Southwire Co. v. Beloit Eastern Corp.

OPINION EDWARD R. BECKER, District Judge. I. Preliminary Statement This suit was brought to recover damages occasioned by the malfunction of a large machine known as a tubular strand- *844 er, which twists together seven or eight wires of aluminum or steel into wire rope. 1 The malfunction occurred on January 23, 1964 and was caused by the failure of a cradle casting which supports the large bobbins of wire that feed the machine. Plaintiff Southwire Company (Southwire), a firm engaged in making electrical wire and cable for distribution in the South, is the owner of the malfunctioning strander which was located at its wire mill in Carrollton, Ga. 2 Defendant Beloit Eastern Co. (Be-loit) manufactured and cast the cradles at its Downingtown, Pa. foundry. Be-loit supplied them to third-party defendant Syncro Machine Co. (Syncro) which manufactured and assembled the strand-ers at its Perth Amboy, N. J. plant. Southwire is the plaintiff in C.A. 42129. The Travelers Indemnity Co. (Travelers) was, at times relevant here, the products liability insurer of Syncro. By reason of damage claims made by Southwire against Syncro, Travelers entered into an agreement on June 15, 1966. Under this agreement, Travelers advanced $95,000 to Southwire in return for Southwire’s covenant not to sue Syn-cro; Southwire also assigned to Travelers the right to sue any person it considered to be liable. Travelers is the plaintiff in C.A. 42130, which it brought in its own name. 3 Beloit is the defendant in both actions and it has joined Syncro as third party defendant in both. The central question in the case is the cause of the failure of the casting. When the casting fractured, the reel and cradle in the No. 6 bay of the strander were thrown through the window of the tube in which they were installed, eaus-ing the damage complained of. Plaintiffs contend that the casting failed because it was defective, porous, weak, and improperly cast. Beloit contends that the failure was caused by: (1) the weakening of the casting which occurred when Syncro welded a counterweight to the casting during the course of assembling the machine; and (2) the inordinate and unanticipated stresses to which the casting was subjected because of improper design and maintenance of the machine. In terms of legal theory, the plaintiffs seek recovery under § 402A of the Restatement of Torts, 4 and alternatively, on the theory of negligence. 5 Be-loit defends on the grounds that: (1) the casting which it furnished was in accordance with specifications and was not defective; (2) there was a substantial change in the casting between the time it was sold and the time of the accident, because of the effect on the casting of the counterweight weld; (3) there was no negligence in the manufacture of the casting; and (4) plaintiffs have failed to prove proximate cause. In particular, Beloit contends that the causes of the accident were: (1) the welding which materially weakened the casting at its most critical area; (2) the stresses and vibrations to which the casting was subjected by (a) the misalignment and excessive vibration of the machine and (b) frequent tangles in the wire. As might be expected in a case of this sort, there developed a battle between experts in the field of metallurgical science. Southwire’s expert was Dr. Thomas F. Talbot, a Professor of Engineering at the University of Alabama *845 and a consulting engineer and metallurgist. Dr. Talbot visited the Southwire plant on the day of the accident, and subsequently examined and tested metal from the fractured cradle. Dr. Talbot expressed the opinion that the casting was not sound and that the counterweight welding did not substantially affect it or cause the accident. Beloit’s expert was Dr. A. W. Grosvenor, a Professor of Engineering at Drexel University and likewise a consulting mechanical engineer and metallurgist. Dr. Grosvenor expressed the view that the casting failed because it was substantially weakened and embrittled from the counterweight welding and because of the high stresses to which the casting was subjected. He expressed his expert opinion that the casting did not fail due to any defect in the cradle as originally cast. Each of the expert witnesses brought a superlative background and imposing qualifications to the witness stand. Each demonstrated encyclopedic knowledge of his field, together with a sharp acumen. We found the testimony of each impressive and they rendered the question before us close indeed. And while we found Dr. Talbot’s testimony to be earnest and well-reasoned, it did not overcome the equally earnest and more convincing testimony given by Dr. Gros-venor. Our detailed findings of fact and conclusions of law now follow. We note that it is the factual and not the legal issues which are crucial in this case. We have made extensive findings of fact in the wake of which the legal issues become relatively simple and straightforward. Suffice it to say at this juncture that, after careful consideration of all the evidence, we have concluded that plaintiffs must be denied relief since they failed to carry their burden of proving their claim by a fair preponderance of the evidence. II. Findings of Fact A. The Machine The strander in question was sold to Southwire by Synero pursuant to a purchase order dated February 2, 1962, and Synero’s acceptance dated February 5, 1962. This was one of two wire strand-ers purchased by Southwire in order to take care of its top product range, the production of electrical wire and cable. Southwire and Synero jointly designed the strander; however Southwire’s part in the design consisted merely in prescribing the sizes of the parts in order to be sure the strander would fit in the space that was available. The strander had the capacity to make seven or eight wire strands of either aluminum or steel and was composed of seven bays, each of which received a bobbin of wire through an opening. The bobbins for the strander in question were 31 inches in diameter. The bobbins were carried by shafts which rested on the cradles. The shafts had bobbin locks which prevented the bobbin from coming off in the event it rotated or turned over. The locks or bars were secured through blocks which were bolted to the side of the cradle. The bearings provided a means of supporting the cradle without a friction grip between the cradle and the spiders which were connected to sections between the tubes or bays and constituted the main support for the tube. The bobbins were held in the same plane by the cradles while the tube rotated at approximately 500-550 r.p.m.’s The rotation of the tube gave a pre-twist to the wire as it came off the machine onto the capstans which were two grooved wheels or drums. The tube and capstan were geared together so that as the tube revolved the capstan moved a pre-determined distance. The individual strands of wire from each bobbin traveled through guide rollers and wire guides on the individual strands of wire in a uniform configuration so that they would join together in a uniform and controlled manner. The individual wire strands were twisted together as they passed through a closing die. From the closing die, the stranded wire or cable was pulled onto the capstans. From that point, the wire was taken up on a ship-out package or take- *846 up reel. The wire produced as an end product by the strander was used as electrical conductor, messenger wire, and steel supporting cable. The strander was operated on a 24-hour, 7-day a week basis, except for down-time to change spools