I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
394
The special transformer appearing in the circuit of Figure 1 is replaced by a voltage-divider arrangement supplying grid bias and the reduced voltage necessary to prevent the gas-filled phototube from glowing. The Type 6557 amplifier tube has suitable plate current to operate the type of relay which was obtained. Time delay is int,roduced by the inductive-resistive effect of a variable resistor in series Tvit'h the relay. +4three-position switch provides for automatic or hand operation and the two pilot lights indicate whether the relay is open or closed (red indicates valve closed and green valve opcn). The relay and photocell are wired into the circuit in such a way that the valve will be closed during the period in which the tube is warming up, and also under such extraordinary circumstances as failure of the light source, sudden ebullition because of increased vacuum, etc. The light source consists of a 32-candlepower automobile headlight bulb fitted with a lens to focus an image of the filament onto the photocell. The solenoid valve was manufactured by the Minneapolis-IIoneywell Regulator Company and requires 0.25ampere st,eady current a t 110 volts alternating current. If the feed enters at. atmospheric pressure, the valve should be designed for a pressure of not more than 15 pounds per square inch. If its rated pressure is too high, the valve Kill not close properly. METHODOF OPERATION.Place the three-position switch in the automatic position and the "time delay" control a t maximum. Cover the photocell opening with the hand and (e) move the "operating point" contrd to the right unt'il the green pilot light turns on; ( b ) slowly move the same control in the opposite dircction until the red pilot light just turns on.
Vol. 16, No. 6
The adp-tnient 13 critical and should be repeated if the icxla~ chatter> or fail> to opcn and close. This unit has been in operation for over a year and has proved very satisfactory. -4nyone with a little radio experience can easily build a control of this type at a total cost of from $35.00 to $50 00. PARTS FOR P H O T O C E L L R E L A Y
1 SPDT senbitwe relay, 1 to 2 ma. to close, contacts rated a t 5 amperes noninductive A.C. 1 Type 6SJ7 receiving tube (Type 6SJ7GT may be substituted) 1 Tvue 930 gas-filled Dhototube 56,bOO-ohm wire-wdund linear potentiometer 2000-ohm wire-wound linear potentiometer 2000-o11m, l0-watt wire-wound resistors lOOO-ohm, 10-watt mire-wound resistor 20-meyohm, 0.5-watt carbon resistor 0.0005-mfd. mica condenser 8-mfd. 450-volt electrolytic condenser transformer, 110 to 6.3 volts at 6 amperes 110-volt, 6-rp. pilot lights, candelabra base 6- to 8-volt, 32-(.p. Mazda bulb, No. 1132 SI'DT toggle switch, center off position 1 2-polr round Bakelite receptacle LITERATURE CITED
(1)
Reich, "Principles of Electron Tubes", p. 317, New York, McGraw-Hill Book C o . , 1941.
New Design of Humidity Cabinet for Corrosion Testing FLOYD TODD, Quaker Chemical Products Corporation, Conrhohocken, Pa.
T
,
HE corrosion-testing cabinet described herein was designed to give an accurate and reproducible compsirison of the relative efficiencies of corrosion preventives which are to be used under indoor storage conditions, including intermediate protection in the process of manufacture. These conditions differ in kind as well as in degree. For indoor protection the rust-preventive coatings should possess sufficient solubility in suitable solvents to be very easily removed. For outdoor protection it may suffice that the compound be roughly removable, for example, by wiping. Furthermore, for outdoor protection any appreciable water-solubility of the protective agent is very objectionable, as rain would remove any watersoluble ingredients, while for indoor protection it may even be desirable, so long as it, does not make the product hygroscopic or susceptib!e to dew or condensation. Testing an outdoor type of compound under indoor conditions, or an indoor compound under outdoor conditions, may lead to adopting materials inadequate for the purpose intended, or to discarding products which would have been most advantageous. This article presents a new testing device for accelerated corrosion tests under extreme indoor conditions (high temperature, high humidity, air exchange, condensation of moisture without washing-out, effect, radiation, and such added chemical corrosive influences as may be desired). This cabinet is based on experience accumulated over more than a decade with all types of corrosion-testing devices, and provides : a more accurate ' zcontrol of z the variables than any other type known to the author. Cabinets previously used include the conventional type, which consists of a clopcd space, provided with a spray or bubbling type of humidifier, means for circulating the Rir, and electrically actuated heater and thermoregulator (4-8). Other types provide for sealed cabinets, operating at high humidity in alternating cycles of higher and lower t,emperature (Ball Bearing Engineers' Committee). The "weatherometer" is a well-known apparatu3 adopted for corrosion tests under outdoor conditions (1). Saltspray cabinets are used for accelerated testing of the relative corrosion proofing efficienries of protective coatings for metal3 which may be subjected to sea air esposuw (2, .3).
