FILE COpy ~ o· t

REPORT No. 88

/

PRESSURE DROP IN RADIATOR AIR TUBES

NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS

11:1 POut.:,,~, TO: lOti' mo'. Ttl:: . !'oS F

NATIONAL ~OVISORY CO', ~ITTEE FOR AERO '"unes LAt1GlEY AH • !JTlCAL L,LIB R TQ:\Y l.lINlilEY F. <.D, • ~h PlOH, vIRGIN I,'

n~QU!:STS fO • f J l :) ~~. U) BE ADliR SS!:" ·Fe Il) •

r- ,.." '~.I'"

WA.SHINGTON

GOVERNMENT PRINTING OFFICE

1920

,

;

•

https://ntrs.nasa.gov/search.jsp?R=19930091151 2020-04-24T23:25:27+00:00Z

REPORT No. 88

PRESSURE DROP IN RADIATOR AIR TUBES

NATIONAL ADVISORY COMMITTEE

FOR AERONAUTICS

WASHINGTON

GOVERNMENT PRINTING OFFICE

1920

- ~ --~-------~--------~-~~----,

REPORT No. 88

PRESSURE DROP IN RADIATOR AIR TUBES

BY

S. R. PARSONS Bureau of IStandards

3

./

REPORT NO. 88.

PRESSURE DROP IN RADIATOR AIR TUBES.

By S. R. PARSONS.

RESUME.

This report describes an investigation of effects of pressure drop in radiator air tubes, con­ducted for the ational Advisory Committee for Aeronautics at the Bureau of Standards.

A small steel tube-0.04 inch (1 mm.) in outside diameter and 20 inches (51 em.) long, with a static pressure opening near the center-was stretch ed through an air tube of a radiator and used to measure static pressure in the stream of air passing through the radiator tube. The measurements lead to the following conclusions:

1. The drop in static pressure in the air stream through a cellular radiator, and the pressure gradient in the air tubes, are practically proportional .to the square of the air flow, for a given air density. The observed values of skin friction agree approximately with those found by other investigators for long pipes. These facts appear to indicate that the air flow is turbulent, even in the short tubes of the radiators.

2. The difference between head resistance per unit area and fall of static pressure through the air tubes of a radiator, noted by various observers, is shown to be apparent rather than real.

3. Radiators of different types differ widely in the amount of contraction of the jet at entrance.

4. F rictional resistance is found to be two-thirds of head resistance for one type of H­inch (0 .87 cm.) circular cells, 5 inches (12.7 em.) deep; and one-half of head resIstance for one type of -r\-inch (0.79 em.) square cells, 4.8 inches (12.2 em.) deep.

5. Supplying heat to the radiator increased the pressure gradient in the tubes of one type, of ~-inch (0.64 cm.) circular cells, 4 inches (10.2 em.) deep, by about 15 per cent for a mean temperature difference of 1100 F . (61 0 C.) between the water and the entering air.

INTRODUCTION.

It has been noted in the course of investigations of aircraft radiators that the drop in static pressure in the air tream between the front and rear faces of the radiator seems not to be equal to the head resistance per unit area of the se tion, as mea ured on an aerodynamic balance. In some of the earlier investigations both in this country and abroad, an attempt was made to measure this pressure drop, and the results obtained were assumed to be equal to the head re­sistance per unit area of the radiator for the same air flow. But as soon a aerodynamic bal­ances became available for the work, and actual head rcsi tance was measured, a considerable difference was found between observed head resistance and head resistance computed from pres­sure drop, and no satisfactory explanation of thi difference was at once apparent. Since the air emerges from the radiator in a turbulent condition, its static pre sure mu t be mea ured under unfavorable conditions, and it was natural to question the reliability of the measure­ments taken. An attempt to measure dynamic pressure before and behind the radiator was made, but with no better success.

The investigation described in this report was accordingly undertaken, in order to make independent, and if possible more reliable, measurements of static pressure at various points in the air stream; and to throw light on the difference, if any exists, between pressure drop and head resistance per unit area.