and except for one 8-hour shift per week during which the strander was serviced and maintained. It produced stranded wire at a rate of 130 to 500 feet per minute, depending on the product and size of cable being produced. For the first few months of its operation, the strander was realigned several times to compensate for vibration. As soon as vibration was detected, the machine was shut down for realignment and service. Although the gross misalignment problem was thus corrected, the problem of misalignment remained to some degree — at least %dths of an inch. Dr. Talbot so testified, although he concluded that misalignment did not cause the accident in question. Each cradle was keel-weighted on the bottom so that it would stay in the same plane with a bearing at each end of the cradle and not rotate with the rotation of the barrel and spider. Occasionally a wire would cross over so that it would be pulled off center and tend to either swing the cradle or turn it around. The bobbin shaft had a rod (called a bobbin lock) that prevented the bobbin from coming out in the event it did rotate or turn over. The strander frequently had tangles which could pull the bobbin completely over and rotate the cradle and bobbin; tangles could be and often were caused by improper wrapping of the wire on the spool or reel. 6 B. The Casting and The Counterweights The cradle casting in question was manufactured by Beloit. Its dimensions were 51%" in overall length with an overall depth of approximately 10 inches. It can best be described as a flat slab of cast iron, tapered and curved at each end, with a semi-circular depression referred to as a “saddle” in the top side of the middle of the casting. Two mating castings were used in each tube or bay, bolted together at either end and forming a hollow, roughly oval, bottomless cradle which could be likened to a bottomless hammock. Between the sides of the cradle sat the reel or bobbin which contained the wire rope. This bobbin turned on an axle, the ends of which were seated in the saddles. Each end of the axle was held in place by a locking bar. To avoid the lateral rotation of the bobbin and cradle, which would tend to snag the wire as it was pulled off of the spool, the ends of the cradle fitted over bearings installed in the revolving main shaft which turned the spider and tube. The cradles were cast by Beloit pursuant to the specifications and a blue print (# SB-3672) of Syncro. 7 Syncro specified that the cradles be made of cast ductile iron, a material developed by In *847 ternational Nickel and which simulated cast steel. The cradles were to have characteristics known as “60-45-15,” which means: 60,000 p.s.i. minimum ultimate strength (i. e., the point at which the casting would break); 45,000 p.s.i, yield strength (i. e., the point at which the easting would be changed permanently and not return to its, original shape); and 15 per cent ductility or elongation (i. e., the length the material might elongate as it is being stressed). The cradle sections were to be symmetrical and “free of blowholes.” They were also to be annealed, so that the structure would be relieved of any stress which may be created by rapid cooling and so that it ultimately would be more uniform throughout its cross section. 8 The casting was poured by Beloit at its foundry in Downingtown, Pennsylvania. The casting was made in accordance with Beloit’s usual procedures. Each pour was tested so as to avoid the dross which collects at the surface in the early part of the pour and to avoid the sludge and sediment which is found toward the end. Certification of the test results were sent to Syncro from each ladle from which the castings were poured; if any report indicated deficiencies, the castings involved were to be sent back to be remelted. As it was manufacturing the strander, Syncro welded metal balance weights to the side of the casting which tended to rise due to the rapid turning of the main shaft of the strander (500-550 rpms.). A large rectangular balance weight was welded directly under the saddle, and a somewhat smaller square balance weight was welded to the left. The counterweights were not of a uniform configuration; instead they were individual to the cradle they were intended to balance. The procedure was done by means of arc welding. In the are welding process, a piece of metal made of iron and nickel (known as a bead) is melted along with the adjacent surfaces of the casting and the balance weight so that all three flow together and harden. The molten metal is then cooled by conduction of the heat into the metal of the casting. The heat transfer is primarily into the metal; a small part of the heat, perhaps 10%, radiates into the atmosphere. Depending on the cooling rate, either a ductile or an extremely brittle zone results in the casting; the more rapid the cooling, the more brittle the zone. This welding procedure had been utilized by Syncro for many years without any known adverse effects. The welding was done by Syncro without the knowledge of Beloit. Syncro was aware of the fact that the welding of the counterweight to the cradle could weaken the casting. However, it was expected that, by virtue of the stability of the strand-er, the casting would hold the weight of the spool in an almost static condition so that the deep casting section would be subject to only limited stress. The safety factor which Syncro took into consideration in specifying the cradles was “five,” which meant that the cradle could accept a stress five times the maximum stress expected to be applied. In view of these facts, Syncro did not deem it necessary to order that the easting be tested by x-ray or magnaglow, 9 or by the “destruct” type method 10 to determine any defects in the casting including subsurface porosity. It is clear that the specifications required that the castings be free of blowholes. However, the industry custom was that such a specification meant that the degree of porosity shall be low enough that there will be no sinks or drop in the surface of the casting. Since the surface of the casting in question appeared acceptable for machining, *848 Synero was satisfied that the blowhole specification was met. Moreover, the test procedures referred to above are extremely expensive, and it is the custom in the industry not to perform such tests unless specified by the customer. The customer generally does not so specify unless the metal is expected to be subjected to critical stress, such as aircraft parts. In any event, no tests were ordered by Synero or Southwire; all concerned seem satisfied that they were unnecessary. After the accident, the casting in question and all other castings in the strander were x-rayed and magnaglowed so that the interior of the castings could be susceptible to visual inspection. In addition, the fractured casting was subjected to additional tests by Dr. Talbot. The visual inspection and the other tests revealed numerous defects therein. While there was substantial dispute between the experts as to whether the tests were properly conducted, and as to the type of fracture involved, the most significant testimonial dispute between Drs. Talbot and Grosvenor related to the cause of the accident and the role played by Syncro’s welding of the counterweights to the casting. We shall discuss and make findings on these critical aspects of the matter after making our finding on the happening of the accident. C. The Accident On January 23, 1964, at approximately 4:00 p. m., the strander malfunctioned when the bobbin, the bobbin mechanism, a fractured