You are on page 1of 8

'd. No.

64C
er
*The performance of drained joints
k,
BY D. BISHOP,
le
rs
is
I J.o.WEBSTER,
AND M. R. M. HERBERT,
c\9
rg MI}I-ISTRY OE TECHNOLOGY, BIIILDING
lr
RESEARCEI STATION, TIN-ITED KTNGDOM.
al
ls

Introd,uct,ion Performance ol the hori.zontal joints


There was (and is) very little factual information about The performance of the horizolxtal joints is assessed
the performance of drained joints between external clad- from observations made from June 1960 to July 1962.
in ding panels and the amount of water penetrating into During this period rain feil on 213 days of the 595 days
ty them. I{ence in 1960 the Building Research Station erec- the rig was commissioned. Subsequent to this period
ted on a site at East Kilbride in Seotland a rig desig:ted modifications to the rig prevented"further use of the
a, to explore the performance of joints under conditious horizontai joints.
1s of natural exposure in a simple and direct way Each joint took the form of a p.v.e. tray divided into
l- (Fig. 1) 1r1 and it is the data collected from this rig three sections (Fig. P) and it was intended that the
t- that forms the basis of this report, An alternative app- trays should be completely sealed at the back to prevent
IS roach by Trygve Isaksen (2) (3) employed laboratory infiltration. In the event the sealing of the back of the
)y test rig has yieided much importaut information. The joints was iess than perfect. The profiles of the five
data derived from these different methods is broadly joints are shown in Fi,g, 3a and, b and it will be seen
complementary. that the joints took the form of slots 3 ft (0.91 m)
long . 3/a in. (19 mm) high in t}te face of the wall. In
the control joint a weir at the plane of the wall face
ensured that water entering the joint was trapped and
caught in containers. The remainilg four ioints pro-
tected their weirs to different degrees'

)n
e

Ie
le

Fi.g. 7. Test ualt of rig (the oertical rain gauge atd, the
horizontal, joints can be cleady discerneil).

The si,te
The rig was erected 650 ft (208 m) above sea level
on a site about 8 miles (13 km) south-south-east of Eig. 2. P.V.C. trag forming horizontal ioint uith' S in.
Glasgow, and 5 miles (8 km) from the river Clyde. The upstand,.
exposure was moderately severe, with a 'driving rain'
index (a) (6) of 6mz sec-1 and the rig received, on The observed performance of the horizontal joints
average, rather more than six storms a yearr lasting clearly indieates the marked ieduction in water pene-
for more than one hour and producing a mearl hourly trating'the weirs protected by sioping or by upstands.
rate in excess of 21 m2 of wall surface- An occurrence The frequency for catches of a range of quantities is
which confirms the severity of the actual exposure on shown at Fi.gs. 3a ar:d b.
this site (5). A detailed examination of the data revealed ano-
malies: for instance the penetration of joiats (C) (D)
and (E) did not always oceur on days when excep-
The rig tionally heavy falls of rain were recorded: also a day-
The rig was d timber-framed structure mounted on a by-day examination of the records shows that the
concrete base and set so that the test wall faced catches in the three coutainers within any joint varied
approximately south-west, the direction of the prevailing considerably on some days. Il is thought that these
wind. It consisted of a series of timber frames suppor- discrepancies arose either because the seal at the back
ting the ply-faced, timber-framed panels comprising the of a joint became temporarily dislodged or because of
test wall. In order to produce rouSyrly the sme surface <streaming>. so that water on the face of the wall
characteristics as concrete, the test wall and the sur- ran in rivulets. l But hazards of this kind oeeur in
faces of the joints were paint-harled (that is' sharp practice and the performance of actual ioints is llkely
sand was sieved over recenlly painted surfaces.) to be equally variable.
A modified Croisset pot-type vertical rain gauge was At the end of this period of two years the rig was
built into one panel of the rig, with its centre 5 ft extensively modified, including replacing the cill with
(1.52 m) above ground level. An anemometer and a a 3 in upstand by one with a 2 in upstand' blocking
directioial wind gauge lvere set about 20 ft (6 m) the two outer sections of each horizontal joint aud
directly in front of the rit. the provision of a perfect seal at the back of the re-
g4L
i
t
FACE OF WALL,
\+-s%- toI the features of the desip. being investigated: (in the
i following figures the divisions of the trays are irxaticated
-l
i
by broken lines and the letters, a, b, etc. denote the
'J individual compartments, a at the front).
7
-o 2oi
The rig was modified several times to examine, iu all,
c the behaviour of seventeen vertical joints. The pertor-
mattce of these Joints will be discussed in terms of the
daily catches, Eor convenience the catches in the control
joints are expressed as inehes of rainfall on a vertical
WEIR IN PLANE OF WALL. surface: inches of rainJallrttfg:gms catch in 11/16
in (19 mm) joint. No attempt has been made to fit
other than..straight llDes to these data for two reasons:
+7,{ |
the actual values measured varied from joint to joiat,
hence refinement in analysis is Dot justified: also at-
tempts to flt a curve to pass through the origin (a
63MM
necessary condition in practiee) would give less weight
to the fewen and more important catches in heavy
storms.