172642--20----2 5

6 ANNUAL REP ORT NATIONAL ADVI ORY COMMITTEE FOR AERONAUTICS.

EXPERIMENTAL METHOD.

1. Ordinary measurements of pressure drop . The bulkhead tunnels used for ca.lorimetric tests of the r adiators were fitted with piezo­

meter rings before and behind the test section.! In tho tunnel inclosed in the steel tank these rings were 134 inches (3.2 cm.) from the face of the mdiator and in the "steam tunnel" there were two pairs of rings, IX!' and 6 inches (3.2 and 15.2 cm), 1'0 pectively, from the faces of the section. R eadings from the 134 inch and the 0 inch rings were prac tically identical for a num-1 er of section .

2. S peciaZ mea urements of pressure drop . The special mea urements of pres ure drop indicated in the attached curves were made in

two ,:ind tunnels: The" team tunnel " or closed tunnel, in which the radiator core completely fills the channel ; and the open tunnel, which i 54 inche. in diameter, and represents conditions in free air. The mea m oments were obtained witj1 the use of a small steel tube, 0.04 inch (1 mm.) in outside diameter and 20 inch es (51 cm.) long, with 0 e end clos d, and a static pressure open­ing near the center . This tube will be referred to below as "the pressure tube." It was passed through an air cell near the center of the radiator and mo,ed forward or backward to obtain the pressure at different points. One side of an inclined water gauge was connected to the rear end of the pressure tube and the other side to the static pressure tube of the Pitot used to meas­ure the velocity of the air stream in the channel.

The pressure tube was supported by two pieces of piano wire which were attached to the ends, passed over cros bars set in the closed tunnel, and rings h eld by wires in the open tunnel­at some distance before and b ehind the radiator-and passed through holes in the tunnel Hoor, to facilitate movement forward and backward. One of the supporting wires was viound arOlmd a spool held by a ratchet, and both were kep t taut by a eight of about 4 pounds (1.8 kg.) hunO' on the other wire. Tho various Iositions of the pressure tube were indicated by marking a point on one of the supporting wires and measuring its distance from some convenient point, such as the floor of the tunnel. In the open tunnel t he positions were checked by frequen t measurements (inside the tunnel) of the distance of the pressure opening from one face of the radiator; but in the closed tunnel such measurements could not be made, a~ld only r elative positions were obtained, the actual position of the radiator on the plot being estimated from the form of the curve after the latter had been drawn. Preliminary trial showed that consistent results could be obtained with only ordinary care in centering the pressure tub e inside of the r adiator ail' cell. In the different sections used the pressure tube occupied between 1 and 3 per cent of the area of the cell through which it passed.

In most cases water was not pa sed through the radiator sections, and they were at the arne temperature as the air; but in one case- type 0-9, 34-inch (0 .64 cm.) circular cells 4 inches

(10.2 em.) deep-after the section had been used at room temperature hot water was pumped through it as in regular calorimetric tests, and a mean temperature difference of about 1100 F. (61 0 C. ) was maintained b etween the water and the entering air.

3. Oomputation. In the closed tunnel the airflow was expressed i pounds per second per square foot of

tunnel section (or of radiator core), and pressure difference was expressed in pounds per square foot. Previous work in a wind tunnel under partial vacuum has hown that pre sure drop b etween piezometer rings before and behind the r diator is inversely proportional to the air density at the front ring, for a given mass flow of air; and proportional (very nearly) to the sCluare of the mass flow of air, for a given density. These two relations were used to r educe the observations to a common density, and to correct for small variations in air How. Obser,ations were taken at from three to six ail' velocities on ach section.

In the open tunnel pressure drop was ex ressed in pounds per square foot, as befor e, but the air flow was expres ed in miles pel' hour of the stream through the tunnel, and the corre­sponding ma s flow of air through the radiator was computed from the r elation between these

I The two tunnels are described in detail in Technical Report No. 60.

