent
structural merits or demerits. The data given are
taken from the following reports of t he Material Section
McCook Field:
M- 734. Metal ribs from Huff-Daland Co.
M- 771. Metal ribs from Airships, Inc.
M- 77la. Metal ribs from Eberhardt Steel Products
Co.
M- 876. Metal ribs from Boeing Airplane Co.
M- 906. Metal ribs from Aeromarine Plane & Motor
Corporation.
M-907. Metal ribs from Douglas Co.
All of the ribs t ested, except one set, were made of
duralumin. The results of the tests on the one set
of steel ribs will be discussed first, after which the
duralumin ribs will be taken up according to the t ype
of member or t ype of trussing used . In all seven
different t ypes were tested.
Type 1.- WELDED STEEL TUBULAR RIBS
8ix welded steel tubular ribs of the t ype shown in
Figure 49 were submitted by the Boeing Airplane Co.
Four of them were tested in medium incidence and
two in high , the results of the t ests being sho\.vn in
the following table :
Loading Rib
Maxi­mum
load
Weight
- ---:-- -- ---
De- 1 Loa~_ in p_ou_11_ds 1 · i .
flee- Weight in ounces Fa lure Ill
t1on 1 · J
. 1'- - ---
cases all six were tested, three under the standard M edium
high-incidence loading condition, three under medium i n c i-
Pounds O·unces fnch es
1, 100 68. 1 0. 160 16. 15 Member de.
incidence. The loading used for the high-incidence deuce.
condit ion is represen ted by a smooth curve from the
leading and t railing edge (see fig . 49), the center of High in­pressure
being a t 30.5 per cent of the chord. The cidence.
2
3
4
5
l,300
IJOO
1,200
1, 375
66. 3 . 157 19. 60 Do.
. 101 n45 Do.
. 140 Ji. 90 Do.
. 140 20. 55 Do.
medium-in cidence loading is represented by a triangular I --------'-----------'------- - - - --
load curve having its vertex at 25 per cent -of the chord 1 Deflect.ion of upper chord midway between the spars.
l.100 . 130 16. 40 Do .
(14)
15
The weight of the rib was obtained without the spar
sections shown in Fig. 49. These ribs had %-inch
0.035 gauge top and bottom chords with i\r-inch- 0.028
compression and -h-inch- 0.028 tension members in
the web system.
In the medium-incidence loading the four ribs sus­tained
an average load of 1,125 pounds at failure, or
approximately 175 per cent of the required. In high
incidence the average load was 1,240 pounds, or 165
per cent of the 750 pounds required. These ribs were·
very stiff, as is indicated by the small deflection, but
the average weight was unduly high and the strength
weight ratios were low for metal ribs of this chord and
depth.
The failure in each of the ribs occurred in member de.
It is probable that these ribs could be reduce<! slightly
in weight if the compression members adjacent to the
spars were made smaller and changed from compression
to tension members.
This rib is not very satisfactory as a whole, as it is
much too heavy, and the fact t hat most of the tubes
used in it are of the minimum sizes practicable for
use in a large rib precludes any great reduction in
weight with revision of the design. These ribs are
very stiff, however, which would have the advantage
that the airfoil section would not be changed under
load, but this is ·a somewhat doubtful advantage, as
t he distortion of the a irfo il under normal flying loads
would probably not have any material effect on t he
qualities of an airplane.
Type 2.- DURALUMIN TUBE RIB, RIVETED
JOINTS
Figure 50 shows the type of ribs submitted by the
Huff-Daland Co. These ribs were constructed of
one-half inch_ 20 gauge duralumin tubing held together
by one-eighth inch dural rivets at each joint. The
upper chord member was spliced at the panel point
just behind the front spar and the lower chord at the
lower end of vertical UV. These joints were riveted­butt
joints, a solid aluminum plug being inserted in the
tubes to carry the stress across the j oi nt. This plug
failed in some of the ribs during test, although such
failure did not occur until about 100 per cent overload
had been obtained. They did, however, show up as
weak points in the ribs, and it is suggested that care
be taken in designing such r ibs to locate the splices at
points of minimum stress.
The following table gives the results of the tests on
these ribs. It will be noted that the test specimens
are designated as of class A or B, the difference being
in the method of making the joints in the web members.
The ends of the web members of the class A ribs were
simply split, flattened, and riveted. Difficulties were
encountered in this process due to the split creeping
up the tube while it was being flattened and riveted,
so a change was made on three of the six ribs submitted.
The last three, the class B ribs, were made by annealing
the tubes first b.v dipping the ends iu melted zinc, after
which they were split and worked with very little
difficulty.
TABLE !.- Summary of tests on 132-inch duralumin tu be ribs
Ribs, class, I ~:x and No. tests
Load
!l'fe­dium
inci­dence
High
inci­dence
Dellec- . Load,
tion 1 Weight' Weight Failure '.~Iteration
-~---1--- ----~~ --- --- --- 1------------------1------------~
Pounds Pounds Inch Ounces
AL I 1 800 0. 114
2 1,230 . 260
1, 000 . 202 43. 9
1,300 . 260
A-2__ __ ___ __ f I 700 .135 } 43 g {
l 3 I, 250 . 243 .
A-3... ___ ___ { ~ 700 I, 585 ·--~~~~_I} 44. O {
B-1... . . • .•• { ! 700 ··1;110· : ~ I} 43. 5 {
{
1 700 - · · ·-- - - . 137 } {
B-2___ __ __ __ 2 . I, 230 - · ··---- · 280 43. 9
3 1,075 ·· ·· - · - · · ··- ·-· ·
{
1 - - . . . . . . 800 . 132 } {
B-3.. . . . . . .. 2 1,360 . . . . . . . . -·----- · 44. o
I, 275 ••• ••••• - - - - - -·-
18. 2
28.0
22. 8
29. 6
15. 9
28. 5
15. 9
36. 0
16. I
39. 3
15. 9
28.0
24. 5
18. 2
30. 9
29.0
None; carried to design load ___ ___ ___ ______ __ _
Shear of rivet at junction of member XY and
upper chord.
Sameasinsecondtest ______ ___ _______________ Ys inch steel rivets substituted in
Tension in one of the split ends of XY near up­per
chord.
None; required load __ _____ ____ _____________ _ Filler piece at joint in member DT adjacent to
panel point.
None; required load . . · -- - -- --- -------- ------­Rivetattaching
clip oftope-off spar to the rib ..
None; required load . _-- - - -- --- --- - - --- - ----··
Filler piece at Joint in member DT .. .. ... . .. . .
None; required load ._ -- -- --- -- -- -------- -··· ­Shear
of rivet at junction of member XY and
upper chord.
$hear of XY at upper chord _____________ __ ___ _
None; required load ___ _____ ___ __________ ____ _
each encl of member XY.
Same as above (third test) .
Ys-inch steel rivets substitntecl in
each end of member XY.
Shear of rivet at junction of member XY and
upper chord.
Filler piece at joint in member DT .... .. ...... Ys-inch steel rivets substituted
each end of member XY.
1 Deflection of upper chord midway between wing beams. 'Weight of rib including wing-beam clips.
These ribs were quite consistent as to weight and
also as to strength, and appear to offer a reasonable
solution of the problem of designing a rib of this size,
though they did carry a load considerably greater than
the required. It is interesting to note that little
difference in strength was to be found between the
84520-26--3
class A and class B ribs, but it is felt that in this regard
these tests did not s imulate service conditions. It is
practically certain that the unannealed tubes would
crack under vibration in service and prove generally
unsatisfactory, while those having the annealed ends
would probably prove quite satisfactory.
