Balder v. Haley

OPINION CRIPPEN, Judge. Michael Balder was injured when a water heater exploded in his mother’s home. He brought an action against the seller of the water heater, a repair worker who had serviced the water heater, the gas company that provided gas for the water heater, and the manufacturer of the water heater. Also included in the lawsuit was Honeywell, Inc., which had manufactured the water heater’s gas control unit. The jury found Balder’s mother 45 percent at fault for the accident, the repair worker 35 percent at fault, and Balder 20 percent at fault. Honeywell, Inc. was exonerated by the jury. Balder appealed, claiming that Honeywell was at fault as a matter of law. We reverse and remand for a new trial to determine whether Honeywell’s inadequate warnings and instructions caused Balder’s injuries. FACTS On July 23,1980, appellant Michael Balder was severely injured in an explosion of gas that leaked from a control on a water heater in the home of Josephine Pirkl, Balder’s mother. The water heater was manufactured by Republic Heater Company and installed in 1970. It utilized a Y5130 gas control unit manufactured by Honeywell, Inc. Michael Balder and his wife Zita brought suit against Republic Heater Company, Honeywell, a repair *858 worker who had attempted to fix the water heater control, and the gas suppliers. Josephine Pirkl was brought into the case as a third party defendant. Balder contends the jury should have found Honeywell to be at fault and then determined whether Honeywell’s misconduct proximately caused Balder’s injuries. The facts of the case require an explanation of the water heater control and the malfunction of the unit that might lead to an explosion. A. Nomenclature 1. Control Knob The user’s main contact with the water heater was through the control knob on the V5130 gas control. The knob, which is mounted on the reset shaft, is marked to indicate three control functions: off, pilot, and on. The knob was designed so that it cannot be rotated from the off position to the pilot position without depressing the knob on a shaft; it cannot be rotated from the pilot position to the on position without removing the pressure on the knob. 2. Gas Cock Situated beneath the control knob is the gas cock, a ported valve which controls the flow of gas. Protrusions from the knob fit into notches on the shaft at the top of the gas cock and allow the gas cock to rotate when the knob is turned. When the gas control knob is in the off position, the port of the valve is positioned to prevent the flow of gas. When the knob is turned 90 degrees to the pilot position, the port is turned to a position that permits a small flow of gas to the pilot chamber for the lighting of the pilot. When the knob is turned another 90 degrees to the on position, the valve permits a large flow of gas to the burner. 3. Reset Shaft The control knob is positioned on a shaft, called the reset shaft, which is housed vertically in the center of the gas cock. When there is no downward pressure on the control knob and the shaft, a safety shutoff valve remains sealed shut, which prevents the flow of gas. When the control knob is depressed, the shaft is depressed, causing the safety shutoff valve to be opened and the seal to be broken. The breaking of the seal allows the flow of gas into the gas cock. 4. Safety Shutoff Valve The safety shutoff valve is connected to the lower part of the reset shaft. It utilizes a system of O-rings to seal and prohibit the flow of gas into the gas cock. When the control knob is depressed, the O-ring seal is broken. 5. Spring Mechanism Attached to the reset shaft is a spring mechanism that is designed to force the reset shaft back to its original position when pressure is removed from the control knob. Thus, if pressure is removed from the control knob and the pilot is not lit, the spring mechanism should place the reset shaft back and close the safety shutoff valve. Added pressure for closing the safety valve comes from a spring directly beneath it. 6. Thermocouple When the pilot on the water heater is lighted, the flame heats the unit’s thermocouple, which in turn creates electrical current. The current creates an electromagnetic force. So long as the thermocouple circuit is complete, the electromagnet holds the valve open, permitting the continued flow of gas into the pilot chamber. This allows the pilot to remain lit even after pressure is removed from the control knob. If for any reason the pilot should go out, the thermocouple would cool, the circuit would be broken, and a properly functioning valve would close, preventing the further flow of gas. B. Function of the Unit When the gas control knob is in the off position, the flow of gas is blocked by the safety shutoff valve and by the position of the gas cock’s port. When the gas control knob is depressed, the safety shutoff valve’s seal is broken and gas flows into the gas cock’s chamber. Rotating the knob *859 and the gas cock 90 degrees to the pilot position causes gas to flow through the gas cock’s port into the pilot chamber. This allows the pilot to be lit. The operator must maintain pressure on the knob for a minute or so, during which time the pilot light heats the thermocouple. After