cradle half and the bolts that secured it flew out of the window of the No. 6 bay of the strander. Remains of the bolts which had held the casting within the locking mechanism were still in the casting, but the complete locking bar and mechanism were missing therefrom. The cradle half which remained in the sixth bay was still secured by the locking bar and mechanism. Immediately prior to the accident, the strander was observed to be operating in normal fashion, and its operator was in the process of replenishing the bobbins. The bobbins in bays 1 and 2 had been replaced, but the bobbins in the remaining bays had not quite run out. Subsequent to the accident, an inspection of the strander revealed that the window of bay No. 6 was bent considerably along the edge. There was also an abrasion inside the barrel of bay No. 6, which indicated that the barrel made one or two revolutions before the objects were thrown clear. The bobbin which was ejected from bay 6 was found some distance from the machine. This bobbin was almost empty and had approximately an inch and one-half of aluminum wire in it. The bearings which secured the cradle were still free and intact. The foregoing factors indicate that the accident was not caused by malfunction of the locking bar and mechanism or by seizure of the bearings. We find (and there is no real dispute about it) that the casting was the first part to break, thus precipitating the strander’s malfunction. 11 We thus turn to the cen *849 tral issue in the ease, i. e., the cause of the casting failure. D. The Cause of-the Casting Failure The testimony of Drs. Talbot and Grosvenor as to the cause of the casting failure was highly technical. Our findings will be understood better and the issues will be highlighted if we first summarize the testimony of the two experts. Then we will make the crucial findings, which are limited to matters in dispute. 1. The Testimony of Dr. Talbot Dr. Talbot visited the Southwire plant on the day of the accident, observed parts of the strander lying around on the floor and had a number of photographs taken. He observed that the half-cradle which had fractured had come from the No. 6 bay. He cut two pieces of the casting from the fractured cradle. He found these pieces to be nodular; i. e., they had low ductility and were basically brittle. Dr. Talbot found porosity and a very poor surface — voids where metal is not hot enough to have complete melding. Indeed, he stated that some 50 to 75% of the castings were non-uniform and contained some porosity. He found center line shrinkage, and said that the casting did not appear to be sound in the center. He also found dross 12 or slag and said that there was “an oxide or something that shouldn’t be there.” He stated his opinion that the whole half of the cradle casting was the first part to break in the sequence of events that occurred; he concluded that it had failed and allowed the bobbin to come out of its position. Dr. Talbot testified that he had all of the twenty-eight eastings of the two stranders (two castings per bay, seven bays per strander) magnaglowed. (See n. 9, supra.) The magnaglow process showed many small surface cracks and one large crack in the vicinity of the counterweight weld. Several castings displayed poor surfaces; in fact, every easting had defects. All the castings were x-rayed to determine whether they should be discarded or could continue to be used safely. Specimens were cut from the fractured casting and tensile tested at Cormet Metallurgical Labs. Dr. Talbot admitted that these tests were not performed in the manner approved by ASTM, in that the surfaces of the test bars were not machined to remove the surface dross which inevitably forms on the surface of a pour of molten metal. Moreover, the tests were made after he had attempted to simulate the conditions which existed at the site of the weld by performing electric arc welding on the areas from which the test bars were taken; but we find that he could not duplicate the conditions. We also note in this respect Dr. Talbot’s testimony that welding altered the composition of the metal with beneficial and detrimental stresses. In sum, we find that his results are not the same as the results that would have been obtained in an original testing of the bars by ASTM approved methods. 13 According to Dr. Talbot’s testimony, the results of the Cormet tests showed that: (1) one sample failed or ruptured at a very low point; (2) not one of the samples tested reached the specified 60,000 p.s.i. ultimate tensile strength; (3) the sections which were welded by Dr. Talbot for purposes of simulating Syncro’s weld were more ductile than the non-welded sections; (4) none of the samples tested showed over 3 per cent ductility, whereas the specifications for. the eastings *850 called for 15 per cent ductility; and (5) a test section taken from the area where the counter-weight was welded to the casting at Syncro broke at 15,500 pounds, but it did not break in the heat affected zone of the weld. 14 Dr. Talbot testified that specimens 3, 4, and 5 were so bad that he felt that the foundry had used either improper gating 15 or improper risering. 16 He said that improper gating could cause turbulence or would fail to skim off the dross, and that a riser would have voids in it but the casting would not if you remove a riser from the casting. He also found what he called an old crack found at the bottom of the casting. He described the operation of the safety limit switches, stated that they were not installed on the strander in question, and expressed the opinion that neither switch would have prevented the accident. Dr. Talbot expressed the opinion that the poor quality of the casting which fractured resulted in the failure of the machine, and that the fracture allowed the bobbin and an accessory part of it to be thrown against the tube wall. He admitted that the only way to control porosity is to obtain x-rays, as the Mag-naglow process merely disclosed surface cracks and not voids beneath the surface of the metal. He also said that the design of the casting determined the number of risers, but he admitted that the design was materially changed by addition of plates welded on the side of the casting. Dr. Talbot testified that the weld marks were about Yi" wide, and that the casting was more brittle at the point of the greatest heat, becoming less brittle as one moves away from the weld mark or zone. On the larger rectangular welding, directly under the saddle, Dr. Talbot pointed out the fracture line running from the bottom of the casting upward about one-quárter of its distance, then traversing diagonally across a raised rib to the left, and continuing upward alongside the white bead mark at the right side of the smaller square welded block. He affirmed that “except that traveling diagonally across from one weld mark to another, [the] fracture was almost squarely on the bead of one weld and the bead of another weld.” He then said that he “felt” that the lower portion of the fracture was a fatigue fracture, from the appearance of the surface of the break. “That is where I would anticipate the fatigue crack existed, if it did exist.” (N.T. 318.) Again he said that he did not think the area where the cracks started was a heat-affected zone though it was within a half inch of it, or even less, but he was not sure. Dr. Talbot confirmed that, in his report to Southwire, he had stated that the strander had been in service approximately 60 weeks and was operating at speeds of 500 to 550 rpms for 120 hours a week, and that his