Auburn University Libraries

1111111 1111111111 111111111111111 1111111111111111111111111111111111111111 11111111

3 1706 025 84954 3

• File C71.6 / I 52 McO OOIC F'I ELD R E PORT , SERIAL No. 1893

AIR SERVICE INFORMATION CIRCULAR

(AVIATION)

PUBLISHED BY THE CHIEF OF AIR SERVICE, WASHINGTON, D. C.

Vol. IV September l , 1922 No. 366

EMERGENCY LANDINGS FROM LOW iiLTITUDES--MINIMUM

ALTITUDE REQUIRED TO TURN BACK INTO FIELD IN
CASE OF ENGINE FAILURE AFTER TAKE-OFF

( FLYING SECTION REPORT NO. 83 )

Prepared by Louis P. Moriarty, First Lieut. A. S.

Engineering Division, Air Service
McCook Field, Dayton, Ohio
March 23, 1922

Ralph Browr P•'.'.'.':JllO

usr. , ·:
APR ,.,L d. ., ,..J13
C

Non-De1>.:< .•, 1
AubUrn University

WASHINGTON
GOVf.HN M ENT PR INTING O FFICC
1922
 C E RTIFICATE: By direction of the Secretary of \Var the mat ter contained h erein is publish ed as administrative
information and is req uii·ed for th·e proper transaction of the pub lie business.
. (2)

t
EMERGENCY LANDINGS FROM LOW ALTITUDE IMUM
ALTITUDE REQUIRED TO, TURN BACK INTO FIELD IN
CASE OF ENGINE FAILURE AFTER TAKE-OFF.

OBJECT OF INVESTIGATION. the airplane coefficients. Th en the altitude lost was

measured in actual flight.
An examination of records of accidents discloses the fact Table 1 gives only altitudes which have been checked
that a large percentage is due to the efforts of the pilot to in flight. The agreement between the measured and
turn back into the field when his engine fails on th e take- computed values, however, is so close that estimates may
·off, without sufficient altitude to complete the turn . For be made for other types by means of the chart on the last
each airplane there is a minimum altitude below which a page of this report. There was a discrepancy in the case
complete 180° turn can not be made. The object of this of the Thomas-Morse, the comp uted value being almost
investigation is to determine this altitude for various types. 19 per cent lower than the measured value. In all other
cases the check was well within the limit of error in
SUMMARY OF RESULTS. instrument readings. A more complete discussion of the
method of computation is given below.
TABLE 1.- Minimum alti tudes for various types.
MATHEMATICAL APPENDIX.
T 0 t 1 M'nim
1 un Most I B est Radius SYMBOLS USED.
Type. ·a l efficient angle or
weigh t.' altitude. airspeed. I bank. or tnrn.
- - - - - - 1- - - - -·_ __ _ I_ __ __ _
IPounds. Feet. M. P , 75If.
DH--4 . .. . ... .
SE - 5 ....... . ....... .
Curtiss J N4-H . ..... .
T homas-Morse MB-3
XBl-A ............. .
4,297
2 058
2: 200
? 548
3'.679
45°
340

~1
270

300
~
,
Feet.

78
73
I
45°
45°
45°
45°
380
330
240
400
360
H
V
V"

V11
=
=
=

=
Altitude lost in 180° tum.
Forwar!l velocity of airplane in feet per second .
Vertical component of forward velocity in feet
per second .
Horizontal component of forward velocity in feet
1 Full military load used iu each case. If airplane is fl own without
full load, the a ltitude loss will be proportionately less. per second.

The altitude given for each type should be taken ~s an 8 = Angle of flight path with horizontal.
absolute minimum for a complete turn of 180°, and can A = Area of wings in square feet.
only be obtained by following fairly closely the air speeds
and angles of bank which are reco=ended. Both theory , d = Density of air in pounds per cubic foot.
and experiment point to the fact that a reasonable devia-
tion from these conditions does not greatly increase the Ky = Lift coefficien t of whole airplane (absolute).
loss in altitude, and , with average piloting, an airplane Kx = Drag coefficien t of whole airplane (absolu te).
can be tui'ned back with safety at the altitudes shown in
the table. On the other hand, even with exceptional r = R adius of turn in feet.
piloting, th ese altitudes can not be appreciably decreased L = Total lift in pounds.
if a complete 180° turn must be made.
There is only one part of the maneuver in which a gain D = Total drag in pounds.
can be mad e, namely, the take-off itself. The pilot should
so '' play his field " on the take-off that a complete half turn 'N = Total weight of airplane in pouuds.
will not be necessary . · ANALYTICAL SOLUTION.

