MILLER, Judge.
This appeal is from a decision of the Patent and Trademark Office (PTO) Board of Appeals affirming rejections of claims 1-18 under 35 U.S.C. § 102(b) and claims 17 and 18 under 35 U.S.C. § 103 in application serial No. 832,542, filed May 12, 1969, which is a continuation of application serial No. 670,189, filed September 25, 1967, for “Piezoelectric Transducer Sensor.” The controlling issues involve the questions of whether the invention was in public use or on sale within the meaning of 35 U.S.C. § 102(b) more than a year prior to the filing date of the parent application, and whether the subject matter of claim 17 would have been obvious to one of ordinary skill in the art. We reverse in part and affirm in part.
The Invention.
The invention is a load sensing piezoelectric transducer used to prevent overloading of power press members subject to stress. The transducer is attached to [1395]*1395the press member so that force applied to the press member is transferred to the piezoelectric element, which generates a proportional voltage. An excessive voltage energizes a machine monitoring means. Fig. 1, reproduced below, is described in the specification as follows:
the piezoelectric transducer unit mounted on a force carrying member 2 of a press. The force carrying member 2 is illustrated schematically. The piezoelectric transducer 3 is shown in cross-section as consisting of a piezoelectric ceramic crystal 4 positioned between a top terminal block 5 and a bottom terminal block 6 with the top and bottom faces of the crystal in contact with and against the surface of each of the respective terminal blocks. The terminal blocks and crystal are surrounded by a metallic sheath 7 which serves as a magnetic and electric shield. The sheath is slightly larger than the terminal blocks and crystal to provide space for a surrounding layer 8 of an encapsulating and insulating plastic layer 8. Leading from the terminal blocks through the insulating layer 8 and the metallic sheath 7 are terminals 9 and 10 with suitable connections for attachment of leads to the electronic circuit.
Claim 1 broadly recites the combination of press, piezoelectric transducer, and electronic circuit:
1. In a machine having two work members movable with respect to each other and subject to repetitive loading forces for transmission to a work piece between said members, at least one force carrying member carrying and transmitting said loading force to one of said work members, and an electrically responsive strain sensing device positioned on said force carrying member, the improvement in combination therewith wherein said sensing device is a piezoelectric transducer,
said piezoelectric transducer comprising a piezoelectric element, a first terminal block positioned against one face of said element and a second terminal block positioned against the opposite side of said element,
a mounting means attached to and extending from the force carrying member, said mounting means comprising a first and a second bracket spaced from each other, one end of each bracket being attached to said force carrying member, the other ends of said brackets having contact means for engagement with said terminal blocks, at least one of said contact means being electrically nonconductive,
said piezoelectric transducer being positioned and clamped between the said other ends of the brackets with the contact means of said first bracket engaging said first terminal block and the contact means of said second bracket engaging said second terminal block so that the clamping force of the brackets is substantially parallel to [1396]*1396and aligned with a line normal to the interfaces between said element and terminal blocks, and
an energized electronic circuit electrically connected to the terminal blocks which circuit is predeterminedly activated by the output signal of the piezoelectric transducer when stressed during a work cycle of the machine to activate a machine monitoring means.
Claims 2, 3, 13-16, and 18 add sheath 7. Claims 17 and 18 limit the brackets to threaded members that are screwed into internally threaded apertures of the force carrying member. The remaining claims are dependent on claim 1.
Public Use Proceedings
Appellant and two individuals named Budraitis and Harding formed a joint venture based on the invention. After an application for patent was filed, they had a falling out, and Budraitis and Harding petitioned for public use proceedings. These were instituted and the testimony of witnesses was taken by deposition. Rule 292(a), 37 CFR 1.292(a).1 The pertinent evidence is summarized below.
