Honeywell International, Inc. v. Hamilton Sundstrand Corp.

378 F. Supp. 2d 459 (2005)

HONEYWELL INTERNATIONAL, INC. and Honeywell Intellectual Properties, Inc., Plaintiffs,
v.
HAMILTON SUNDSTRAND CORP., Defendant.

No. Civ.A.03-1153 GMS.

United States District Court, D. Delaware.

July 5, 2005.

Thomas C. Grimm, Karen Jacobs Louden, and Leslie A. Polizoti, of Morris, Nichols, Arsht & Tunnell, Wilmington, Delaware, and Jonathan F. Putnam, Lee Ann Stevenson, Vickie Reznik, and Michael Strapp, of Kirkland & Ellis LLP, New York, New York, for Plaintiffs.

Richard D. Kirk, of The Bayard Firm, Wilmington, Delaware, and Mark L. Levine, Chris J. Lind, and Brian C. Swanson, of Bartlit, Beck, Herman, Palenchar & Scott, Chicago, Illinois, for Defendant.

OPINION

SLEET, District Judge.

I. INTRODUCTION

The above-captioned action is a patent infringement suit in which plaintiffs Honeywell International, Inc. and Honeywell Intellectual Properties, Inc. (collectively, "Honeywell"), the owners of U.S. Patent No. 6,035,626 ("the '626 patent"), accuse defendant Hamilton Sundstrand Corp. ("HSC") of infringing several claims of the '626 patent. The invention of the '626 patent "relates to a control system for regulating the amount of torque applied by a starter/generator to a gas turbine engine during startup." '626 patent, col. 1, ll. 9-11. The accused device in this case is HSC's APS 2300 APU (Auxiliary Power Unit) with a starter generator, "which is used in aircraft to provide compressed air for starting of the main engines and control of the environmental conditions within the aircraft as well as electric power." (D.I. 158, Ex. 2 at 5.) Presently before the court are two outstanding pre-trial matters: (1) HSC's request to preclude Honeywell from presenting a damages calculation based on sales projections of the accused product that did not exist at the time of the hypothetical negotiation (D.I.156); and (2) HSC's motion on prosecution history estoppel (D.I.158) The court will address each matter in turn.

II. DISCUSSION

A. HSC's Request to Preclude Honeywell From Presenting a Damages Calculation Based on Sales Projections of The Accused Product That Did Not Exist at The Time of The Hypothetical Negotiation

Honeywell and HSC both manufacture a wide array of equipment for the aerospace industry. As the two giants in that industry, they are fierce competitors and frequent fliers in this court. Among the customers for whose business they compete is Embraer, a large regional jet manufacturer in Brazil. In 1999, HSC was awarded a contract to supply Embraer with the accused APUs for approximately 20 years. (D.I. 156, Attach. B at 11.) Roughly contemporaneous with that award, each company prepared its own sales projections for the HSC-Embraer contract. HSC projected in 1998 that it would sell 609 APUs to Embraer by 2017. (Id., Attach.E.) Similarly, in 1999, Honeywell projected that HSC would sell between 348 and 600 APUs to Embraer through the life of the contract. (Id., Attach. F; Attach. D at 250:6-12.) However, because the APU was still in development, the first sales were not scheduled until 2003. (Id., Attach.E.) In the interim, on March 14, 2000, the '626 patent issued.

According to Honeywell, the unforeseen and tragic events of September 11, 2001, caused an unexpected increase in the "importance of the large regional jet market." (D.I. 159 at 2.) Consequently, Embraer's demand for HSC's APUs increased dramatically. The most recent projections, calculated in 2004 and 2005, predict that HSC will sell 1,001 APUs to Embraer by 2017. Honeywell estimates that these new projections increase the damages to which it is entitled from Embraer sales alone by roughly $11 million. (D.I. 156, Attach. C at 6-7.) More specifically, Honeywell contends that if it had engaged in a negotiation with HSC when infringement began, i.e., March 14, 2000, HSC would have accepted a 10% royalty rate to license the '626 patent until its expiration.^[1] (Id., Attach. B at 24-25, Attach. C, Ex. 4a.) Furthermore, Honeywell asserts, HSC would have agreed to pay for the license as a lump sum, discounted to net present value as of March 2000.(Id.) Using the 1998-99 projections as the royalty base yields roughly $17 million in damages flowing from the HSC-Embraer contract. On the other hand, using the 2004-05 projections (in addition to actual sales thus far) as a royalty base yields roughly $28 million in damages. (Id., Attach C at 6-7.) Not surprisingly, the parties disagree on whether the more recent projections may be used as a royalty base to calculate Honeywell's damages.

HSC objects to Honeywell's use of the 2004-05 projections because they would not have been available to hypothetical negotiators in March 2000. According to HSC, the case law has universally rejected the factfinder's consideration of post-negotiation sales projections in the context of the hypothetical negotiation construct, particularly where pre-negotiation projections exist. (D.I. 156 at 2-3.) Honeywell answers that HSC's argument ignores the "book of wisdom" concept set forth by the Supreme Court over seventy years ago in Sinclair Refining Co. v. Jenkins Petroleum Process Co., 289 U.S. 689, 53 S. Ct. 736, 77 L. Ed. 1449 (1933), and adopted by the Federal Circuit in Fromson v. Western Litho Plate and Supply Co., 853 F.2d 1568 (Fed.Cir.1988). (D.I. 159 at 3.) According to Honeywell, post-negotiation projections are indeed relevant under this precedent. (Id. at 4.) Honeywell further points out that HSC's damages theory also relies on post-negotiation information, namely, actual sales data from 2003 and 2004. (Id. at 5.) In reply, HSC argues that the "book of wisdom" cases stand "for only the undisputed proposition that some post-negotiation evidence, such as actual sales data, may be considered in some circumstances. They do not support a blanket rule that all post-negotiation evidence is always relevant. To the contrary, the Federal Circuit has made clear that if sales projections existed at the time of the hypothetical negotiation, post-negotiation projections may not be used to determine the amount of a hypothetical up-front lump-sum royalty payment." (D.I. 163 at 1 (emphasis in original).) Thus, the question presented to the court is whether sales projections that were unavailable at the time infringement began may be used as a royalty base to calculate damages.

Upon a finding of infringement, the plaintiff is entitled to "damages adequate to compensate for the infringement, but in no event less than a reasonable royalty for the use made of the invention by the infringer." 35 U.S.C. § 284 (2001). "A reasonable royalty calculation envisions and ascertains the results of a hypothetical negotiation between the patentee and the infringer at a time before the infringing activity began." Integra Lifesciences I, Ltd. v. Merck KGaA, 331 F.3d 860, 869 (Fed.Cir.2003). Therefore, "[t]he first step in a reasonable royalty calculation is to ascertain the date on which the hypothetical negotiation in advance of infringement would have occurred." Id. at 870. It is important to note, however, that the ascertainment of this date does not rigidly foreclose the factfinder from considering subsequent events. To enforce such rigidity would be to ignore a limitation inherent to the hypothetical negotiation method. "In a normal negotiation, the potential licensee has three basic choices: forego all use of the invention; pay an agreed royalty; infringe the patent and risk litigation." Fromson, 853 F.2d at 1576. In a hypothetical negotiation, however, the factfinder "presumes that the licensee has made the second choice, when in fact it made the third." Id. The unwanted effect is that the factfinder may be tempted "to pretend that the infringement never happened." TWM Mfg. Co. v. Dura Corp., 789 F.2d 895, 900 (Fed.Cir.1986). Thus, the hypothetical negotiation approach "must be flexibly applied as a device in the aid of justice." Id. (internal quotations omitted).