:
:
:
:
In all prior apparatus the heating has been centralized in too small an arca. As a result, convection currents and radiation effects have occurred within the cabinets, causing uneven exposure conditions and consequent serious irregularities in corrosion test results. Furthermore, many humidity cabinets are square or rectangular in shape. Such a design is conducive to undesirable channeling of air and moisture in those cabinets which operate undq the usual dynamic exposure conditions. In any thermostatically controlled apparatus, whether humidity or salt-spray cabinet, the temperature fluctuates within a i'ange of 1" to 3" F. (0.5556" to 1.6668" C.), because of the lag in the regulator. These variations may be accentuated by convection currents. Even when the temperature fluctuates within constant limits, the actual corrosion temperature ratio may vary greatly, depending on the quality of insulation and the temperature and ventilation of the air surrounding the apparatus. T h k is illuKtrated by the following curvec, which show the temperature charts in the same cabinet, in which the regulator shuts off t,he heat at 122" F. and starts heating 120",a better than average range :
'z
IZO'
R O O M TEMPERATURE 70'F:
' IZO'
:
ROOfl
TCPIP. 70'F - - W I N D O W OPEN 8 F T FROfl C A B I N E T
ROOM T E M P E R A T U R E
92'F
On the up\vai,d ,slope 011 the temperature curve the air i n the c-abinet i* less than saturated wit.h moisture; on the downward .slope it i h supewat,urated, and condensation may occur. This wndenration i,- irregular and emphasizes any slight surface irregularity of the -amples. Moreover, the type of cycle profoundly affects the corrosion behavior, leading to appreciable discrcpancie.; i n c,ori.o*ion time and particularly in corrosion
June, 1944
ANALYTICAL EDITION
types, dependent on external factors not coIltrolled in xny ra,binet specifications with which the author is familiar. humidity cabinet for corrosion testing has therefore been de-;igned in order t o eliminate the disadvant,agesof prior cabinet'$. Its drsign gives additional advantages inhrrcnt in only this type of c.nbinet.
395
The walls and bottom are heated ovt'r a preponderant area. This heating is effected by means of the vapors of a conatantboiling liquid and i$ therefore ab-olutely uniform and lagless. currents are completely eliminated, Still more important, the temperature "curve" is here a straight line, the cycles of recurrent sub- and supersaturations are eliminated rompletely, hrc~auxrthe temuerature is that of the boiling point of a constant-boilink liquid, arid no mechanical
H\
temperature and ' humidit'! conditions within the r a b net give accurate conipari;;om of the relative effiriencies of corrosion preventive>, platings, varnishes, or other protective coatings. These recjults can be accurately reproducrd. By using cyclopentanr as the heating liquid a teniperature of 120" F. (49' C.) is maintained, which is :I commonly specified condition, By selecting the heating liquid, practically any temperature desired can he maintained exactly. The air ust:d in the cabinet is preheated to the temperature of the rabinet and prehumidified to 10070 before it enters the cabinet chamber, regardless of t h e operating temperature of the cabinet or of outside conditions. .iir and watrr entering the cabinet are automatirally and accurately controlled and are not affected even b y relatively wide variations i n the laboratory or
J
or &harpangles, which trr!tl to cause irregularititla in
air currents within thtb rahinrt. The only variable, that of the effect on boiling point of variations of atmospheric pie.sure, ha5 proved prwtically insignificant.