DISCUSSION OF METHODS USED. Figure 1 shows the forces acting on an airplane in a

gliding turn. The drag D is balanced by th e component
In the solution of this problem, reco urse was had to both of the weight in the direction of the flight path W sin 8 .
theory and experiment. Briefly, the method of attack The resultant, R , of the lift, L , and the component of the
was as follows: The minimum altitude lost in the t urn , and weight perpendicular to the flight path, may be considered
the best combination of air speed, angle of bank, and as an unbalanced fo rce produ cing an acceleration toward
radius of t urn to give this minimum were computed from the center of the t urn .
5469- 22 (3)
 4
'f'he following relations are apparent from the figure: For simplification, the following notation is in trod ucecl:
(1) sin 8= VJV a=(d A ,r)2
(2) D=W sine , b= (d A) 2
(3) L2- W2 cos2 8=R2 c=W2
Equation (10) becomes
i
(4) W si n 8 = Kx C, A y e
(11)
"
1·, v•
. '-g
i'\ 2- d?
(o-) w·,
, - cos-, e+ r 2 = K y- g2 A0 V4
? - differnn tiating
51{2 4 a(br2 ky 2 - c)K/r 3 - 2 abr5kx2 ky 2
(12)
3 br (br2ky 2-c )2 0
(6 ) V (I = Kx cl ()"A V
oH 2 2 a(br2 ky 2 -c)r4 kx-2 abr"h2 kyky'
" (13)
i>K, (brzkyz-c)z 0
Combining (4) and (5), and dividing by V4, gives
Combining (12) and (13), and simplifying
., d ? A? " 72 cl?
(7 ) K x + -v,_v_ = K 2 - \.2
, g:! tauz e g2 r2 y g2 1 (14)
now
Neglecting r2 kx ky 2 and 2 ~ kx, which are relatively smal I,
H=,r r tan 0.
or and solving for r, we have
}I2 w
(8) tan2 o~ 7r2 r2 (l 5 ) r= (b (ky 2 -k: ky ky's; dA (K /-kx ky ky' )t

L cos u

W cos 8
w cos e w
Sect . A- A
FIG . 1.

Substituting in (10), and simplifying

(16 ) I-I2=1o2w2 . K xz .
d 2A.2 (kx ky" ky' -kx2ky 2 ky12 )
or This equation is of interest, as it brings out th e fact that
K/d2 A21r2r4
?
for a given combination of airplane coefficients, the
(10) H- K· 2a2 "l\.2 r2- vv2
· y altitude lost in a 180° turn is directly proportional to the
For a given airplane, flying in a given d en sity, d, A and wing loading and inversely proportional to the d ensity.
Ware constant. The variables entering into equation (10) Differentiating with respect to K,, equating to 0, and
are therefore r, K, and Ky. Of these three, r and Kx simplifying, we have
may be taken as independent variables, K y being con- (17 ) K/ky' -3 kx ky'ky 12 +2 kx 2 ky ky' 3 -kx ky 3 ky 11
sidered as a function of K, . In order that H may be a + 2 kx 2ky 2ky1 ky11 = 0.

I
minimum it is only n ecessary th at - · We will now assume that ky can be expressed as a fun ction
oH = O of lex, of the form
or (18 ) Ky=a + b kx+c kx 2
and While such an expression will not hold true over the whole
"~- =0 range of values of K,, it will approximate th e curve very
oK, closely over the range where t h e minimum value of H
A solution of these two equations simul taneously will occurs.
therefore give the radius of turn at which the least altitude Differentiating (18 )
will be lost as well as the best angle of incidence at which (19 ) l~y 1 =b+2 C Kx
to fly during the turn. (20 ) K/' = 2 C.

'
 5
b uustituting (18), (19), and_(20) in (17) and collecting terms. The recovery.-It has been found that th e excess speed
(21) 12 c4kx7 +25 bc 3kx6 + (16 b2 c2 +8 ac3 ) kx 5 +(3 b3c+3 abc2) in the turn, which is usually from 15 to 20 miles per hour,
kx 4 - ( 4 ab 2c+4 a2c2)kx3-( ab 3+5a2 bc)kx2 +a3b=O can be used in coming out of the bank without loss in
altitude. It is only necessary, then, to allow an altitude
In this equation the constants a, b, and c can be found for above the ground equal to one-half the span of th e air-
any airplane whose lift-drag curve is known, and the equa- plane in order to allow the lower wing to clear the ground.
tion can be solved for kx . This value of kx can then be The average correction will be about 12 per cent of the
substituted in equation (16) and the minimum value of total altitude.
H can be found . A further arbitrary addition of 10 per cent will be added
For example, in the case of the DH- 4 as a safety factor to allow for inaccuracies in piloting,
a= - .186 mal-ing a total correction of 32 per cent.
b= 15.27 With this correction the value of H for the DH- 4 is 343
C=-84.4 feet.
At first sight this method of correcting for conditions at
Substitutin g th ese values in equation (21 ) and solving, we the beginning and end of the maneuver may seem very
find approximate , but allowing an error as large as 25 per cent
Kx= .07
in each of the corrections, the probable error in the total
which can b e substituted in equation (16), giving a value altitud e frou{ this cause will be less than 5 per cent, which
of H=260 feet . means an average error in altitude of only 15 feet. The
A correction must now b e made for the conditions at the final test of any assumption is the accuracy of the resul ts
beginning and end of the maneuver. which it gives , and except in the case of the Thomas Morse,

r-"