Justice Cardozo's opinion for the Supreme Court in Sinclair Refining lends further insight on the importance of maintaining a flexible approach to the hypothetical negotiation. In that case, the plaintiff loaned to the defendant "an experimental still for cracking petroleum oils to produce gasoline." 289 U.S. at 690, 53 S. Ct. 736. As part of the consideration for the loan, the defendant agreed to assign any improvements it made on the still to the plaintiff. Id. Subsequently, the defendant successfully applied for a patent to which the plaintiff later claimed an interest pursuant to the loan agreement. Id. at 690-91, 53 S. Ct. 736. The defendant refused to assign the patent to the plaintiff, effectively forcing the plaintiff to sue the defendant for damages under a breach of contract theory. Id. at 691, 53 S. Ct. 736. During discovery, the plaintiff was denied information "as to the number of crackling stills constructed by the defendant under [its] patent, as to the extent and time of operation, and as to the amount of gasoline and other petroleum products yielded thereby." Id. The district court sided with the defendant in part because "the value of the patent[] has no relation to the sales of the patented device, and that evidence of such sales would be inadmissible if offered." Id. at 691-92, 53 S. Ct. 736. In other words, the district court held that since damages for breach are assessed as of the time of the breach, the value of the patent at that time was unrelated to any greater value it later obtained. Thus, the district court deemed evidence of post-breach commercial success to be irrelevant and inadmissible.

The Supreme Court disagreed. Justice Cardozo explained that, because a "patent is a thing unique," "[t]here can be no contemporaneous sales to express the market value of an invention that derives from its novelty its patentable quality." Id. at 697, 53 S. Ct. 736. Nevertheless, "the absence of market value does not mean that the offender shall go quit of liability altogether." Id. If trial is shortly after the breach, "the only evidence [of the patent's value] may be that supplied by testimony of experts as to the state of the art, the character of the improvement, and the probable increase of efficiency or saving of expense." Id. at 698, 53 S. Ct. 736. However, if trial is years after the breach, "[e]xperience is then available to correct uncertain prophecy. Here is a book of wisdom that courts may not neglect. We find no rule of law that sets a clasp upon its pages, and forbids us to look within." Id. Justice Cardozo further explained:

An imaginary bid by an imaginary buyer, acting upon the information available at the moment of the breach, is not the limit of recovery where the subject of the bargain is an undeveloped patent. Information at such a time might be so scanty and imperfect that the offer would be nominal. The promisee of the patent has less than fair compensation if the criterion of the value is the price that he would have received if he had disposed of it at once, irrespective of the value that would have been uncovered if he had kept it as his own.

Id. at 699, 53 S. Ct. 736. Thus, "[t]o correct uncertain prophecies in such circumstances is not to charge the offender with elements of value non-existent at the time of his offense. It is to bring out and expose to light the elements of value that were there from the beginning." Id. at 698, 53 S. Ct. 736.

Over fifty-five years after Sinclair Refining, the Federal Circuit in Fromson adopted the Supreme Court's rationale for flexibility — the "book of wisdom" — and applied it to the hypothetical negotiation method of calculating damages under § 284:

The [hypothetical negotiation] methodology encompasses fantasy and flexibility; fantasy because it requires a court to imagine what warring parties would have agreed to as willing negotiators; flexibility because it speaks of negotiations as of the time infringement began, yet permits and often requires a court to look to events and facts that occurred thereafter and that could not have been known to or predicted by the hypothesized negotiators.

853 F.2d at 1575. Indeed, the flexibility offered by the "book of wisdom" is as important in the context patent law as it is in the context of contract law because it discourages infringement. If the hypothetical negotiation could not be informed by post-negotiation information, then prospective infringers might perceive "that blatant, blind appropriation of inventions ... is the profitable, can't-lose course." Id. at 1575. In other words, prospective infringers might rationally conclude that, at worst, upon a finding of infringement "a license can be compelled, probably at the same royalty that would have been paid if the patentee's rights had been respected at the outset." Id. at 1574. Moreover, the "book of wisdom" prevents the hypothetical negotiation method from determining a reasonable royalty at a point in time before the patent has proven its worth. In doing so, the "book of wisdom" concept protects the quid pro quo arrangement underlying patent law by ensuring that the patentee will be adequately compensated for infringement. See id. at 1575.

Importantly, the plain language of the statute clearly supports the flexible "book of wisdom" approach by ensuring the recovery of "damages adequate to compensate for the infringement, but in no event less than a reasonable royalty for the use made of the invention by the infringer." § 284 (emphasis added). It is axiomatic that "the use made" will not be known until after infringement. And since the date of the hypothetical negotiation is "before the infringing activity began," Integra, 331 F.3d at 869 (emphasis added), information not available as of that date must necessarily be considered by the factfinder. Furthermore, § 284 "does not mandate how the district court must compute [the reasonable royalty], only that the figure compensate for the infringement." TWM Mfg., 789 F.2d at 899. Thus, the "methodology of assessing and computing damages under [the statute] is within the sound discretion of the district court." Id. at 898. The only discernable limitations on that discretion are: (1) that the damages be no less than a "reasonable royalty for the use made of the invention by the infringer," § 284; (2) that the damages adequately "compensate for the infringement," id.; see also Integra, 331 F.3d at 870 ("Royalties, like lost profits, are compensatory damages, not punitive."); and (3) that the reasonable royalty be based on "sound economic and factual predicates," Riles v. Shell Exploration and Prod. Co., 298 F.3d 1302, 1311 (Fed.Cir.2002).

Given this flexibility, the court would clearly be acting within its discretionary limitations by permitting Honeywell to calculate damages using the 2004-05 sales projections as a royalty base. These most recent projections may approximate what will end up being "the use made of the invention" by HSC more closely than do the 1998-99 projections. If the factfinder were to accept that proposition, then using the 2004-05 projections as a royalty base would be an economically and factually sound basis on which to adequately "compensate for the infringement" without acting punitively.

HSC makes several arguments in opposition. First, HSC contends that using post-negotiation projections to calculate damages is impermissible under the Federal Circuit's decision in Interactive Pictures Corp. v. Infinite Pictures, Inc., 274 F.3d 1371, 1384 (Fed.Cir.2001). In that case, the plaintiff presented evidence at trial of sales projections taken from the defendant's own business plan, which was created just two months before the date of the hypothetical negotiation. Id. at 1384. The defendant objected that the "projection of [its] future sales was speculative, as it was based on an outdated business plan and its optimistic assumptions of future revenue growth." Id. The Federal Circuit disagreed:

[The] business plan and its projections for future sales were prepared by [the defendant] two months before infringement began. Thus, rather than being outdated for purposes of the hypothetical negotiation, those projections would have been available to [the defendant] at the time of the hypothetical negotiation. The fact that [the defendant] did not subsequently meet those projections is irrelevant to [the defendant's] state of mind at the time of the hypothetical negotiation. Nor does [the defendant's] subsequent failure to meet its projections imply that they were grossly excessive or based only on speculation and guesswork. Instead, [the defendant's] subsequent failure to meet its projections may simply illustrate the "element of approximation and uncertainty" inherent in future projections.