N
CONSTRUCTION OF CABINET
IN
-
The cabinet, shown ill vertical section in Figure 1 and in horizontal section in Figure 2, consists of a vaportight jacketed cylindric,al container, V , with a conical bottom and top. The chamber is made of KO.22 B. & S. gage Monel metal. The annular spaces and lid, W , are insulated with loosely parked rock wool or glass wool. Fourteen feet (120 cm.) of 0.25-inch (0.6cm.) copper tubing are bent into a spiral and tarked uniformly around the bottom of the cabinet chamber. The top of the spiral tubing is bent abruptly downward in the central part of the cabinet with its orifire at". An overflow tube, 0, iq wed to keep a constant
INDUSTRIAL A N D ENGINEERING CHEMISTRY
396
Vol. 16, No. 6
%
are heated by means of the hot vapors rising from a t sucha rate as to produce slight refluxing in condenser B. ecause of the good insulation around the cabinet the "low heat" of a small hot plate, T, was found t o be sufficient.
H
fine brifiee a t the bottom
~~~~~, ~~., if0 . 'The ~ .~ water on leaving A 1s used ~Y
~~
~~~
V I
~
~
~
hour. The excess air is a h w e d to escape a t the top of tube 1. This air-pressure regulator tube will keep the rate of air flaw into the cabinet constant regardless of relatively wide variations in laboratory or plant air pressures. The metered air and water in tube K descend and enter the bottom of the cabinet a t P, then travel a long upward spiral path through the submerged copper tubing. Durinc this passage the
W ~ervestodistribute the emerging i i r in a uniform Gsnner throurbFisure
HorizontaI
2.
Section of Cabinet
r
ing on the hd to return to the bottom of the cabinet by way of the center. I n this Nay the test panels will not be wet by water A_._. urupa.
water level in the cabinet as SI by the horizontal dqtted li ne wooden support, U,IS used cabinet is mounted in a 01.5(1.25-em.) plywood containe:r. heating flask, Q, is a N i t e r ro bot,tozwd Pyrex flask in &r bath, S. The test panel supports C? nsist, of 4 concentric rings whieh ate divided into quadrants. ICach quadrant is a separate un it and may be removed separately f i1r panel inspections. Each quadrant slides into its own vertical suPPmt sleeve which is fastened to the center bottom of the cabine!t derate. Even with the very low hoiline evrloprntnne (120" P.) the normal 1os4 i s less than f p r n t week!;
~~~~
OPERA1101I ( (
The ealinet is filled with watL. "- "_ " " ~ ,yl.c rxrrcired t o sash tlrc caliinct thoroughly t o remove acid soldering 0uxt.s before ~iiiiir.)W,ir ,water and the walls of the cabinet yl
...YIY
y1
Figure
3
~
lune, 1944
397
A N A L Y T I C A L EDITION
The fundamental difference between the cabinets is most clearly apparent from the two last items in the tabulation. If rust preventive "Ql" in its experimental stage had been evaluated on the basis of the conventional cabinet alone, it would have been discarded as no more effective than the sodium sulfonate or a fatty acid solution, and only half a8 effective as the older rust preventive az". Thus a pr,oduct highly meritorious in its field of application would have 1ieen discarded. :n + ).a - nonohima+. n h n m d _" it. i n he a t ____, 1eS.d -_-_ yIyu " yl.. ~ "~ However, )Lo +art 7 t o 10 times more efficient than the petroleum sulfonate or the fatty acid, and a t least 3 times more efficient than Q2, under conditions resembling ordinary factory storage, where a frequent flow of water over the surfaces of objects stored is out of question. Corrosion resistance is a complex phenomenon, and a corrosion preventive which is better under certain test conditions is inferior under other test conditions, and vice versa. The same is true of actual storage under practical conditions. A corrosion preventive which is most potent in actual use where Continuous condensation of humidity takes place may be far from the optimum for storage in even extreme humidities without actual condbnsatition; and a corrosion preventive which is best for the corrosive but not extremely humid atmosphere in steel plants is not best for storage under tropical conditions which are characterized hy high temperature and humidity but where corrosive fumes are absent. A detailed analysis of the factors entailed would take us beyond the frame of the present subject; suffice i t to my that any accelerated corrosion test must he adaDted t o the natural conditions in view. The cabinet here described will give an accurate reproducible measure of storage resistance under indoor conditions, in the absence of continuous precipitation or of corrosive vapors. The influence of these latter factors should he measured separately where they play a part, and should not be introduced where they do not enter into the applications envisaged.