PRESSURE DROP I RADIATOR AIR TUBES. 7

two quantities preliously found in the regular tests of the radiator.2 Correction for air density was made as explained in Technical Reports Nos. 60 and 63 , for measurements in the open tunnel, and corrections for small variations in velocity was made on the assumption that pres­sure drop, like h ead resistance, is proportional to the square of the air velocity. Observa­tions were taken at only one velocity on each section, except that in the case of the Spery type 10 velocities were used with each of two positions of the pressure tube, and the results showed the assumed relation between pressure drop and velocity to be approximately true.

It was found, however, that the pressure gradient along the tube of the radiator seemed not to be the same in the open and closed tunnels for the same mass flow of air and the same den ity, and this diiJ'erence was interpreted as indicating an error in the measurement of either the pressure gradient or the air flow. The measurement of the air flow was known to be sub­ject to error as high as 3 to 5 per cent in different radiators, and it appeared reasonable to regard the pressure gradient within the tube as a good indication of the air flow. Accordingly, for comparison of data obtained in the t,,-o tunnels, the results were reduced not to the same air flow as indicated by the usual measurements, but to the same pressure gradient in the tubes (tpis gradient being proportional to the square of the air flow), and this condition was taken to repre ent equality of air flow in the two tunnels. The differences between the air flows previ­ously computed and those gi,en by this procedure were from 5 to 8 per cent.

DESCRIPTION OF CURVES.

In the accompanying curves pressure drop in pounds per square foot is plotted against distallce along the axis of the air tube of the radiator, the distance being measured in inches, forward and hachvard from the r ear face of the section. The location of the two faces are indicated by heavy dotted line ; and for each of the three sections tested in both the open and closed tunnels, head resistance corresponding to the indicated air flow is shown by a solid vertical line marked "R." •

Plots 1- 5 show re ults of tests in the closed tunnel, and in plots 6-8 results of tests in the open tunnel are sho,,'n, with the closed tunnel curves of equal pressure gradients superposed upon those of the open tunnel.

Plot 2 shm s the effect of imparting heat to the air as it goes through the radiator. In this case th mean temperature difference between the water in the radiator and the entering air was ahout 1100 F. (61 0 C.).

Plot 3 gives a comparison of pre ure drop in two radiators whose tuhes are of very nearly the same dimen ions, but with dillerent conditions of surface. In one (C-10) the cooling sur­faces (walls of the air tubes) 'were lery rough, and in the other (C-9) they were somewhat smoothed, though not highly polished.

It will be noted that plots 5 and ,representing the S ery type (illu trated in the photo­graphs, fig .9 and 10), are plotted on twice the scale u ed for the other sections, in order to show clearly the loops in the curves as they follow the four constrictions in the air tubes.

DISCUSSIO OF RESULTS.

1. The I ressure measUl' ments were of coul' e made under unfavorable conditions, and the results obtained should be used with a little caution, but the consi tency of the observations seems to indicate a fair degree of reliability, and a comparison of pre sure gradients in the three radiators with circular cells shows a reasonable agreement with values of skin friction found for long tubes by Stanton & Pannell, Saph 8:; Schoder , and others, the values here obtained being slightly higher than those of the other obserrers, for smooth tubes.

2. The drop in pre " re in the air stream through the radiator and the pressure gradient in the tubes are practically proportional to the square of the air flow, for a given air density. This fact and the approximate agreement of ob erved skin friction with that found by other illvc~ tigators appear to indicate that there is turbulent flow in the short tubes of the radiators.

2 The regular mcasurement of mass /low of air was made with a special air Venturi meter, described in Technical Report No. 60.

~

() ~:::::=:. :::~::::: ::='::::=.:.:.:.~~ .. :.~:::J::~:.::: :::'::'j:~:'::'~ ::::.:~ :: :::-_ .. _ 5 \ /'1=4.60. R I

I \ I B ._ ,

10 1 , ----... V I I ~ It

IS ' i-'- - /'1 =6.50. I

~20 L, i , PI -... ~ 1 1

! i'''~ II -~~h.-.-.... -.-.. -e~ I~ '" ! ...... 1"-- (v ~.XI I i'--.. I

~ ! "'~ i7 I"RESSUREDROPINRAOIAlVl1 q; i I """ : / in' Closed Tunnel.