Id. at 1385 (emphasis added). Then, in rejecting the defendant's argument that future projections "must later bear a close relation to actual sales revenue," the court explained that "[s]uch a proposition would essentially eviscerate the rule that recognizes sales expectations at the time when infringement begins as a basis for a royalty base as opposed to an after-the-fact counting of actual sales." Id. (emphasis added). Thus, Interactive Pictures contains language that seems to erect a barrier between pre-negotiation information and post-negotiation information.

In spite of the Federal Circuit's language, it must be reiterated that "[c]ases should not be cited for mere words. What counts is what the court did in a cited case." Fromson, 853 F.2d at 1578. In Interactive Pictures, the court upheld a jury's damage award that was based on the defendant's overly-optimistic sales projections created shortly before the hypothetical negotiation date. Had the negotiation actually taken place, those projections clearly would have been a part of the defendant's calculus. As such, the royalty was based on "sound economic and factual predicates." Riles, 298 F.3d at 1311. Furthermore, because the royalty was based on sales projections that exceeded the defendant's actual sales, the plaintiff was awarded an amount in excess of a "reasonable royalty for the use made of the invention by the infringer." § 284 (emphasis added). Moreover, the damages were not likely punitive because the jury awarded an amount within the lower of two ranges proposed by the plaintiff's expert. Interactive Pictures, 274 F.3d at 1384. Therefore, although language in Interactive Pictures suggests the Circuit's disapproval of using post-negotiation information to construct the hypothetical negotiation, nothing "the court did" prohibits the discretionary use of such information in future cases. Rather, the most that can be inferred from Interactive Pictures is that it is not an abuse of discretion in those circumstances for a district court to permit the jury to disregard post-negotiation information.

HSC also cites Riles in support of its position. In Riles, the plaintiff was the owner of a patent relating to a method of securing a fixed, offshore oil-drilling platform to the ocean floor without using something known as a "mud mat." 298 F.3d at 1305-06. The jury found that the defendant's oil-drilling platform infringed the plaintiff's patent, and awarded damages in an amount corresponding to a percentage of the platform's value at the time of trial, plus a percentage of the first year of revenue generated by the platform. See id. at 1311-1313. In remanding the damages portion of the trial back to the district court for reconsideration, the Federal Circuit explained:

A reasonable royalty determination for purposes of making a damages evaluation must relate to the time infringement occurred, and not be an after-the-fact assessment. Clearly, [the plaintiff's damages expert's] models did not reflect what royalty rate a hypothetical negotiation between [the defendant and the plaintiff] would have yielded at the time the infringement began. Instead, the models reflected [the plaintiff's damages expert's] assessment of the worth of [the defendant's] oil rig at the time of the trial. [The plaintiff] did not provide any evidence or testimony to show that [its damages expert's] models reflected what the parties might have agreed to, at any time, particularly at the time the infringement began.

Id. at 1313 (citations omitted) (emphasis added). As in Interactive Pictures, the language in Riles also seems to create a barrier between pre-negotiation information and post-negotiation information.

But again, looking to what "the court did," it is clear that the case was remanded because the district court impermissibly allowed a damages theory that was not based on "sound economic and factual predicates." Id. at 1311. Specifically, the plaintiff's damages expert based his calculations on the assumption that the plaintiff would have been able to enjoin the defendant from using its platform altogether in the event that negotiations would have proven unsuccessful. Id. at 1311-12. However, his assumption was not supported by the record. Id. at 1312. The plaintiff's expert also failed to quantify the value of the patented feature relative to the value of the entire platform and the revenue it generated, failed to account for non-infringing alternatives, and failed to account for the its own past licensing practices. Id. at 1312-13. Thus, in the overall context of the case, the court's disavowal of "after-the-fact assessments" was merely an observation that the plaintiff's overly-simplistic damages theory was based on the present market value of the platform as a whole, and little else.

Finally, HSC points to Integra, where the plaintiff charged the defendant with infringing a patent involving pharmaceutical technology from 1994 through 1998. 331 F.3d at 870. The jury awarded damages to the plaintiff based both on the defendant's "1995 expectations of obtaining FDA approval" related to the patent, and on the consideration paid in other licenses executed by the defendant with a party unrelated to the litigation in 1990. Id. The Federal Circuit remanded the case back to the district court for several reasons. First, the evidence was ambiguous as to whether infringement began in 1994 or 1995. If it began in 1994, then the date of the hypothetical negotiation must also have been in 1994, and the defendant would not yet have expected FDA approval. Thus, "an earlier date [would] change the risks and expectations of the parties." Id. Second, the court explained that the record did "not show that the [1990] licenses occurred under similar scientific or economic circumstances to the hypothetical ... license" proffered by the plaintiff. Id. at 871. This was particularly true if 1994 was the proper hypothetical negotiation date because the parties' ability to assess risk as of the earlier date may have been dramatically different than it was for the parties to the 1990 licenses. Thus, the 1990 licenses were not necessarily probative in the case at bar. Id. Finally, the court instructed the district court to consider on remand numerous other relevant factors that did not appear to have been considered the first time, including (1) a more analogous 1995 license agreement involving similar technology, (2) the fact that the damages totaled approximately 75% of the price the plaintiff paid in 1996 to purchase the entire company that owned, among many other things, the patent at issue, (3) the point in the drug development process at which the patented technology would have been used by the defendant, and (4) the number of other patent licenses necessary to develop the drugs at issue. Id. at 871-72.

Compared to Interactive Pictures and Riles, Integra appears to erect a much more solid wall between pre-negotiation information and post-negotiation information. Clearly, the court placed very heavy emphasis on the information available to the parties as of the hypothetical negotiation date. On the other hand, the court also directed the district court to consider information from 1995 and 1996, which obviously occurred after the hypothetical negotiation date. Thus, Integra appears to acknowledge that while some post-negotiation information can be considered for purposes of the hypothetical negotiation (i.e., the analogous 1995 license, the 1996 purchase), other information cannot be considered (i.e., the 1995 expectations of obtaining FDA approval). One might argue that, under Integra, post-negotiation information can only be considered if it is used to test the reasonableness of the assumptions about the hypothetical negotiators' states of mind. Cf. Franconia Assoc. v. United States, 61 Fed. Cl. 718, 766 n. 87 (Fed.Cl.2004) ("subsequent events may be used to test the reasonableness of assumptions used in projecting damages"). However, that is an elusive distinction because the ability of post-negotiation information to test the reasonableness of pre-negotiation assumptions depends, at least in part, on the foreseeability of intervening events. For example, no one would argue that the hypothetical negotiators in this case could have foreseen the events of 9/11, but perhaps the negotiators could have foreseen a general decline in the commercial airline industry. If so, then post-negotiation information might be an accurate test of the reasonableness of the assumptions. If not, then the information might not provide an accurate test. Either way, measuring the foreseeability of intervening events is not a proper role for the court.

