Armor plate makers sometimes use the copper ball or Siemens' water pyrometer because they can place a number of the b.a.l.l.s or weights on the plate in locations where it is difficult to use other pyrometers.

One of these pyrometers is shown in section in Fig. 109.

SIEMENS' WATER PYROMETER.--It consists of a cylindrical copper vessel provided with a handle and containing a second smaller copper vessel with double walls. An air s.p.a.ce _a_ separates the two vessels, and a layer of felt the two walls of the inner one, in order to r.e.t.a.r.d the exchange of temperature with the surroundings. The capacity of the inner vessel is a little more than one pint. A mercury thermometer _b_ is fixed close to the wall of the inner vessel, its lower part being protected by a perforated bra.s.s tube, whilst the upper projects above the vessel and is divided as usual on the stem into degrees, Fahrenheit or Centigrade, as desired. At the side of the thermometer there is a small bra.s.s scale _c_, which slides up and down, and on which the high temperatures are marked in the same degrees as those in which the mercury thermometer is divided; on a level with the zero division of the bra.s.s scale a small pointer is fixed, which traverses the scale of the thermometer.

[Ill.u.s.tration: FIG. 109.--Siemens' copper-ball pyrometer.]

Short cylinders _d_, of either copper, iron or platinum, are supplied with the pyrometer, which are so adjusted that their heat capacity at ordinary temperature is equal to one-fiftieth of that of the copper vessel filled with one pint of water. As, however, the specific heat of metals increases with the temperature, allowance is made on the bra.s.s sliding scales, which are divided according to the metal used for the pyrometer cylinder _d_. It will therefore be understood that a different sliding scale is required for the particular kind of metal of which a cylinder is composed. In order to obtain accurate measurements, each sliding scale must be used only in conjunction with its own thermometer, and in case the latter breaks a new scale must be made and graduated for the new thermometer.

The water pyrometer is used as follows:

Exactly one pint (0.568 liter) of clean water, perfectly distilled or rain water, is poured into the copper vessel, and the pyrometer is left for a few minutes to allow the thermometer to attain the temperature of the water.

The bra.s.s scale _c_ is then set with its pointer opposite the temperature of the water as shown by the thermometer. Meanwhile one of the metal cylinders has been exposed to the high temperature which is to be measured, and after allowing sufficient time for it to acquire that temperature, it is rapidly removed and dropped into the pyrometer vessel without splashing any of the water out.

The temperature of the water will rise until, after a little while, the mercury of the thermometer has become stationary. When this is observed the degrees of the thermometer are read off, as well as those on the bra.s.s scale _c_ opposite the top of the mercury.

The sum of these two values together gives the temperature of the flue, furnace or other heated s.p.a.ce in which the metal cylinder had been placed. With cylinders of copper and iron, temperatures up to 1,800F. (1,000C.) can be measured, but with platinum cylinders the limit is 2,700F. (1,500C.).

For ordinary furnace work either copper or wrought-iron cylinders may be used. Iron cylinders possess a higher melting point and have less tendency to scale than those of copper, but the latter are much less affected by the corrosive action of the furnace gases; platinum is, of course, not subject to any of these disadvantages.

The weight to which the different metal cylinders are adjusted is as follows:

Copper 137.0 grams Wrought-iron 112.0 grams Platinum 402.6 grams

In course of time the cylinders lose weight by scaling; but tables are provided giving multipliers for the diminished weights, by which the reading on the bra.s.s scale should be multiplied.

THE THERMO-COUPLE

With the application of the thermo-couple, the measurement of temperatures, between, say, 700 and 2,500F., was made more simple and precise. The theory of the thermo-couple is simple; it is that if two bars, rods, or wires of different metals are joined together at their ends, when heated so that one junction is hotter than the other, an electromotive force is set up through the metals, which will increase with the increase of the _difference_ of temperature between the two junctions. This electromotive force, or voltage, may be measured, and, from a chart previously prepared, the temperature determined. In most pyrometers, of course, the temperatures are inscribed directly on the voltmeter, but the fact remains that it is the voltage of a small electric current, and not heat, that is actually measured.

There are two common types of thermo-couples, the first making use of common, inexpensive metals, such as iron wire and nichrome wire.

This is the so-called "base metal" couple. The other is composed of expensive metals such as platinum wire, and a wire of an alloy of platinum with 10 per cent of rhodium or iridium. This is called the "rare metal" couple, and because its component metals are less affected by heat, it lasts longer, and varies less than the base metal couple.

The cold junction of a thermo-couple may be connected by means of copper wires to the voltmeter, although in some installations of base metal couples, the wires forming the couple are themselves extended to the voltmeter, making copper connections unnecessary.