From 1959 to 1963 appellant was employed as an electrical engineer for Verson All-Steel Press Co., where he was involved mainly in designing strain gauge load sensing equipment for controlling large presses. By 1961 appellant had conceived the present invention, utilizing a piezoelectric transducer, as an improvement over strain gauge control devices. With the aid of Budraitis, a draftsman at Verson, he proceeded to develop and test his conception in the basement of his home. However, Budraitis testified that commercially available strain gauges were accurate to better than plus or minus two percent and that the necessity of achieving at least that accuracy was recognized. From his work at Verson, appellant knew that to be properly tested for accuracy and stability under various conditions the invention would have to be installed for a long period on a large press costing at least $10,000, which he did not have at his disposal. Consequently, petitioner Harding, a full-time vacuum cleaner salesman, was invited to join the venture to assist in interesting several companies in installing the invention on their presses. In order to encourage purchase of the invention, the quoted price was one-fourth the going rate for prior control devices.
In the fall of 1963, after a number of unsuccessful attempts, a quotation to Starline, Inc. was accepted. The agreement provided that payment was subject to the device’s operating to the satisfaction of Starline for sixty days after installation. The joint venture (Industrial Standard Manufacturing Co.) represented to Starline that the device would operate to an accuracy of plus or minus two percent. However, after about three months, “test data” showed that the device was experiencing errors of fifty percent or more. Because of the inaccuracy of the device, it would sometimes shut down the Starline press at loads below the maximum permissible load setting. To prevent this, the operator would adjust the load setting to above 100 percent. It was appellant’s opinion that overall redesign was necessary to overcome the problems.
[1397]*1397Starline’s plant manager, Meany, who was responsible for authorizing payment for the device, believed it to be operating satisfactorily. However, Meany was not aware of the magnitude of the errors occurring. He testified that he had the “feeling” that the device was experimental as far as both Starline and the joint venture were concerned, and that the device purchased by Starline was the first of its kind. Although appellant could not recall whether he advised Star-line that the device was experimental, he testified that it required testing before it would be ready for commercial sale.
No changes were made in the device during the approximately three years it remained on the press. Neither appellant nor petitioners personally profited from the sale to Starline because their expenses exceeded the selling price. Appellant and petitioners agreed to withhold as much information about the device as they could. Third parties were not informed that the basis of the invention was a piezoelectric element, and this could not be easily discovered since the element was completely encapsulated in epoxy resin. The device installed on the Starline press differed in several respects from that disclosed in the patent, including the absence of metallic sheath 7. Appellant admitted, however, that the Starline unit meets the language of claims 1 and 4-10.
In August 1964, a quotation to Alpha Products, Inc. was accepted. The device installed at Alpha also lacked the metallic sheath, but differed in several respects from that installed at Starline and was designed to determine the load on the press rather than to automatically stop the press. Appellant testified that the primary purpose of the Alpha installation was experimental, that the device was the first of its kind, and that he had no idea how it would act. However, the plant manager at Alpha testified that Alpha did not consider it experimental. Like the Starline device, the Alpha device was sold for substantially less than comparable commercial equipment, produced no personal profit for appellant or petitioners, was subject to the purchaser’s satisfaction, and operated erratically. Information about the piezoelectric element and other features was withheld from Alpha. Appellant testified that erratic performance could have been from any of a number of problems and that redesign of the device was required. Alpha paid for the device and used it for about a year. Neither appellant nor Budraitis ever informed Alpha about the unsatisfactory “test data.”
From the time of the Alpha installation and continuing into 1966, appellant further experimented and modified the invention. In August 1966 an agreement was reached with Ford Motor Co. for delivery and installation of a modified device, which was the first to include the metallic sheath. Ford’s manager of plant engineering testified that the device was purchased for experimentation to determine the feasibility of protecting presses from overloading and that personnel from Industrial Standard participated in the experimentation. Appellant testified that his primary purpose with respect to the Ford installation was to test the redesigned invention.
The date of installation of the device on a Ford press is not clear. Parts required for installation were purchased only after the agreement, and Ford’s manager of plant engineering testified that Ford had experienced difficulty finding time to make the installation after it had received the device. He was not sure of the installation date of Sep* tember 25, 1966, set forth in an evaluation report dated January 19, 1967. A Ford intra-company communication dated March 20, 1968, indicates that the device was calibrated on August 27, 1966, and, after having his memory refreshed by this document, petitioner Budraitis testified that he believed August 27 was the date he and appellant first went to Ford to do the installation and calibration.