HSC proposes that Integra (as well as Interactive Pictures, and perhaps Riles) stands for the proposition that post-negotiation sales data can be considered for purposes of the hypothetical negotiation, but that post-negotiation sales projections cannot be considered. For support, HSC points to LinkCo, Inc. v. Fujitsu Ltd., 232 F. Supp. 2d 182, 188-90 (S.D.N.Y.2002), in which the Southern District of New York drew precisely that distinction. Although LinkCo was a trade secret case, the court analyzed the same general body of case law discussed here. The court concluded that because post-negotiation "estimates do not reflect the parties['] perceived value of the trade secret during the negotiation," "sales projections are only relevant in a reasonable royalty calculation when they are available before the time of the misappropriation and would have been considered by the parties." Id. at 189. On the other hand, according to the court, post-negotiation sales data is relevant because, in certain cases, no pre-negotiation projections would have been available to the negotiators. Id. at 190. With all due respect to the LinkCo court, its distinction is unpersuasive because it fails to explain what it is about the nature of post-negotiation projections that is less reliable than the nature of post-negotiation sales data. Furthermore, the court's distinction fails to account for the situation in which only post-negotiation projections are available. Such a situation could easily arise where a defendant has made infringing devices but has not yet made any sales at the time of trial. In that event, under LinkCo, potentially crucial information would be withheld from the factfinder for no discernable reason. Thus, the court declines to follow LinkCo.

Nevertheless, the court is unable to draw a meaningful distinction between the facts in Integra and the facts in the present case. In Integra, the Federal Circuit reversed the district court partially on the basis that it may have permitted the factfinder to consider post-negotiation information that would have "change[d] the risks and expectations of the parties." 331 F.3d at 870. The 2004-05 sales projections fall squarely in that category. On the other hand, under Fromson, post-negotiation sales projections also fall squarely in the category of "events and facts that occurred thereafter and that could not have been known to or predicted by the hypothesized negotiators." 853 F.2d at 1575. Therefore, the court is presented with two cases, both of which are binding, that dictate opposite results.

HSC suggests that the court should resolve the conflict by following the more recent case, Integra. (D.I. 163 at 2 n. 2.) The court disagrees. The recent vintage of a case has no bearing on how well it was reasoned. In this case, the court is persuaded that the result dictated by Fromson is the most sensible. First, it promotes flexibility in damage calculations by not erecting an unnecessarily rigid barrier to relevant post-negotiation information. Second, it discourages infringement by placing the risk of success on the infringer. Third, it protects the quid pro quo underlying patent law by preventing a premature valuation of the patent. Finally, it permits a damage award more in keeping with the plain language of § 284 by adequately compensating the plaintiff for the "the use made of the invention" by the defendant. This list is probably not exhaustive, but is, in the court's view, sufficient to demonstrate that Fromson should control in this case.

As to HSC's remaining argument that the 2004-05 projections should not be considered because they conflict with the 1998-99 projections, that is an argument best reserved for the jury. Therefore, the court will deny HSC's request and permit Honeywell to use the 2004-05 sales projections for its royalty base.

B. Prosecution History Estoppel

The second motion presently before the court is HSC's motion on prosecution history estoppel. HSC argues that certain limitations in claims 1 and 3 of the '626 patent are subject to prosecution history estoppel and, therefore, that Honeywell is barred from asserting the doctrine of equivalents with respect to those limitations. In particular, HSC contends that Honeywell should be so barred with respect the claim limitations italicized below:

1. A method of starting a gas turbine engine along a preselected start schedule comprising the steps of:

providing a starter/generator for driving said engine;

sensing engine speed and generating a first signal thereof;

sensing elapsed time of the engine start;

producing a second signal, in response to said elapsed time, indicative of a predetermined speed for said engine;

combining said first and second signals to form an error signal; and

adjusting the torque output of said starter/generator in response to said error signal so that said starter/generator only delivers the torque necessary to keep the engine accelerating along the schedule.

Id., col. 5, ll. 26-40 (emphasis added); and

3. A system for starting a gas turbine engine along a preselected start schedule, comprising:

a starter/generator drivingly coupled to said engine;

a sensor that senses the speed of said engine;

a first signal produced by said sensor;

a timer for determining elapsed time from the engine start;

a second signal produced by the system, corresponding to the elapsed time determined by said timer, that is indicative of a predetermined speed for the engine;

an error signal formed from said first signal and said second signal; and

a controller that adjusts the current flow from the starter/generator to adjust the torque output of said starter/generator based on said error signal, so that said starter/generator only delivers the torque necessary to keep the engine accelerating along the schedule.

Id., col. 5, l. 44 — col. 6, l. 11 (emphasis added).

"The doctrine of equivalents prevents an accused infringer from avoiding liability for infringement by changing only minor or insubstantial details of a claimed invention while retaining the invention's essential identity." Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 234 F.3d 558, 564 (Fed.Cir.2000) (Festo I), rev'd on other grounds, Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U.S. 722, 122 S. Ct. 1831, 152 L. Ed. 2d 944 (2002) (Festo II). "When, however, the patentee originally claimed the subject matter alleged to infringe but then narrowed the claim in response to a rejection, he may not argue that the surrendered territory comprised unforeseen subject matter that should be deemed equivalent to the literal claims of the issued patent." Festo II, 535 U.S. at 733-34, 122 S. Ct. 1831. Known as prosecution history estoppel or file wrapper estoppel, it is a "rule of patent construction that ensures that claims are interpreted by reference to those that have been cancelled or rejected." Id. at 733, 122 S. Ct. 1831 (citations omitted). "The logic of prosecution history estoppel is that the patentee, during prosecution, has created a record that fairly notifies the public that the patentee has surrendered the right to claim particular matter as within the reach of the patent." Festo I, 234 F.3d at 564-65. "Questions relating to the application and scope of prosecution history estoppel ... fall within the exclusive province of the court." Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 344 F.3d 1359, 1369 (Fed.Cir.2003) (Festo III).

For prosecution history estoppel to apply, it must first be established that the amendment in question narrowed the literal scope of the claim. Id. at 1366. Once this is done, "the second question is whether the reason for that amendment was a substantial one relating to patentability." Festo III, 344 F.3d at 1366. If that question is answered in the affirmative, it is presumed that "the patentee has surrendered all territory between the original claim limitation and the amended claim limitation." Id. at 1367. "[I]f the patentee fails to rebut the Festo presumption, then prosecution history estoppel bars the patentee from relying on the doctrine of equivalents for the accused element." Id. If, however, "the patentee successfully rebuts the presumption, then prosecution history estoppel does not apply and the question whether the accused element is in fact equivalent to the limitation at issue is reached on the merits." Id.

1. Claim 1

In the original application, claim 8 (which became claim 1 in the issued patent) read as follows:

8. A method of starting a gas turbine engine through a starter/generator comprising the steps of:

sensing engine speed and generating a first signal thereof;

sensing elapsed time of the engine start;

producing a second signal, in response to said elapsed time, indicative of a predetermined speed for said engine;

combining said first and second signals to form an error signal; and

adjusting the torque to said engine in response to said error signal so that said engine starts along a predetermined schedule.

(D.I. 158, Ex. 1 at HNY020469.) However, the examiner rejected claim 8 in a September 1993 office action as anticipated by prior-art references Jesrai, LEMD, and Cronin. (Id. at HNY020488.)

Honeywell responded that Jesrai and LEMD were distinguishable because both of those references taught "the use of an air turbine starter," not "a starter/generator to drive an engine during a start." (Id. at HNY020494.) Furthermore, in the invention of claim 8, "it is the output torque of the starter/generator that accelerates the engine during startup and ... the torque output is adjustable in response to a signal indicative of the engine's speed and/or the elapsed time of the start." (Id. at HNY020494.) Thus, Honeywell distinguished Cronin on the basis that, although it taught the use of a starter/generator, it failed to disclose "the means, if any, by which the torque output of the motor/generator is controlled."^[2] (Id. at HNY020494-95.) Honeywell also amended the claim as follows:

Claim 8 (Amended). A method of starting a gas turbine engine [through] drivingly coupled to a starter/generator comprising the steps of:

sensing engine speed and generating a first signal thereof;

sensing elapsed time of the engine start;

producing a second signal, in response to said elapsed time, indicative of a predetermined speed for said engine;

combining said first and second signals to form an error signal; and

adjusting the torque output of said starter/generator [to said engine] in response to said error signal so that said engine starts along a predetermined schedule.^[3]

(Id. at HNY020493.)