16
Type 3.- DURALUMIN TUBE RIB, WELDED
JOINTS
The Douglas Co. submitted six ribs, of which the
chord members and the compression diagonals were
made from one-half inch 22 gauge duralumin tubing,
and the tension diagonals of three-eighth inch 2.2 gauge.
The joints were welded . Figure 51 is a photograph of
this type of rib.
The following table gives the data obtained during
the t est of these ribs :
Loading conditions Rib Maximum
No. load
Pounds
Medium incidence ___ __ ____ _____ _ 1 !)60
2 1, 225
3 750
High incidence __ --- - - --- - - - - --- - 4 890
5 930
6 880
1 Exclusive of spars and bands for spar connections.
" feight I
Ounces
37.15
31. 05
30. 94
37. 60
31.18
34. 40
Deflection Load in pounds
Weight in ounces Failure
I nches
0.382
. 400
. 757
.163
. 250
. 250
25. 8 T ension in de at point of weld.
39. 5 Do.
24.2 Do.
23. 7 T ension in bf near panel points.
29. 8 T ension in de at point of weld .
25. 6 Tension in kl at point of weld .
No alterations were made on any of these ribs to
improve their strength, owing to the type of joint used.
These ribs are extremely variable, as a study of the
above results will show. The weights vary over a
range of about 20 per cent, while the strengths and
deflections show even greater variations. The fact
that all but one of the ribs failed adjacent to a weld,
three of the failures occurring at the same place but at
widely va.rying loads, led to the belief that the strength
of the welded joints must have been highly irregular.
Tests were therefore made on a number of the joints
with check tests on the tubes to determine the efficiency
of these joints. It was found that the welding affected
the tubes so that the joints developed only from 20 to
80 per cent of the strength of the original tube, with no
method apparent for telling just how much of the tube
was injured.
Phys'ical properties of clear and welded tubing from Douglas ribs
Location of test specimen in rib, see Fig. 51
Ultimate Efficiency Diameter of
strength tubing,
lb/sq . in . pereen tage 1 inch
E-g ___ · ---· -------- · --·-- -- -------- - --- -· -·--·-- - ---·- ------ -- -- -- -
E-k. __ · - - --- - -- · - - --·- · ------------- - - --- -- - -- · ---- - -----: ____ ____ _
Straight clear tubes, not affected by e-L - - ---------- - ----·- - --· --------- - - -- - --- ----·-··--- ·-- - -- ------ -
heat of welding. f-h __ __ ---- -·- ________ : ____________ ·- - - --·--·-- · - --- -·- ---
g-h ____ ._ - - d-e ____ ______-_-_ _· _--_-__-__-__-_-_-__-_·_-_·_-_-_-_-_-_ _-_·_· __-__- ·-. - -------· ·--- ·--- - •--. ·---·-·-- !Through joint at e-f and f-g to lower chord __ ____ _ _
Straight tubes with truss membeis Through io!nt at g-h, h-i, and i- k to lower cho1 cL _
welded to side. Through iomt at f-g and g-h to upper chord _____ _ . ---·------
Through joint at i-k and k-1 to upper chord__ ___ _ __ . .. ·- --·- __
Member k-i pulled from joint at lower chord __ ___ ______ ___ ________ _
!i-k at joint to upper chord·-- --- ··--·--------- ----- - ----·-- --·- -- - --
a-bat joint to lower chord ___ ___ ___ ·--- ___ ------ --- -- - - - _ -- · - - · --· -
b-c at joint to upper chord ___ · - - __ ___ ·-·- -- --·---- ------- ---- - --- - -
Members pulled from welded joints ___ m-n at joint to upper chord __ __________ ___ _____ _ , ____ _____ -------·
n-o at jomt to lower chord ___ --·- - - - -·- - -- - ---- ____ __ ___ ·--- ____ ·--
p-q at joint to lower chord __ __ -- -- - - - - ----------- - ----------------
<>-P at joint to lower chord ______ ____________ _______ - - - ----- _____ _
52, 100 --- ----- ---- o. 500
43, 800 --- --------- . 500
51, 600 --- ---- ---- - . 500
51, 300 ---- --- --- -- . 500
""· 700 --- --- ---- -- . 375
58, 900 ------ --- --- . 375
33, 500 G7. 4 . 500
31, 200 62. 8 .500
42, 300 85. 1 .500
40,360 81. 2 . 500
23, 470 41. 3 . 375
JG, 750 33. 7 . 500
26, 500 46. 7 '375
35, ()()() 61. 6 . 375
12,480 22.0 . 375
38,900 G8.5 . 375
26, 900 47. 4 . 375
20, 820 36. 7 . 375
1 Based on {49, 700 pounds per square inch for ~ inch tube} h. h 1 56, 800 pounds per square inch for Ys inch tube , w ic va ues are the averages from the tests on the unwelded tubes.
The strength to weight ratio in medium incidence
was about average for this size rib; in high incidence it
was low. Owing to the unreliability of the welded
joints and the .weight of these ribs they are considered
unsuitable for use in airplanes requiring ribs of this
size.
It is possible that the weight and strength of these
ribs could have been materially reduced by using a
lighter-gauge duralumiri for the built-up members, but
difficulties would probably have been encountered in
riveting the joints and in preventing crinkling failures
in the members.
Type 4.-DURALUMIN RIBS, HOLLOW
RECTANGULAR MEMBERS
Another type of rib (see fig. 49), made up of rec­tangular
members formed from sheet duralumin, was
submitted by the Boeing Airplane Co., and one of
the six was tested under the medium-incidence loading.
This rib was extremely heavy, 120 ounces, and so strong
as to break the test jig at 2,700 pounds load, or 415
per cent of the design load, without apparent injury
to the rib. It was obviously useless to test the rest
of these ribs, so no further studies were made on this
type.
It is not thought that these ribs offered a fair test
for this type of construction, and no conclusions as to
its merits will therefore be attempted.
Type 5.- DURALUMIN RIB WITH FLANGED
CHORDS, TUBULAR WEBS
The Aeromarine Plane & Motor Corporation sub­mitted
six duralumin ribs having chord members
which were flanged from fiat 0.035 duralumin sheet to
an L shape. The tubular web members, which were
flattened at the ends and riveted to the chords as shown
in Figure 52, were al\ % inch duralumin tubes ex-
17
cept members hi and op, which were Y2 inch alumi­num.
The leading edge was stamped from 0.025 dural
sheet.
Three of these ribs were t ested under the medium­incidence
loading and three under high. The weights
of the ribs were very uniform, averaging about 35
ounces, and the loads at failure in both conditions were
quite consist ent and quite high, so that the strength
to weight ratios were good. In high incidence the aver­age
load at failure was 1,250 pounds, or about 167
per cent of the design load. In medium incidence
failure occurred at 985 pounds, or 154 per cent.
The results of the tests are given in the following
table:
. Rib Maximum Load in pounds
Loading conditions No. load Weight t Deflection We ight in ounces F ailure
Pounds Ounces Inch
Medium incidence __ ____ _________ I 875 35. 2 ------ ------ 24. 9 Tension in lg at spar attachment.
2 J, 125 35. 2 0.300 32. O I Compression at op at npper chord .
3 950 35. 5 . 225 26. 8 Do.
' 3 l, 100 ---------- -- . 240 --- ---- - -- --- ---- - Tension at Qr at spar attachment.
High incidence __ ------··--·----- 4 1, 250 35. I • 296 35. 6 1 T ension in lg in flattened portion .