WEIR PROTECTED BY LEVEL CILL.

-. i
t.r9MM
I
\
t. T-l

WEIR PROTECTED BY INCLINED CILL,

Fig. 8a. Profil,es and perlormance oI llorizontat, joints.


ply foce
maining section each joint. During the next 18
of
months water penetrated onlg ci[ A (the control cill),
vvhich continued to behave much as before.
Edgcs verticol
This was an entirely rrng>.pgsfsd result, especially as trom
previous experience had shown the high potential water
ioad .on.horizontal joints.
When the strips sealing the backs of the horizontal
joints were removed both cill B (with a horizontal cill)
and cill C (with a sloping cill) admitteit water in times
of heavy rain.

Th,e performance ol Dertical joi,nts


X'ig.{ illustrates the general arrangernent of the vertical
joiDts, all of which were unfilled and scaied at the portmcnts
back. At the foot of each joint a series of trays were
provided to trap water that had entered the joint, the Fng. 11. Diagrammetric sketch o! oertical loint, protected,
arrangement varying from joint to joint, depending on agadnst sitl,e flow, the sketch cut-a7oau to show com,part-
,nents oI the col,lecting tray.

Belataoe perlorn'Lance ol nsertical joints of iilenti,cal


wiiltlt,
o
; In this section the comparison wrll be between the three
z control joints used at different times during the experi-
B ment (some of these times overlappeat), and the vertical
rain gauge which remained in one position throughout
the period of the tests.
The performance of joint 6, the control joint which was
WEIR PROTECTED BY3IN 76MM, UPSTAND in the rig for a little more than 2afz !e*q is showa at
fi,gure 5. In the first seven months of exposure its per-
206
formance differed from that in the next 22 about
t/e months, but the difference is not significant 2 and the
combmed performanee is shown (line A).
At the end of the secoud period plugs in the barriers
-
; at the back of the joint were removed in order to
z permit air movement through the joints, In these con-
o ditions the perdormance of the joint 6 altered signifi-
cantly as indlicated by line B.
Two other control joints were included in the rig
during other periods and the performance of both these
Fig, 3b. Proliles anil perlornlamce of horizontal joints. joiats differed significantly from that of joint 6. Whitst