a minute, the thermocouple mechanism operates to hold open the safety shutoff valve. Pressure can then be removed from the gas control knob and gas will continue to flow into the pilot chamber. Protrusions on the control knob prevent rotating the valve to the on position while the knob is depressed. Rotating the control knob another 90 degrees to the on position will allow a large flow of gas through the gas cock’s port into the burner chamber. The pilot flame will cause the burner to ignite. If the pilot light should be extinguished, a properly functioning spring mechanism will close the safety shutoff valve and gas will cease to flow. If the pilot goes out and the reset shaft is jammed down in some way, the spring mechanism will not reset the firings of the safety shutoff. Thus, gas will continue to flow and instead of being consumed by the burner’s flames, will flow out of the heater control. Liquid propane gas is heavier than air, so instead of. dissipating in the air, the gas will collect near the floor. The collection of gas creates a risk of explosion. C. Manufacturer’s Problems with the V5130 There is convincing evidence that the gas control knob used on the V5130 is a defective product. The knob wears out and becomes useless over a period of time because it was made of a low grade plastic. Once the knob wears out, the user could operate the control only by depressing and rotating the reset shaft with fingers or a tool. This can cause a dangerous condition: without the presence of the control knob, the shaft is more likely to be wiggled or bent. That can cause gas to leak through the safety shutoff valve and out from around the reset shaft. In addition, if the knob is not in place, the shaft can be depressed while the valve is in the on position, which would cause serious gas leakage if the pilot is not lit. Moreover, without the knob, it is less visually evident if the valve is in the on position or if the shaft is stuck for any reason in a depressed position. Honeywell acknowledges the importance of having the knob in place. A two inch by three inch sticker positioned on the heater near the gas control unit says “Danger,” “Explosion,” and states this information: “Red knob on this valve is a critical safety feature. If red knob is loose, broken or missing, shut off gas at manual shut-off and call gas company.” Although the control knob wore out quickly and was a critical safety feature, Honeywell acted slowly to remedy the problem. The defective design of the control knob was evident from the time of 1960 tests, which indicated that the knobs did not meet a predetermined durability standard. As evidence mounted on the poor quality of the knobs, a proven and inexpensive alternative design was rejected for the American market because of a theory that Americans used the knobs less frequently than consumers in other markets; distributors in other countries demanded a better knob. Honeywell received numerous customer complaints on the control knob, and beginning in 1975, explosions linked to defective knobs were reported. Although the company destroyed records of explosions occurring before 1978, evidence indicates company knowledge of a dozen explosions involving broken or missing knobs on Y5130 water heater controls. Finally, in January 1985, Honeywell began an “urgent” effort to reach all owners of the controls, seeking to replace the knobs and to affix warnings against improper use of the control. Recall advertisements notified owners of the danger of “explosion from damaged knobs on the Honeywell V5130 water heater controls.” Customers were advised: “Improper use and servicing can result in explosion or fires.” Newly affixed labels warned against improper use. Dealers were told to *860 act in the recall because their customers’ “lives may depend on it.” D. Use of Product Within several years after installation of the Pirkl water heater, the control knob malfunctioned. There was critical evidence the control unit was harmfully used after the knob failed. The jury heard evidence that the control assembly was disassembled at one time and improperly reassembled. Evidence indicated the reset spring had been replaced with a ball point pen spring, causing the spring mechanism to be less efficient. The control was operated without the knob, despite the warning that the unit knob was a “critical” part. A gas leak developed around the reset shaft as evidenced by a tiny flame burning from the shaft housing, the top of the gas cock valve, when the pilot light was burning. A tenant who occupied the house for several years first spotted the leak and attempted to fix it by melting dental wax around the shaft. Later, a repair worker saw the leak and stopped it by applying industrial adhesive around the shaft. The adhesive dried to a rubbery consistency and may have held the shaft firmly in place while preventing the leak of gas. There was evidence that if the shaft had been depressed while surrounded by the adhesive, the spring mechanism could not have reseated the safety shutoff valve’s O-rings when pressure was removed. Thus, gas would be permitted to flow and collect on the floor surrounding the water heater. There was also evidence of foreign material within the control