Homeland Housewares, LLC (“Homeland”) petitioned the United States Patent and Trademark Office Patent Trial and Appeal Board (“Board”) for an inter partes review of claims 1-16 of U.S. Patent No. 7,581,688 (“ ’688 patent”), which is assigned to Whirlpool Corporation (“Whirlpool”). The Board did not construe the key term “settling speed” found in the claims and determined that the claims were not invalid as anticipated by prior art reference U.S. Patent No. 6,609,821 (“Wulf”). Homeland appeals. We reverse.
Background
The ’688 patent relates to household blenders. The invention claimed in the ’688 patent is a pre-programmed, automated blending cycle designed to blend items “quickly and reliably—by repeatedly dropping to a speed slow enough to allow the blender contents to settle around the cutter assembly, and then returning to a [higher] speed suitable for processing the contents.” Appellee Br. 4.
As admitted in the ’688 patent itself, it was well-known that a user could manually pulse between a high speed and a low speed to “achieve[ ] ... a pattern of movement that introduces the entire contents of the reservoir into contact with the rotating blades” for efficient mixing. ’688 patent, col. 1 11. 20-23; see also id. at col. 6 11. 46-50. Thus, the claimed automatic blending routine was, in the prior art, done manually. There were also blenders on the market which allowed “preprogram[ing] ‘on-off se-quencefs] [to] enable[] hands-free operation of the blender.” Id. at col. 1 11. 38-39.
An independent claim at issue, claim 1, provides:
*1374A cycle of operation for a blender comprising a motor, a container for holding items for processing, and a cutter assembly located within the container and operably coupled to the motor whereby the motor effects the rotation of the cutter assembly, the cycle comprising: automatically controlling a rotational speed of the cutter assembly to effect a pulsing of the speed of the cutter assembly wherein each pulse comprises:
(A) a constant speed phase, where the operating speed of the cutter assembly is maintained at a predetermined operating speed,
(B) a deceleration phase, where the speed of the'cutter assembly is reduced from the operating speed to a predetermined settling speed indicative of the items in the container having settled around the cutter assembly, which is less than the operating speed and greater than zero, and
(C) an acceleration phase, where the speed of the cutter assembly is increased from the settling speed to the operating speed.
Id. at col. 7 11. 4-23 (emphases added). Claim 1 is representative and there are no patentability distinctions offered here with respect to the other claims of the ’688 patent.1
On June 2, 2014, Homeland petitioned the Board for an inter partes review of claims 1-16 of the ’688 patent, seeking a construction of “settling speed,” and arguing, inter alia, that the claims at issue are invalid due to anticipation by Wulf. Like the ’688 patent, Wulf noted that it was well-known that manually “[p]ulsing the motor ... at high and then low speeds permits the material being blended to fall back to the region of the cutting knives[,] thereby improving the blending or mixing of the material.” Wulf, col. 111. 36-39. Wulf notes that this manual “process can be very frustrating,” id. at col. 2 1. 20, and thus teaches “a blender ... that is programmed to [automatically] accomplish predetermined [blending] functions and routines,” id. at col. 211. 25-27.
In its Final Written Decision, the Board declined to provide a construction of “settling speed” and concluded that Homeland had “not shown, by a preponderance of the evidence, that any claim of the ’688 patent is anticipated by Wulf.” J.A. 7,14.
Homeland appeals. We have jurisdiction under 28 U.S.C. § 1295(a)(4).
In appeals from inter partes reviews, we review the Board’s conclusions of law de novo and the Board’s findings of fact for substantial evidence. Microsoft Corp. v. Proxyconn, Inc., 789 F.3d 1292, 1297 (Fed. Cir. 2015). Claim construction is an issue of law that we review de novo where, as here, there is no relevant extrinsic evidence. Teva Pharm. USA, Inc. v. Sandoz, Inc., — U.S. -, 135 S.Ct. 831, 841, - L.Ed.2d - (2015).