242
he Table 7
ed
he Catch in joint (em)s- 95%
Joint No. of observations 1118 [a + b (in rain in confidence limits
:ll, vertical rain gauge)l for b
,r-
he 6
:ol (A) (plugs in) 146 T6 :0.15 + 3.58 (v.r.g.) 3.12--4.03
raI
6
16 (B) (plugs out) 72 To :0.09 + 6.61 (v.r.g.) 6.16-7.06
fit Te :0.71 * 4.22 (v.r.g.) 2.42-$.02
18:
9 19
lt, 11 19 T11 :0.15 + 1.81 (v.r.g.) 1.00-2.64
lt-
(a N.B. Periods A and B sip.ificantly different.
ht
ry the catch in joint 6 varied between 3 and 4 times that in the shielded joint, ligure 6, Iine D. The two sets of
in the vertical rain gauge (neglecting the constant results obtained from the contr6l. and shielded ioints
term), this ratio for the other control- joints was 4.2 (Table 2) confirm that the water eutering joints frotn
and 1.8 respectively both results being based on com- the face of adjacent panels aceounts for about 80 Va
paratively few observations. of the total, and that this proportion and the total are
Despite these differences, it is ciear that the catch in both increased when the air barrier is imperfect.
vertical joints is considerably gteater than would be
obtained by multiplying the rainfail on an adjacent verti- Table 2
cal surface by the face area of the joint. In other words Bummarv of results: catchee in control iodnt.
only a proportion, and perhaps a small proportion, of
water entering a joint does so directly as rain, an I

I Catch i:r control joint


hypothesis examined in the next section. I

Sid,e llous 7st contpd"i,som


The effect of side flow was studied by comparing control joint versus
catches in the control joint (figure 5/ with those in vertical rain gauge.
joiats shielded from side flow (ldgure 6/. Because of the plugs in 3.6 . vertical gauge
variable performance of apparently identical joints, how- plugs out (control joint 6.6 . vertical gauge
ever, it ig now realised that it would have been better only)
if a single joint had been used and the performance gnil, comparisom
before and after fitting deflectors compared with the
control joint versus shiel-
vertical rain gauge.
year ded joint
Figure 6 displays the data collected during a 2 1/z
period, and the best line C, fitted to the data. If the plugs in 5.5 . shielded joint
constant is neglected it will be seen that the catch in plugs out (both joints) 8.3 . shielded joint
the control joint was almost six times that in the shiel-
ded joint. When the main trial had beeD completed,
plugs were removed from the backs of the joints. The Comparison showed that ttre ratio of water caught in
catch in.the control joint then rose to nine times that the shielded joint to the control joint was roughly
similar to the ratio of the catches itr the rear compart-
ments of these joints. A day-to-day comparison showed,
>t however tlrat although the catches in the back com-
l1 8in, I

r 2o3mm. partments of both joints were of the same order at


times of low and moderate rainfall, the catehes in the
rear compartment of the control joint were much
larger than those in the shieldbd joint in times of
d heavy rainfall (Fig. 6/. This suggests that in all but
:- t
i,rz s hqavy rainfall only w"ind-driven ralr penetrates deeply
iato joints: and that in heavy rainfall some of the
water entering a joint iadirectly also penetrates to the
back of joints, and considerably increases the water-
o, b,c,d, denots componment5 load on the air barrier.
in collection troys
Thus when the back of an unfilled vertical joint is
sealed and there is lo possibility of air movement
tlrough tJre joint:
5
,/. -
o a) about 20 Vo of the total catch enters as wind-driven
'6>' / rain: some of this penetrates to the back of a joint,
6\
z-
o-
!

r
I
I
^:b) dl but tlie majority strikes the sides of a joint and
drains to the foot:
-oi ;! I l-'/ I
b) about 80 Vo of. the total catch enters a joint as water
.EE from the face of the adjoining walls. Hence any
:.}
UJ
z
r< o- fofm of sufface treatment tooliog, argins or the
;g like would be expected to- alivert water away from
.'6 I
I
- anal sigrrificantly reduce the waterload.
a^ :4 I

l I
ajoint,
o o.t o.2 0.3 0.4 0.5 0.5 0.7 0'8 0'9 t.o in./doy
Rorn in verticol rorn gouge
When the backs of the Joints are not effectively sealed
the catch, for a given rainfall, is iacreased by roughly
. odlt ..!.h rh.n Joi^t 5 ...1.d .l h!.k
tr ol Doct
150 to 200 7o abolt 15 % of. the total catch entering
' not taolt{
as wind-driven rairq the remainder entering from the
Fig. 5. Perfonno'nce of conttol ioint, face of adjoiDing waIs.