and evidence that a foreign object could lodge between the O-rings to hold open the safety shutoff valve. Josephine Pirkl had been warned of the dangerous condition of the gas control unit. She did not heed the warning and failed to replace the unit or its critical parts while knowing of the defective condition that existed. On July 28, 1980, Michael Balder and his two children went to Pirkl’s home to cut firewood. While Balder and his children were outside, Pirkl was fixing lunch in the house. She called Tom Haley, a repair worker, when she discovered that she had no hot water. Haley told her to check the water heater’s pilot light. Pirkl called to Balder, who went inside. After observing that the pilot light was not burning, he discovered that the valve could not be turned with the defective knob. Pirkl reported this to Haley who told Pirkl to shut off the gas. Balder claims that when he went into the basement to do so, the water heater exploded, causing severe injuries. There is some evidence, however, that Balder was handling the heater controls or the gas supply line when the explosion occurred. E. Trial At trial, Balder attempted to prove that Honeywell was liable for sale of a dangerously defective product. Balder also sought to show that Honeywell was aware of the defective condition of the gas valve knob and failed to remedy the defect through recall or warnings. Honeywell presented evidence that even if it was at fault, Josephine Pirkl’s and other parties’ superceding fault caused Balder’s injuries. In a five-sixths verdict, the jury found that Honeywell was not at fault under theories of product liability or negligence. The jury found that Haley, the repair worker, was 35% at fault and Pirkl was 45% at fault. Balder, the jury found, was 20% at fault. The jury found that. $2.25 million would fairly compensate Balder and his wife for their losses. The trial court ordered Haley to pay the Balders $1.8 million plus costs and disbursements, and the court ordered Pirkl to pay Haley $1,012,500 plus his costs and disbursements. Appellants moved for a judgment notwithstanding the verdict, or in the alternative, a new trial. These motions were denied. The Balders appeal, claiming that Honeywell was at fault and that Michael Balder was improperly held to be 20 percent at fault for his injury. ISSUES 1. Did the jury err in finding Honeywell was not at fault for distribution of a defective product? *861 2. Did the jury err in absolving Honeywell from fault for inadequate warnings and instructions for users of its product? 3. Could the jury absolve Honeywell from fault for negligence due to fault of Michael Balder? 4. Did the jury err in finding Michael Balder 20 percent at fault? ANALYSIS This case was presented to the jury upon theories of defective design and negligent failure to warn. On both theories the jury exonerated Honeywell, finding that Honeywell had not sold a product that was dangerously defective and was not negligent. The trial court properly instructed the jury regarding products liability and negligence, following directives of the supreme court. See Bilotta v. Kelley Co., 346 N.W.2d 616, 621-22 (Minn.1984). Appellants seek upon appeal to set aside the finding of the jury and to obtain a new trial. We will not upset a jury decision unless it is manifestly contrary to the evidence viewed as a whole in a light most favorable to the verdict. Hudson v. Snyder Body, Inc., 326 N.W.2d 149, 155 (Minn.1982). We must sustain the verdict “if it is possible to do so on any reasonable theory of the evidence.” Hestad v. Pennsylvania Life Insurance Co., 295 Minn. 306, 311, 204 N.W.2d 433, 437 (1973). 1. Defective design To show that a manufacturer is liable in tort under a products liability theory, a plaintiff must show that the product was in a defective condition unreasonably dangerous for use, that such defective condition existed when the product left the hands of the defendant, and that the defect was the proximate cause of the injury. O’Laughlin v. Minnesota Natural Gas Co., 253 N.W.2d 826, 831-32 (Minn.1977). A manufacturer of a product has a duty to use reasonable care in the manufacture, assembly, inspection, and testing of the product to protect those who will use the product from unreasonable risk of harm while the product is being used for its intended purpose. See Holm v. Sponco Mfg., Inc., 324 N.W.2d 207, 212 (Minn.1982). Defective design of a product will constitute a defect for the purposes of product liability. See Bilotta, 346 N.W.2d at 621. A manufacturer’s duty of care in designing a product is stated as follows: [A] manufacturer is obligated to exercise that degree of care in his plan or design so as to avoid any unreasonable risk of harm to anyone who is likely to be exposed to the danger when the product is used in the manner for which the product was intended, as well as an unintended yet reasonably foreseeable use. Id. (quoting Holm v. Sponco, 324 N.W.2d 207, 212 (Minn.1982)). The definition of “reasonable care” required of a manufacturer varies from case to case and involves a balancing of the likelihood of harm, and the gravity of harm if it happens, against the burden of the precaution which would be effective to avoid the harm. Id. Thus, the