purpose in mentioning this was to try to determine the amount of fatigue it had endured. He said the normal endurance level of steel *851 was 10,000,000 cycles, and that “if it goes 10,000,000, you assume that is past the endurance level.” (N.T. 320.) He added that if the casting did not break within 10 million cycles, there would be a decreasing possibility that a fracture due to a defect would occur; indeed it “probably won’t.” (N.T. 332.) Dr. Talbot also testified that he would expect a faulty casting, such as the one which failed and was tested by him, to last without a fatigue crack “anything from 100,000 cycles to 10,000,000 cycles. He had experimented by cutting bars out of the same specimen and testing them. He had found that one broke at 54,000 cycles, and one went for several million cycles and did not break. He concluded that one could run ten identical tests and get as much as a three to one difference “when you get a fatigue failure or maybe even more.” He then said that in his report he calculated that from the number of hours per day, with uninterrupted use except for maintenance and changing of bobbins, and operating at 500 to 550 rpms a minute, this casting had held up for two to two and a half million cycles. However, during a recess Dr. Talbot recalculated the number of cycles, and he found in his notes that 216,000,000 cycles had been completed by the strander which first broke, and that it should have been over 250,000,000 in his report for the strand-er now in question. (N.T. 327.) Dr. Talbot also testified that the normal stress of each cycle would be higher if the machine tended to get a little out of alignment. He made it clear that he was speaking of revolutions of the barrel rather than revolutions of the bobbin, as the bobbin is just feeding wire, and there should be very little dynamic forces applied through the bobbin other than the static weight. Dr. Talbot conceded that he did not recommend welding. He conceded that most of the cracks he found in the fractured cradle were found in the weld areas. The hardening was also in the weld areas, and hardening causes brittleness. He concluded, “It was my estimation that the fracture did not start in the weld bead but it did occur in the weld bead after it was initiated.” (N.T. 335.) Dr. Talbot testified that the static weight of the spool would never break the casting. He thought that if this had been a sound casting, it would not have failed as it did. He further stated that if it had not been welded, it (the failure) might have occurred “next week” (although it has completed some 1,440,000,000 additional revolutions in the past eight years since the accident), but he went on to state that: “the welding did, in essence, cause a heat-affected zone, which the crack was in this area and it had to play a part in it." (N.T. 349) (emphasis added). Moreover, on rebuttal, Dr. Talbot acknowledged the existence of a high-tensile-stress-affected zone outside of the heat-affected zone resulting from the welding, but could not say how high the stress or the magnitude of the stress, or whether the stresses were detrimental or not. (N.T. 507.) 2. The testimony of Dr. Grosvenor Inter alia, Dr. Grosvenor testified about the effect of porosity on tensile strength. He said that center-line or interior porosity will reduce the strength in direct proportion to the percentage of voids, and that a void on the surface of the casting would not have an appreciable effect on the strength of the casting. The expression center-line, he explained, refers to anything below the surface. He then said that the process of welding ductile iron is liable to reduce the strength of a casting, because adjacent to the weld where the temperature was reheated into a red-hot temperature range the mass of the casting adjacent to it would tend to cool rapidly. Hardening and embrittlement will result unless the casting is annealed after the welding. Dr. Grosvenor stated that the 60-45-15 strength requirements prescribed in this case made the casting many times stronger than was required for its expected use, namely, to hold the weight of *852 the spool in an almost static position. He described the casting as “tremendously strong.” (N.T. 450.) He made rough calculations of the margin of safety and found that a casting meeting the requirements specified would be twenty-five times the required strength to carry the weight of the 1500 lb. reel. From his observation of: (1) the x-rays showing certain voids, (2) the sample P-15(b) of the fractured surfaces indicated on Exhibit P-56, and (3) microscopic examination and micro-hardness tests of samples from 15(a) and 15(b) shown on-the same Exhibit P-56, he testified that it was his opinion that the tensile strength of the casting which fractured, without regard to the weakness caused by the welding, but judging from the porosity alone, was not reduced by more than thirty percent. He went on to say that even if the porosity of this casting had reduced its tensile strength by fifty percent (which he did not concede), the casting still would never have broken from the foundry defects when loaded from above with the weight of the bobbin. Dr. Grosvenor testified that the Bri-nell hardness had gone up to about 400 (from the normal of 170) at the closest point he could measure, with even harder regions close to the surface where he could not measure it precisely but could get an indication of the hardness through microscopic examination. He said that it had gone way over “file hardness”, i. e., the degree of hardness at which it could scarcely be scratched with a file, and that this hardness had embrittled the structure. Dr. Grosvenor said that hardening causes expansion — • sizable expansion, tugging on adjacent areas, setting up residual stresses of considerable magnitude. He testified that the fracture as shown in the photograph P-39 was in the heat-affected zone, an area of high stress, quite close to the weld, and that the heat affected zone spread out for perhaps According to Dr. Grosvenor, the significance of the fracture adjacent to the weld, rather than in some other porous area, was that: (1) the casting originally had a reinforcing rib in the form of a three-sided square, and the weld left a valley between the weld and the rib — a thinner area surrounded by two thick areas — which created a structural notch; and (2) the residual stresses set up in this valley (between the balancing weight and the rib) resulted in triaxial stresses which created a more dangerous concentration than would prevail if there were a simple tension or biaxial stress. In Dr. Grosvenor’s opinion, the crack started and ran up that valley, then branched out, crossed the rib, and went up another weld heat-affected zone on the outside, on the other side where there was still a chance for triaxial stress. Addressing the matter of Dr. Talbot’s tests, Dr. Grosvenor said that they did not compare with the ASTM test requirements, which prescribe machining the dross off the test bar before testing for ductility. He would not accept the tension test made on an original cast surface. And he said that Dr. Talbot “did not” reproduce in his tests the stress pattern which took place when the accident occurred. (N.T. 472.) It was Dr. Grosvenor’s opinion that there was no evidence of fatigue failure indicated in the photographs or in his examination of the two pieces of the casting placed in evidence and shown on the photograph P-56. He testified that this appeared to be an overload fracture rather than a fatigue fracture which would show progressive