P48
ru,
I
2O4MM

2i,tz.s

lo .}OMM
2O3 mm. )6 9.5MM

o.b.c.d, denote comportments in coll€ction th;',t7.sMM


troys.
t".25.4MM

c m.3

rooo
o
! 800
!
E
E
,9 b_1 too& Fig.'1. Elfect ol ioint uitlth.

, c*" )+9 ,.& except for the position of the vertical grooves. fireir
performatrce was observed to determine the proportion
/ (,"" of water caught in the compartments in front of the
=o
o 7
4rooo \
.tront edge of the vertical grooves, in the grooves, and
between the grooves aJod the backs of the joints. The
performance of joints 11 and 9 B was compared to
o 2000 3000 4000 5OOO cml
determine the effect of vertical aual incli.neat gxooves
Doily cotch - control Joint 6 (washboard grooves) ou the amount of water pene-
.
trating deeply into joiats. Finally the catches in the
D€ily .olch rh.n ,oi6t 6 r.ol.d ot bd
o . r r aotxoladdbcl compartments towards the back of the five Joints were
examined to determine the waterload at the air barriers
]: at the back of the jobrts.
looF The total catches were compared with that in the
control joint to verify that these joints were exposed
I to roughly similar waterloads (Tablg 3).
t q
!
o 'oI ffi a .hi.td.d ,oinr
"o,.n ?able 3
'oi I co,.tr in Gonrrol ,oi6t
a
) ,of Joint Catch in joint (cms; 95 /6 confidence
t :a -l b (eatchjoi:rtG) limlts for b
L ,"f
4 T12: 14 + 0.50 T6 4.4H.53
o 9..9- 1q T a: 105 + 0.80 T6 0.7H.84
O.lroe t-9 lo-2r 30-59 60-99 OO-t99 2OO.299 sO-35OcD! 9a Fge:,, 104 + 0.77 T6 0.71---0.83
iOapori:on ol cotcb.r ot baak (in coiportiant d) or
ona rhi.ld.d roint.
conkol 11 Tir: -i- 60 + 0.84 T6 0.79-0.89
Fig. 6. Pdrforrnance of shi,eliled ioint.
Only small catches (typically less than 200 cm3/day)
were recorded on days wittr low rainfall compared with
Width of joints the catches (typically more Uran 1000 cm3/alay) on
Early ia the trials four shielded joints of the design days with high rainfall and, although the former were
shown at Fig. 7 were studied to obtain irformation more numerous, they had little effect on the regres-
about the relationship between the width of a joint and sions. OnIy one joint of each type was studied and the
the amount of rain directly entering: it. The data ob- effect of minor differeuces in joints of the same tJE)e
tained were not very reliable, probably because of the is not known, hence no attempt is made to assess levels
detailed desiBx of the joiDts &nd the lgading to the of sigreificance.
contalners. The catches were i! proportions 0,52 | O.7O I The catches irr three zones of joint 4, together with
0.64 :0.80 for 7s in,, art in., ]al,l. in. and 1 i]I., wide straight lines fitted to the date are shown at ligure 9.
joints respectively. Similar reiationships were calculated for the remaining
Of these, there is no significant difference in the joints, neglecting the constant terms which are signifi-
inner two proportions. Hence the observatioDs suggest cant only for small catches aud were not important, aIrd
oniy that the effeet of altering the width of a joint the proportion of the total catches in each compartmetrt
from % ir-, is to i:lcrease the amount of driving rain are summarized at J?igs. 70a) and b) .
entering it by roughly 50 Vo. When this conclusion is
combined with that of the previous section that the The performance of joints L2,4,9a aDal 6 (the control
indirect component is predomlnant - be seen
it will joint) are illustrated at ?ig,8a),9 a:t.d 10. It will be
- have a critical
that the width of the joint should not noticed:
effect on the waterioad a joint must withsta:rd. (r) The praporLion oI Water caught towards the front
of the three grooved joints was higher than that
Eflect ol uertica,l and inclineil grooDes on thc perfor- in the same seetion of the control joint.
tnance of Dertical joints (II) Joints 12 and 4 per{ormed similary; joint 9 b hadl
The effect of the desip of the profiles of the panels performance intermediate between joilts 12 and 4
on either side of a vertical Joint was studietl by obser- and the coatrol joinL
ving the performance of the five Joints shown at Fig. 8 (IU) The catches in the compartments immediately
a) and b). below each of the vertical grooves were propor-
These joints were only half the depth of the orig"inal tionally smaller thall iD the compartments in
verticai joiDts and joints 12, 4 and g a were similar either side of them. This refutes the assumption