.j:J , "'- V C-8, %z"circuiorcell, Ydeep.1--I ...... /'1-Airflowmlbs./sq.II./~ec. ; '" AirDenstiy, o'D75Dlbs./cuff. l--

4() I ""', R · lIeodResislonceinlbs/sq.fl.

.f.5 Jr-:- -Rod;o~or-~. I I t I I} (j 4 . Z 0. 2 4 (i

I'l0l-1: ~e fromReorroce -irK/1es,

0.

..5

If)

" ~/S -!:?. .; ~20. ~

~ ,~ ,,2..1 § '" ~JO .t:

~ M'

--..j

I

, , I

I

i~ I I I

! ~ i ~

I

I

I I I 18 /1=4.60.

l.t ...:.-<

~ ~ WI /1'6.

150.

!f -- '1. z~o. ~ t ~ V-

"-~ Ir t .... -' ~f OJ ,

~ :// I"RESSUm: DRo.l" ,W RAD//ITDR

" I' in Closed li./nnel.

I I

I

~ , ! '~ ~;

C oS, '/4 "circulor cell, 40'...ep. H JS

40

-'1S 8

Ptoli!. 6

/'1= Airflow Inlb.s./sqfi/sec.

i /lirOer /slly,o.D75Dlbs!cu. rl.

~ Rodialor Cold· t--

I .. Itol · ----I- /?odialor I

4 Z 0. Z -I (j Dis/once frol77 Reor /Oce- incnes.

Q

.5

/0

~ 15 l).

..!:!. .; ~2IJ , \}

C§ 2.5

" ~ ~.Jo. <t

JS

40

-'1510. Plolj.

o

5

/0.

~ 1.5 l).

..!:!. . ~20.

2-c§25 ~ ~ ~Jo. ~

3'.5

40

4.58

Plol4

IJ

~. 9=

6

.. .. .. Ii ....

I

I

~ I . ~ , I

\ I \ , \ I 6

I :/ -<>- -4:qi . I~ 11/ I' .... ~ !# V- /'1~7.60. , ~ ~ I; ----" -!~ 1'0', '-J i!/ i ~ r--.. 1 I "\ "'- ''' 111 I"RCSSURE OROI"I/'I RADIATDR I 1"-, '"

In Closed Tunnel.

I ''. .-c-9, '/4'circular cell, 4" d"",p.

"J Smoothed Tubes

I ')", e ·C-lo.,'/4-Clrculorcell, -I"deep.

! . HoughTubes-------.

~ I'1·Alrfiowul lbs./sq.fl./sec .

r-1odla(or- Air Oensi/y-o.D7.5D Ibs./cu. fl.

-4 2 4 (j

Di:slcnce fromRe, 0. 2

'Face-inches

::::::::: . .... ~::~:::r:::::·::

.. --.-- ........ .. --~ ~:-::: ~.:::::: .. .. J .. .---......

I

! II I /'164.60. 1

1\,00 f--.. ha. I 8

1 . I ~ Ff I I 1'1.~§. . .5q Ii

I~ ~ It" . _~..J

~ I ~ ",I7. - -, /. i"-:.- ~~J:~!L .. _-- ~--, J.P "- if I t-....

- II ,.?

1\ II I PRESSURE OI/OP 1/'1 RAOIATDR

I, t'n ClosedTLlnnel. J-.-I 100, -S/J6-sqvare cell, 4.8aeep.

/'1-Alrrlowinlbs./SIj1.fl./sec_ I I AIr Oensily- ao.150. Ilis/cu.fl. f--

~ -Rf,*o~or - R 1f-lieodResislonceinl/Js;/sq.ft.

6 4 . 2 0 2 4 6 OisloncelromRearrace-inche.s..