5 1, 300 35.1 . 272 37. 0 Compression in op in flattened portion.
6 1, 200 34. 0 . 263 35. 3 Tension in lg at upper chord .
t The weight includes vertical members er and rs, which are really parts or the spar.
' Rib 3 was repaired by replacing op with a heavier member and t est ed again to det ermine the second point of weakness.
The limiting members are seen to be either the ten­sion
members adjacent to the spars or the compression
member op, although both are greatly overstrength as
regards the required load. In ribs of this construction
the full strength of the diagonal or vertical members
can not be developed because of the strength lost in
the process of flattening the ends. Some of the tubes
developed cracks in the flat portion, but in any case
this part of the member is not as strong as the unde­formed
tube.
It would be somewhat difficult to reduce the weight
of these ribs very much, due to the fact that minimum­size
members are used at the present time. Heavy as
it is, however, this rib is quite satisfactory, as it is
rela tively stiff and it gives a structure of quite con­sistent
strength .
-
Type 6.- STAMPED DURAL RIB WITH FLANGED
CHORD AND WEB MEMBERS
Six ribs were built by the Eberhardt Steel Products
Co. which were stamped out of 0.037-inch duralumin
sheet. The edges of all members were rolled or turned
back to form V-shaped channel members, and as all
flanges were on one side of the rib the web members ·
were not concentric with the flanges, with the result .
that bending was imposed on the gussets, causing
them to buckle. Figure 53 shows the back view of
this rib and gives a good idea of this type of construc­tion.
Four of the ribs were tested in medium incidt>nce
and two in high, the results being given i11 the following
table :
Rib Maximum I Weight DeHeetion Load in pounds No. load Failure Weight in ounces
Loading conditions
--
Pounds Ounces Inches
Medium incidence. __ . ___ . _____ _ 1 1,500 62. 3 I 0. 493 24. 1 Member hi.
2 1, 830 62. 8 . 4J5 29. 2 Bending in flange just in front or rear spar.
3 1,425 62. 3 . 270 22. 9 Member hi.
4 1, 900 62. 4 . 472 30. 5 Do.
5 1, 900 62. 4 . 282 30. 5 Do.
6 1, 700 62. 6 1. 343 27. 2 Do.
It may be seen from the above data that these ribs
carried about two and a half times the required load
in each condition . They were, however, quite heavy,
. load
but due to their great strength the resultmg w~e1g ht
ratios are about average in medium incidence though
somewhat low in high incidence.
In all cases but one failures occurred in the gussets
at the end of member hi, indicating that such gussets
are a source of considerable weakness in ribs of this
t ype, since it is impossible to form wide, stiff flanges
on them. In the one case where failure occurred near
the rear spar it is thought that this was caused by
twisting of the rear spar, which was held by one clamp
whereas two clamps were used in all of the other t est s.
The deflection of these ribs was small for the load
carried, being less than one-half inch in all cases except
one, in which it was noticed during test that the spar
support in the testing machine had loosened slightly.
- -- - -
If this rib could be reduced in weight somewhat it
would be admirable for use on a large airplane, especi­ally
on a quantity production contract. The cost of
the dies for stamping would, however, preclude its use
in small quantities.
Type 7.-STAMPED DURALUMIN RIB
Two sets of ribs were submitted by Airships, Incorpo­rated,
which were stamped from 0.015 duralumin sheet.
As will be seen in Figure 54, these ribs differed from
the other types in that they were composed of two
trussed chords, one in the plane of the upper surface
of the airfoil, the other in the plane of the lower, which
were not connected with each other except in the leading
and trailing edge portions. The first set of six were
formed by stamping triangular holes in flat dural
sheet, and flanging over the remaining metal to form
the truss members. The result was a remarkably
light but remarkably flimsy rib, which had little or no
18
lateral stiffness and which failed under about half the those..-in t he first ribs, as is evident in Figure 54, the
· d I d fl a nges reqmre oa . were formed in .a much better. wa·v , a nd the
The second set of ribs, which were revised and sub- ribs were co nsiderably stiffer, but hea:-ner ..
mitted by the designer to replace t he .first set, were Omitt ing the first set from c?ns1dera t 10n_ as im­made
up of cha nnel sections formed from 0.015 dural~-I practicable stru ctur~s the foll?wmg table gives the
min and riveted · with three-thirty-second-duralumm results on tests of this t ype of n bs .
rivets. These channel members were much wider t han
Loading condi t ions Rib T est Mmauxmi- Weight Detif ol enc - Load in pound Failure Alteration No. No. Weight in ounce load
---- -- ---
Pounds Ounces Jnclles Shear in rivets at members I- 2, None.
l I 360 30. 2 0. 236 11. 9
2-3, in upper t russ.
2 425 30. 2 14. I Shear in rivet s at members 14- 15, -h steel rivets substi tuted for ---- --- - ---- 15-16, lower truss, 13- 14 , 14-15, rivets that failed in first test.
upper t russ.
members 2-3, St eel rivets subst ituted for those
Incidence. ___ __ - - ---- 3 400 ::: : \-----~~~- ---- 13. 2 Shear in rivets at 3-4, upper truss. failing in second test .
Medium incidence. _ 4 550 18. 2 Shear in member 2-3, u pper truss . Steel rivets for those failing in ---- third test .
6 l 450 31. 5 .413 14.3 Shear in member 2- 3, upper rivet Ys dural rivet s substituted for -?.
in 12-13, 13-14. at all weak points shown in
other tests.
High incidence ____ __ 2 I 575 31. 3 --- ----- 18. 5 Shear in rivets at 1- 2, 2- 3, upper None. and lower trusses.
Ys d ural rivets substituted in
-- -- 2 275 31. 3 --- -- -- - -------- ------ -- - Shteraurs sin. 1 ivets at 2- 3, 3-4, in upper failures in first test .
3 650 31. 3 . 370 20. 8 Shear in members 2- 3, in lower Ys dural rivets at members 1-2,
--- - t russ. 2- 3, 3-4, 4-5, in both t1u sses .
5 1 600 31. 0 . 420 19. 4 Shear in rivets at members 13- 14, None.
14-15, in upper, at 1-2, 2-3 , both
None of t hese ribs held t he full design load, even
when altered before test. All were well p roportioned
so far as having a st ructu re of uniform st rengt h is
concerned.
The substit ut ion of a la rger rivet a t one point of
weakness would change the location of the failure , but
very little more load would be carried before the failure
occurred at some other point either as a rivet failure or
shear in one of the members.
Due to this fact t he ribs could not be modified
sufficient ly to carry t he required loads without a com-·
plete redesign of all members, so the t ests were stopped
after the data on the four ribs given above ha d been
obtained.
One objection to this t ype of rib is the multiplicity
of small parts which are held together by single rivets,
the failure of any member or rivet resulting in failure
of the rib.
Ribs of this t ype have a serious source of weakness
in the fact that the compression chord of one or the
other of the trusses, depending on the flight condition,
receives no support against lateral buckling from the
fabric covering of the wings. In normal flight con­ditions
this would be the lower chord of the upper t russ,
which, unfortunately, is the most heavily loaded and
would therefore require t he addition of the most ma t e­rial
to prevent buckling. Stringers might be used to
tie the ribs together and prevent buckling, but such
a procedure involves a considera ble increase in weight
without a corresponding gain in strength.
It is possible that by using a heavier-gauge material
and increasing t he length of some of the members,
using fewer of them but making each compression
member heavier , a satisfactory rib of t his type could
be built, but · it is doubtful whether sufficient lateral
stiffness could be obtained even with this cha nge wit h•
out running into excessive weight.
u pper and lower .