244
GROOVE

98MM
3/e
ioo
N
ao
60
l?.5 MMrl o bg d a, JorNT l2 l- ao
z
[,o
{o
U b
L

6ROOVE
F
r roo N
0
lBo
(
l7,5MM obcgef JOINT 4 U@
i 40
/20
u
t-o
t,
'57
t4t' {
i
{-

L
o GROOVE
N
60
tJ)
J GO
l7 5MM,l q b c ac Jott'lr 9o {
F 40
o
F

l3 ,,/4'l
0
:F
IOMM. I6MM. 16 MM.I6MM. I6MM. I6MM.IO
{
)
l co
I 80
l
U 60
l7.5ut4,t o b, c d e I JorNr ll 40
_t_
o

FAONT I )N ? rN. slN. AACK


OEP?H OF JOINT
Fig. 8a. Perlormance ol joints wdth oertical, or incldned, grooltes.

made in the interim paper (1) viz., that the of the groove has little effect. However, the com-
vertical grooves act as drairqge channels, collec- paratively low waterload on joiBt 12 must be
ting water flowfurg across the face of the Joints. remembered and, on account of this, the per-
The relatively small catches in the compartments formance of joint 4 was probably the more signi-
below the grooves may have been caused by the po- ficaot.
sitions of the divisions betrveen the compartments
(see broken lines on Fig. I a) which may have Duin6 the experiment, joiut 9a was replaced by joint
directed water draining from the edge of grooves I b (Fig. 8 b/, similar in size and arrang:ement to joint
away from the compartments immediately below 4, but with inclined (washboard) grooves in the faces
them. of the joint, both behind and in front of the verti-
(rv) Although similar proportions of water were cal groove. The amount of water caught in this joint
caught ijl the first r/, in., (12.7 mm) of each
(To.:-+31 + 0.13T^) was sig:riflcantly smaller tlan
grooved joint, from thereon the 'rete' of catch ls thdt caught in the Sther grooved joints and was, in
steepest in the portion of the joint ia front of
fact, smaller tlan ttrat caught in the joint shielded
the vertical groove. This suggest tbat water enter- from side flow. Thus it appears that the washboard
grooves returned the majority of water entering the
ing a joint clings to the froat edge and is cons- joint to the face of the wall.
trained from extendi:rg towards the back of the The performance of joints 11 (with six vertical groo-
joint by the edges of the vertical gtooves. In prac-
ves) and I b (with washboard grooves) are compared
tice the extent to which this effect occurs
can be easily demonstrated would depend- onit with joints 4 (vertical groove, 2 in., from front of joint)
and 6 (control joint) at Fig. 8a) a;o.d b), 9 alad. 70-
the waterload, the roughness- of the surfaee and Bearing irx mind the dlfference in waterload on the
?
ts-
the sharpness of the iaterrupting edges; matters joints, it will be aoticed that:
for separate studies. Both joiots 12 and 4 arrested
about 95 per cent of the water entering them at (I) The proportion of water caught towards the frout
a depth of 2 in., (51 mm), this suggestlng trrat of the grooved joints was higber than that ln the
beyoud tllis range, the depth of tIIe joint i! front control joiat.