degree of care required will depend upon the degree of potential harm, both qualitative and quantitative, and the cost of measures required to avoid the harm. In sum, if the manufacturer failed to use reasonable care in designing its product, the product is in a defective condition unreasonably dangerous to the user. Appellants in this case claim Honeywell was liable in tort for the defective design of the knob on the gas control unit. They argue that there is overwhelming evidence of the poor quality of the gas control knob. Such evidence, appellants claim, requires a conclusion that the jury’s determination is manifestly contrary to the evidence. Appellants further contend that Honeywell’s emphasis on misuse of the product constitutes only a “cause” defense. They concede that a jury finding of “no cause” may have been justified, but claim that the “no cause” evidence was not a proper basis for a finding that Honeywell was without fault. Appellants claim that Honeywell was at fault as a matter of law for distribution of a defectively designed product. *862 Honeywell disputes appellants’ analysis of the case. It claims misuse relates to the ease in two ways. First, Honeywell observes, the misuse was unforeseeable and constituted an intervening, superceding cause, preventing a finding that Honeywell’s conduct could have proximately caused Balder’s injuries. Second, Honeywell contends, absent the risks due to owner misuse, there was not sufficient evidence that the defective knob created an explosion hazard. The trial judge agreed with respondent’s analysis. Explaining denial of appellant’s posttrial motions, the court said that use of the valve without a knob did not “in and of itself create a danger.” Thus, appellant is confronted with the contention that Honeywell’s “no cause” evidence bears directly on the issue of faulty product design, justifying the jury decision that never reached the question of cause. This evidence, it is asserted, explains away much of the evidence of product hazards, materially reducing the evidence of hazards that trace to a defective part on the V5130 unit. Appellant’s position on appeal adds weight to respondent’s analysis. In briefs and arguments, appellant has not asserted that the evidence compels a finding of Honeywell’s fault independent of owner misuse. Nor has appellant asserted that there is compelling evidence that the misuse of the control was foreseeable by the manufacturer. While the evidence here would support a finding of Honeywell’s distribution of a defectively designed product, we cannot say the evidence permits no other finding, recognizing that we must view all the evidence in the light most favorable to the verdict. See Hudson, 326 N.W.2d at 155. In addition, like the trial court we fin'd authority for the contention that intervening cause evidence bears materially on the initial determination of whether a manufacturer is at fault for a defectively designed product. We disagree with the trial court’s summary of evidence that the knob defect alone may have created no danger, but the point has merit when limited to the danger of injuries from a major explosion. A manufacturer or supplier is liable only for defects existing at the time the product leaves the control of the manufacturer or supplier. Magnuson v. Rupp Manufacturing, Inc., 285 Minn. 32, 38-39, 171 N.W.2d 201, 206 (1969) (quoting Restatement (Second) of Torts § 402A, comment g (1965)). As a result, mishandling by the consumer after the product has left the control of the manufacturer or supplier may defeat proof of cause but may also defeat proof of the ingredients of fault. Id. at 44-45, 171 N.W.2d at 209 (quoting Prosser, The Fall of the Citadel, 50 Minn. L.Rev. 791, 840 (1966)). Likewise, only those defects that pose an identified “unreasonable risk” are actionable under a products liability theory. Bilotta, 346 N.W.2d at 621 (quoting Holm v. Sponco, 324 N.W.2d at 212). Whether a particular hazard is an unreasonable risk will depend upon the qualitative and quantitative degree of the risk and the cost of eliminating the risk. Magnuson, 285 Minn, at 42, 171 N.W.2d at 207-08. Evidence of foreseeability goes directly to the question of whether an actionable defect exists. See Magnuson, 285 Minn, at 40-42, 171 N.W.2d at 207-08 (user’s awareness of risk relates to determination of the “magnitude of the risk” and is associated with the question of whether a defect is actionable). Although the issue in Magnuson was a user’s awareness of a defect, the court’s language and its rationale also affect the question of consumer misuses. Damaging use is- a relevant consideration when determining which risks are unreasonable and foreseeable and thus deciding the scope of a manufacturer’s or supplier’s duty. See id. The evidence in this case permits a finding that the harmful use and repair of the water heater gas control unit was such that it was unforeseeable. This observation is not openly disputed by appellant. It is tenuous to argue that Honeywell should have been able to foresee that a user would melt dental wax around a gas leak or that the function of an essential safety feature would be impaired by use of an adhesive. Nor is it apparent that Honeywell had to *863 