----- - -
RATING OF THE VARIOUS TYPES
It iR practically impossible to. rate t hese ribs in any
way, owing to t he fact t hat nearly all of them were so
far over the required strength that it would be impos­sible
to say what modifications could be made to reduce
each type in both weight and strength to a point where
·it just sustained the r equired load.
The type 5 ribs submitted by the Aeromarine Plane
& Motor Corporation were really the most consist ent
of the t y pes t ested , both as to weight and strength.
Moreover , t heir ratios of strengt h to weight were good
iu both high and medium incidence, and their deflection
was not great. They are rela tively easy to build,
and with some r evision to stiffen the ends of the web
members where they are fla ttened a t the joints should
give a very satisfactory rib.
The t ype 2 ribs submit t ed by Huff-Daland also
showed up fairly well, although they were somewhat
heavy . They gave quite consist ent results and are
of a type which is r eadily built in large or in small
quantit ies, as no expensive dies or jigs are required.
Their strength-weight ratios are good, and if care is
taken to ·prevent the splitting of t he ends of the
diagonals a reasonably light, dependable rib should be
obtained with t his t ype of construction.
Of the other t ypes of ribs little need be said, since
all have objectionable features which milita te against
their use. The type 6 ribs, which are stamped from
duralumin sheet, are heavy and would be quite costly
to produce in small qua ntit ies. The welded st eel
tubular and the built-up duralumin types submitted
by t he Boeing Airpla ne Co. are heavy, the built-up
type being too strong and heavy for consideration here.
The welded duralumin tube rib is impractica l because
of the lack of dependa bility of the welds. And finally ,
the ribs submitted by Airships, Incorporated , are lack­ing
in lateral stiffness and would probably prove to be
19
heavy when used in an airplane wing, due to the ex­cessive
amount of material necessary to keep them from
buckling laterally.
o one type of rib can therefore be rated very far
ahead of the others, but it is thought that of the seven
tested, types 5 and 2 show up better than the others,
both from the standpoint of weight and dependability.
No effort will therefore be made to give a definite
rating to the various types or to say definitel~- which is
the best, it being left to the judgment ·of the future
designer of ribs of this size to decide, after weighing the
good and bad features of each type as emphasized by
these tests, as to what type he will use in any given
I case.
FIG. !-Sketch showing method of supporting and loading experimental test spars, !ateral supports 45 inches apart.
Loads placed 5, 10, and 20 feet from inner point of support; outer point of support 15 feet from inner as shown. Max­imum
overall depth of spar allowed by specification, 15 inches
FIG. 2
D
'Q
D
20
1..9'-4Q ---------------...;
IS '-o 11 -----------.i
F10. 3.- Huff-Daland test spar
___________________ ,20 ,_ 5 2. " ____________ _______ __
8
· FIG. 4.-Huff-Daland test spar-Second article
FIG .. 5.-Comparison of observed and computed loads in Huff-Daland Experimental Spar-Second Article.
Upper figures are observed values, lower are ohtained from the stress analysis
21
•
~
!ltW ~~
r;rs:
OM ~"'
-¥: ~l--1 r-. !"'--
""" ..-::. ...-- ... ~ .........
k ----:::::::: ~ ~~ io-""' (:..; "
""' ~ ............. ,..,.._ a ·~· .IJ. .;,-; 1':1 r. lfY. i ............ , ~
~ ·-"'-'! ""' ~ ~ .... ~
"'"' ~ ')'..... ~ ..... ~ ~ k:: c:.,... ~·
~ -.......... ,~ v r- ... i-' "
" '' ~ '"' \ ~ .... \ ~
""' 1'1< " '· ,_ J 1L!. Ct.. r.:VE OF c~ - .lJEJ ;i..;,.Cz 'UI) '~ aBTA!rt.'.:J.J .ilt ,, "IL!. u~ 'C OF A WlL.LEor _,'MG~
>-- - THE CUR-.£' OF aBSERl'ED ./JEFL.EC770N8 Ii/AS' am'.AINED FRON A SECatlLJ I'ES T Af4IJ£ O/Y THE FIRST -
A ~lCL.E WITH A R EBUIL T' CA!i7'1LEVER TIR THE OBSERVED .JJEFL.ECT/ONS IYERE MEASURELJ TO· -
>-- - 0.01 INCH HEIYCE THE PREC/S.10/Y OF THE READINGS PERMITS OF Ali ACTU'U. Pt:IS/TIO/f OF' TUE ·-
.POlltr AIYYNHERE Wl'Ofllf THE CIRCLES .DRAWIY ABOUT T'HE 0/38E.'RYED A?IM'S.
1 .DEFLECTION OF EXPERIMENTAL SEMRS
COMPARISON OF CONPOI'ELJ ~ QBSER VE.O f--
.DE FLE CTIONS IN S PAN OF HOFF .LJALAl'/.IJ
...._
WEL.DE.D 10.20 $TEEL TOBE S f-'AR. -
..D; PD IM:; ~ lr.h O}: L~ .,.
rur -UJl:"J /If /~8
' ry ?n .< () 40 ,;b t!b n t o _qh tdo 11b r.i,, 1.ii '"° 15b ,Ji d? Jo
FIG. 6
.£JEF.l£CTIOIYS OF £.YP£'RINENJ;4L SR4RS
CQVPAR.0'"0.IY OF F.IR8T AND .8£COIYLJ ARrECLE8
HUFF .LlALAIYLJ J.,Z'LOELJ 1020 ST.CCL TO.BE SPAR
TY.PE-/
- - - '---- - ·- · .. 1.. .T IT::.r _I
- WT. OE' F/RST ART/CL.£ 75. 9 LBS OF .S£COIY27 ARTICLE 7¢ LBS - - M4/0R ..DIFFERENCE .BET/v'££/Y 884.RS:- THE FITTll>GS AT THE A?ffr'TS - - - OF SVPPORT NERE .57"RENTh'EN£.L7 ON THE SECO/UJ ARTIC.l~ ANLJ - SOME OF rHE rVBE SIZES A/ERE CH4/YGE£J. - ........
""
,,
..... ~ - ~ ~ ~ r'.;;: - - ·'- " Alll .
,, ~ I'\ '" -
inN !i.1 9: '\ ,, -8&. ,.,a vn A l?r. ('r
,,., ~ ;;:; ' ~ "'-': r.' \ \
~ \J I\. F, PS .. ~D' .,,.., r
,.,,., ~ ~
:::: t\l
,,., J
i
·->-
""'" l.o.l 'rl7]< WC ~ ~ w ll'i. l'YE. II? "tl 'Y7 a"'.5 /A R:U 'r -
~ < () 4b ,,..., Ah II VJ /~b fa r:J /MJ /HJ ..2 'Yl :z,,.., <>.itJ!
F IG. 7
22
FIG. 8
3
TU.B/E ..S/.Z-cS
SJ"'#BOL 0.0. Jlf/ALJ...
A 2 . .082
fi 1.87 .067 .SECT/O# X::.Y
c /.S .On/
,0 /.37 .067
.OJS
E /. .065
,F . 7.5 .045
FIG. 9.-Enginooring Division steel spar
23
.DEFLECT/OH OF £rPERJMEN7'AL SPARS
ENGJUE.&P/IYG .DMS/OJY IVEL.LJE!J C'HROHE HOLYBJJEJYOM
flTEEL · 8E4R.