245
!.eel.
98 mm.
lNcLtxED 6RoovEs lNcLrNEo gRooY!s

E
z t00
\J
g0
d
.T -- 60
il,; lotHr 9b

I
/16 tl

ll J fr|il 40
-_l t-
I 20
o
l
0 t) e f
o
d. i(oNr BAcK
u 3x
F
DEPrH oF JorN?

(o
207 aa.
t00
ts
o
F 00

U
:
F
60

CoutRor- Jorur 6
-r
llt t) +0
/tL t7-5 ffi. id b d )I
20

0
U I

!R0 r'lI
l' Zl

DE PtH oF JoINT

Fig, 8b, PerJormance oI iodnts uith Dertical, or incltneil, groooes.

(II) There was no appreciable difference in the per-


formance of joints 4 and 11; apart from the pro-
c.! portion of water caught il the first g"oove. Al-
a c g." I though as far as precast concr€te constructioa is
r500 concerned joint 11 is rather impracticable, joints
,9- with a groove very near the front edge could be
made in timber, pla.stics or extruded. metal.
@)
-r ^'"/
".rr

(III) The distribution of the water eaught in joiat 9 b


800 clearly demonstrates the way in which the i-nclined
1y a-' BTooves drain water to the front of joints. ID this
4C0
,v joint the vertical g?oove divided t}re joint into two
, .Ar
I

I reg'ions; all the water penetrating to the vertical


.r* l
groove draining to the front compartment, all the
400 800 i20a t600 2000 2400 .t2oo *oo 10oo
ToT r rArcd water penetrating more deeply draining to the
compartment next to the vertical groove. Although
18 per ceat of the total cateh in joint 9 b pene-
? ,,! trated more deeply than the vertical groove the
f
1
4ooo
a+ b+c ) actual quantities involved were of the same order
! tzoo as those in the other grooved joints. It is probable
I

s9 2- that more of the water entering this joint would


; 2100
I I
have been drained to the front compartment had
:,'-
)2 the inclined grooves not been interruped by the
vertical gTooves, a feature which argues for the
ili
o t600
,{ omission of the vertical grooves, aJld heDce the
] eoo baffle or the use of a cruciform baffle not requi-
ring a vertical groove to secure it in place.
O rCO 800 t21o $oa 2ooo 24OO 28tu a2}o ,600 lo,o cra
To:AL cAtcH
Waterload at the back of joints
Thus far, the discussion has been concerned with res-
tricting the penetration of water into joints by desig.D
of the joint. ft is important to study also those occasions
1e+f)
in which water penetrates deeply into joints, beeause un-
F6 less a baffle is provided it is this water which must be
excluded by the a,ir banier.
Individual catches in each joint were plotted and
:: 150
,P ')/ "9- I
regression analyses showed that the quantities caught
were sip.ificantly related to the totat daily rainfall,
=o
but not related to the intensity of rain this was
surprising because the latter relationship -would seem
go

.>a
0
to be more likely (Fig. ,r). Eor the present purpose it
0 {00 e@ Q@ t60o 2ooo 21oa ,eil t2oa $00 4A00 ea' Seems more iinportant to consider the actual catches,
TorAL ca_cP because these are the waterloads joints must resist,
Fi.g.9. Catches i,n cornpartn ents of joint 4. rather than comparing the waterload with the total
6, A.
catch or with any-other measurable factors. Catches From.Fig. lPo it will be seen that all joints catches
peuetratilxg more deeply then the vertical gEooves are sxggsding 20O cms occurred on oaly 5 per cent or fewer
compared at Fig. 72 a and the catches in the compart- of tlre days;.on which rain fell. A-lso smaller quartities
me[t next to the back of each joint at Fig. 72b. of water penetrated beyond the vertical grooves set 2 in.
from ttre front of joiats 4 antl 9b, than penetrated
POSITION OF G ROOVES
beyond 4 in. into the coutrol joints (it is interesting

roo
@ru
t24
w,
9q
to note that the presence of washboard grooves did not
malre any significant difference to this aspeet of per-
formance). Joint 12 had a less satisfactory, as would
90 be expected and joint 9 a a better performance thau
ao joint 4.
[-
7
u 70 40
U rRoL (6)
e
J. o9
u I
G
50 lr
,0
40 1
( ['r5h
;s grootcs)