Discussion
I
Anticipation is a two-step analysis. The first step is properly interpreting the claims. Beachcombers v. WildeWood Creative Prods., Inc., 31 F.3d 1154, 1160 (Fed. Cir. 1994). The second step is determining whether the limitations of the claims, as properly interpreted, are met by *1375the prior art. Id. The Board determined that Wulf did not anticipate the ’688 patent because its disclosures did not meet the “settling speed” limitation. J.A. 14. However, the Board did “not adopt any explicit construction of the term for [its] Final Written Decision,” J.A. 7, even though the parties disagreed as to claim construction.2 Just as district, courts must, “[w]hen the parties raise an actual dispute regarding the proper scope of ... claims, ... resolve that dispute,” 02 Micro Int’l Ltd. v. Beyond Innovation Tech. Co., 521 F.3d 1351, 1360 (Fed. Cir. 2008), the Board also must resolve such disputes in the context of IPRs. See CSR, PLC v. Skullcandy, Inc., 594 Fed.Appx. 672, 678 (Fed. Cir. 2014) (holding that “[t]he Board erred by failing to construe ‘threshold value’ as it is used in claims 1-6 before finding that [prior art reference] Smith failed to disclose a ‘threshold value’ ” in anticipation). Given that the Board did not rely on extrinsic evidence here as to claim construction, we can determine the correct construction of “settling speed” and then determine whether the Board correctly held that Wulf does not meet the limitations of claim 1. Teva, 135 S.Ct. at 841.
“[T]he claim construction inquiry ... begins and ends in all cases with the actual words of the claim.” Renishaw PLC v. Marposs Societa’ per Azioni, 158 F.3d 1243, 1248 (Fed. Cir. 1998) (citations omitted). Here, the relevant language of claim 1 provides that during pulsing, “the speed of the cutter assembly is reduced from the operating speed to a predetermined settling speed.’”688 patent, col. 7 11. 15-17 (emphasis added).
Whirlpool proposes that “a predetermined settling speed” means “a speed, greater than zero, that indicates that items have settled around the cutter assembly.” Appellee’s Br. 43. At times on appeal, Whirlpool argues that empirical testing is required to establish a settling speed. Whirlpool recognizes that empirical testing would require determining the settling speed for each individual blender and its content load, “[b]ecause so many factors affect the settling speed.” Appellee’s Br. 9; see also id. at 45; Oral Arg. 18:16-25 (skilled artisans looking at the ’688 patent would “perform tests to determine ... at what point in time [the blender ingredients] settles to arrive at the predetermined settling speed”).3 We conclude that a construction that would require empirical testing is incorrect. Indeed, the dissent also does not endorse a claim construction that requires empirical testing.
• The words of a claim are generally given their ordinary and customary meaning. Phillips v. AWH Corp., 415 F.3d 1303, 1312 (Fed. Cir. 2005). In some cases, the ordinary meaning of claim language may be readily apparent and claim construction will involve little more than the application of the widely accepted meaning of commonly understood words. Id. at 1314.
Here, it is undisputed that the plain meaning of “predetermined” is to determine beforehand. This plain language defi*1376nition does not require that a predetermined speed be empirically determined for each use, depending on the particular blender or the individual contents of the blender.
Claims must also be read in view of the specification, of which they are a part. Phillips, 415 F.3d at 1315. While the specification refers to an embodiment of the invention in which “a predetermined settling speed” is empirically determined and varies depending on blender use, the process for empirically determining a settling speed is neither taught in the specification nor a part of the claims. The claim language only requires “a predetermined settling speed,” and does not require empirically determining a particular settling speed for a particular blender or a particular blender load.
Whirlpool argues that “a predetermined settling speed” should be defined in light of the specification because the specification suggests that settling speed can be empirically determined. The specification states that the
predetermined settling speed ... [is a] function! ] of the blender motor size, the cutter assembly configuration, the container size and configuration, the properties such as hardness and viscosity of the items to be processed in the blender, and the like.... Thus, these speeds and time periods will vary for different blenders, and must be determined empirically for a particular blender.