foresee the owner’s failure to repair or replace worn-out equipment that the owner knew was faulty. Because the misuse was unforeseeable, Honeywell’s duty was limited to the facts in the case that exist absent evidence of misuse. Thus, the only question that remains is whether Honeywell breached its duty based upon the residuum of facts. Did the trial court correctly assess the evidence showing hazards of the defective Honeywell knob “in and of itself?” There was convincing evidence in this case that the gas control knob was defectively designed and manufactured. It is not so evident, however, that use of a broken knob would jam open the gas valve so that a large quantity of gas would leak. It was shown that if the knob was placed on its stem backwards, depressed, and turned forcefully, the valve could be jammed into the on and depressed position and thus cause a major gas leak. The jury, however, could have reasonably concluded that such an awkward use of the knob was unforeseeable. Likewise, the jury could have reasonably concluded that there was not a specific, unreasonable risk of jamming the valve in a depressed position through use of the control without a knob. There was evidence that a spring necessary for reclosing the safety shutoff valve was missing. Appellants claim the spring fell off when the control knob was removed, but the jury could reasonably have concluded that internal parts of the gas control unit would not have been misplaced absent an unforeseeable and improper dis-assembly and reassembly of the control unit. There was evidence that a missing control knob could make more likely a jarring of the reset shaft and subsequent gas leak caused by a breaking of the safety shutoff’s seals. The jury could reasonably have concluded, however, that such a gas leak would be too small to produce an explosion. Although Honeywell admitted that the presence of a gas control knob was a “critical safety feature,” that is not compelling evidence that the absence of the knob creates demonstrable hazards of an explosion. Similarly, evidence of other explosions was not conclusive proof of a particular defect that would produce a major gas leakage. 2. Negligent failure to instruct and warn The trial court instructed the jury on the law of negligence to deal with evidence on Honeywell’s instructions and warnings for users of the V5130 control. See Hauenstein v. Loctite Corp., 347 N.W.2d 272, 275 (Minn.1984). Appellant addresses both the negligence and product liability findings of the jury in asserting that respondent’s position on causation could not justify a finding of no fault. 1 His contention has merit on the negligence issue. 2 *864 Negligence is the doing of something which an ordinarily prudent person would not do or the failure to do something which an ordinarily prudent person would do under like or similar circumstances. State v. Munnell, 344 N.W.2d 883, 886 (Minn.Ct.App.1984). An act or omission is not negligence if it could not be reasonably anticipated or foreseen that some harm would result. Logan v. Hennepin Avenue Methodist-Episcopal Church, 210 Minn. 96, 98, 297 N.W. 333, 334 (1941). If a manufacturer or supplier knew or reasonably could have discovered the danger involved in the use of the product, and if the product is not accompanied by adequate warnings or instructions, then the manufacturer or supplier is negligent. See Johnson v. West Fargo Manufacturing Co., 255 Minn. 19, 24, 95 N.W.2d 497, 501 (1959). A manufacturer or supplier must provide adequate warnings of dangers inherent in improper use of the product, if the use is one that the manufacturer or supplier could reasonably foresee, and must provide adequate instructions for the proper use of the product. Id. A manufacturer that provides advice on the use of a product assumes the responsibility for giving both accurate and adequate information. Id. The warning will be adequate if heeding of the warning will make the product reasonably safe for use. Magnuson, 285 Minn, at 39, 171 N.W.2d at 206 (quoting Restatement (Second) of Torts § 402A, comment j). Methods of communicating warnings must be reasonable under the circumstances to protect those whose safety is threatened. Sterling Drug, Inc. v. Yarrow, 408 F.2d 978, 994 (8th Cir.1969). In addition, the trial court instructed the jury: The court also told the jury that a manufacturer is “obligated to keep informed of scientific knowledge and * * * discoveries in their fields.” We agree that the duty to warn and instruct is continuing, not based alone on facts evident at the time of sale. See Hill v. Wilmington Chemical Corp., 279 Minn. 336, 342-44, 156 N.W.2d 898, 903-04 (1968). Where a manufacturer learns that previously distributed products pose dangers to users, its duty is mixed: the manufacturer must give additional warnings adequate for reasonable safety of users, or must take other remedial steps to the same end. Braniff Airways, Inc. v. Curtiss-Wright Corp., 411 F.2d 451, 453 (2d Cir.), cert. denied, 396 U.S. 959, 90 S.Ct. 431, 24 L.Ed.2d 423 (1969). If after a product has been sold, the manufacturer becomes aware of dangers in the use of the product, it has a duty to provide users with adequate warnings of the danger or instructions