00

~ !2\ o ~ ~ t=J '"d o ~ !2\ ~ H o !2\ ~ (;

Ul o ~ o o ~

S t=J t=J ":j o ~

~ E5 !2\

S H o Ul

o

10

2.5

zo

IS

PRESSURE DROP IN RADIATOR AIR TUBES. 9

I i AI ... .-_ ... .. _- . .... _ .. .-..... ~ .. - . .. _- -_ . .. . _.--..... ~.

-.--. --.. ---- _.h. _ --_ .. _- ·:·:::·::r.:::::: ---_._.- . __ .- -

PIi'CSSURC MONN I?AOIATOIf --;

1\\ th Closed TUnnel. 0·1, IrreJtlQrcel1, coSle"deep. 11-Air "/OWff/lbs./sg-!t/sec. _If

l\\ AirOensily.Q07.501. 5./CU.TI.

I R'lieod Resislonce Inlbs.jsqft II I

I W· ~ I I .. __ .. ._ .. -

I • I

I -- ... _ ... -I />1.6.!J!- .......

~ f\ !f ~' 1100 i ~ .....

I V!1 f\ YI '''1' )80 I I

IJ fI (\' fI if /

.--I fI

I i V' jl 'iJA' -II! I I

A J ?y ! ! V

~ Ii i r ~ fI( ! I ·

~ I--- Rod;otor(. \11 !

f 2 /0 0-

~istance from ReorfOce·inches-

, I ! i I I I I

AI .. ... -.- . -.. - .... -. ...... J ... .. __ ... ......

If

... - ..

20

1.5

..---

IS

20 10

Plot 6.

... --- .... V V ~ l I I

6

I l r I: ~ i I ~ ,V ......

~ ....• _- .... ......... .h ___

.~.

I h> IT I "\( I I Pl?cSSUl?cOli'O?lNI?AOIATOI?

! ! x·Closed Tunnel..,. Open7iJnne1. I 100, S/t6-squQrecell, 48oeep.

i + Ali-now- 5.70/bs./sqft/sec.

i I Av-Oens/fy- a0750Ibs./cu.fl.

H -Ii'r$.o~or - ---j R-/feadl?esisloncelnlbs/sqf/

420 2 tJlstan~e tromli'earF"ace -Inches.

4 6

......-

r

I ! ! i

I I

Ai I -~~ .. -.. . _ .. - 1-- .... --- ··_··_·r·_··- -- --- --.- _._- ··1 ..-0"""" I

!\ .! R

i I

i ~. Iff . -.-. ... _ ._l.

! ~ j ,

...... PRCSSURC MOPIN RAOIATOR .

! ".J1 X' Closed Tunnel. ~ ·Openlimnel C-il. '1&£ cirr:ular ~ell, S "deep.

rl Airflow- 5. 70/bs./sq. fl./sec.

r-----Ii'fdlol,r -AlrOensify.o.07.50Ibs./cu.ff. R-lieQd Resislonce in/bs./sq.11.

4 2 0 2 4- 6 Di$tan~e from.Rearrc~e -1i7~hes.

If)

! T Ai .... --1-...... ~. ------ ----., .. ------ .----. - .. --- ----- . .

~ T /' PflCSSURC MOf'lN RAtJlATOII

~

1\ .-Closed Tunnel. ,,·Open 1iInnei. - ,-,-O-/, Irrequlorcell, c-'lfJ"deep. AlrF'low- 6..50Ibf!cib ''j..sec. R

~ AirOenstly- 0.07. 0. z,s. cu.ff. - --r-

I Ii' -fleod Res':slonce In Ibs./sqfl. I I

s

j \ I

I I --- -_._---- -. _ ..

! I ..-.. i I :.--I

1\ f\ I!#' I I

v

!i y I 'J ./\ I! I

I V 1: : 10

·1 T -IS"

PlotO.

I -llodiol0'l

.:t 2 I 0 I ~i$lon~e from Rear fO~e - Inche.$.