small steps with markings similar to beach marks or clam shell marks. Furthermore, Dr. Grosven- or noted that a fatigue fracture surface is usually quite smooth, but no smoothness was indicated in the photographs. (See N.T. 52 through 55.) On the other hand, the surface produced by an instantaneous break was rougher and more jagged. Dr. Grosvenor testified that the load alone — the weight of the loaded spool— was so small that barring some untoward incident, such as an impact, they may go on forever, but that if they last *853 ed for 60 weeks and didn’t break, they didn’t fail by fatigue. He stated that “the welding very seriously damaged the strength of the cradle casting that broke.” (N.T. 448.) In his opinion, the casting failed because of the high stresses and the embrittling from the welding combined. He demonstrated the high stress pattern. He examined the sample 15(b) by microscope and did a micro-hardness test. It had a hardened zone and therefore a volume change which had to produce stresses, and the line of crack ran along the side of the weld bead, crossed over and ran up along the side of another weld bead. As he said, “There is evidence aplenty of a high stress pattern because this is a part that is lightly loaded in ordinary service.” (N.T. 473.) Dr. Grosvenor gave the opinion that the stress pattern of the weld, per se, caused the rupture, and he testified that he believed that the bending started on the outside of the surface shown on P-39, not on the inside surface shown on P-56. Dr. Grosvenor also pointed out that Dr. Talbot had testified that in cutting the samples it was found that they were hard in the heat-affected zone next to the weld; that he (Talbot) said that he might have made micro-hardness tests, but did not; and that Talbot did not find the striations or beach marks usually found in a fatigue failure. We asked Dr. Grosvenor whether, if Dr. Talbot’s tests were proper by ASTM standards, and his tensile strength yield strength and elongation figures were accurate, he (Dr. Grosvenor) would then agree that the cause of the fracture was fatigue and a poor casting as opposed to the effect of the welding. He replied: No, not at all. Because even with his test method, with which I disagree completely, he has with the exception of two bars, he has strengths which run up to at least two thirds the required specified strength. He has two that are low. And if I were to take these as indicative of the character of the material, then the strength of the material is reduced, let’s say, to not more than half of 60,000 is 30,000. the thing is twenty-five times or more, twenty-five or more times as strong as it need be, and to take half of it would make it twelve or thirteen times as high. It still has a safety factor of twelve or thirteen if we base it on his tests. (N.T. 497-98.) We then observed that Dr.- Talbot’s tensile and yield strength figures were far better than the elongation figures. Dr. Grosvenor replied: A. In the normal use and the stresses to which this is subjected, his percent of elongation has no effect because in the normal operation of this that saddle is never deformed by Vio of 1 percent. Q. So it is the tensile-strength and yield-strength figures that are critical ? A. The only reason we are interested in having 15 percent or more elongation is in case of an accidental blow. If somebody drops a reel on it, they might break it. If something came loose and rattled around inside the drum and struck a blow. If the operator didn’t latch the axle in tight. Something of that sort. Then — which is not normal use. This is abnormal — it would tend to cut down on the damage done. But percent elongation is not important after we get the thing assembled. (N.T. 498-99.) 3. Our Findings of Fact as to the Cause of the Casting Failure Notwithstanding this lengthy recapitulation of the expert testimony, it is apparent that there are many areas of agreement between Drs. Talbot and Grosvenor as to some metallurgical principles and some physical findings. It is unnecessary that we restate the areas of agreement. Instead, we shall concentrate upon what is our critical task as the fact finder: resolving the areas of disagreement and uncertainty *854 and making the ultimate findings upon which our decision must turn. First. We find that, before being installed in the strander, the casting which ultimately fractured was substantially altered and materially weakened by the arc welding of the counterweights. We further find that the alteration and weakening was a substantial factor in causing the accident. We credit Dr. Grosvenor’s testimony in this regard, and we note that Dr. Talbot was equivocal at best on the point (see pp. 849-851 supra). By way of subsidiary findings, we note that, inter alia: (1) there was significant hardening and additional stresses and ultimate embrit-tling of the metal in the heat affected zone; (2) the left weld bead of the larger weld was along the line of greatest stress produced by the weight of a.loaded spool; and (3) the fracture emanated from the heat affected zone close by the weld bead. 17 Even if we were not to make a positive finding on this subject, as we do, we would be obliged to find plaintiffs had failed to establish by a fair preponderance of the evidence that the casting was not substantially changed or materially altered. Second. Plaintiffs have not proved that, even aside from the alteration, that Beloit cast a defective cradle which was unreasonably dangerous and which proximately caused the accident. Plaintiffs have proved that the casting was significantly porous, but all castings are porous to some degree. Dr. Talbot estimated that 50% to 75% of the castings were non-uniform and contained some porosity, but his tests cannot be fully credited because they departed from accepted ASTM procedures, since they were cut from finished castings and the surfaces were not machined to remove dross. Moreover his simulated welding did not match the actual conditions of the counterweight weld. Hence, we can draw no conclusions from the fact that Dr. • Talbot’s test bar did not break in the heat affected zone of the welding. Syncro and Southwire did not specify, nor expect, a completely non-porous casting, and did not order x-rays or other tests to check the porosity. The “free of blowholes” specification related only to surface defects. We credit Dr. Grosven- or’s testimony that the casting was very strong. It lasted for over 250 million cycles, well beyond the expected endurance limit, 18 and not under static conditions, before it failed. Had it been subjected to a practically static load, as some seemed to think it would be, it might never have failed. But even under conditions of stress (see infra), the plaintiffs did not prove that the casting, had it not been altered by the welding, would have failed anyway. Put differently, Dr. Talbot did not overcome Dr. Grosvenor’s testimony that even if the porosity of the casting had reduced its tensile and yield strength by 50%, and its elongation by still more, it would not have broken from foundry defects. The evidence as to what would have happened had there been no counterweight welding is not conclusive. At best for the plaintiffs, it is in equilibrium, hence we must render on this point what is called, in a legal colloquialism, a “scotch verdict,” which means “not proven.” Third. There is one other factor which may have contributed to the accident: undue stress from the machine. In the early months of its operation, the strander had a serious vibration problem which was ultimately