concerning methods to minimize the dangers. Appellants in this case claim Honeywell was negligent in failing to warn of the consequences of misuse of the gas control unit and in its failure to warn against any use of the unit without the control knob. The evidence permitted a jury finding that Honeywell might reasonably fail to foresee the specific harm done by others to Josephine Pirkl’s water heater control. Like the claim of design defect, the claim of negligent warning failures could properly be examined “in and of itself,” without regard for what others did to the Pirkl control. As to the negligence issue, however, the residuum of evidence leaves an inescapable picture of neglect on the part of Honeywell. While others were also at fault, it is undisputed that Honeywell had known the control knob was inadequate since 1960, that it disregarded inexpensive opportunities to correct the defect and delayed in its recall, and that since at least 1970, when the Pirkl unit was sold, it recognized explosion dangers associated with use of the unit after a knob failed. It is evident that known risks included some gas leakage from the valve if users attempted to operate the control without a knob. User difficulties in coping with a broken knob or gas *865 leaks were foreseeable. The company became aware of explosions that related to the control knob defect. The only warning given by Honeywell fails to instruct on the single safe course of action for users of the faulty knob: the abandonment of any use of the control after a knob fails until it is replaced. In other words, the warning does not instruct users that they should never remove the knob or use the control without a knob. The warning does not address the danger when the shaft without a knob is depressed or stuck while in the on position, or that without the knob a user cannot observe that the shaft was depressed or tell that it was in an on position. The warning does not magnify the hazard of gas leakage and explosion. It does not make it evident that a gas explosion will likely occur under some circumstances if the unit is operated without a knob in good working order. The warning does not disclose that Honeywell had known for many years that the knob was not designed for frequent use and that there was a hazard the knob would fail. More significantly, the warnings and instructions do not give information on the replacement of knobs. Looking beyond the evidence on the warning label actually used, the record includes additional evidence of Honeywell’s negligence. In spite of accumulated knowledge spanning two decades, Honeywell failed to provide updated warnings to any owners or users of installed units, failed to specifically advise anyone that the unit should never be used without a good knob, failed to provide any owners with adequate information for replacement of defective knobs, and failed to provide for any inspection or recall of V5130 control units. In marked contrast with its conduct during the preceding 25 years, steps taken by Honeywell in 1985 are significant because of the urgency attached to them, the specificity of instructions against use with defective knobs, and the thorough effort by Honeywell to contact all dealers and owners and to replace all faulty knobs. In sum, the evidence on warnings and instructions “in and of itself” shows negligence of Honeywell, Inc. While the Balders may not have proven a specific dangerous defect, independent of misuse by others, the evidence does establish that Honeywell knew there were risks of explosion connected with defective or missing control knobs and they failed to act reasonably to protect users from those hazards. In other words, difficulties in proof of specific problems did not hamper proof of Honeywell’s knowledge of risks of catastrophe. Thus, for example, there is mixed significance in Honeywell’s acknowledgement that the control knob was a “critical safety feature.” As we have said, that evidence does little to prove a specific product defect. The evidence is highly probative, however, of knowledge of risks such that warnings and instructions must be sufficient to make the product safe for use. There is mixed significance in Honeywell’s acknowledgement that the control knob was a critical safety feature. As we have said, that evidence does little to prove a specific product defect. The evidence is highly probative, however, of knowledge of risks such that warnings and instructions must be sufficient to make the product safe for use. Honeywell took the position at trial that this case did not involve issues of warnings. Respondent offered no testimony to contradict appellant’s evidence on the deficiency of warnings. Again on appeal, respondent failed to address the warnings issue. Contrasting with Honeywell’s approach to the case, we conclude that the jury could not properly disregard the evidence of negligence in the withholding of warnings and instructions related to a known faulty part of the V5130 control. We are mindful of the degree of deference to be given to the fact finder. We more commonly confront the question in negligence eases where summary judgment or directed verdict favors the defendant. The issue here is the same, however, no more and no less significant. Where evidence overwhelmingly preponderates in fa *866 vor of a claimant, even if there is some evidence in favor of the defendant, fault must be found without deference to a jury decision. Frey v. Montgomery Ward & Co., 258 N.W.2d 782, 786-87 (Minn.1977) (affirming a directed verdict of negligent failure to warn on use of a space heater). See J.N. Sullivan & Associates, Inc. v. F.D. Chapman Construction Co., 304 Minn. 334, 336-38, 231 N.W.2d 87, 89-90 (1975) (affirming directed verdict on negligence in construction of a sewer project, notwithstanding some evidence for plaintiff). 