10 ANNUAL REP ORT NATIONAL ADVISORY COMMITTl~E FOR AERONAUTICS.

3. The results seem to indicate that the differences for orly observed between pre sure drop and head resistance per unit area were appar nt rather than real, for the present measure­ments fail to show convincing evidence to prove that pressure drop is unequal to head resist­ance per unit area, provided the former is measured at such sections acrOs the air stream as will eliminate the effects of changing velocity at ntrance and exit of the radiator. On the other hand, the curves show clearly that if pressure drop is measured between two piezometer rings 1>i inches (3 .2 m.) from the faces of the radi tor (as in the tunnels described), the result will be a value somewhat higher than the head resistance per unit area, and the discrepancy will be greater for some types of core than for others; for in some radiators the high-velocity jets issuing from the tubes persist for a greater di tance than in others, before uniform distribu­tion of flow across the channel is regained. In the earlier wod, the air flow through the radi­ator was used as a basis for comparison between pressure drop and head resistance, and errors in the measurement of air flow through the core in the open tunnel are responsible for some of the discrepancies noted; because the measurement of air flow is attended with some difficulty, and is subject to errors that appear to run as high as 8 per cent, and which lead to errors in head resistance about twice as great, since head re istance is proportional to the square of the air flow.

4. A comparison of the curves for the typed and o. 100 indicated a marked difference between radiators in the amount 01 contraction of the jet at entrance and exit. The type 0-8 is made of circular cells expanded to hexagonal form at the ends, and the transition from one form to the other furnishe a very crude streamline form, but one that i much better than that of the type o. 100, which is made with square cells, the ends of the water tubes being merely pressed together to form a joint, with no approach to a streamline form. The difference in the form of the curves for these two types is dOubtle 's due ill part to this difference in form of en­trance, and also in part to the fact that one ha circular and the other square cells. If, in the curves for No . 100, the loops at entrance and exit are interpreted as representing the contrac­tion of the jet, and if it is a sumed that the air stream fill the tube in the part for which the pressure curve is straight, Table II shows that the tube is filled for only about 82 per cent of its length. Whether thi condition indicate that 2 per cent of the waUs of the tube is scoured by the air stream, and the remaining 18 per cent covered by air that is turbulent but not rapidly changed, is a que tion that must be postponed until more i !mown about the conditions of turbulent flow in the radiator tubes.

5. If it is a 'sumed that the loss of head in the part of the tube for which the pre. sure curve is straight is due to skin friction, Table I show::! that the frich n alloss is two-thirds of the head resi tance for 0-8-H-inch (0.87 cm.) circular cells, 5 inches (12.7 em.) deep-and one-half of the head resistance for o. 100-!o-inch (0.79 cm.) square c lls, 4.8 inches (12.2 cm.) deep.

6. The effect of supplying heat to the radiator is to increase the press 'e gradient and the pI' sure drop between the front and rear faces, the gmdient being increased by about 15 per cent in the ca e of the type C-9->i-inch (0.64 cm.) circular cell::!, 4inche (10.2 em.) cleep­for a mean temperature difference of 110° F . (61 0 C.) betw en the water in the radiator and the entering air.

T e statement has been made by other investigators th t increase in head resistance due to temperature difference can be computed from the increa e in momentum of the air as it is heated and consequently made less dense; and it may be ell to point out that while such reasoning may apply to pressure drop for a given ail' flow through the radiator, there is some question about its applicability to head resi tance at a given airplane speed.3 The expansion of the air while it is being heated in the radiator tube tends Lo do two thillgs: To push the ail' out from the rear face at a higher velocity than it had at entrance; and to de\elop a back pres­sure, acting against the pressure that drives it through the tube. This back pressure tends to retard and reduce the air flow, and by so doing, to decrease the skin friction and consequently

• In considering tbis problem, it must be borne in mind tbat thero are two speeds concerned-tbe rate of Ilow of air through the radiator, which may be expressed in pounds per second per square foot of lrontalar n; and the linear velocity with which the radiator may be regarded as passing through still air. Tbe comparisons made throughout this report are based on equal rates olllow through the radiator.