From the foregoing, it may be seen that accurately to measure the temperature of the hot end of a thermo-couple, we _must know the temperature of the cold end_, as it is the _difference_ in the temperatures that determines the voltmeter readings. This is absolutely essential for precision, and its importance cannot be over-emphasized.

When pyrometers are used in daily operation, they should be checked or calibrated two or three times a month, or even every week. Where there are many in use, it is good practice to have a master pyrometer of a rare metal couple, which is used only for checking up the others. The master pyrometer, after calibrating against the melting points of various substances, will have a calibration chart which should be used in the checking operation.

It is customary now to send a rare metal couple to the Bureau of Standards at Washington, where it is very carefully calibrated for a nominal charge, and returned with the voltmeter readings of a series of temperatures covering practically the whole range of the couple. This couple is then used only for checking those in daily use.

Pyrometer couples are more or less expensive, and should be cared far when in use. The wires of the couple should be insulated from each other by fireclay leads or tubes, and it is well to encase them in a fireclay, porcelain, or quartz tube to keep out the furnace gases, which in time destroy the hot junction. This tube of fireclay, or porcelain, etc., should be protected against breakage by an iron or nichrome tube, plugged or welded at the hot end. These simple precautions will prolong the life of a couple and maintain its precision longer.

Sometimes erroneous temperatures are recorded because the "cold end" of the couple is too near the furnace and gets hot. This always causes a temperature reading lower than the actual, and should be guarded against. It is well to keep the cold end cool with water, a wet cloth, or by placing it where coal air will circulate around it. Best of all, is to have the cold junction in a box, together with a thermometer, so that its temperature may definitely be known.

If this temperature should rise 20F. on a hot day, a correction of 20F. should be added to the pyrometer reading, and so on. In the most up-to-date installations, this cold junction compensation is taken care of automatically, a fact which indicates its importance.

Optical pyrometers are often used where it is impracticable to use the thermo-couple, either because the temperature is so high that it would destroy the couple, or the heat to be measured is inaccessible to the couple of ordinary length. The temperatures of slag or metal in furnaces or running through tap-holes or troughs are often measured with optical pyrometers.

In one type of optical pyrometer, the observer focuses it on the metal or slag and moves an adjustable dial or gage so as to get an exact comparison between the color of the heat measured with the calor of a lamp or screen in the pyrometer itself. This, of course, requires practice, and judgment, and brings in the personal equation. With care, however, very reliable temperature measurements may be made. The temperatures of rails, as they leave the finishing pa.s.s of a rolling mill, are measured in this way.

Another type of optical pyrometer is focused on the body, the temperature of which is to be measured. The rays converge in the telescope on metal cells, heating them, and thereby generating a small electric current, the voltage of which is read an a calibrated voltmeter similar to that used with the thermo-couple. The best precision is obtained when an optical pyrometer is used each time under similar conditions of light and the same observer.

Where it is impracticable to use either thermo-couples or optical pyrometers, "sentinels" may be used. There are small cones or cylinders made of salts or other substances of known melting points and covering a wide range of temperatures.

If six of these "sentinels," melting respectively at 1,300, 1,350, 1,400, 1,450, 1,500, and 1,550F., were placed in a row in a furnace, together with a piece of steel to be treated, and the whole heated up uniformly, the sentinels would melt one by one and the observer, by watching them through an opening in the furnace, could tell when his furnace is at say 1,500 or between 1,500 and 1,550, and regulate the heat accordingly.

A very accurate type of pyrometer, but one not so commonly used as those previously described, is the resistance pyrometer. In this type, the temperature is determined by measuring the resistance to an electric current of a wire which is at the heat to be measured. This wire is usually of platinum, wound around a quartz tube, the whole being placed in the furnace. When the wire is at the temperature of the furnace, it is connected by wires with a Wheatstone Bridge, a delicate device for measuring electrical resistance, and an electric current is pa.s.sed through the wire. This current is balanced by switching in resistances in the Wheatstone Bridge, until a delicate electrical device shows that no current is flowing. The resistance of the platinum wire at the heat to be measured is thus determined on the "Bridge," and the temperature read off on a calibration chart, which shows the resistance at various temperatures.

These are the common methods used to-day for measuring temperatures, but whatever method is used, the observer should bear in mind that the greatest precision is obtained, and hence the highest efficiency, by keeping the apparatus in good working order, making sure that conditions are the same each time, and calibrating or checking against a standard at regular intervals.

THE PYROMETER AND ITS USE

In the heat treatment of steel, it has become absolutely necessary that a measuring instrument be used which will give the operator an exact reading of heat in furnace. There are a number of instruments and devices manufactured for this purpose but any instrument that will not give a direct reading without any guess work should have no place in the heat-treating department.

A pyrometer installation is very simple and any of the leading makers will furnish diagrams for the correct wiring and give detailed information as to the proper care of, and how best to use their particular instrument. There are certain general principles, however, that must be observed by the operators and it cannot be too strongly impressed upon them that the human factor involved is always the deciding factor in the heat treatment of steel.