The examiner considered Honeywell's arguments, but deemed them moot in light of his determination that claim 8 was an obvious combination of LEMD and Cronin. (Id. at HNY020501.) In a February 1994 office action, the examiner explained that "LEMD teaches a system for accelerating a gas turbine engine" comprising: (a) a means for sensing the speed of the engine; (b) a torque regulator responsive to the speed sensing means; and (c) a "predetermined schedule of engine acceleration." (Id.) Although LEMD lacked "explicit teaching of a motor ... also being able to function as a generator," the examiner continued, it "would have been obvious to one of ordinary skill in the art to have incorporated the starter motor/generator of Cronin into the invention of LEMD for the benefit of saving the weight of an additional generator in the aircraft, a primary concern in airplane design." (Id.)

In June 1994, Honeywell responded to the examiner's obviousness rejection by amending claim 8 "to more clearly define the subject matter of the present invention." (Id. at HNY020511.) Honeywell explained that the "amendment was not presented earlier because the need for clarification did not become apparent until the last office action in which new grounds of rejection [i.e., obviousness in light of LEMD and Cronin] were asserted." (Id.) Amended claim 8 read as follows:

Claim 8 (Twice Amended). A method of starting a gas turbine engine drivingly coupled to a starter/generator comprising the steps of:

sensing engine speed and generating a first signal thereof;

sensing elapsed time of the engine start;

producing a second signal, in response to said elapsed time, indicative of a predetermined speed for said engine;

combining said first and second signals to form an error signal; and

adjusting the torque output of said starter/generator in response to said error signal so that [said engine starts along a predetermined schedule] said torque output tracks the torque-speed characteristic of said engine.

(Id. at HNY020509.) According to Honeywell, this amendment was intended "to clarify that the torque output of the starter/generator is controlled so that it tracks the torque-speed characteristic of the engine during startup." Honeywell referred the examiner to Figure 3 of the original application (id. at HNY020477),^[4] "which shows that the output of the starter/generator vs. engine speed 96, 98 has the same shape as the torque-speed characteristic 92, 94 of the engine even as inlet conditions change" (id. at HNY020511.) Honeywell then distinguished LEMD and Cronin on the basis that neither reference disclosed its "torque output characteristic." (Id.)

In an October 1994 office action, the examiner once again rejected claim 8.^[5] (Id. at HNY020530.) In the examiner's view, the newly-added limitation "said torque output tracks the torque speed characteristics of said engine" was fatally indefinite under 35 U.S.C. § 112 ¶ 2 because "[n]o definition or explanation of an output ``tracking' a characteristic can be found in the specification." (Id. at HNY020531.) The examiner was also unpersuaded by Honeywell's June 1994 arguments because the "meets and bounds" of the new limitation were not clear, and because it appeared "that the teachings of cited prior art lay within the scope" of the new limitation. (Id. at HNY020534.) Furthermore, the examiner pointed out that "Cronin teaches a controller ... for controlling torque applied by the starter/generator." (Id.)

Honeywell responded in February 1995 with two additions to the specification and an amended claim 8:

The advantage of the present invention is that for a particular start schedule, the starter/generator delivers the optimum torque. That is it only delivers the amount of torque necessary to keep the engine on the schedule. This is shown in FIG. 3, where the difference between a dashed line and a solid line for a given inlet temperature is that amount of torque necessary to drive the engine along the preselected schedule.

(Id. at HNY020545.)

In contrast, with the control 60 the torque output of the starter/generator is optimum for a preselected acceleration schedule. The excess torque or inefficiency associated with prior art DC starters and air turbine starters is eliminated.

(Id.)

Claim 8 (Three Times Amended). A method of starting a gas turbine engine [drivingly coupled to a starter/generator] along a preselected start schedule comprising the steps of:

providing a starter/generator for driving said engine;

sensing engine speed and generating a first signal thereof;

sensing elapsed time of the engine start;

producing a second signal, in response to said elapsed time, indicative of a predetermined speed for said engine;

combining said first and second signals to form an error signal; and

adjusting the torque output of said starter/generator in response to said error signal so that [said torque output tracks the torque-speed characteristic of said engine] said starter/generator only delivers the torque necessary to keep the engine accelerating along the schedule.

(Id. at HNY020546-47.)

Honeywell explained that these amendments were intended to overcome the examiner's § 112 ¶ 2 objection to the "tracks" phrase. (Id. at HNY020548.) As to the examiner's objections under 35 U.S.C. §§ 102 and 103, Honeywell argued that neither LEMD nor Cronin taught "only delivering that torque necessary to keep the engine accelerating along a preselected schedule." (Id. at HNY020549-50.) Furthermore, Honeywell pointed out that neither reference taught the use of elapsed time to determine a predetermined speed for the engine. (Id. at HNY020550.) Without further explanation, the examiner allowed claim 8 in April 1995. (Id. at HNY020552.)

The first question the court must answer is whether Honeywell narrowed the last claim limitation of claim 1 by amending the final clause from "so that said engine starts along a predetermined schedule" to "so that said starter/generator only delivers the torque necessary to keep the engine accelerating along the schedule." HSC observes that the latter form of the limitation is narrower than the former because the former "does not specify how close the torque delivered must be to the predetermined schedule."^[6] (D.I. 158 at 3.) Honeywell responds that the amendments were merely clarifications of material that was already present in the original application. (D.I. 166 at 3-5.)

A side-by-side comparison of the plain language of the limitations reveals that HSC's observation has at least colorable merit. However, an amendment to a claim limitation is not necessarily narrowing, even when a side-by-side comparison makes it appear to be. For example, in Interactive Pictures, a claim limitation originally containing the phrase "output signals" was amended during prosecution to contain the phrase "output transform calculation signals." 274 F.3d at 1377. The Federal Circuit first looked to the context of the entire claim limitation, which in part read, "image transform processor means ... for producing output transform calculation signals...." Id. at 1374. Since the word "transform" in the phrase "image transform processor means" described what the processor did, the court reasoned that the insertion of "transform calculation" to modify "output signals" was duplicative. Id. at 1377. The court then looked to the specification and found that the function of the "image transform processor means" was to "derive[] equations by which the transform is calculated," which was further evidence that "transform calculation" did not add anything new when it was used to modify "output signals." Id. at 1378. Finally, the court pointed to a passage in the prosecution history where the patentee had used the phrases "output signals" and "output transform calculation signals" interchangeably. Id. Given this evidence, the court concluded that the amendment was not narrowing because it "did nothing more than make express what had been implicit in the claim as originally worded." Id. at 1377.