[1398]*1398Problems were encountered during installation2 and calibration, and calibration was not completed as of October 6, 1966. In December an error in calibration was discovered and corrected, following which Ford indicated in a letter to Industrial Standard that it considered the device to be operable. However, the letter also indicated concern over the long term operational characteristics and stability of the device and requested complete wiring diagrams, complete sensor description and mounting instructions, and complete calibration instructions.3 Subsequently Ford purchased a second device and subjected it to extensive testing. Although neither Ford installation was perfected, a device sold in Canada in 1967 and subsequent devices achieved an accuracy of plus or minus two percent.
Rejections
On the basis of the evidence presented in the public use proceedings, the examiner held that there had been public use and sale of the claimed invention. Accordingly, he made an ex parte rejection of all claims under 35 U.S.C. § 102(b). Dependent claims 17 and 18 reciting screw-in brackets, which the examiner apparently agreed were not present in the devices sold before the critical date,4 were also rejected under 35 U.S.C. § 103 on the basis of the public use and sale in view of either of two prior art patents.
The board affirmed these rejections, holding that the experimental use exception to the public use or on sale bar was not applicable because appellant’s intent was to penetrate the market and to gain a commercial advantage rather than to experiment. It noted that there were numerous attempts to sell the invention over a period of three years before the critical date; that appellant hired a salesman, Harding, to sell the device; that no offers to sell contained language showing that the sale was incidental to experimental use or requiring the purchaser to report data to appellant; that purchasers were not required to agree to maintain the device in secrecy; that, according to the testimony of Budraitis, few changes in the device resulted from the experimentation; and that the device was paid for in each case, which it interpreted as an indication of customer satisfaction. The board said:
We do not doubt that appellant’s motive in part was to try out the equipment and perfect his invention, but we are also of the opinion that the hiring of a man trained to sell, is some evidence that market penetration was a goal. Such a motive prohibits a finding of experimental use as would extend the one year statutory grace period provided by 35 U.S.C. 102(b).
If we were to point to a single legal principle that is determinative of the result in this case, it is that commercial exploitation of the product is fatal to a claim of experimental use
A bona fide experimental use involves no commercial exploitation . . [Emphasis added.]
Concerning all claims, the board stated that all that is required by section 102(b) is substantial identity between the device on sale or in public use and the subject matter claimed. In regard to claims 2, 3, 13-16, and 18, which require sheath 7, the board acknowledged that there was a question whether the sale to Ford took place prior to the critical date. However, it said that, even though payment was delayed, the sale took place [1399]*1399because an executory contract of sale creates a bar even with delivery after the critical date, citing Robbins Co. v. Lawrence Manufacturing Co., 482 F.2d 426 (9th Cir. 1973).
OPINION
A classic case on public use or sale and the experimental use exception is Elizabeth v. Pavement Co., 97 U.S. 126, 24 L.Ed. 1000 (1878).5 There the inventor had laid a seventy-five foot length of pavement on a toll road in which he was a stockholder. This pavement was his invention, and the seventy-five foot length was laid for the purpose of testing its durability under heavy traffic. After six years a patent application was filed. The Court held the use to be experimental, noting that the nature of the invention required public experimentation and explaining that use by the inventor or another under his direction to perfect the invention is not a bar. The Court stated that an invention may be used on the premises of another and inure to the other’s benefit without loss of right to a patent if it is used under the inventor’s surveillance for the purpose of testing and making any necessary changes. Recognizing that it is in the public interest that the invention be “perfect and properly tested, before a patent is granted for it,” the Court added that public knowledge is irrelevant if the use is experimental, i. e. “a bona fide effort to bring [the] invention to perfection, or to ascertain whether it will answer the purpose intended”; however, “[a]ny attempt to use it for profit, and not by way of experiment,” would bar a patent. Id. 97 U.S. at 137. And “if the inventor allows his machine to be used by other persons generally, either with or without compensation, or if it is, with his consent, put on sale for such use, then it will be in public use and on public sale, within the meaning of the law.” Id. at 135.