corrected. But, despite attention to the matter, the problem of misalignment was never completely cured. Moreover, from time to time (indeed, “frequently,” according to Roger J. Schoerner, the Vice President of Southwire), the strander would en *855 counter tangles caused by improper wrapping of the wire on the spool or reels which might pull the bobbin over and rotate the cradle and bobbin. Beloit has argued from Schoerner’s testimony that a snag in the wire probably pulled the locking mechanism from the cradle, bending the bobbin shaft at the time and causing the failure of the counter-weld and of the weakened casting. Be-loit also points out that seven bobbin shafts were bent in the two stranders and had to be replaced. (See testimony of William R. Benson, Southwire’s Cooper Division Manager and Project Engineer with respect to the stranders.) Be-loit submits that the frequency of the bobbin shaft bending supports this theory. To this argument, Beloit adds Gros-venor’s testimony that high stresses were a factor in the accident, as well as the weakened casting. We find that the casting was subjected to more stress than was originally contemplated by Syncro or Southwire and that the stress affected the casting in some degree, the extent of which is unknown. We make no finding that the machine-induced stresses which we have described caused the accident. We discuss and make findings on these factors because the plaintiffs have not negated stress as a possible cause of the accident. (See Discussion, infra,.) The presence of these factors further weakens the plaintiffs’ contentions that the accident was caused by the cradle as originally cast by Beloit. Again, a “scotch verdict”: “not proven.” Fourth. Plaintiffs offered virtually no evidence that Beloit failed to exercise due care in the manufacture of the casting. Dr. Talbot testified that test specimens 3, 4 and 5 were so bad that he felt that the foundry has used either improper gating or improper risering. However, Dr. Talbot was not an expert in the pouring or foundry process as such. No effort was made by the plaintiffs to establish any impropriety in Beloit’s methods or to counter the testimony of Jerre Lender, who supervised the “pour” of the casting in question. This is understandable since the plaintiffs have not pressed their theory of negligence, preferring to rely upon § 402A. Based on the evidence presented, we find that plaintiffs did not prove negligence. We turn now to a discussion of the applicable principles of law. We have made no findings of fact with respect to Beloit’s claim over against the third party defendant Syncro because of our disposition of the principal claim. 19 III. Discussion Our discussion of the applicable legal issues need not be lengthy, for the foregoing findings of fact are essentially dispositive. We must, however, discuss the legal issues so as to elucidate the manner in which we have instructed ourselves in the law, sitting, as we have, without a jury. We will confine our discussion to plaintiffs’ claim under § 402A of the Restatement (Second) of Torts (1965), for plaintiffs did not establish any negligence by Beloit in the manufacture of the casting, nor press the negligence claim. Two threshold issues may be quickly resolved. First, we note that *856 under Pennsylvania law § 402A applies to property damage claims. See Mac-Dougall v. Ford Motor Co., 214 Pa.Super. 384, 257 A.2d 676 (1969). Second, Pennsylvania has applied § 402A to manufacturers of component parts. In Burbage v. Boiler Engineering & Supply Co., Inc., 433 Pa. 319, 249 A.2d 563 (1966), the Pennsylvania Supreme Court held that a manufacturer of a valve, which was incorporated into a boiler by another party, was liable to plaintiff when a boiler exploded due to a defect in the valve, thus rejecting the contentions of the component part manufacturer that the liability shifts to the assembler or manufacturer of the product who incorporated the component part. Plaintiff must prove certain essential elements in order to recover under § 402A. In pertinent part, section 402A provides: § 402A. Special Liability of Seller of Product for Physical Harm to User or Consumer (1) One who sells any product in a defective condition unreasonably dangerous to the user or consumer or to his property is subject to liability for physical harm thereby caused to the ultimate user or consumer, or to his property, if (a) the seller is engaged in the business of selling such a product, and (b) it is expected to and does reach the user or consumer without substantial change in the condition in which it is sold. Restatement (Second) of Torts § 402A (1965). Although it is not articulated as such in the section, it is the law in Pennsylvania that the plaintiff must prove that the defective condition of the 'oduct was a proximate cause of the ac-dent. Colosimo v. The May Department Store Company, 466 F.2d 1234 (3d Cir. 1972). These essential requirements of 402A frame the principal legal issues in this ease. (1) Did Beloit sell a defective product, unreasonably dangerous to the user or consumer? (2) Was that product expected to and did it reach the user or consumer without substantial change in the condition in which it was sold? (3) Was a, defective product a proximate cause of the accident in question? Consistent with the order of our findings of fact, we will address the second and third questions (which are essentially two sides of the same coin) before taking up the first one. In view of the plethora of eases which have been brought under § 402A, it is surprising that so little has been written on the subject of § 402A (l)(b), which deals with substantial changes in the condition of the product. It seems to us that one essential element of a plaintiff’s case under § 402A is proof that the product is expected to and does reach the user and consumer without substantial change in the condition in which it is sold. While no reported case that we can find contains any discussion of the point of burden of proof (or, more precisely, risk of non-persuasion), there are several cases which mention it in passing. Judge Spaulding, in dissent in the case of Kuisis v. Baldwin-Lima Hamilton Co., 224 Pa.Super. 65, 301 A. 2d 911 (1973), wrote that it is the plaintiff’s burden to show that there was no substantial change in the defective part of the product between the time it was sold and the time of the accident. 20 Assuming that plaintiff had the burden of negating all reasonable causes for the malfunction except for the manufacturing defect relied on in a § 402A suit, Judge Staley, in Greco v. Bucciconi Engineering Co., 407 F.2d 87 (3d Cir. 1969), spoke of the substantial change concept in terms of intervening su *857 perseding cause. 