---~- - - '----~----------

", 1

~1Ul. -~~ ~ ~ . f.' to1 ,..I V ~ ~ ~ ~ ~l~~ ~ lI-

~~ . ~ [l~ ~~f',;

"' ;t Ii. ~ . 4 I' f\ ~.1

;"I I&! iIt~ II ~,f\ ' j ~i I:! ~I l) ~J r' [\ . j ('1 ~.

~.( ~( ,~ [,( ~ 1'1t_

r( ~{ ~ ( [ ( t:( I, r;.( ~( fi I' I} I) l" 1 ~/l

rw l' D If ~ i~ r'l ~~l ~

~ , ~ 1.-' I; 1\ ~~ ~~. r' ~d~ ~ i,i I

~~ ~ ~~ ~ ; \1.

rr~~,~ :~ . ~~ ·'~l~'~ ~ .,-FIG.9. FIG. 10.

Spery radiat or. See plots 5 and 8.

PRESSURE DROP IN RADIATOR AIR TUBES. 11

the head resistance. On the other hand, a decrease in air flow through the radiator requires that a greater part of the approaching air shall be deflected around it, and this condition tonds to increase the head resistanee. l "ltil experimental evidence is a,ailable, the question of the elIect of temperature difference on head re:3istance will be left open.

•

TABLE I.-Mass flow of air is in lb. per sec. per sq. ft. frontal area.

Radiator. Mass

flow of air.

Gradi· ent,

Ib./ft.' per in.

Length over

which gradient

is uni· form, in.

Friction Joss,

Ib./ft.'

Head resist­ance, Ib./ft.·

Friction loss as %01 n·.R.

1----------1·-----------------------0-8 ......................... 4.60

1.

02

1 4.3j 5.70 1.57 6.50 2.16 7.00 2.46 7.80 3.02

4.4 6. 75 9.3

10.6 13.0

6.9 10.5 13.7 15.9 19.7

64 64 68 67 66

Mean.................. .......... ........ .. .......... .. .... .... ......... . 66

No.1oo .. ....... . ..... ..... . . 4.60 5.70 6.50 7.00 7.80

1.

02

1 1.58 1. 95 2.27 2.60

3.6j 3.7 5.7 7.0 8.2 9.4

7.4 11.3 14.7 17.0 21.1

50 50 48 48 45

Mean... .. ............ ........ .. .. •....... •. .•...... .......•.. ...... .• .. 48 C-9(cold)................... 4.60 1. 19 1 j 4.3 ...... ...... .. ..... .

S.70 1. 79 6.45 ................. . . . 6.50 2.40 1 3.6 8.65 ..... . . ...... ......• 7.00 2.67 9.6 .... .. . ...... ...... . 7.80 3.60 13.0 ................... .

0-10.................... . . .. U8 ~:g~ /} 35{ 18:g :::::::::: :::::::::: 7.00 3.46 . 12.5 .......... .. . . ... .. . 7.80 4.00 14.0 . .......... .. ...... .

TABLE II.

Deptb of radiator . .. ........ . . ...... ... . ....... . . .... ....... . . inches .. Length of air tube ...........................• . ................. do .. .• Length over which gradient is uniform ......................... do ... . Per cent of tubeiength over which gradient is uniform ............... .

TABLE IH.-Effect of heat, radiator C-9.

Gradient.

Air flow. Percent 1---'1---1 increase.

1 _______ I_c_O_id_._ .~ __ _

4.60................ 1.19 1 1.35 13 5.70.... ............ 1. 79 1. 97 10 6.50....... ... . . .... 2.40 2.78 16 7.00........ .. .. . ... 2.67 3.18 19 7.80.... ... ......... 3.60 4.14 15

Mean ...•.••..........• ·1···· ····· ·1--1-S-o

0-8. No. 100.

5.0 .4.6 4.3 94

4.8 4. 4 3.6 82