A contrasting case, in which a patent was invalidated because of public use, is Hall v. Macneale, 107 U.S. 90, 2 S.Ct. 73, 27 L.Ed. 367 (1883). The patent covered safe bolts that were hidden from view when used. On the issue of experimental use, the Court noted that the invention was complete and said (id. at 97, 2 S.Ct. at 79):
It was capable of producing the results sought to be accomplished, though not as thoroughly as with the use of welded steel and iron plates [a subsequent development in safes] . . . . [I]t is not shown that any attempt was made to see if the plates of the safes could be stripped off ... to prove whether or not the conical bolts were efficient. The safes were sold, and, apparently, no experiment and no experimental use were thought to be necessary.
In Smith & Griggs Manufacturing Co. v. Sprague, 123 U.S. 249, 8 S.Ct. 122, 31 L.Ed.. 141 (1887), the Court held invalid for public use a patent on a machine for making buckle levers. Rejecting the patentee’s argument that the use was experimental, the Court pointed out that the machine was useful for the purpose for which it was designed and was used for profit in the regular conduct of an established business. Although the inventor was also engaged in the invention of improvements, the alterations were not vital to the organization of the machine, and it was commercially successful without them. However, the Court said:
[1400]*1400A use by the inventor, for the purpose of testing the machine, in order by experiment to devise additional means for perfecting the success of its operation, is admissible; and where, as incident to such use, the product of its operation is disposed of by sale, such profit from its use does not change its character; but where the use is mainly for the purposes of trade and profit, and the experiment is merely incidental to that, the principal, and not the incident, must give character to its use. The thing implied as excepted out of the prohibition of the statute is a use which may be properly characterized as substantially for the purposes of experiment.
Id. at 256, 8 S.Ct. at 126. The Court also held that, once public use or sale before the critical date has been established, the burden is on the patentee to prove that such use was experimental by “full, unequivocal, and convincing” proof.6 See also Electric Battery Co. v. Shimadzu, 307 U.S. 5, 59 S.Ct. 675, 83 L.Ed. 1071 (1939); Root v. Third Ave. R. R., 146 U.S. 210, 13 S.Ct. 100, 36 L.Ed. 946 (1892); Manning v. Cape Ann Isinglass & Glue Co., 108 U.S. 462, 2 S.Ct. 860, 27 L.Ed. 793 (1883); Egbert v. Lippmann, 104 U.S. 333, 26 L.Ed. 755 (1881); Coffin v. Ogden, 85 U.S. (18 Wall.) 120, 21 L.Ed. 821 (1874).
In In re Bertram, 88 F.2d 834, 24 CCPA 1073 (1937), this court held an invention unpatentable because it had been operated publicly solely for the purpose for which it was intended and profit had resulted. In affirming a Patent Office rejection, the court in In re Josserand, 188 F.2d 486, 38 CCPA 994 (1951), indicated that if the purpose intended or designed for the invention requires severe tests over a considerable period of time and no attempt is made to use the invention for profit during the testing period, it is not a public use.
The Second and Tenth Circuit Courts of Appeals have stated that use is experimental only if that is the primary purpose and commercial exploitation is incidental. Cali v. Eastern Airlines, Inc., 442 F.2d 65 (2d Cir. 1971); Universal Marion Corp. v. Warner & Swasey Co., 354 F.2d 541 (10th Cir. 1965), cert. denied, 384 U.S. 927, 86 S.Ct. 1444, 16 L.Ed.2d 530 (1966). Even though an inventor may claim to have a bona fide experimental purpose, he is not permitted to commercially exploit his invention for more than the one year period provided by 35 U.S.C. § 102(b) before filing his application. Pickering v. Holman, supra note 5; Amerio Contact Plate Freezers, Inc. v. Belt-Ice Corp., 316 F.2d 459 (9th Cir.), cert. denied, 375 U.S. 902, 84 S.Ct. 191, 11 L.Ed.2d 143 (1963).