21 And, in an opinion -just filed, Judge Weis wrote: Thus the plaintiff failed to prove a critical prerequisite that the product was in the same condition, so far as was relevant, on the day of the accident as it was at the time Bethlehem acquired possession. Schreffler v. Birdsboro Corporation, 490 F.2d 1148, at 1153 (3d Cir. 1974) (footnotes omitted). In effect, these cases imply that plaintiffs’ failure to negate substantial changes is the same as saying that they failed to prove proximate cause, since they failed to negate the break in the causal connection between the original defect and the ultimate injury. Rarely do judges comment, except in charges to the jury, on who has the burden of proving proximate cause, since it is such a fundamental element of plaintiff’s case that it is assumed sub silentio to be within plaintiff’s burden of proof. See Restatement of Torts (Second) § 433B (1965). Similarly, we believe that plaintiffs have the burden of proving no substantial change in order to prevail under § 402A(l)(b), although, as indicated above, we can find no case that directly reaches this holding. The plaintiffs’ brief suggests that the question of proving substantial changes is an affirmative defense. They cite no case or commentator to support their proposition, and we have not found any authority or persuasive reason for their conclusion. To be sure, there are reasons for requiring the defendant to carry the burden of coming forward and alleging a substantial change. In many cases, the alleged substantial change is perceived only by the defendant who should, therefore, have to allege it and maintain the burden of going forward with the evidence on the point. The plaintiff is accordingly put on notice to try to prove no substantial changes were made. And as a general rule, rather than requirng a plaintiff to negate an infinite number of possible changes, it seems more reasonable and in keeping with our adversary process to expect the defendant to allege the substantial changes he expects a plaintiff to try to disprove. But once the defendant has come forward, we believe that the burden of proof (or again, more precisely, *858 the risk of non-persuasion) must remain with the plaintiff; if the scales remain in equilibrium on the point, plaintiff is the loser. To impose on the plaintiff the burden of proving both proximate cause and no substantial change is consistent with the underlying concept of § 402A. In the seminal case of Greenman v. Yuba Power Products, 59 Cal.2d 57, 27 Cal.Rptr. 697, 377 P.2d 897 (1963), Chief Justice Traynor articulated the rationale behind § 402A as follows: The purpose of [strict] liability is to insure that the costs of injuries resulting from defective products are borne by the manufacturers who put them on the market rather than by the injured persons who are powerless to protect themselves. 22 While § 402A is meant to require manufacturers and sellers to bear much of the responsibility and cost of injuries to consumers resulting from their defective products, it is not meant to impose upon each manufacturer and seller an absolute liability as insurer for all injuries to consumers, regardless of the relation of plaintiff’s injuries to the particular defendant’s product. In the instant case, we have found that plaintiffs failed to negate the substantial change alleged by Beloit and similarly failed to demonstrate the proximate cause element necessary for recovery. Indeed, we have found that Beloit proved (although it had no such burden) that the casting failed, not as a self-contained component part with a defect unreasonably dangerous at the time it left Beloit, but due to the changes that were made in it by the counterweight welding. Accordingly, plaintiffs cannot recover. Plaintiffs also have failed to meet their burden under the first condition of § 402A — proof of a defective product unreasonably dangerous to the user or consumer. It is not enough that the casting in question be proved porous, for even if a porous casting is considered defective because it is a less than perfect casting, it must also, under § 402A be “unreasonably dangerous.” The latter concept has been variously defined. Comment (i) to § 402A states: The article sold must be dangerous to an extent beyond that which would be contemplated by the ordinary consumer who purchases it, with the ordinary knowledge common to the community as to its characteristics. Frumer and Freedman, in their treatise, Products Liability, in § 16A[4][e] at pp. 3-317 to 3-325 suggest that “unreasonably dangerous” is equivalent to the older warranty notion of fitness for the ordinary purpose for which it was manufactured. 23 In the instant case, even though plaintiffs have shown that the casting was porous, they have failed to prove that it was defective and unreasonably dangerous in the sense that the porous casting without substantial changes could not have withstood the stress of normal operations for its expected useful life. We have found as a fact that the casting in question had been subject to over 250 million cycles, far more than the 10 million cycles during which any original defect was expected to show up, according to Dr. Talbot. It was also subject to welding and considerable strander-in-dueed stress which plaintiffs failed to negate as possible causes of the accident. 24 Plaintiffs then have failed *859 to meet their burden of proof and judgment must be entered for the defendant Beloit. The third party claims thus become moot. Footnotes [1] The suit was tried without a jury. This Opinion constitutes our findings of fact and conclusions of law pursuant to Fed.R.Civ.P. 52(a). [2] Southwire’s damage claim, which totals approximately $290,500, encompasses claims for: (1) damage to the strander and to equipment and fixtures nearby; and (2) consequential damages for loss of production, sales, profits, customers and good will. [3] Actually Travelers has brought both actions, electing, as is its right under the loan receipt agreement, to bring one suit in its own name and one in that of Southwire. The cases were consolidated for trial. [4] The parties agreed prior to trial that the law of Pennsylvania should govern, and Pennsylvania has adopted § 402A. Webb v. Zern, 422 Pa. 424, 220 A.2d 853 (1966). [5] While a claim based upon breach of warranty was originally asserted, it was abandoned prior to trial when it became apparent that the defendant would prevail on that issue by its defense of statute of limitations. [6] On February 22, 1963, eleven months prior to the accident, strander No. 4329 broke down. No. 4329 was the “sister” strander purchased at the same time as the strander in question in this case (# 4328). One of the castings on No. 4329 was thrown out of an open “window” of the tube in which it was installed. Three of the tubes of the strander reared up, then dropped to the floor, and extensive property damage resulted. Following this incident, Syncro delivered to Southwire certain limit switches for installation in the stranders. These switches were designed to prevent excessive rocking of the cradles while the stranders were in operation and to cut off the motor to prevent damage in the event of any such malfunction of the machine. In particular, they were to prevent rocking in excess of twelve degrees and were tied to the circuitry to make certain that the bobbin lock was on and that the locking bar held the bobbin in a locked position. These switches were not installed in either strander up to the time of the incident in question. William R. Benson, manager of Southwire’s Copper Division, testified that they did not install the limit switches because they performed the same function as two other circuits on the machine (although he did not describe the other circuits or indicate how they operated). [7] Actually, a total of 14 cradle sides were specified for seven bays of the strander in question. [8] Annealing is a process involving heating and slow cooling of the metal. [9] Under this process, sometimes known by the tradename “Magnaflux,” the surface of the metal is magnified and made to glow, revealing surface cracks not visible to the naked eye. [10] In the destruct method, samples from the castings are destroyed so that a cross-section can be examined to determine more readily if subsurface defects exist. [11] Dr. Talbot’s opinion that the cradle was the first part to break is supported by the following. The