3. Open and obvious danger ? Responding to posttrial motions for appellant, based singularly on the issue of warnings and instructions, the trial court did not rest solely on the analysis of Honeywell’s fault “in and of itself.” The court added that the jury was entitled to conclude that warnings were not needed because “the danger of removing the knob was open and obvious.” That was an erroneous analysis of the circumstances here. There is no duty to warn if the user knows or should know of the potential danger. Minneapolis Society of Fine Arts v. Parker-Klein Associates, Inc., 354 N.W.2d 816, 821 (Minn.1984) (verdict on negligent failure to warn unjustified where “overwhelming” evidence that purchaser’s architects should have known a product could be unsuitable for exterior use and that the architects were in fact advised to consult with the manufacturer before making exterior installation of the product). Special awareness of a risk may explain a finding of no fault. See Dishington v. A. W. Kuettel & Sons, Inc., 255 Minn. 325, 332, 333, 96 N.W.2d 684, 689, 690 (1959) (a hospital maintenance worker was properly denied a negligence verdict where he was “well aware” of an “obviously” dangerous risk in a construction area and actually knew of the risk); Magnuson, 285 Minn, at 36, 42, 171 N.W.2d at 204, 207 (the general verdict against an injured snowmobile owner may stand where he was “well aware” of a dangerous protrusion on his machine, because awareness is associated with the question of the manufacturer’s fault); cf. Magnuson, 285 Minn, at 48, 171 N.W.2d at 211 (Rogosheske, J., concurring specially) (suggesting that the circumstances of a particular user are more aptly considered as a matter of comparative fault). Ordinary and foreseeable use of the V5130 control included use by visitors or new owners or tenants, persons not familiar with the condition of the unit. There is insufficient evidence here to support the view that the problems with the control unit were open and evident to an ordinary user or especially open and obvious to Michael Balder. Moreover, there is no evidence to support the trial court’s observation that danger in “removing the knob” was open and obvious to ordinary users. In fact, there is no evidence that Michael Balder ever operated the control without the knob; Balder testified that when he first checked the unit he used the knob to try turning the valve. It is important to note that the injured user here was Michael Balder, not his mother or one of the people who had worked on the water heater earlier and knew its condition. It is evident that Balder saw the control on the day he was injured and he said he initially attempted to turn the shaft by using the deteriorated control knob. His comments after being injured may indicate that he touched controls on the heater or the gas line on a second visit to the basement where the explosion occurred. There is not evidence, however, that he had prior knowledge about leakage of the valve or about the advice of repair workers to replace the unit. He could see a substance on the shaft, but there is not additional evidence that he was well aware that this material posed a danger of major leakage and explosion. Cf. Magnuson, 285 Minn. 32, 171 N.W.2d 201; Dishington, 255 Minn. 325, 96 N.W.2d 684. Balder’s mother had smelled gas, but there is not additional evidence showing that Balder knew of a gas leak or that the evidence of leakage was substantial. The jury properly found Balder partially at fault, but allocated to him only 20 percent of fault for his injuries. *867 The findings of the jury do not specifically address the view that Honeywell is fully absolved of fault for negligence because of Balder’s awareness of risks. The jury instruction on Honeywell’s duty to warn dealt with the foreseeability of risks, but the jury was not instructed on the limits of Honeywell’s duty based on hazards that are open and obvious to Michael Balder. Rather, the circumstances of Balder were appropriately treated by the trial court and the parties as one of comparative fault, the approach suggested by Justice Rogosh-eske’s concurring opinion in Magnuson, 285 Minn, at 48, 171 N.W.2d at 211. Based on the preceding discussion, we conclude that Honeywell’s negligent fault cannot be wholly discarded based on the view that hazards with its unit were sufficiently open and obvious. It is apparent, as appellant acknowledges, that the fault of others might lead to a verdict that Honeywell’s negligence did not cause Balder’s injuries. It is improper, however, on the evidence here, to fully absolve Honeywell from fault in connection with the dangerous condition of Josephine Pirkl’s water heater. Given evidence showing limited fault of the plaintiff, together with almost undisputed evidence that Honeywell knew for over two decades that a part of its product would often fail, and knew serious risks were associated with use of the product after a failure, Honeywell is among others who share fault in the circumstances here. Absent a finding that its fault is to be disregarded because of some superseding cause of Michael Balder’s injuries, it is among those liable for those injuries. 