:
"'
~ 11 r.•IH
i 1./
_,,V"
v I',
' 'r--
h .~;
I

I : IA YIJ Id Pl" V lclci lb' I'- 4 ,ul',,. I

I
.• V r...
V
,:. . ;, V ,. ,--
:
1:i'1n• r ! ~v -1-" ....--
-1--
" I'\ n
;

r l ·•
~
r--. "
.--,_ -1--""
-~ ,..._ A11i!l
I'\
I'\
I

I/ "
bf I\., '
·r
C
::: r- r--.. Lo.-" ...... "" I", '
... I/ -1-" !"'I I'\ C

--
1;'
~
QI
• ~
1V
0 l
'-- I-
0 2 0
I'-
~ 0 4 0 5 O 11>-" 0.1 0 8 O 9
I'\.
l 0
,.,.,._ ;
·-·-·'"• I'-
_'O
i
~

..:
I AllllUJa
-r-- ,-
cc el pr, t:' On
-- ~ ......
ct s r-....

FIG. 2.- Latcrn.l control, Xl31- A. S11 owing angular vclociti~ and acceleration s (luring a bank of 45 degrees.

The slurt.- Immediatelv after the failure of the engin e as note·d a9ove , the agreement with the res ult s of alc Lua
('there is an interval during which the pilot must nose over tests has been very close.
to prevent stalling. During this time' there is always a The above method is perfectly general, and can be
drop in air speed, which varies with the alertness of the applied to any airplan e whose weight , wing area, and lift-
pilo t and the characteristics of the ajrplane. Tests were drag curv e are known. It is, however, very laborious, as
mad e to determine this drop in air speed under conditions it involves the solution of a seventh-degree equation. The
simulating sudden engine failure, and the average drop following graphical method, while not as general, is much
for the DH- 4 and the XBl- A was 5 miles per hour. more easily applied and can be used for types eq uipped
The al titude lost in regaining this speed was computed as with R. A. ];. 15 or sections of approximatel y similar
14 feet for the DH- 4, or 5 per cent of the total altitude. characteristics.
In banking, there is a further loss of altitude. The time
of bank from O to 45° was computed for the XBl- A by GRAPHICAL SOLUTION.
the method of step-by-step integration, using a computed
lateral radius of gyration of 5.5 feet. The curves of angular Equation (18 ), between lift and drag coefficients will be
velocity and angular acc eleration are shown in Fig. 2. used as the basis of the graphical s,o lution:
The computation gave 0.96 second to 45° , wh ile a rough
measm ement in flight gave 1.1 seconds. Half of this time (18) ky=a+b kx+c kx2
may be consid ered as not producing any turning. The
For the DH- 4
to tal time for the turn being 11 seconds, the loss is 5 per
cent of the total. (22) lcy= - .186+15.27 kx -84.4 lex'
5469- 22- 2
 6
A plot of this equation is shown in Fig. 3. It will be The value of K d epends upon the p arasite areas of the
noted t hat the agreement with the experimental results is two types.
very good, except in the region of maximum ky. At (29) K=0.64 [ (~),-(~ ) ]
h=. 07, where the minimum value of H occurs, t he
empirical curve agrees in slope and location with t h e actual Ae=equivalent flat plate area of parasite resist-
curve. ance.
For any oth er given t ype equipped with R. A. F. 15 For the DH- 4,
section it is assumed that t h e difference between its drag Ae=l4
and the drag of the _DH-4 is a' constant for all values of ky , A =44.0
t hat is, Ae
A = .0318
(23 ) kx=lcx+ K Information is available on the eq uival ent parasite area
of most types which h ave been flight-tested at McCook
(Subscripts refer to any given type other tha_n DH-4.) Field. It is more convenient, however, to use fineness

']TrG. 3.

:IBq1iatio n (18) will be of the form : . o f Ar.