..-
Figure 4
Test panels were suspended in different parts of this cabinet to check its uniformity; regardless of where they were suspended the corrosion was the same. Furthermore, because of the uniformity of the corrosion effects it was found that 1 X 3 inch test panels could be used with results fully as relishle as those obtainable with the usual 2 X 4 inch test panels. These smaller test panels permit a considerably larger testing capacity. Table I illustrates differences in performance between the present cahinet and a cabinet of the conventional type. Both cabinets were operated a t 120' F. and 100% relative humidity. I n the conventional cabinet moisture frequently condensed on the t,hr panels, nanals. and dripped drinned down from them. IIn n the cahinet here described same. oondensation took d a c e when the cold panels were introduced into the cabinet. TI& initially candensed water remained 8.8 a dew on the panels, since evaporation could not take place at loo'% relative humidity, but there was no continual condensation. nor anv flow of water over the surfaces of the test panels. follows; Test The panels used in the test were prepared as follows; panels of of 1 x X 3 inches of of S.A.E. 1025 sheet steel. B. Q S. ,gage 24 were cleaned by washing in benzene and wiping dry with a clean cloth. Completely Comdetelv new and uniform surfaces were exDosed Danels bv carefully buffing on t~ 010th buffing ~ o s e don the test danels'bv ~
~~~~~~~~
D
~
~~
~
-...
~1
__
I
Table I., Cabinet Performance e pq+d !or Firat 3 nyBs *Ix.&S 10 appear O D Surface Of Iron Paneis
p
Material Tested
Conventional osbinet
New oabioet
10 min. 4 24 24 28
25 mjn.
120 144
236 328
120
No oorroeion in
240
840
Blank Paraffinic mineral oil S.A.E. 40 1 0 4 but 1riciml+ in Stoddard solvent 10% metlyi +ate 1x1 Stoddard solvent 10 wood r(hsm ~nStoddard solvent 20% sodium (petroleum1sulfonate in Stoddard solvent 10% &io aeid,in 40 Viscosity plineral.oil 20% oomrneroial rust preventive Q1 m Stoddard solvent c~mmemialr u t preventive 92 in Stod20$ ard soivent
66 80 65
2624 houm
ACKNOWLEDGMENT
The author takes this opportunity to express his sincere ap preciatiou for the interest and cooperation which Johen Bjorksten, chemical director, Quaker Chemical Products Corvoration. has given in the preparation of this paper.
~~~
troit, Mich. Care waitaken'during chis buffing operation t o remove all sharp burrs a t the edges of the test panels. The test panels were again wiped with a clean cloth to remove any traces of polishing grit and were then used. Care was taken not to touch the prepared test panels with the fingers.
A.S.T.M. Non-Metallic, Constructional 11, ~ 1 3 3 (1942). 5 Ibid., 1020.
Amv-Navv Aeronautical Smoifieation AN-00-S-91 (1938). ..
~~
All determinations were made in triplicate, and the figures shown are the average ,values. The variations of the triplicate determinations were within a range of 10% for the new cabinet and 28% for the Conventional cabinet.
H&R. 0.;U.S. Patent 1,969,606(1934). Jameson, C. W..U.S. Pstent 1,870,512(1932) Naylor, R.B.. U. S. Patent 1,327,838(1920). P r w , H.,and Gregg. J. L..A.S.T.M. Proe.. 41,758 (1941). Tesehner, B.S., U. S. Patent 1,831,980(1931).
