DENISON, Circuit Judge.
The court below thought that the patent sued upon was invalid for anticipation, or for lack of invention, or both, and dismissed the infringement bill brought by the Northern Company against the McDonough Company, based upon patent No. 1,148,483, issued July 27, 1915, to the plaintiff, upon the application of Andrews, for a “method of feeding water to boilers,’’ filed in October, 1913. As the water in a steam boiler is converted into steam and passes over to the engine, its place must be supplied by other water to be fed in, hence called “feed water.” The level of water within the boiler was commonly indicated by a vertical water tube gauge upon the outside, connected at top and bottom to the interior, and thus maintaining in the visible tube the same water level. The desirable interior water level having been determined upon, _ the fireman, informed by the gauges, maintained the level by hand manipulation of the inlet valve. The obvious defects of this method were early met by devices for automatic regulation. The simplest of these was the float system. A float was carried on the surface of the boiler water, and, as it fell, it would operate mechanism which opened the inlet valve; through this greater opening the water would rush in and at once restore the water level; and the converse operation was. similar. The whole theory of operation of this type of automatic regulator was to maintain substantially a constant level, and a very slight variation in either direction brought immediate restoration, though the phrase, “a constant level” is not one of precision. There was necessarily a constant commotion in the water, making an irregular andi unsteady surface.
Another type of automatic regulator, known before this invention;, was the thermostatic. In its simplest form this consisted of a tube placed horizontally, outside the boiler, with one end connected to the low or water part of the interior, and the other to the upper or steam part. The tube was made of some substance changing excessively by a change of temperature. As the water level fell, the water-filled tube would be emptied and in turn be filled with steam. This, being relatively hotter than the water, would cause the tube to expand. The expansion, through compounding leverage, would open the inlet valve, the water level would rise again, the steam in the tube be displaced by water, and this intermittent operation would continue. Such thermostatic-regulators also were essentially of the constant level type. A slight change in the water level brought the compensatory action. It is now to be seen that such a tube might perhaps have been so connected with the interior, by separating vertically the points of connection, that the-action upon the inlet valve would have been greatly retarded; but, even if this could have been done successfully, it was not. Generally [490]*490speaking, this was the state of the art when Andrews made his invention ; but there was another phenomenon which had then been observed, even if not so well understood as it later was, in addition to that mere agitation of the water which made an unsteady surface.
The first happening was what the experts now call the “surge.” To use more concrete terms for illustration, we may suppose that the engine in its normal action is demanding and receiving 10 units of steam per second. If the demand is suddenly increased to 15 units, and this amount rushes out of the boiler, other water will turn into .steam rapidly, and the water level, if the inlet valve is untouched, will •fall. This is the normal and inevitable result, but it is not immediate. The instant effect is that the water rises somewhat, or “surges” up. This is because the steam pressure on top of the water is relieved, •and the contained steam or air bubbles in the water are permitted to .•expand, thus raising the water level, though the weight of the water is not increased, and there would be little response in the outside tube to such a change of level. However, this surging effect is only for the instant, the steam pressure is quickly restored, or nearly so, and the water proceeds to fall as it theoretically should. It is not thought that this fleeting surge has much, if any, substantial effect as to the automatic regulation, though it makes for an initial delay in the desired valve effect; but it accounts for otherwise unexplainable abnormalities in the water level, and introduces an element of inaccuracy into the reading of the water levels, which makes them unsafe guides, within close limits.
The other happening is that, when the steam demand is increased, the surge has subsided, and the water level starts to fall, there will be an interval of time before the effect of the fall will be registered in the opening valve and the fall be checked. The delay caused, by the surge, the expansion of the tube, and the transmission of the closing impulse was called the “lag.” The existence of this lag was well understood, but it was regarded as an evil, the effect of which was to be overcome by such a quick and wide opening of the valve as would prevent much fall of the water and would promptly restore the desired constant level.