30
F o9a
I 20
0 q
l to Joirl 4 2' lron
(J
o ? sAcK oF 12.4,9a Joinl 6 coITRoL)
FAONT
{ ru
OEPTH OF JOINT
t0

nl tz (q'fv. 1'1roa iacc)


t- .t:_-:.
{ (o) Jorsrs wrrH gINGLE GRoovEs Joinl 9a ..1

i .0
, jrot
0 r00 200 t00 100 4o
u "n!
o .u (ruclrNeo GRcovEs ) (a)
il( N,4ULTI . GRO(>V e) CArcHEs rN coMpARTMENTS.BETwEEN THE vEATIcAL 6Roovs5
Atr0 THE BAC|(S 0F Jottrr9

o 4
J t0
eo
F
t- o '7()
f- \
)- bU ;q.
t- rnor- (a) ti ry9r
Joirt
ri
is
F
:s ( 40
J
i \ lr-.
)e J t0
l )U loiil' ioint
b l i.'6r')l-. "12
-{{_-_
U 20
d
is 0
,o 0 100 200 100 loo ioO c.l
rl F RONT I 2 5 6ACK CF 4,9b,ll
1b1 .-crrcxes rN coMpARtMtllrs NExr ro IiE llcKs oF JorNrs.
.e DEPTH OF JOINT.
Fig. 12. Open. ioints: catches at the back ol the 1)ertical
:e
h
( b) Jorxrrs wlrH
MULTr - GRoovES grooues.
)- Fig- L0. Comparatl,ae perlormance of grooued, joi,nts.
e Comparison of the catches in the compartments next
r to the back of the joints @dg. 12b) demoastrates the
e effectiveness of the washboard grooves in leading water
al away from the back of joint 9 b. No catch in joiDt 4
i ':t6 ,r7s;. a b c g e , exceeded 100 cms, the next most satisfactory ioint, The
I importance of these results must be emphasized and it
e 'arould seem that the waterload at the back of drained
joints, especially those protected by washboard g?ooves,
3
F- is so small that intermediate baffles could be omitted
in all but tall buildings, or builctings erected in very
c^? exposed positions. Naturally the omission of baffles does
!00 reduce the factor of safety implicit with all drained
joints incorporating baffles and slightly increases the
;80 ,,v demand for careful workmanship by operatives fixing
/' air barrlers.
i60 ,"/ The performance ol jodnts proteoted by ballles
,"$l
.. +o
Y The interim report (1) recomrnended the use of loose
baffles in vertical joints and this was foilowed by en-
s
Y Vr!urr rMrtM qUmeS ff0m matry arChitects wishi:a6 to include similar
joi:rts in fqll fuuildings. Therefore three vertical joints
o
:97
-T
!68
?a)
400
with various baffles were included in the rig for a
period of almost one year (Fig, 13). These joints ad-
-0
0
/ )1.2
il-6
145
2r0 mitted little water to the compartments at the back of
I 2 3 I I 6 ? 3 a'n the baffles; aetuel joints will, of course, have to contead
Darly ra:niall ' vuLicaL ran qauje with water flowing from the walls above them, and
Fig, 11. Open ioint with grooue 2$2 in. lrorn lront rain- accommodate dimensional toleranees (including varying
lall (oersus) aatchet at baclr oI Derticd,l groooe. rvidths within any Joint) and vagaries of workmanship.