’688 patent, col. 5 11. 18-27 (emphasis added). But this suggestion cannot define the scope of the claim, since it provides no meaningful definition of an empirically de: termined settling speed other than with respect to a single example (relating to crushed ice). Of course, “particular embodiments appearing in the written description will not be used to limit claim language that has broader effect.” Innova/Pure Water, Inc. v. Safari Water Filtration Sys., Inc., 381 F.3d 1111, 1117 (Fed. Cir. 2004). And Whirlpool in its brief at one point appeared to concede that settling speed does not require empirical testing. See Appellee Br. 45 n.8 (“Although Homeland argues that Whirlpool’s construction requires that the settling speed be determined empirically for a particular blender, Whirlpool’s construction does no such thing.” (quotation marks and citation omitted)).
The definition proposed by Homeland is also incorrect. Homeland proposes that “settling speed” means any comparative low speed less than the operating speed. But not every lowering in speed will cause settling. By way of example, as the appel-lee points out, if the operating speed is 6000 rpm and it was lowered to 5900 rpm, the contents will not necessarily settle.
Under these circumstances, we, of course, may adopt a definition not proposed by either party that best fits with the claim language and specification. See Exxon Chem. Patents, Inc. v. Lubrizol Corp., 64 F.3d 1553, 1556 (Fed. Cir. 1995) (“[T]he judge’s task is not to decide which of the adversaries[’ constructions] is correct. Instead the judge must independently assess the claims, the specification, ... and declare the meaning of the claims.”).
The broadest reasonable construction of “a predetermined settling speed” is a speed that is slower than the operating speed and permits settling of the blender contents. This is consistent with the ordinary and customary meaning of the words of the claim, as discussed above, and with the specification. Phillips, 415 F.3d at 1312, 1315. For example, this is in accord with the specification, which refers to the illustrative settling speed at 1120 rpm as a “speed that slows significantly enough to allow the contents to reach the ‘settled’ condition to be processed again.” ’688 patent, col. 6 11. 16-18. This is also the correct construction because it is “the broadest *1377reasonable interpretation consistent with the written description.” In re NTP, Inc., 654 F.3d 1268, 1274 (Fed. Cir. 2011) (quoting In re Baker Hughes, Inc., 215 F.3d 1297, 1301 (Fed. Cir. 2000)). Here, unlike in Exxon, 64 F.3d at 1556, our claim construction is not new, but simply represents a midpoint between the two opposing constructions now urged by the parties. Significantly, this construction is also practically identical to the construction utilized by the Board in its Institution Decision, see J.A. 314, and the overall construction urged by the patentee’s expert, see J.A. 406, 412, apart from the requirement for empirical testing.
II
Based on this construction of “settling speed,” we conclude that the Board erred in finding that Figure 25 of Wulf does not anticipate the ’688 patent.
Figure 25 of Wulf discloses the following automated blender routine:
[[Image here]]
The Board found that Homeland’s anticipation theory “is not without appeal,” but also found that Homeland failed to present evidence showing Figure 25 of Wulf anticipates the ’688 patent. J.A. 13-14. In order to evaluate whether Figure 25 anticipates, we must determine what Figure 25’s “low” speed means. For this, we look to Wulfs specification, which teaches two relevant characteristics of “low” speeds. “[W]hen a patentee uses a claim term throughout the entire patent specification, in a manner consistent with only a single meaning, he has defined that term ‘by implication.’ ” Bell All. Network Servs., Inc. v. Covad Comm’ns Grp., Inc., 262 F.3d 1258, 1271 (Fed. Cir. 2001).
First, in Wulf, the term “low speeds” appears only once in the specification, where the background teaches that “[p]uls-ing the motor on/off or at high and then low speeds permits the material being blended to fall back to the region of the cutting knives thereby improving the blending or mixing of the material.” Wulf, col. 1.11. 36-39. This is consistent with the testimony of the Whirlpool expert, who stated that “slower speeds will tend to allow items to settle, while higher speeds •will tend to ... keep[ ] items suspended above the cutter assembly.” J.A. 409-10. Because Wulf uses “low speeds” to refer to speeds at which blending ingredients fall back to the cutters, we conclude that Figure 25’s use of that same term should be understood in the same manner.