'I'YPE .2
WT. OF CHROME HOL.Y..BD.EJYUM SPAR 60 LBS.
h ?.h lYO SECOM? ART/CLE M48 TESTED FOR 7"Hl8 SPAR
AS THE F'/RST ONE HELD 100 A/YD /iJ4/LED ll'.JYLJER
h11 h 10.5 PERCEtf7" OF THE .DE.S'~/Y LOAD.
~
hl'l; b~
':>.
~r- r-.- i...--
i,....-- ........
~ !'... -
hi/
~ + ' ~ \
~ ~ r--
h~. D~ I\.
~ ' h<11 rs \
'l \
~Lt.fl
\
Vl ...:h '
.£, ~. ~~ ~ l~( PM L.t. 'E'2 " ~ VJ ~a ~
( ~~ ..:IO .!'In fib /llO 1:. ~ ;,,IO !~10 It.VJ 2. bO f) t20 2. tto
FIG. LO
. >A
, __
- r
, __ ,
... " v ,...- I>- . ~ ...
' '"'~ - .. ' ~ v~ ./
"~
s ...... ::::-. r.t IM} 'Ul 'f:D f r:Jf'l ~r. rm \[t; .....::: ,;"'
IS r-- t-.. / / ..:.~
,,~ ~ ~ ... v _...I<:; r.;
h ...... ~ 1---" '/ ~ £..
·~, k1 §: .... - ..., ~ v t=i
~ ~ ......, _.. IDF. PL CJ r&i 5' ....
"~·" i..-
""'
-
~
~ ~ -,. " IVrE :- rHE CORllE' or COMPCl'rEIJ .DEFLEC77aNS N45' OBrAINELJ Er llSE OT A ML.,l/or LJMGRAM. ~ THE CURVE OF ODSER//Ell .DE?'L.EC7'.IONB M4S OB7'A/IY£LJ FROM .L7A7'A OB7"AllY£1J .DORIN(; 71IE BR4R J'E.11' - j- THE OB!lERIZ.P .OD"£EC77QIV'$ 10!:/CC HCASVRED ro 0.01 //YCH HENCE THC .PRECIS/ON OF rHE Re4WNGS
I PCRM.IT.8 OF AH AC7U4L P0817'/0N OE' 7'HC .POINT ANYWHERE IN Th'E CIRCL.£5 ../JRAWIY ABour THC
OES.•'DVE'.lJ ~.
I I
.OEF£.CC'7'/0N ur ~zz:;
WG&OtY or OBSERYE..lJ AtY.ZJ CavPOr&J ..OPU'C'7'/0H
~;.· 7:4 'YC, ... } wa l.f u "7 Sl PA pe; ,. .4 V.& fe'J ~.!
""' sPAIY CJtF' ..!:W~ .LJ/K M!ZPLZJ CMeAtU/lS!Z'& £9.w.
l I? ,.., <l'J .M1 .j 1 ti~ ?P c::.1? .SI? 1dti 1.b 1Jo 1Jo /~ 1$ 14o l fO I 180
FIG. 11
25
i------5~0-' ---t-----.. 5 !o "'- --------:.5'-O" ___ ,.,.-1• 0------5 : o "----i
t
SYMBO~
2'.00 .o 2 1.25 . 65
B 1·75 .()47 /./25 .0.35
C /.So .045 .ozs
D / • .3 '.S .04S
F IG. 14.-Boeing experimental wing beam
WI'. OF FIRST AR7'.ICL.e .54 LBS. OF S.t:-Ca'Y.O ART/CL& 62j LBS.
Jt.£4JQR .LYFF..eeENCE ..lJETWK£.N SPARS.·- CH11YCE.D .FA'O?ef A
'C4PERE.D 7"0 A PARALLEL CE:I0/2l) SPAR A/YLJ SOME MEMBER.!!!
/NCREA8C.D JJY &3'£.
,) ~ -"
II'> r,,.., / \
~ I / r\.
I,,"~ '~~ I I \
-~ i'.. I/
I . Q;1 \\
h::> hr~ , .............. / I ~ '\
51 ' ...
!'-... ./ / / ~ \
,., .. ,~ " \ ..... ~i-- ......... / ~ \ \
~
...... ,.. ...- ,,,.- ." \ \
,., .d , ">.. ~ ,
~ -1. DEFLECT/ON OF &YPER/MEN£4L SPARS ~ \
~ ~ Cl:W'PARISOtY OF FIRST AND 8£COtY.0 AR7YC£.E \ \
~ BOE/NO .STEEL TOBE AN.D IV/RE SPAR \ '
~ ~ 7'YPE-G' l-?.~ SEi t"nA '/) fH 7C::: •!\ 'I-~
~/.A' J7 ,,,,,. Tl~ 'L.E '~ 4-i)
11'1 ""
\
I
VJA) \
.DJ, 5"7,; !thl :.!' F. 2a. '1" IN. i'YEl'C? r-tl ~ a ·"'""'
..,,.., "r -· -
{ "lt'l .d h 60 ,;;() /£0 d'l J<J0 It '/J /AD ~":\') :?. PL? 1:1. 117
F IG. 15
26
27
. Fm. 18-Experimental Duralumin and Steel Spar Results or Extensometer Readings on First Article.
Upper value is from tbe extensometer stress reading. Lower value is from the analysis of the spar.
Chord members are duralumin tubes modulus of elacticity assumed 10,500,000 pounds per sq uare inch.
Web members are mild carbon steel tubes, modulus of elacticity assumed 28,000,000 po11nds per square inch.
SECTIO!f A-:4.
SECTION OF SC/PPORT' F/'I'TINGS
Fw. ISA - Kerber-Boulton test spar- first arti cle
.DURAL TU.OE
NEE HEMBEJZS 1 qD. X.0.15 M'UL
STEEL 7'U.BE
FIG. lSB.- Kerber-Boulton test spar-second article (rebuilt)
TUBE
SA!JDLES
.075.STEEL
28
.DEFL.ECr/ONS Sh'OJVN-ARC OlJTA/NElJ FRON .Jv7£L/or LJl4GRAM A5'..5VM/M7 m PLAY UY PIJ'i JOJ/'ITS -
.DEF.lEC1'10/i5' Sh'Olv'N.,. AS,80HE aooo //YCH PL.AY Ar £ACfl ,/01/YT. OR ao1 INCH OH EACH 1J£4GON4L. -
Ill II
LJEFZ.EC770NY SHOJ.'H 4 ASSl/NE QOIO /NCH PL.AY AT EACH JO..w7,; OR 002 INC¥ O/Y EACH 1J£4~ - - OJJSeRYElJ .IJEFL.EC770bS ARC .LJ£;5'1GM4T'ELJ .BY C/2CLES. -
,>,.-.-,,- - -
_j~ ,-- .. •:::- .... ~
"'" v rq,.
-
~ '-· _ ... i-- ,,/' [.A ~/ l!l;
.£! ~~ [', - c 'M. '°' '!:'n lJ• 'F. v' / .. It!,., -· ,• / ./" ~
~ I\." +' ........ \. -/ ~ .. / J. I~
VJ/~ ;:_ 't'' ...... \ :\. -"'· / ~ ·' v / !:!. ".>..
" "- " -~- . .... ... v - ' !"-.... ... ...
Vli.~ ~ "'" ~" ~-- f--• ~ -- \\ ,/ /' - - v §
ll:; ' ... " ,.,.. .. ~/ ~ '<.
lllll "~ '·t / \ / /, ,.,,, .,,.., V}'IJ , nrF'L. C7 llOl\'S'
~ '" ....... -:...' / IL --
""'' ~ '· - - ~ ,.. ,,.'
i: ' ,......._
... / '-" /
"'" E ' ·-. __ ,,..