The Federal Circuit came to a similar conclusion in TurboCare Div. Of Demag Delaval Turbomachinery Corp. v. Gen. Elec. Co., 264 F.3d 1111 (Fed.Cir.2001). In TurboCare, the patented technology involved a seal in a turbine situated circumferentially around the turbine shaft such that the seal moved closer to the shaft as pressure increased, and farther away as pressure decreased. Id. at 1113-15. In the issued claim, the inward movement of the seal to the smaller-diameter position was explicitly limited by contact between two certain surfaces. Id. at 1116. In other words, the inward movement of the seal stopped upon contact between the two certain surfaces. However, the original claim made no reference to contact between the two certain surfaces as a means of stopping the inward movement of the seal. Id. at 1125. The original claim merely referred to a smaller-diameter position. Id. The defendant argued that the claim had been narrowed by the inclusion of the reference to "contact." Id. The court disagreed because the original specification explained that "the seal ring segments ... move radially inward until restrained by contact at surface 17." Id. at 1125-26 (emphasis in original). Furthermore, the specification also described a figure in which the two surfaces were visibly touching as the "small clearance condition." Id. at 1126. Thus, the Federal Circuit held that the insertion of the "contact" limitation only "redefined" the limitation without narrowing it. Id.

In the present case, while it is arguably true that the issued claim limitation is facially narrower than the original, HSC has not established an actual narrowing of the limitation in light of the prosecution history. First, the original specification submitted with Honeywell's March 1993 application reveals in several passages that the object of the invention is to minimize the expenditure of energy by automatically adjusting the torque output of the starter/generator, which is precisely what the present version of the limitation describes:

Accordingly, a need exists for a control system for a starter/generator that maintains a desired acceleration of the engine with minimum expenditure of energy, and which automatically adjusts the amount of torque to account for changes in engine drag due to cold soak or any other conditions to which the engine is exposed.

(D.I. 158, Ex. 1 at HNY020459 (emphasis added).)

An object of the present invention is to provide a control system for a starter/generator that maintains a desired acceleration of the engine with minimum expenditure of energy.

Another object of the present invention is to provide a control system for a starter/generator that automatically adjusts the amount of torque to account for changes in engine drag.

. . .

The present invention achieves the above-stated objects by providing a control system that electronically controls the torque applied to the engine by a starter/generator so that the engines [sic] accelerates along a predetermined schedule programmed in the control system.

(Id. at HNY020460 (emphasis added).)

Thus, the present invention provides a control system for controlling the torque applied by a starter/generator to a gas turbine engine during startup that accelerates engines according to a predetermined acceleration schedule with minimum expenditure of energy, and which automatically adjusts the amount of torque applied to account for changes in engine drag due to cold soak or any other conditions to which the engine is exposed.

(Id. at HNY020466 (emphasis added).) Perhaps the most revealing aspect of the original specification is that Honeywell equated minimizing the expenditure of energy by automatically adjusting the torque output with its goal "that the engines [sic] accelerates along a predetermined schedule." (Id. at HNY020460.) That phrase is nearly identical to the original limitation: "so that said engine starts along a predetermined schedule."^[7] Moreover, the application was also submitted with Figure 3 (id. at HNY020477), which Honeywell explained during prosecution "shows that the output of the starter/generator vs. engine speed 96, 98 has the same shape as the torque-speed characteristic 92, 94 of the engine even as inlet conditions change" (id. at HNY020511). Thus, it is clear that from the time the application was first submitted, the limitation was directed to only delivering "the torque necessary to keep the engine accelerating along the schedule."

The series of office actions and responses detailed above also support the court's conclusion. In February 1994, the examiner explained his opinion that, except for a combined starter/generator, LEMD contained all the elements of claim 8:(a) a means for sensing the speed of the engine; (b) a torque regulator responsive to the speed sensing means; and (c) a "predetermined schedule of engine acceleration." (Id. at HNY020501.) Honeywell responded in June 1994 by amending the limitation from "said engine starts along a predetermined schedule" to "said torque output tracks the torque-speed characteristic of said engine." (Id. at HNY020509.) In the October 1994 office action, the examiner rejected the new limitation both because it was indefinite under § 112 ¶ 2, and because it was still obvious in light of the prior art. (Id. at HNY020534.) When Honeywell responded in February 1995, it cleared the § 112 ¶ 2 hurdle by amending the limitation to its present form, and adding language to the specification that amounts to no more than further explanation of the substance that was in the specification from the beginning. (Id. at HNY020545-47.) Importantly, Honeywell also argued that claim 8 was allowable because neither LEMD nor Cronin taught "the use of elapsed time." (Id. at HNY020550.) Without further explanation, the examiner allowed the claim. (Id. at HNY020552.) Thus, the prosecution history contains no evidence that Honeywell ever convinced the examiner that LEMD did not teach the same method of applying only the torque necessary to accelerate the engine along a predetermined schedule. It is entirely possible that the claim was allowed because of Honeywell's final observation that neither LEMD nor Cronin taught "the use of elapsed time." In other words, the prosecution history contains no evidence that the limitation was ever effectively narrowed in the eyes of the examiner. Of course, the examiner's opinion is not binding on the court, but it is persuasive. Therefore, the court holds that Honeywell did not narrow the last limitation of claim 1 during prosecution. As such, prosecution history estoppel does not preclude Honeywell from asserting the doctrine of equivalents with regard to that limitation.

In so ruling, the court is cognizant of the Supreme Court's holding in Festo II that estoppel may apply even if a narrowing amendment is "only for the purpose of better description." 535 U.S. at 737, 122 S. Ct. 1831. The Court believed that it was improper to conflate "the patentee's reason for making the amendment with the impact the amendment has on the subject matter." Id. at 736, 122 S. Ct. 1831. In essence, the Court's overall concern was preserving the public-notice function of preventing patentees from recapturing objectively surrendered subject matter in subsequent litigation. See id. at 733-35, 122 S. Ct. 1831. However, that rationale does not translate into a rule that an amendment ostensibly narrowing a limitation always narrows the subject matter of the claim. Indeed, given the inability of language to always fully "capture the essence of a thing in a patent application," id. at 731, 122 S. Ct. 1831, it is often very difficult to discern whether an amendment is merely an alternative way of describing the same thing, or whether it actually surrenders subject matter. Thus, it would be inaccurate and imprudent to conclude that an amendment narrows the scope of a claim by merely giving the original and amended limitations a cursory, side-by-side comparison, without first exploring the entire prosecution history.

Here, HSC presents a colorable argument that the amended limitation is narrower than the original limitation. However, the amendment did not change something in the nature of a neat numerical range. See, e.g., Warner-Jenkinson Co. v. Hilton Davis Chem. Co., 520 U.S. 17, 117 S. Ct. 1040, 137 L. Ed. 2d 146 (1997) (claim amended to include "a pH from approximately 6.0 to 9.0" in order to avoid prior art). Instead, the amendment changed the description of an extremely complex technology. Consequently, it would be disingenuous for a lay court such as this one to proclaim an ability to discern in the abstract the relative subject matter encompassed by alternative descriptions of such a complicated device without resorting to anything but the language of the claim itself. In recognition of this, the court very diligently analyzed the "impact the amendment has on the subject matter" of claim 1 by studying the entire prosecution history. For the reasons set forth above, and after careful and extensive deliberation, the court concludes that the public-notice function underlying the Supreme Court's decision in Festo II is protected by this court's holding because, in its view, no subject matter was objectively surrendered in the course of amending the last limitation of claim 1.