Andrews, by the invention of the patent in suit, discarded the entire •constant level theory. He concluded that the level ought not to be constant, but that, within the limits of safety, it ought to be varied. He observed that, when the demand for steam increased, it could best be supplied, having due regard for the fuel problem, by lessening the volume of water which was being converted, and establishing a lower level to be kept constant while this higher demand continued. Further, to apply the former illustration, he observed that when the normal water level was adjusted to meet the normal demand for steam, and then the demand for steam increased from 10 to 15 units, it was better that the water level should fall, say 5 inches, and be maintained at the lower level, as long as the 15-unit output was needed, and that the converse was true when the steam demand fell below normal and the water level was raised. The benefits of this so-called variable constant theory of water level involve technical considerations not necessary to con[491]*491sider. Its advantages are now generally conceded, although the constant level theory also has relative advantages and has its advocates. _
_ If any one before Andrews both observed and intelligently appreciated the factors involved in the variable constant theory and worked out his ideas into concrete form, it is not disclosed by this record, as we understand and interpret the testimony. Such prior appreciation of the theory as there was, if any, was vague and abstract. He next observed that his desired result — the variable constant — could be obtained by exaggerating and prolonging’the lag, by loading an artificial lag upon the natural one. It is plain enough, after it is once conceived, that if the effect of the fall of the level upon the inlet valve is delayed until the water is fallen, say 5 inches, and then the valve is opened just far enough to supply the water needed to meet the 15-unit steam demand, its lower level will be maintained while that demand condition continues, ánd so all the way up and down on both sides of the arbitrary normal.
Having this conception of the new method of operation, it was not difficult to devise means for applying it. Andrews in his patent claims the method very broadly, and does not in these broadest claims restrict himself to any special means. The particular plan he adopted was to incline his thermostatic tube as far away from the horizontal and toward the vertical as was practicable.
Free access — add to your briefcase to read the full text and ask questions with AI
DENISON, Circuit Judge.
The court below thought that the patent sued upon was invalid for anticipation, or for lack of invention, or both, and dismissed the infringement bill brought by the Northern Company against the McDonough Company, based upon patent No. 1,148,483, issued July 27, 1915, to the plaintiff, upon the application of Andrews, for a “method of feeding water to boilers,’’ filed in October, 1913. As the water in a steam boiler is converted into steam and passes over to the engine, its place must be supplied by other water to be fed in, hence called “feed water.” The level of water within the boiler was commonly indicated by a vertical water tube gauge upon the outside, connected at top and bottom to the interior, and thus maintaining in the visible tube the same water level. The desirable interior water level having been determined upon, _ the fireman, informed by the gauges, maintained the level by hand manipulation of the inlet valve. The obvious defects of this method were early met by devices for automatic regulation. The simplest of these was the float system. A float was carried on the surface of the boiler water, and, as it fell, it would operate mechanism which opened the inlet valve; through this greater opening the water would rush in and at once restore the water level; and the converse operation was. similar. The whole theory of operation of this type of automatic regulator was to maintain substantially a constant level, and a very slight variation in either direction brought immediate restoration, though the phrase, “a constant level” is not one of precision. There was necessarily a constant commotion in the water, making an irregular andi unsteady surface.
Another type of automatic regulator, known before this invention;, was the thermostatic. In its simplest form this consisted of a tube placed horizontally, outside the boiler, with one end connected to the low or water part of the interior, and the other to the upper or steam part. The tube was made of some substance changing excessively by a change of temperature. As the water level fell, the water-filled tube would be emptied and in turn be filled with steam. This, being relatively hotter than the water, would cause the tube to expand. The expansion, through compounding leverage, would open the inlet valve, the water level would rise again, the steam in the tube be displaced by water, and this intermittent operation would continue. Such thermostatic-regulators also were essentially of the constant level type. A slight change in the water level brought the compensatory action. It is now to be seen that such a tube might perhaps have been so connected with the interior, by separating vertically the points of connection, that the-action upon the inlet valve would have been greatly retarded; but, even if this could have been done successfully, it was not. Generally [490]*490speaking, this was the state of the art when Andrews made his invention ; but there was another phenomenon which had then been observed, even if not so well understood as it later was, in addition to that mere agitation of the water which made an unsteady surface.