4. Causation Finally, we have considered the question of whether the fault of others may have constituted an intervening cause as a matter of law. The evidence does not permit that conclusion. There is evidence sufficient to support a verdict that the inadequate warnings by Honeywell contributed independently to produce the explosion that seriously injured appellant. See Medved v. Doolittle, 220 Minn. 352, 357, 19 N.W.2d 788, 791 (1945). What we have said as to product misuse while discussing Honeywell’s liability for defects shows the evidence supporting the jury verdict for Honeywell on that issue, but it is a much different thing to say that the misuse superseded fault of Honeywell as a matter of law. More particularly, on the evidence here the jury could find that deficient warnings could have brought about the misuse of the control by Pirkl and others, or could have produced Balder’s injuries even without that misuse. See Carlson v. Fredsall, 228 Minn. 461, 467, 37 N.W.2d 744, 748 (1949) (mere fact that one party was more negligent than another does not release the less negligent party from liability). On the evidence here, the jury faces a panorama of fault on the part of Honeywell, Inc., Josephine Pirkl, and Tom Haley, and there is ample evidence to support a verdict that does not disregard Honeywell’s fault as one of the causes of Michael Balder’s injuries. 5. Balder’s fault The jury found that Michael Balder himself was 20% at fault in causing the accident. Appellant claims this finding was manifestly contrary to the evidence since he had no knowledge of any defects and checked only to see if the pilot light was burning. We disagree. While Balder’s awareness was not such that Honeywell could be absolved of any fault, we are satisfied the evidence permitted a finding of his comparative fault. DECISION The evidence supports a finding that Honeywell, Inc. has no fault in this case on a theory of defective design. However, the evidence does not permit a verdict that respondent Honeywell, Inc. was not negligent in withholding warnings and instructions on malfunctions of the Y5130 control unit. There is sufficient evidence to support a finding of appellant’s comparative fault. *868 Appellants are entitled to a new trial to determine the merits of their claim that Honeywell’s negligent failure to warn and instruct caused their damages; in the event of a finding that Honeywell’s neglect was causative, there must be a reallocation of fault among Honeywell, Haley, Pirkl, and Balder. In sum, upon the new trial of this ease the jury must decide the issue of cause, but with a directed determination of Honeywell’s negligence. Reversed and remanded. Footnotes [1] The jury also responded negatively to this question related to a claim of fault under the Consumer Product Safety Act: "Before July 23, 1980, did Honeywell, Inc. obtain information reasonably supporting the conclusion that the V5130 valve contained a defect which could create a substantial product hazard?" That verdict has not been made an issue in post-trial or appellate proceedings. [2] We are not at liberty to speculate about the jury’s thoughts on an unanswered special interrogatory, the question whether negligence of Honeywell, Inc. directly caused Michael Balder's injuries. Thus, it would be wrong for us to decide the case on an episode of record that we find troublesome. Before the jury deliberated, it heard a proper instruction on superseding cause, related to seven special interrogatories dealing with direct cause. During the course of its deliberations, the jury delivered a question described by the trial court as "concerning the relationship between superseding cause and liability." After a telephone conference with counsel, the court repeated to the jury its earlier instructions on superseding cause. Twice, however, during this contact with the jury, the trial court gave an added instruction that "there is no liability without direct cause.” We decline to speculate whether the added instruction may have prompted confusion between "liability,” a concept not included among special verdict interrogatories, and the various questions of fault posed to the jurors along with separate questions on direct cause. This speculation is also stymied as a matter of fact by the question and comment of the only juror who addressed the court at the time, a juror who was one of the ten who later joined in a verdict; this juror asked that the repeated instructions be reread, and he showed singular and appropriate interest in the relationship between superseding cause and di *864 rect cause. The jury returned a verdict on the second day following this contact with the trial judge.