2. Claim 3

Two limitations of claim 3 are at issue in HSC's motion. The court will first address the last limitation of the claim, which reads in relevant part: "so that said starter/generator only delivers the torque necessary to keep the engine accelerating along the schedule." In its opening brief, HSC argues that the court should dispose of this limitation in the same way is disposes of the corresponding limitation in claim 1 — that is, the court should bar Honeywell from asserting the doctrine of equivalents with regard to this limitation. (D.I. 158 at 5-6.) The court will accede to HSC's request, but with the opposite result. Thus, the court holds that prosecution history estoppel does not preclude Honeywell from asserting the doctrine of equivalents with regard to the last limitation in claim 3.

Addressing the other limitation at issue requires the court to review a small portion of the prosecution history. In March 1999, claim 10 of the application (which became claim 3 in the issued patent) read as follows:

10. (Four Times Amended) A system for starting a gas turbine engine along a preselected start schedule, comprising:

a starter/generator drivingly coupled to said engine;

an engine speed sensor that senses the speed of said engine;

a first signal produced by said engine speed sensor;

a timer for determining elapsed time from the engine start;

a second signal produced by said timer that is indicative of a predetermined speed for the engine;

an error signal formed from said first signal and said second signal; and

a controller that adjusts the torque output of said starter/generator, based on said error signal, so that said starter/generator only delivers the torque necessary to keep the engine accelerating along the schedule.

(Id., Ex. 1 at HNY020670.)

The examiner rejected the claim in June 1999 on two grounds. First, the examiner rejected the claim as "unclear" under § 112 ¶ 2 because the claim ostensibly assigned conflicting functions to the timer: the fourth limitation assigned it the function of determining elapsed time from engine start, whereas the fifth limitation assigned it the function of outputting a signal indicative of the predetermined engine speed. (Id. at HNY020686.) Second, the examiner rejected the claim as obvious under § 103 in light of the combination of prior-art references Lafuze, Pollak, and Rozman. (Id. at HNY020687.) Lafuze disclosed an engine starter with a current-controlled starter generator, but did not disclose its torque characteristic or the use of a timer. (Id.) However, Pollak disclosed a starter that sensed the engine position to control the torque delivered to the system (id. at HNY020687-88), and Rozman disclosed a timer and speed sensor that determined whether the speed was low too long (id. at HNY020688). Thus, the examiner concluded:

It would have been obvious to one of ordinary skill in the art at the time the invention was made to have modified the controller of Lafuze to operate according to preselected torque characteristics as taught by Pollak and to modify Lafuze to control the starter according to a sensed timing as taught by Rozman et al. because Pollak teaches starting torque of the engine should be controlled to start the motor efficiently using a torque feedback loop, and because Rozman et al. teaches that if an engine is stalled then the system should be disabled for protective purposes.

The phrase "an error signal formed from said first signal and said second signal" is met in that the torque is produced from the sensed speed and from a modified speed signal. These signals are converted to an actual torque and a command torque. Since both the speed and [sic] signal and a modified speed signal are used in forming the final torque error this limitation is met.

(Id.)

In response to the examiner's objections, Honeywell amended claim 10:

10. (Five Times Amended) A system for starting a gas turbine engine along a preselected start schedule, comprising:

a starter/generator drivingly coupled to said engine;

[an engine speed] a sensor that senses the speed of said engine;

a first signal produced by said [engine speed] sensor;

a timer for determining elapsed time from the engine start;

a second signal produced by the system, corresponding to the elapsed time determined by said timer that is indicative of a predetermined speed for the engine;

an error signal formed from said first signal and said second signal; and

a controller that adjusts the torque output of said starter/generator by controlling the current flow, based on said error signal, so that said starter/generator only delivers the torque necessary to keep the engine accelerating along the schedule.

(Id., at HNY020693-94.) Accompanying its amendment was Honeywell's explanation that claim 10 was amended to overcome the examiner's § 112 ¶ 2 objection. (Id. at HNY020695.) Honeywell also asserted that the amendment was made "to further define the subject matter of the invention" in order to overcome the examiner's § 103 objection. (Id.) However, Honeywell failed to explain why the amendment was no longer obvious in light of the combination of Lafuze, Pollak, and Rozman. Rather, Honeywell merely recited the claim limitations and summarily stated that "[c]laim 10 is allowable because it is not anticipated by or obvious over the prior art of record." (Id. at HNY020696.) In fact, that is the identical "argument" Honeywell set forth in its March 1999 response to overcome a previous obviousness objection to claim 10. (Id. at HNY020673-74.) Without further explanation, the examiner issued a notice of allowability for claim 10.

HSC argues that Honeywell's amendment to the fifth limitation narrowed the claim to avoid what is known as a "fail to crank" fault timer disclosed in Rozman. (D.I. 158 at 8.) In other words, HSC contends that because the invention was obvious in light of a combination including the Rozman's "protective mechanism that disables the motor ``if the speed is too low too long,'" i.e., a "fail to crank" fault timer, Honeywell amended its timer limitation to exclude that subject matter. (Id. at 8-9.) Honeywell once again responds that the amendment was merely a clarification of material that was already present in the original application. (D.I. 166 at 7-9.)

The court agrees with Honeywell. First, it is not at all clear that the amended language disclaims "fail to crank" fault timers. As amended, the claim appears to have retained the functional ability it is alleged to have had before the amendment, i.e., the ability to detect when the speed is low too long, because a comparison is still made between the predetermined speed and the actual speed in producing the error signal of the sixth limitation. Second, the examiner did not reject claim 10 because it purportedly employed a "fail to crank" fault timer. Rather, the examiner explained that, in spite of the fact that Lafuze did not disclose "using a timer," it would have been obvious "to modify Lafuze to control the starter according to a sensed timing as taught by Rozman et al .... because Rozman et al. teaches that if an engine is stalled then the system should be disabled for protective purposes." (D.I. 158, Ex. 1 at HNY020687-88.) It was merely the way in which Rozman detected a stalled system, i.e., by using a timer, that rendered the invention obvious. The fact that the Rozman timer was used to detect a stalled system was irrelevant to the examiner's argument. Any attempt to distinguish Rozman would have required Honeywell to eliminate the use of a timer, which it did not do. Finally, the substance of the amendment itself is directed at the examiner's § 112 ¶ 2 objection, not his § 103 objection. In other words, the amendment adequately addressed the examiner's concern that the timer seemed to have conflicting functions, but it did nothing to substantively distinguish the combination of Lafuze, Pollak, and Rozman because the amended claim continued to contain all the limitations of that combination pointed to by the examiner: (1) the starter/generator controller of Lafuze; (2) the torque control of Pollak; and (3) the timer of Rozman. Moreover, Honeywell only made a hollow attempt to distinguish the combination. Nevertheless, the claim was allowed as amended. The fact that the prosecution history contains no indication as to why the examiner did not persist with his § 103 objection is of no concern here. What is important is whether the prosecution history reveals a narrowing of the limitation. Because the prosecution history does not support such a conclusion, the court holds that Honeywell is not barred from asserting the doctrine of equivalents with regard to the fifth limitation of claim 3.

III. CONCLUSION

For the foregoing reasons, the court will (1) deny HSC's request to preclude Honeywell from presenting a damages calculation based on sales projections of the accused product that did not exist at the time of the hypothetical negotiation (D.I.156); and (2) deny HSC's motion on prosecution history estoppel (D.I.158).