The first happening was what the experts now call the “surge.” To use more concrete terms for illustration, we may suppose that the engine in its normal action is demanding and receiving 10 units of steam per second. If the demand is suddenly increased to 15 units, and this amount rushes out of the boiler, other water will turn into .steam rapidly, and the water level, if the inlet valve is untouched, will •fall. This is the normal and inevitable result, but it is not immediate. The instant effect is that the water rises somewhat, or “surges” up. This is because the steam pressure on top of the water is relieved, •and the contained steam or air bubbles in the water are permitted to .•expand, thus raising the water level, though the weight of the water is not increased, and there would be little response in the outside tube to such a change of level. However, this surging effect is only for the instant, the steam pressure is quickly restored, or nearly so, and the water proceeds to fall as it theoretically should. It is not thought that this fleeting surge has much, if any, substantial effect as to the automatic regulation, though it makes for an initial delay in the desired valve effect; but it accounts for otherwise unexplainable abnormalities in the water level, and introduces an element of inaccuracy into the reading of the water levels, which makes them unsafe guides, within close limits.
The other happening is that, when the steam demand is increased, the surge has subsided, and the water level starts to fall, there will be an interval of time before the effect of the fall will be registered in the opening valve and the fall be checked. The delay caused, by the surge, the expansion of the tube, and the transmission of the closing impulse was called the “lag.” The existence of this lag was well understood, but it was regarded as an evil, the effect of which was to be overcome by such a quick and wide opening of the valve as would prevent much fall of the water and would promptly restore the desired constant level.
Andrews, by the invention of the patent in suit, discarded the entire •constant level theory. He concluded that the level ought not to be constant, but that, within the limits of safety, it ought to be varied. He observed that, when the demand for steam increased, it could best be supplied, having due regard for the fuel problem, by lessening the volume of water which was being converted, and establishing a lower level to be kept constant while this higher demand continued. Further, to apply the former illustration, he observed that when the normal water level was adjusted to meet the normal demand for steam, and then the demand for steam increased from 10 to 15 units, it was better that the water level should fall, say 5 inches, and be maintained at the lower level, as long as the 15-unit output was needed, and that the converse was true when the steam demand fell below normal and the water level was raised. The benefits of this so-called variable constant theory of water level involve technical considerations not necessary to con[491]*491sider. Its advantages are now generally conceded, although the constant level theory also has relative advantages and has its advocates. _
_ If any one before Andrews both observed and intelligently appreciated the factors involved in the variable constant theory and worked out his ideas into concrete form, it is not disclosed by this record, as we understand and interpret the testimony. Such prior appreciation of the theory as there was, if any, was vague and abstract. He next observed that his desired result — the variable constant — could be obtained by exaggerating and prolonging’the lag, by loading an artificial lag upon the natural one. It is plain enough, after it is once conceived, that if the effect of the fall of the level upon the inlet valve is delayed until the water is fallen, say 5 inches, and then the valve is opened just far enough to supply the water needed to meet the 15-unit steam demand, its lower level will be maintained while that demand condition continues, ánd so all the way up and down on both sides of the arbitrary normal.
Having this conception of the new method of operation, it was not difficult to devise means for applying it. Andrews in his patent claims the method very broadly, and does not in these broadest claims restrict himself to any special means. The particular plan he adopted was to incline his thermostatic tube as far away from the horizontal and toward the vertical as was practicable. Obviously the vertical tube was best adapted to span wide limits of water level, and the horizontal best fitted to give effective length for maximum expansion. For example, if a fall of 5 inches from the arbitrary normal, and the corresponding rise, were as much as could be permitted within safety limits, the tube could be lengthened to nearly the diameter of the boiler and inclined at 15 degrees from horizontal, and the length of the tube thus subjected to temperature changes would be greatly increased, while at the same time neither the maximum nor minimum water level would go beyond the ends of the tube and thus cease to affect it." By the lengthening of the tube an increased thrust or pull at the free end! would be given, and hence a more positive control of the valve. By using properly computed leverage and balanced valves, perhaps also graduated, he secures very effective manipulation of the inlet, so as to-practice his theory.
With a single exception to be mentioned, the record does not show that the horizontal tube, intended for the constant level method, had been inclined away from the horizontal, excepting in that slight degree that would facilitate drainage out of the tube as the water fell. This did not even approximate the inclination adopted by Andrews, and was not for the same purpose, nor would it give the same result. The commercial success of this patented method has been very great. The plaintiff company and its licensees have very largely occupied the market, and the device has, in a considerable degree, displaced others.