. a Iunct1on
.
as a parameter, fi ne ness 1Jerng /C 'I'h. e lo
. 1-
(24)
lowing tabl e gives th e values oft_, l{ , all([ tl,c coc ffi cieu l~
:-:i ubstiluling (2:3 ) in (24) . of eq uation (24), [or val'ious valu es o[ fine ncRs:
'fABLB 2.
(2fi) l.·y = a, ·+ b1 h + /., 1 K + c, Kx2 +2c1 IG-x+c, K 2
Ac K a, b, c,
<..:011tbiui ug (18) a,nd (25), aucl cq uatiug coe ffi cients o[ Fineness.
1\
like powers of kx, - - -1-
(26) a1 =a-bK+ cK2
90 .. . ............. .
) 00 .. .. . . . • ... . ... • . .. .
llO .. ... . •. ..... . .
.0.5()
. 0326
.022'3
.. Ol16-I
000512
-. 00595
I - . 3754
- .1938
-.0980
17. 23
15. 3r,
14.2, I
-Si.4
-84. 4
-'l4.{
(27) b1 =b-2cK 1120 ........... . .•...... .0153 - . 01055 -.0343 13.49 -84. I
(28) c1 =c
 \

F rG. 4.

l
r FIG. 5.
 . \

8
Fig. 4 shows polar diagrams for various finenesses. It
is realized that these curves are approximations, since H
they are based on the assumption of equation (23) .
This same method of attack has, however, been used in
the Kerber method of performance prediction, 1 and has
given excellent results. A report containing a complete
discussion of the relation between fineness and parasite
area, and of the method of determination of fineness with

l
any air foil is now being prepared by Mr. Kerber, and will
be published within a short time.
The relation between fineness and 1e is shown in Fig. 5.
F rom eq uation (16),
W Kx
H = d A Ky( lex ky k.-y~, -_ ..,K,...x""'
2K=-c
/ =2 )1
A~suming standard density of 0 .0761 pounds per square foot,
.
F tG . 7.

(30)
w
H =4L2A / Rewriting eq u:.ition (4),
wh ere
Kx
f=Ky (lex ky ky' - K/ki/ 2 )1 .
V= W"
A
.l

c~in g)•
K xd
o

Equ,1tio11 (30) may be plotted with H and fas coordi-

or
nates and w . . a sen.es of
A as a parameter, g1vrng 1· .
nl( iating

straight lines, as shown symb olica lly in Figure 6. Section (33)

(2) of Figure 10 is plot ted in this manner.

where

which may be plotted as shown in Figure 8.

(See section 6 of fig. 10. )

FIG. 6.

Rewriting eq uation (15), .

or . I

(31)
. w
r =l3 .1Aj'
F tG. 8.

where Let u= angle of bank

f'=(Ky 2 - kx ky ky') , y2
Combining equations (30), (31) , and (8), . (34) tan u= -
gr
(32) H=r ff' (Act ually , Vh 2 should be used, . but the slight increase
which may b e plotted as shown in Figure 7. in accuracy does not justify the added complication. )
(See section (4) of ch art on page 10. ) S ubstitutj ng (31 ) and (33) in (34),
1 See Air Service Information Circular, Vol. II, No. 183. (35) tan ii=.00237 f'/112 •
 9
The quantities/,!', and.fl/ are functions of the lift and In using the chart, it is only necessary to l"Ilow the wing
drag coefficients. In Table 2, fineness has also been sh own loading and t h e fineness . If the fin eness is not known , it
to be a function of t hese coefficients. That is: can be found from t h e high speed of the airplan e, as out-
lined in Air Service Information Circular, Vol. II, No. 183,
F =!, (K" Kv )
/'./'/2=./~ (K ." K y) f'f''
Ji'=/ 3 (Kx, K s)
We have here three equations in five variables. An
elimination of any two wo uld give one equation in three
variables. For example, an elimination of K x and Ky
woHld givearelation between F,f'f' 12 andj11 , which could
be plotted as sh own in Fig ure 9.
Section 5 of Figure 10 is obtained in this manner, except
that t h e abscissae are angles of bank instead of values of
t he function F .f112 • This same line of reasoning applies
to sections (1) and (3).
It will be noted that plotting by this method wo uld be
an extremely complicated process, owing to fact that no
l
simple analytical expression can be fo und for the relation-
ship of Figme 9. In the actual construction of the chart, I
the computations were greatly simplified by reversing t h e FIG. 9.
process. For each fineness, values of Ky were computed or, if the parasite resistance is known, fineness can be
for various values of K" by use of equation (24) and Table fo und by use of Figme 5. Knowing the fineness and wing
2. The functions/, .fl, etc., were then computed for these loading, and assuming an angle of bank (45 degrees will
values of t h e coefficients, and corresponding values of give best results), H, r, and V can be found as sh own by·
H, r, and V were fo und. the dotted example lin e.

r
 10

I ;
H-

A/""'5P£ED /"r.P.ff.
Fm . 10.- Chart for prediction of minimum altitud e lost in 180° turn.

0
1
l
I
I
,

'