.LFFLECT'/O.IYS OF' £Yi'£RIMEN7:4L SPARS ,.,.,. CCMPARtSON or OlJSER//ElJ AND COHPCl7'£LJ .LJEFL.EC1'/0NS' ON
SECONLJ ARTICLE OF COMB/M4T/CW 87'EEL <f ./Xl/G4L 8E4R
.a 'S'7. f4h ~ .r:: ea "'1" IC~ ~; 6i/R ~ 'T ~
.. ,_ ~- -
( " '" ~n Ah ... ? ,;h ?I? 6~ Oh !('() /lb IA.0 I~') !~to /.''{} lo'.> /)~ /~~
FIG. 19
.DEFLECT/ON OF El'P£R/ME/'iTA£ SPARS
- - COMPAR/SON or r&sr ANO 8ECON.LJ ARr/CLES
- - KEl2.BEIZ-.BOVL7UN S7'EE£8 .LJC/RAL SPAR
- -wr. !I'YP.E: 4 OF FhaS'r Al2r/CLE ol,f LBS., OF S.£t:'~ AR2"2CL.C 50 LBS.
~JOR D.t'rL:e>~E ~ 8PAM:-Plf?AC7"JC4L.LY AL.l - MEHB,£,eg REOES/6/YElJ ANZJ /i?ELJVCELJ IN SIZE.
.nt
",, ,,,,,~ "" ~ 1-
~ "
/ )/ ~
/lUl I':' ........ - - I_..(' )"" ~\ \. / I 't ·, ' ,. / '
,,.,., \ ' /
~ '-.. ... v' I~ \ \
""' ~ ~- I+- v h ·\ l\
~ - ~ \
f\41 ~ \ \
').. \ Ar.r;. >,<;"T' ,d) •'1'11 ~r *'
""' ~ vv
SJ \ '
ll&:/ ~ ll. \
~ \ - !:' ,,, ...,
~
rl7
Irk' bM" /JWA •p ;:,,. !M ~ t. ~ ..... ·-
J "" ,, 11 ~n ~b 11V, 1.ao /4'() Ailn /);i'J ~'l? .z~ .a tO
FIG.20
29
FIG . 21
r
FIG. 22
,Ot/;;>-'fL TOBE cJ/ZES
Al 2 " 0. 0 . f,O 95 ~LL
Bl 2 " o.D. f.065~LL c r /" o.o. I n6s lf/..4LL
30
FIG.23
.SUPPORT FIT7YIYGS
FIG. 24
.!JooGLA S 7Esr ®PAR
_/=JR .ST /!RT/CL E .
31
DEFLECTIOIY OF EIP£R!ME/Yn4L 8PARS
I
.lJOUGLAS WELDED .DURALUM/ff TUBE SPAR
.T'YPE .5
I LI I I I I I I I I I I I I I I I
W'I'. OE' DOUGLAs DURAL SPAR 56 L.B,8. :
THERE ~ NO .8.ECQ!iD AR7"/CLE TESTED FOR
TH/8 SPAR. THE F/Rsr ART/CLE 8HOW.llYG 7'HE
/NPRACJ"/CABILITY OF 7'.H/$ TYPE OF C0.1Y87'Rl/C'T/'ON
V1 /( -------
VJ r:ln ~
.............. v "" 111 //" ~!'....
/ v ~ I'\
~~ ...... ~ v ~ r\
Vl?I ~ ~ \
~ " \ ~
h?J'
.:!': <j \ ~ \
I l!JLV ~ ~ \
~
~ \
In i7~ \
.DJ ;n; ~M Z' FJE (JN JJ WE. ~} ~JI vr OF 8t ~A 7122 ' I
t? 2~ ..d i'J ,:; (J Ai'? / (/() /~ r? 14' !GI? !tb 2, ~ !?. 'fl '!), 'U)
FIG. 25
FIG. 26
SECT/Oli B:-13 S£CI'/O/'f C-C
• 040 GENERAL
, 06.S A 7' PO/IYI'.5' .SECT/ON A-A
OF SCIPPORT
FIG. 27.-L. W. F. test spnr
f' IG. 28
33
FI0. 29
FIG. 30
3-!
! I
- - - - - -~ - ~ ~-
~-- !Jll. ~ t ·tit. . ~- ' -
~ ~ ~~ · ~ . o J I '"
-" I>.;..\ IV. .IJI ·~· l.) ~ !), l'l ~~I ILi ,IJ '- m ,_ IV. :\JI '-'II II.I. <-I IJ_• II)
I
..:IAr-""
.,,,, ""'"· Llr.>J:" 8 llM !'>LT • .... ILJz: loJ:>IT- 1,, ... ,,. - t,-,r ,,,,, , - . I
IUVJ l/L!6b /. 0, :J:;b /(.11 0
"' I 1 J
Ill'.: I~ I +f~ I I
~ I I J
lal~ J l
~ J j I I
70"1 I I I 1
I~ I 7
6ll~ I ,:j ."J e/ ~ I
~ Fil ~ ~ ~ I
~ ~ ~ I "} ' 1
,..._ I I
40~ j I - I 7
~ I I I I ·'
1.10 '-I I ~
I I I s7
120 J . l . I -, l
I I
110
1 I I - -
~~- 7_ r , I - - (rt ~fA -1 .. ~ L
" I 0'1 • 'll'.l "'"'
.,,,,, aiYl /II 0 i ~ AIVJ 1000 , ~m /"' ':\-') ,,,Iv> ; fl!}') I ebo /0~0
FIG. 31
U\/'I
rT 7 j
1 701~ =
I/ I/ I
I
J I
l c.O ['}/ .
so~ . I
I
I/
151 I ; i fl .
j I . , l ~
I I I
FIG. 32
.
7
I
J
l
j
I
I
.~ '- I
, [M,
";"'· 3tl /1 llo
I ~
\
\
\ '~"
·~,\
P\
~
: '
:
_ _,_
2'>0 A h" w, AM lk.)
35
'""
tn
,,,,.. I~ i..- f-" ,....-- ~ .......
~ ~""': ~ - i..- ....- ~ :-- - -- -
~ r', ~~ - ill II - i--
I I' 1-... "" --i.-- .-
lt M I ! ...... .... ' .- t
i ~ -.... ,_ ....... '
,,, ,,, i r.. '"'-.
~ ~ ' lm.i ~£ i't/,!;r:; ')£) "".£. 17.7l ?JY. ~
...
- -
rHE COMPVrrn LJEnECT/ONS WERE OB7'A INED BY C/,8/IYG THE PREC$E £_~7'/0NS FWZ EE.4Jl1S
l/NOER COMB.I/YEO L.OAO/NC7. THE l-KJMEIY7' OF' /lY£27'/A .BE/NG DIKEH AS 7'HA7' OF 7'H£ .SPAR - -
f-- f-- F'L.ANGES a¥L.Y - -
I
I _JJ__J_ LU _ ~
.DEFLEC7'/0IY OF E:YPERll'f£'./Y7'AL. SPARS
COMPARISON OF OBSERPED G( CCIMPl.ITCD
ZJEF'L.EC7'/0IY8 OF ,SECOIYD A.R:r!CLE OF
.BlQL7'-0P .lJl!RALUMl/Y SPAR
.L~ t:4J Vl }~ ~ Ll ~~ ~ ,5 ~ ~ r JIY I~ t't2 ~
' h ,..., .,h ,..,,, • h >il'1 ~ b ~IO -~h " In 111? 1'V1 1< \'J -'< l".l 1•117 bi.? ,,, -~
F IG. 3~
l I I I I I I I I I I I I I I I
LJEFLECrJON OF C.1'PER/NENJJ4L .SPARS ·
COMPAR/SON OF rnasr ANLJ $£COM7 ARTICLE
L..W.F. DUIL7'-UP -1JUJZALl/MIN SPAR
T'YPE-6
,___ ~ . ~ - =r:J - L LJ_ITl I I I J _ L_.I _ _L __ _ - - ,__
WT OF FIRST · ART/CLE 74 j LLJS, OF BECOIY.D ARTICLE 74 / L BS.