There is one instance of public use in Cleveland, antedating Andrews, which requires attention. _ A thermostatic control.tube of this general type was installed, having a total inclination of substantially 3 inches. Remembering that the rise or fall should not go quite to the ends of the tube, to preserve a margin of safety, and allowing one-half inch at each end, there remained 2 inches of effective inclination, thus providing for a fall of 1 inch below the the normal. We think this should be [492]*492treated as of the constant level type. The variation is not much more than would be desirable to cover ordinary agitation in the water, including the surge. There is no evidence that any one using this device understood the Andrews theory of retarding the lag and maintaining a variable constant, or that the inclination was for any purpose, except to establish and facilitate drainage. Even if the inclination was for the purpose of obtaining greater operative length of the tube, and therefore more motion toward the valve, still that would not involve the Andrews invention, unless it had been accomplished with suitable leverage and valve construction, first, to retard the valve action, and, second, to give just sufficient valve action to meet the new level of steam demand; and there is no satisfactory proof that either of these things was accomplished, even accidentally.
It is true that the physical change made by Andrews, as compared with the Cleveland device, is only an increase in the degree of inclination accompanied by appropriate mechanical adjustment to get the desired result; but nevertheless the essential theory of operation was new. Changes in degree, seemingly as great as this, have been thought unpatentable, and there are decisions of this court looking in that direction; but no such strict rule is now approved by the Supreme Court.
In the Minerals Separation Case, 242 U. S. 261, 37 Sup. Ct. 82, 61 L. Ed. 286, it appeared that it had long been known that oil was useful as a separating agent, and that others had used oil for this purpose, and, moved by considerations of economy, had used the least amount of oil which they thought would be effective. The patentee discovered that there was functional merit as well as economic saving in cutting down the amount of oil, and he cut it still further, explaining his theory of operation, and specifying in his claim that he used as the desired quantity “a fraction of 1 per cent.” This patent was sustained, although the patentee’s step- in advance was wholly a change in degree.
A very close parallel to the present case is presented by the Eibel Case, 261 U. S. 45, 43 Sup. Ct. 322, 67 L. Ed. 523. There the patent differed in no respect from the old art, excepting that the inclination of the table over which the water and pulp mixture flowed had been increased. In the old art one side had been elevated 3 inches, in order merely to produce a flow down the table. The patentee discovered that a more rapid flow had additional functional merit, and he increased by 9 inches the table inclination, thereby inducing a more rapid flow and a resulting better treatment of material. This patent was fully sustained, and claims which were in general language broad enough to cover instances found in the old art were restricted by construction, so as to give protection to the actual invention. This treatment, of course, depends upon finding that there is in the language of the claims, a vagueness or ambiguity which permits some latitude of construction.
We think Andrews’ claims in this patent are entitled to this same treatment, when studied with reference to the earlier Cleveland practice. For example, his claim 2, which is given in the margin,1 might [493]*493be read so as to cover and include the Cleveland device. Within the small limits there permitted, the rate of change of the inflow necessarily did lag behind the rate of change of the outflow while the outflow varied, but this was only to the extent that such lag was inevitable. The characteristic artificial lag of the patent, piled upon the natural lag, was neither known and understood to be present, nor was it actually present in any such substantial degree as necessarily to involve and anticipate Andrews’ operative theory.
We conclude, therefore, that the claims in suit are valid; but the proof does not satisfy us that they are infringed. The defendant’s device is, so far as concerns the thermostatic tube, almost identical with the Cleveland device. The total inclination which the defendant contemplates for, or employs in, the installation of the device it makes, is 3% inches. It may be that, ‘even with this difference, there might be such a nice adjustment of leverage and of valve that a level of say one inch lower than the fixed normal would be maintained by a particular increased demand, and that this would be sufficient to be a practicing of Andrews’ invention. That we do not decide, because it would involve a somewhat closer analysis and comparison of the Cleveland device; but the proof is not sufficient to show that condition. Defendant advertises its device as operating on the constant level theory. This well may be its actual method of operation; if it employs the variable constant, it does so in an uncertain way, merely touching the edge of plaintiff’s field. The trespass is not clear.
Plaintiff’s own expert witness says he thinks the Andrews characteristic method is not used unless there is an inclination of at least 4 inches. That, also, we need not decide. It is sufficient to say that the same \iew of the Cleveland device which convinces us that it is not an anticipation of Andrews’ real invention also indicates to us that this defendant does not employ that invention.
For this reason, the decree of the court below is affirmed.