bobbin shaft in bay No. 6 was broken at the pin hole (a hole at the end of the bobbin axle through which a pin is inserted to keep the bobbin axle from rotating). The only way the shaft could have been broken at the pin hole would be if the casting moved relative to the bobbin shaft. The bobbin shaft in bay No. 6 was also bent at a location which would be consistent with the casting twisting relative to the bobbin. Therefore, the condition of the bobbin shaft after the accident verified that the shaft was secured by the hold down or locking mechanism prior to the accident. The bolts which secured the end of the cradle to the strander were sheared off as a result of the accident, and the source of this force was the cradle trying to separate from the bobbin shaft. The bolts failed ductilely, which indicates they came loose or elongated before they broke. The locking mechanism or the hold down mechanism of the bobbin shaft was in place subsequent to the accident. The mechanism was also bent by the movement of the bobbin shaft relative to it. [12] Dross is waste product or impurities formed on the surface of molten metal. [13] Dr. Talbot was recalled in rebuttal. He testified: (1) that he realized that the tests ■were not ASTM tests when he made them; (2) that j'ou get a transformation and a volume expansion in a heat-affected zone that does cause a residual stress ; and (3) that the weld bead that he had welded should have: (a) that same volume expansion; (b) basically the same expansion for the length of the weld; and (c) the same residual stress pattern or maybe not quite as high. He also agreed that because of the dross he would expect a lower reading than if the test had been made in accordance with ASTM methods. [14] The actual results were: Specimen Tensile No. Test (psi at which Ductility 1 56 800 samPle broke) 2% 2 42,100 3% 3 56,800 4 Broke away 0 from welded-affected area at 15,500 5 34,100 6 4,500 0 7 43,200 8 54,500 2% [15] Gating is the wáy one designs the entrance where the molten metal flows into the mold cavity. [16] A riser is the last part of the molten metal to solidify and should then be removed since it will contain voids which would otherwise be present in the cast metal. [17] According to all the testimony on the subject, and particularly the photograph P-39, the fracture line, for about one-third of its length, was on or within a small fraction of an inch of one of the weld beads, then crossed over and for two-thirds of its length was on or within a small fraction of an inch of the other weld bead. [18] Dr. Talbot testified that if it didn’t crack by that time, it probably won’t, at least not due to a defect in the casting. (See N.T. 332.) [19] A most unusual state of affairs arose shortly after the conclusion of the trial. As we were in the course of conferring with counsel about the post trial briefing schedule, Drs. Talbot and Grosvenor were having a conversation about the issues that developed at trial. Several days later, as the result of that conversation, Beloit submitted to the Court an affidavit by Dr. Grosvenor to the effect that in the post trial conversation, Dr. Talbot had advanced a different version of what had occurred than he had testified to. In particular, the Grosvenor affidavit contended that Talbot had conceded that a significant amount of stress was imparted to the easting by misalignment of the strander. The affidavit was immediately countered by the plaintiffs and its substance denied. Under the circumstances, we felt it necessary to hold a hearing on the matter and did so. We shall not recapitulate the testimony. The hearing turned into a “Mexican standoff.” In sum, we do not believe that Dr. Talbot said anything different off the record than he did on the record, hence we do not dwell upon the testimony adduced at the post trial hearing in any significant degree in our findings. [20] Judge Spaulding cited Burbage v. Boiler Engineering & Supply Co., supra, as authority. While we agree with Judge Spaulding’s statement of the law, we note that in Burbage the Pennsylvania Supreme Court merely upheld the jury’s finding that the changes in question did not cause the defect; the court did not confront the burden of proof question. [21] Intervening superseding cause is a concept used in negligence cases and may be defined as an act of another person or force which by its intervention prevents the defendant from being liable for harm to another which his antecedent negligence is a substantial factor in bringing about. Restatement (Second) of Torts §§ 440-53 (1965). See also 27 P.L.E. § 80 at pp. 152-53 (1960). We are not unmindful of the fact that the notion of “substantial change” is difficult to define, except perhaps on an ad hoc basis. A substantial change which negates § 402A liability may well be posited as something less significant than an intervening superseding ease. The rationale behind this argument is as follows. Since a defendant in a § 402A action can be held liable without proof of any fault on his part, this broad liability should not be imposed on a faultless defendant unless the plaintiff can prove a chain of causation linking the defendant’s defective product to plaintiff’s injury without substantial changes that may be less significant than intervening superseding causes needed to negate the liability of a negligent defendant. On the other hand, since § 402A liability is designed to be broader than negligence liability, it would be consistent with the intent of § 402A to require that, for a substantial change to negate § 402A liability, it must be at least as significant a break in the chain of causation as an intervening superseding cause is in negligence law. One could even argue that substantial change should be defined as .the sole proximate cause before negating the § 402A liability of a defendant with a defective and unreasonably dangerous product. In the case at bar, we do not need to adopt any of these approaches, since we have found : (1) that plaintiffs failed to prove that the casting was a defective and unreasonably dangerous product which would have broken had there been no welding or undue stress; and (2) that the defendant has proved, although it did not have to, that a substantial change (i. e., the counterweight welding) was a cause of the accident. In this case then, plaintiffs failed to prove both causation in fact and proximate cause, since they failed to establish an unbroken chain of causation (or repair a bx’oken one) between an originally defective casting and the ultimate injury. [22] In addition to Judge Traynor’s reason, Dean Prosser notes that permitting a plaintiff to sue the manufacturer directly on a theory of strict liability avoids a series of actions which would accomplish the same result, after an otherwise time-consuming and wasteful process. See Prosser, Law of Torts 674 (3d ed. 1964). [23] See also Dorsey v. Yoder Company, 331 F.Supp. 753, 760 (E.D.Pa.1971), aff’d 474 F. 2d 1339 (3d Cir. 1973), (quoting Dean Wade’s formula for determining unreasonable danger in J. Wade, Strict Liability of Manufacturers, 19 Sw.L.J. 5, 17 (1965)). [24] A plaintiff in a § 402A case can establish a defective condition by proving that the *859 product functioned improperly in the absence of abnormal use and reasonable secondary causes. See Greco v. Bucciconi Engineering Co., 407 F.2d 87 (3d Cir. 1969) ; Burchill v. Kearney — National Corp., Inc., 468 F.2d 384 (3d Cir. 1972) ; Hayes v. Pa. Lawn Products, Inc., 358 F.Supp. 644 (E.D.Pa.1973). Plaintiffs have not established liability in this fashion either.