241
il

securely and became di.splaced in gusty weather: aiso


the grooves had rounded edges, a feature which probably
?O5 MM
detracted from their performance. Both joints caught
about 60 per cent of the water caught in the control .. .
rl;#;
joint and therefore received a simiiar waterload to the
other grooved joints. Catches in lhe rear compartments
of these joints were compared with joint 4 (the joint $&
361
with a vertical groove 2 b. (5L mm) from the front), ttr
17 5uv the joint with a baffle in the first pair of grooves had
a slightly better performance ,and the joint with the
FRONT
baffle in the second pair of grooves had a markedly
FACE
MM 4IMM 4I MM
better performance.
The baffle in the third joint was a hollowneoprene
(a) tube Eig. 78c fot eutry into the joiat these tubes are
TRIPLE 6ROOVED JOI}IT WITH BAFFLE collapsed by exhasting the air and are subsequently held
i:r place by friction). A 1 in. diameter tube was entered
into a 11/16 in., (17.5 mm) wide joint without clifficulty
and remained firmly in place; it must be remembered,
gu

20] MM
however, that few actual joints have parallel faces and
that variation in joint widtbs in any one joint may be
greather than the tolerance for a tube of a given dia-
meter. In the event the tube was placed in two joints;
MM
in the first at almost the piane of the wall face; i:r the
second positioned about 1 in., from the front edge of
FRONI the joint: in both positions the neoprene tube effectively
prevented water entering the joints.
(b)
5r[6LE 6RoovED JorNT wrrH BAFFLE.
Relerenees
1. D. Bishop. Weatherproof joiuts between precast con-
crete panels. Ttre Builder, 1962, Jan. 5tb.;202 (6190),
2t4.
25MM.
F RONT 1n
ofots 2. Trygve IsakseD- Modeme husbygg'ingEtekntkk. R,e-
port No. 71, Norwegian Builtling Research Institute,
Oslo 1962.
ll.5
3. Tygve Isaksen. Desip of joints and driving rain
penetration. Report No. 93, Norwegian Buifding
Resea"rch Institute, Oslo 1964.
4. Lacy F., E. and Shellard II. e. An index of driviug
203 ww raitl Meterological Magazine, Lonclon, 91 (1962),
(c) pp.477-184.
JoINT gEALED wrTH NEopF.ENE TuBE
5. Lacy R. E. Driving rain maps aJld onslaught of rain
Fdg. 13. Vertical ioints protectetl, bg baffles. on buildings. (RILEM/CIB Slmposium. Moisture in
buildings, Helsinki, August, 1965).
Two of the joints (Fig. 73a and b) had loose baffles
iocated in,vertical gxooves; in the front pair of grooves
1) Isaksen has shown that rivulets coinciding with air
of joint a, and iro the second pair of g:rooves of joi:rt b.
Unfortunately the grooves were shallow Ieakage are imporlant sources of faiiure.
less than
9/8 in. (10 mrn) deep - not held
so that the baffles rvere 2) level of significance,9S Eo throughout the paper.
-

8,6sum6
Le comportement des joints draiD6s

Les joints draia6s, c'est a dire ceux dans iesquels la pour les joints verticaux:
barridre d. l'inJiltration d'air est situ6e a une certaine
distance derriBre le plan de la surface du mur et Ie la contribution relative i la charge d'eau globale de Ia
joint est ouvert devant et muni d'un 6cran, sont accept6s pluie battante et de l'eau coulant le long de Ia surtace
des panueaux,
pius largement d pr6sent. Cet expos6 examine Ie com-
portement de ces joints, 6tabli d partir d'une instaliation I'effet de la- largeur des joints et de la prdseuce de
rainures dans les faces des joints et de leur situatiou
comprenant
sur la quaatitd d'eau p6n6trant dans les joints,
pour les joints borizoata,ux; les ava:etages relatifs de rs'illures verticeles et iu-
clinees.
I'effect de la trauteur d'6l6vation et d'une imperme-
abtlit6 d6fectueuse sur la.p6n6tration de I'eeu.

You might also like