ORDER

IT IS HEREBY ORDERED THAT:

1. The Defendant's request to preclude Honeywell from presenting a damages calculation based on sales projections of the accused product that did not exist at the time of the hypothetical negotiation (D.I.156) be DENIED; and

2. The Defendant's motion on prosecution history estoppel (D.I.158) be DENIED.

MEMORANDUM

The court will now address Hamilton Sundstrand's motion in limine to preclude any argument that an invention must be commercially viable to constitute a reduction to practice. (D.I. 108.) In particular, Hamilton Sundstrand asks the court to order Honeywell to refrain from arguing to the jury that Hamilton Sundstrand's "demonstration of an auxiliary power unit, or APU, to Boeing in May 1992 could not constitute a reduction to practice under 35 U.S.C. § 102(g) because the APU was not ``commercially viable,' not ``ready to be installed on aircraft in revenue service,' otherwise not ``commercially ready,' or similar argument." (Id. at 1.) Hamilton Sundstrand points to two specific examples in the reports of Honeywell's experts, David Stewart and Gerard Muller. In Stewart's report, he disagrees with the conclusion of Hamilton Sundstrand's expert, Mr. Goolkasian, that there was a reduction to practice at least by May 13-14, 1992:

This is incorrect for two reasons. First, although something was demonstrated on or about May 13-14, 1992, Mr. Goolkasian has neither related whatever was demonstrated to the claims of the '275 patent nor has Mr. Goolkasian established that whatever was demonstrated was a reduction to practice of the invention claimed in the '275 patent. Indeed, Michael Hanson testified on November 30, 2004, that the breadboard he worked on "would require further development, not just more testing, before it could be placed on an aircraft in revenue service."^[1]

(Id., Ex. B at 7.)

In Muller's expert report, Hamilton Sundstrand alleges that Muller "dismissed HSC's demonstration to Boeing because the APU was not at ``a production stage.'" (D.I. 108 at 2.) Hamilton Sundstrand points the court to Exhibit C of its opening brief, however, presumably due to an oversight by counsel, the text of Exhibit C relates only to the doctrine of equivalents. Thus, the court was unable to read the objectionable material first hand. Nevertheless, since Honeywell does not dispute Hamilton Sundstrand's characterization, the court assumes it to be relatively accurate.

Honeywell says it "intends to introduce evidence about the state of development of the APU that Sundstrand allegedly demonstrated to Boeing in 1992 to show that the demonstration did not accurately duplicate actual working conditions of a production APU on an aircraft and that the demonstration APU was not reduced to practice." (D.I. 113 at 3.) Honeywell argues that "Sundstrand seeks to exclude this evidence because of extensive testimony from its witnesses that the demonstration APU was not even at the pre-production stage of development and that there was no way of determining whether the demonstration APU would work for its intended purpose." (Id.) For example, Honeywell points to the deposition testimony of Hamilton Sundstrand's 30(b)(6) witness, Michael Hanson, in which he said that the demonstration APU was in breadboard format, and that it required further development before it was ready for use on an aircraft. (Id. at 3-4.)

Honeywell argues that "in order to show reduction to practice of complex inventions like the start method claimed in Honeywell's '626 patent, tests are required ``that accurately duplicate actual working conditions in practical use.'" (Id. at 2.) Furthermore, Honeywell points to several non-binding appellate decisions from the 1950's that Honeywell argues stand for the proposition that in cases relating to the aerospace industry, tests embracing the actual conditions which the invented devices would meet in actual service on an airplane are required before there can be a reduction to practice. Gaiser v. Linder, 45 C.C.P.A. 846, 253 F.2d 433 (C.C.P.A.1958); Radio Corp. of Am., 232 F.2d 726 (3d Cir.1956); Chandler v. Mock, 40 C.C.P.A. 846, 202 F.2d 755 (C.C.P.A.1953). Thus, according to Honeywell, "evidence that the Sundstrand ``breadboard' APU was not intended to be an actual product and could require extensive changes before it would work for its intended purpose is undoubtedly relevant to the issue of the reduction to practice of that APU." (D.I. 113 at 4.)

Hamilton Sundstrand, on the other hand, contends that "Honeywell's argument is contrary to well settled law, and would threaten to confuse or mislead the jury." (D.I. 108 at 1.) For support, Hamilton Sundstrand points to more recent decisions from the Federal Circuit, which are said to stand for the proposition "that an invention need not be commercially ready or otherwise be a ``commercial embodiment' to constitute a reduction to practice." Taskett v. Dentlinger, 344 F.3d 1337 (Fed.Cir.2003); DSL Dynamic Sciences, Ltd. v. Union Switch & Signal, Inc., 928 F.2d 1122 (Fed.Cir.1991); Barmag Barmer Maschinenfabrik AG v. Murata Machinery, Ltd., 731 F.2d 831 (Fed.Cir.1984). Thus, it says, the evidence is inadmissible under Rules 402 and 403. (D.I. 108 at 1.)

The court finds the reasoning in Taskett, to be dispositive. In that case, the claim at issue was a method for automated purchase of prepaid telephone services, which required "obtaining financial authorization for said request by a central terminal." 344 F.3d at 1339. The question under consideration was whether actual reduction to practice required "real money, from a real bank account at a third-party financial institution [to be] actually debited and actual telephone calls [to be] made based on the printed receipts." Id. The Federal Circuit held that it did not because: (1) the language of the claim limitation did not require on its face or when read in light of the specification that the "central terminal" be located at a third-party financial institution; and (2) a "viable commercial embodiment" was not required, as reduction to practice can be based on tests alone. Id. at 1340-42. The Taskett court also cited with approval its earlier case of Scott v. Finney, where the court explained that tests can be sufficient if they make "a reasonable showing that the invention will work to overcome the problem it addresses." 34 F.3d 1058, 1063 (Fed.Cir.1994). No actual use was required in Scott. Id.

In this case, the court agrees with Hamilton Sundstrand because Honeywell's theory applies the incorrect legal standard. As in Taskett, nothing in the claim language of U.S. Patent No. 6,035,626 requires actual use in aircraft. In fact, a text search of the patent turns up no references to "plane" or "airplane" or "aircraft." Therefore, the court will grant Hamilton Sundstrand's motion to the extent that it seeks to preclude Honeywell from introducing evidence that the 1992 APU was not reduced to practice because it was not a commercial embodiment.

ORDER

IT IS HEREBY ORDERED that:

1. Hamilton Sundstrand's Motion In Limine to preclude any argument that an invention must be commercially viable to constitute a reduction to practice (D.I. 108) be GRANTED.

NOTES

[1] The license also appears to contemplate revenue from aftermarket parts and service extending beyond the life of the patent. (See D.I. 156, Attach. C, Ex. 4a.)

[2] Honeywell seems to have used the terms "starter/generator" and "motor/generator" interchangeably.

[3] The court indicates removed language with brackets and inserted language with italics.

[4] As far as the court can discern, Figure 3 of the application contains only insubstantial differences from Figure 3 of the '626 patent.

[5] For reasons unimportant here, the applicant abandoned the original application and filed a continuing application. (D.I. 158, Ex. 1 at HNY020522.)

[6] HSC also argues that the former wording does not "specify whether the adjustment of torque needs to be when the engine is accelerating or when it is at a constant speed." (D.I. 158 at 3.) However, the latter wording does not make that specification either. Thus, with regard to this argument, the court is not persuaded that the claim limitation was narrowed.

[7] The one obvious difference is the use of "accelerates" in the specification and "starts" in the original limitation. However, that difference is meaningless for present purposes because the point of accelerating the engine is to cause it to start.

[1] The '275 patent referred to in Stewart's report is not the patent at issue in this case.