MAJOR -1J/F.FERENCE ..BETWEEN 8PARS:- A COVER PLArE M4S
t- ADJ)ED ro THE 8ECOJ'iZJ ART/CLE.
'1 ,,,n -- \ /r,../ "' "~ y' b i\.
I'\, • ,/
~ SS [I ' . 'l?t ~' ~~' ~ \,
...... ~- >-:=;j ~ .... ~-< '
~
Vl~
~ \J \\
...... ~ \
Ill&
"': (( ..-- '\,: v
~ .s~c PM ~ ..... ""; '7C. £ \'
Vl ~D r VH IS'T i.4/ •r, CL IE' ' ~, ~~
t \\
Ill ,;,n ~ \
_jJ, ts'.?J ~ 'E .z;;. -:>OJ 11 ..& ~ "'R p( ~lli 7' vF 18l PA !R"' ' "1'1 A ili ,;::; .-, ;:; () /£~ I.ff) N-0 ltli? /~ 0 -2 :tJ ,Z ~o ..z, v
FIG. 34
•
Ill'> ll'>.
~iii
[(lJ. bl:I , ~
;::,~
lnr hr:::\~
r...
ru ti; .... ~
?I!
37
:zo '- o·-------------- - -1
60 --~+------ 60 ---------
----1--~ F/Ll£R PLA'I'E
----+--- ./0.¥,?j_..rj CHAIYJYEL AT' SUPPORT~ LOAD PO/IYTS
T/P/c.4L .S.eCT/O/V
/IATCh'.EP THUS
r2a
FrG. 38.-Aeromarine test spar-second article-all dural, weight, 65 pounds
j
t,..-:: v
,.v
-- / ~""
!"':; ~ ,.013 'Elc V£0 ~f .,,.... i-.71 ?N' _,,/
.-- ~ r- r-G
-~
~
IM '"I ~ r-.:::: -....... I/ ~rt'
I~ ......... 1--NC:: ~ L--..; ;::::--
,, < ') I ~ ..._- ._
-
gs CCr-!A "/T. 'YJ .Dr rfi'L E'c2 '/OJ ~
f-- >---
_._ ._ __ -- -
/YOT£:- TH.C CURVE OF COMPU'r£.O .DEFLECT/QNS N4S OBTAD'IE'.D .IJY THE PRECISE FORMULAS
FOR .BEAM.5 IN COMBINED LOAD/NG'. - - THE CONPC/TElJ A?!Nrs ARE REPRes.£NrELJ .BY CROSSES, THE OBSE'RV£D .BY CIRCL.ES
_L.I. Li J _ L I . l . I _ -'-~- _
.DEF'LEC7'/0N OF £1'P£12/M£/f'n4L SPARS
OOHFJ412J:IOIY OF OBSERVED 8 CONPV7'£D .DEFLECTION
OIY FIRST A.RT/CLE OF AEROMARINE Pl.ATE
GIR.DER SPARS.
- ~ -- .Dl5'7. Ith :£' ,F ~Q 11" u. tr2 .s1'49 -~~ tz:. '!i.. I@ Jfft.? .+.,
( 1IJ <h -.h ,,.h ,VJ ·h ,1-7 ,., <;11 /i la /,VJ t l/l pl!) ~ I !.i? ~IQ ;; t;i /, Vi
FIG. 39
38
I
.DEFLECTION or CJ'PERIHC/Y'rAL 88'-~~·
COMPARISOIY OF F/R$1" ANJJ SECQ'YO AR7'/CL.E8
AEROHARllY£ .bl:'.l.E'ALU"fUV .PL.A7'£ GIRDER 8.R4R
T'YR!:-7
wr. or F/RSr ARr/C,t;E ooj~ OF SECOIY/J ARrlCLE 65 .uJS.
MAJOR ..O/FF£€'~E .D£7WE.E/Y SE4RS:-COYSR .PL.Are Al1/JEf) '
~ - 1V .EOrH roP AIUJ LXJTTOM CHOROS OF SECOND AR'l'JCLE l'iL4R I
Ol/rE.R POI/YT OF SUP~
./l" 0 c ~ - / ......,
~N / ~·
_,
\... ) I/ !-."' ,.,Xr-,.. v /' ~ ,,
!'-.' l't.. / /, ~ ~'
l'l<ll ,,.. r-... IP I ~ \ \
!': ', ....... v , II' ." ' -
~ ', ,
,.,,., ~ ~
~ ~ ...... / I!: \ \
,., ,., ~ \ l
~ 1 -
~ ... !
~ \ ---s. r.r ""lJ A 7l r1 •
"'"' ~ \ I
~ ,,J/,S 17:4, ~c.. ~( w //'/. ~ ' I- ~D'l ~, r, ,..,,. \ '"' -V''-' -·=· '1' -·~
fl .,,., ,,,, ..::h /;'} Vf1 ;;h ;~ta J.b? /,ft') ?.II'> \,[M <>.in
FIG. 40
INEOAl?lJ SVPPORT ../J{,OCK Ollr..BOARLJ SUPPORT ..!Jl.OCK LJIAl=HRAGHS
Fm. 41.- Engineering DivisioD"wood box beam
39
F'IG. 42
1•10. 43
40
I I I I I
.JJt:PLECTIOJ'f OF £W.£Rh"1'KJYrAL S/341c'8
COHPARISO!'I OF FIRST ANJ} SECO/'IIJ Al2T!ClES
EN6'/!iEERl!iG _f}JffiJ!O!i WOOJ.J ..ml'" SB4RS. lYPE-8
- ,__ I I I I I I I I I I I I I I I I I
- - JVI'. OF FIRST ARTICLE GG LBS. OF 12:..ECONlJ A&J"'/CL.£ 60 j LBS.
' M4JOJ2 .DIFFERENCE BEThlEEIY Sptj,fil._:-L/CJHT.ER .DIAPJIRAMS ,1ffD - - I'IT'I'JliGS WITH PLnYaJ.D FILLER .BLOCKS AI' POlllTSOF SUPPORT - - l!i J:_EgJlilJ ART/CLE FLANGES AIYJJ PLYIVOO.D Jv'£BS /VERE l//'ICHANGE.D
'-'11 h
'
h 11h
~t--..,_ ~ ~- ......... ~ .
h II ~ " ~ \
,.,,,, . li'i [\,
~l ~ ' I' ~ 88 ro1v ~k'~ t!CL~ ''
h <V
~j S:: I\\ ~
~ ~ '
'\
lnat ""- I\ ··~ r lk'7 VCl E
s::i ;:s \
h-V ~
~
h&;J ~
~J. ~:?;< IJYC ~ IJ:'la .;M //Jty,!: !? ' ft/IJ f'7' Of' 8{, ~ 'R.I
b ?h 40 ,:b ,C:\'l 11 111 /. In 1'::!'.J n IO mo .2. 0 .2.~ ,ZtuJ
J.' IG . 44
1,, , ~
l111b v 'r--.