The Seventh Circuit held that there was a placing on sale of a device (equipment requiring erection) when it was completed to such an extent as to enable the purchaser to know how it would perform. Julian v. Drying Systems Co., 346 F.2d 336 (7th Cir. 1965).
Claims 2, 3, 13-16, and 18
We are not persuaded that a prima facie case has been established a public use or placing on sale of the invention defined by these claims (which require metallic sheath 7) more than one year prior to appellant’s filing date. The date when this device was delivered and finally installed and calibrated is unclear.7 The board relied on the executory contract of sale, which had been entered into prior to the critical date. However, for an invention of the type involved here to be “on sale,” it must be complete at least to such an extent that the purchaser knows how it will perform. The record fails to show that the invention was complete or that either appellant or Ford knew how the invention would perform at the time it was [1401]*1401installed.8 Indeed, the board acknowledged that there was a question whether the sale took place before the critical date.
Therefore, we hold that the invention of claims 2, 3, 13 — 16, and 18 was not shown to have been in public use or on sale more than one year prior to appellant’s filing date.
Claims 1 and 4-12
Appellant admits that at least some of these claims read on devices sold and installed prior to the critical date. Accordingly, he had the burden of establishing by clear and convincing evidence that such sales were for experimental purposes and that they were necessarily incident thereto. In re Blaisdell, supra note 6.
As related above, there were two sales of appellant’s devices without the metallic sheath. These devices were complete in that they fell within the scope of the claims and performed the function therein recited, though not as well as later improved devices. The first sale, in the fall of 1963, was to Starline, Inc., and the second, in August 1964, was to Alpha Products, Inc. There also were several unsuccessful attempts to sell the device. As noted by the board, no offers to sell indicated that the sale was incidental to experimental use or required the purchaser to report performance data to appellant. The plant manager at Starline said he had a “feeling” that the device was experimental, but acknowledged that he had not been told that it was. Also, he thought that it was operating satisfactorily (not having been told of the errors that were occurring) and authorized payment for it. The plant manager at Alpha did not consider the device experimental. Apparently Alpha considered it satisfactory too (not having been told of the unsatisfactory “test data”) and paid for it.
Appellant’s failure to communicate to any of the purchasers or prospective purchasers of his device that the sale or offering was for, experimental use is fatal to his case. Robbins Co. v. Lawrence Manufacturing Co., 482 F.2d 426 (9th Cir. 1973). Moreover, his failure to disclose what he regarded as unsatisfactory operations of the device to Star-line and Alpha supports an inference that he was seeking to penetrate the market for commercial purposes. Although selling the devices for a profit would have demonstrated the purpose of commercial exploitation, the fact that appellant realized no profit from the sales does not demonstrate the contrary. Also, that the basis of appellant’s invention (the piezoelectric element) was not disclosed to the purchasers and, being encapsulated in epoxy-resin, was unlikely to be discovered does not preclude a finding of commercial exploitation. Hall v. Macneale, supra.
Accordingly, we hold that appellant has not sustained his burden of proving that the sales were for experimental purposes and were necessarily incident thereto.
Section 103 Rejection of Claims 17 and 18
Since the rejection of claims 17 and 18 under § 103 depends on the loss of right under § 102(b), the rejection of claim 18 must be, reversed. The correctness of the rejection of claim 17 under § 102(b), based on the board’s statement that “[substantial identity is all that is required between the device on sale or in public use and the subject matter claimed for there to be a bar,” need not be considered, because we affirm the rejection of claim 17 under § 103.
[1402]*1402Appellant does not dispute the propriety of combining the teachings of prior art references with a device that was in public use or on sale prior to the critical date. The sole issue, therefore, is whether the evidence here is sufficient to show that the subject matter of claim 17 would have been obvious to one of ordinary skill in the art. In re Blaisdell, supra note 6, 242 F.2d at 781, 44 CCPA at 849. The references show that the use of screws to attach devices to a press was well known. We agree with the board that this would have been an obvious substitution for welding.
The decision of the board is affirmed on claims 1, 4-12, and 17, and reversed on claims 2, 3, 13-16, and 18.
Modified.