A pyrometer is merely an aid in the performance of doing good work, and when carefully observed will help in giving a uniformity of product and act as a check on careless operators. The operator must bear in mind that although the reading on the pyrometer scale gives a measure of the temperature where the junction of the two metals is located, it will not give the temperature at the center of work in the furnace, unless by previous tests, the heat for penetrating a certain bulk of material has been decided on, and the time necessary for such penetration is known.

Each a.n.a.lysis of plain carbon or alloy steel is a problem in itself.

Its critical temperatures will be located at slightly different heats than for a steel which has a different proportion of alloying elements. Furthermore, it takes time for metal to acquire the heat of the furnace. Even the outer surface lags behind the temperature of the furnace somewhat, and the center of the piece of steel lags still further. It is apparent, therefore, that temperature, although important, does not tell the whole story in heat treatment. _Time_ is also a factor.

Time at temperature is also of great importance because it takes time, after the temperature has been reached, for the various internal changes to take place. Hence the necessity for "soaking," when annealing or normalizing. Therefore, a clock is as necessary to the proper pyrometer equipment as the pyrometer itself.

For the purpose of general work where a wide range of steels or a variable treatment is called for, it becomes necessary to have the pyrometer calibrated constantly, and when no master instrument is kept for this purpose the following method can be used to give the desired results:

CALIBRATION OF PYROMETER WITH COMMON SALT

An easy and convenient method for standardization and one which does not necessitate the use of an expensive laboratory equipment is that based upon determining the melting point of common table salt (sodium chloride). While theoretically salt that is chemically pure should be used (and this is neither expensive nor difficult to procure), commercial accuracy may be obtained by using common table salt such as is sold by every grocer. The salt is melted in a clean crucible of fireclay, iron or nickel, either in a furnace or over a forge-fire, and then further heated until a temperature of about 1,600 to 1,650F. is attained. It is essential that this crucible be clean because a slight admixture of a foreign substance might noticeably change the melting point.

The thermo-couple to be calibrated is then removed from its protecting tube and its hot end is immersed in the salt bath. When this end has reached the temperature of the bath, the crucible is removed from the source of heat and allowed to cool, and cooling readings are then taken every 10 sec. on the milli-voltmeter or pyrometer. A curve is then plotted by using time and temperature as coordinates, and the temperature of the freezing point of salt, as indicated by this particular thermocouple, is noted, _i.e._, at the point where the temperature of the bath remains temporarily constant while the salt is freezing. The length of time during which the temperature is stationary depends on the size of the bath and the rate of cooling, and is not a factor in the calibration. The melting point of salt is 1,472F., and the needed correction for the instrument under observation can be readily applied.

It should not be understood from the above, however, that the salt-bath calibration cannot be made without plotting a curve; in actual practice at least a hundred tests are made without plotting any curve to one in which it is done. The observer, if awake, may reasonably be expected to have sufficient appreciation of the lapse of time definitely to observe the temperature at which the falling pointer of the instrument halts. The gradual dropping of the pointer before freezing, unless there is a large ma.s.s of salt, takes place rapidly enough for one to be sure that the temperature is constantly falling, and the long period of rest during freezing is quite definite.

The procedure of detecting the solidification point of the salt by the hesitation of the pointer without plotting any curve is suggested because of its simplicity.

COMPLETE CALIBRATION OF PYROMETERS.--For the complete calibration of a thermo-couple of unknown electromotive force, the new couple may be checked against a standard instrument, placing the two bare couples side by side in a suitable tube and taking frequent readings over the range of temperatures desired.

If only one instrument, such as a millivoltmeter, is available, and there is no standard couple at hand, the new couple may be calibrated over a wide range of temperatures by the use of the following standards:

Water, boiling point 212F.

Tin, under charcoal, freezing point 450F.

Lead, under charcoal, freezing point 621F.

Zinc, under charcoal, freezing point 786F.

Sulphur, boiling point 832F.

Aluminum, under charcoal, freezing point 1,216F.

Sodium chloride (salt), freezing point 1,474F.

Pota.s.sium sulphate, freezing point 1,958F.

A good practice is to make one pyrometer a standard; calibrate it frequently by the melting-point-of-salt method, and each morning check up every pyrometer in the works with the standard, making the necessary corrections to be used for the day's work. By pursuing this course systematically, the improved quality of the product will much more than compensate for the extra work.

The purity of the substance affects its freezing or melting point.

The melting point of common salt is given in one widely used handbook at 1,421F., although chemically pure sodium chloride melts at 1,474F. as shown above. A sufficient quant.i.ty for an extended period should be secured. Test the melting point with a pyrometer of known accuracy. Knowing this temperature it will be easy to calibrate other pyrometers.