L/,.. Ir'\ J r'\ " "'" '"' I/ .... -<-' f" '-v - .... _
~
;ti " v ./-- !;:
'"" ~ !ii ........ _,..v v ~
~ !): ~~r'\ nf:'1". ./' / ,- r--., ~
hM >- t;; '~ .;- -- ... /,- v '-V ~
~ ~ --~ ~/ /(../ ~ ,- r--., ~ · ~
h/l< 1;l ~ -, ~ - --"I-~ --~ ~
5 ~ - '7"'\ ----I'_ '-V ~
hl\J '-l .... _, I'- (]fl '>E. 'W 'n DE. Z" lr.7' v,v1 ... -
- - NO'rE:- THE Cl/Rl/E OF CQ"1POT£LJ ..OEF£ECT.!O/'SS A/AS OBJ/11.l'fED .BY USE OF THE PRECISE FVRMVLAS I"OR .BEANS JIY COMEIJYELJ LOALJ/IYG.£
~ - THE aBSERVELJ ..O£F£ECTIO/Y.S WERE N'Edtf.ORELJ 'TV 0 .01 INCH h'ENCE THE J'REC/810/Y OF THC READINGS
>-- - PCRMIT8 OF AN A CTlML POSITION OF THE POINT AJYYlt/h'ERE Wl'l'HllY THE CIRCLES LJRAMY ABOUT
THE ~f:FVELJ fil~Z:.8.
n.G"FLEr.7'/0/Y OF EJEJ.w.l.tY~ALI 81Al.s 1- I I l
~v ,N. OOSERVELJ AliD CONPC/T£0 .LJEFL.ECTIONS
JN SPAN t_v· ENG/, 'R/nv- uv IYIOn
~ - - - ~ D..l: t2J4 tyc; ~ IFic ~JI-; .c b<'. 'S Pl w.,.. h v .h V'C) ~ I I -
I h "' • h Ah Al? . " ?11 J;t'.l so /00 /JP ,,ao do 140 1..t? i.io i1o 16'0
f l (). 45
41 •
COMPARISO/'/ or J)£Fr.£CT!a>YS CY' EXPER/HENrAL SAtlRS ·- F/Rsr Al?r!'"L£S U/'fOER so PO< CENr or .OCSIGN LOAO •' CON~~~ ~f!f[f:;r;c:;~o~o ~:Jc~~G~~!w - I - '
·- ~-· - ..... - '-. I /
~ ... --~- ,,,_.,, ~ •• v [/ ~ I
,
~ -- r~ "" -·- ;::.-·...,.._
'1 -~ L/ "' ·' - ~-:' ~,...
' I'~ . ~ \\ \ \ ~ \ ~ - n/ \ ~~ "
" ·- / w I "' ' ,, !'\ - \ 1\11\ / I In • ,. '/ " \ I'\ !-7'.PL I
' ' ,\ ,x ' - ·'- · ~-'- ·- , f7 •. \ \ r .... •O
':'"/'?£ ~"' ./"?£ (J.r COl'iJ'r.R. iw.-ar lNF-' '. ,\v
' l I P"J.51 lS-• .:'Al? YLW 1 "'(!<'\ ·-· ,,__ --,_ I,? ,, ,,
1 ~ 1 - l/ \ \ !>-WP. ·3
.. h , - j.Vl'L0.:.-0 ,aa? S7"££L. ;lf.'J 008 Pr •, 1 \ ·;, In"
' 2 -· ~c .s~ ARr1c. l£.< I I , __ \
.-h 3 -- j;.17" /'i/Cl<EL $1'£.EL s..;o 0 29 1 ' '"" ! " ··- lS'T£.!'L Ji@ J)(JRAL 6 1.S 0 12 I ~ "''" 1''P£
1;,,.,,.. WT.OF' ilrfl'.." \ Vl11"f''<
: arsn. TY?ECY"~. SPAR 11iY .,
'
·- ~ 0 _, W£UJ£D Oll.RAL μ-•. o OJ7 \ ,., '; ' I WELD£0 1112081'££[ 140 OM I ~
I 6 _ ..
IButL r VP ..Olll2AL -;;3 026 I 2 -~ iwELDED CHl10LY.5TtEL. 60.0 007
7 ltx;,.24!. PlA1"£' G/PNI. 60S 0.24 I ,. 3 - -HT. NICKLE sr£El 62.5 a35 \
,_ _ 8 ---lsPRUCE .DOX Sl34R 6, .0 o.x 4 - - srEEL Al(1J .Ol//UL. so.o 028 \
" I i
, __
s --.M SECO/'iO ARTICLE rz:si .. n I
!ii I 6 - -- IBu!L1' I.IP 01/RAL 74.7.S 0.24 I I
1.;:1 I i '"" 7 --~PLAT£G/Jlll!J, 6.S.O oss
' \ 8 ---- ll-PRl<T~SPARS •as 007
, .. Ill 1,,,, I
<; I I I I I I
, 21 ,, , I
~ ~ · I , I
'''" I
luo
1--
. ~
I
1, . , I I I I " ... ' I I I ' ,,
I I I
"' I I I <
I l.2/j.."'J; w:.c 'n ?,_, L 'FJ £"P1 l?f<~ WL~) FS. i 101$n "' " r.~ ,., ~ ""'"' or lh\ h t'f"'
A~ ao .:, '"' '" ·- ' .L
--
'0 " ,~ <o ,, ,, 0 "'h "" .h ···- ~,,,
FIG . 46 FIG. 47
COMPARJSOIY or OE:rLECT/tJIYS or QPERINE:ffTAL SPM?S .....
S£rt1MJ Ali?TCL.E:S 11/YOER .Jlf P£RC£/YT La4lJ
_..• r/- "'-- / V £ '-'f
·~ \ :::.. - / - ' ~>- 1::~--- I/ /1 1/ "\\' - --- -
\ .o ""J' ~-
\ I/ I,, II ,\ \ fib£ ~
'\\ ' ', I 17' · l VP ::;
I I I/ 11 v:w. !E· li
·~- \1~. (,/ , \.' """ I / II I , ,1 Im '£ ~
I\ .,, I; ; I r;
( '~ "'--1-
' ' ' \•
,., ) ! 1·, ' lo ii
: 1'." ' ,, ii II\ \ ', II I
; I r~
19>1'1&
J,l['OJ '"""'~- . 11
OF n'PC ar CO/'JST.R.
Bfl<R IN l<PAR BAY I
I -- WEL.bEJJ !OZOsrEZ. 74:(!. O.O!) rn ,_, 2 -· 'Wf.WED CMCu'sr. 600 0.1.s '.
:J - -'iT NCKrL S1'££L 625 06S ,,
tin, 4 - ··- sr££t. AtilJ .lXIRAL JO.v 0.S4
s h'O&.i tco1>21 ARTCLE T£. i'r£D \
1,,, "
_ ..
iaulLT lfP 17lllil4L :t'Af 0 . .56 ·,[\\
T --Lro m111 PLATE (j/:>/Jrl MO 0.65 1 1 I
I-l>
PRU:-£ .lJOX SPAR I • O tJ ---- 60.S O!:J
.. I i I I i ,. '' : : ' \
·•h I I I I
I I '
I
l mr, lw -..rk r S' Jo~r.~.,. ,,(, 1-vr , I
,I d,, . .h , i, .h ,., ~h .1o "" FIG. 48
• 42
F IG. 49
FIG. 50
FIG. 51
43
FIG. S2
FIG. 53
F10. 54