Second, Wulfs specification uses the term “low” or “lower” in the context of motor speed discussions in four additional places. See Wulf, col. 1 11. 27-31; col. 1 11. 41-43; col. 15 11. 32-46 (discussing Figs. 28-30); col. 19 11. 57-60. In each discussion, Wulf makes clear that a “low” speed is discretely and significantly different from a “high” speed. For example, Figure 30 shows that there are only two buttons— low and high—that a user may press. J.A. 153. In another example, Wulf teaches that “speeds in a low range [in the prior art] are obtained by applying only half ... voltage to the motor.” Wulf, col. 111. 28-29. Therefore, we conclude that Figure 25’s *1378use of “low” should also be understood to indicate a speed that is significantly different from a “high” speed.
In light of these teachings from the Wulf specification, Figure 25 discloses a settling speed limitation consistent with our construction, as well as the other elements of the pulsing cycle in claim 1. Specifically, claim l’s “constant speed phase, where the operating speed of the cutter assembly is maintained at a predetermined operating speed,” is found in Wulf Fig. 25’s “forward high speed 15 seconds.” Claim l’s “deceleration phase, where the speed of the cutter assembly is reduced from the operating speed to a predetermined settling speed indicative of the items in the container having settled around the cutter assembly, which is less than the operating speed and greater than zero,” is found in Wulf Fig. 25’s “ramp down to low, low for 5 seconds.” And Claim l’s “acceleration phase, where the speed of the cutter assembly is increased from the settling speed to the operating speed” is found in Wulf Fig. 25’s “ramp to high, high for 15 seconds.” Figure 25 thus contains every limitation found in the ’688 patent’s claim 1.
Ill
The Board also found that Homeland “left the testimony of Patent Owner’s witness, Mr. Faerber, unrebutted. Under such circumstances, [it was] unwilling to discount Mr. Faerber’s testimony that Wulf’ does not anticipate. J.A. 14. However, we must disregard the testimony of an expert that is plainly inconsistent with the record, NantKWest, Inc. v. Lee, No. 15-2095, 686 Fed.Appx. 864, 874-75, 2017 WL 1735330, at *9 (Fed. Cir. 2017), or “based on an incorrect understanding of the elaim[s],” Cordis Corp. v. Boston Sci. Corp., 658 F.3d 1347, 1357 (Fed. Cir. 2011). That is the situation here, where Faerber makes several incorrect statements with respect to the record, and in one respect, adds an additional claim requirement.
Faerber first states that ‘Wulf gives no indication of whether any of these speeds ... would cause items to settle around the cutter,” and that Wulf “only discloses items settling when the motors stops.” J.A. 430-31. This is clearly contradicted by the record, as Wulf teaches that “[p]ulsing the motor ... at high and then low speeds permits the material being blended to fall back to the region of the cutting knives.” Wulf, col. 111. 36-38. s
Faerber then states that with respect to the- routine disclosed in Figure 25, the “low” speed cannot be the settling speed because it “is maintained for at least five seconds ... [and] there is no reason to maintain a settling speed for so long.” J.A. 430. The ’688 patent claims do not contain any limitations with respect to how long the settling speed needs to be maintained. The specification teaches that these “time periods will vary for different blenders, and must be determined empirically for a particular blender.” ’688 patent, col. 5 11. 25-27. In fact, in the sole example provided, “4 seconds represents a predetermined deceleration time period during which solid particles have accumulated into the ‘settled’ condition.” Id. at col. 4 11. 23-25. Faerber thus incorrectly imports a limitation into the claims.4
Faerber finally states that because Figure 25 is entitled “Powdered Drinks,” “even very low speeds will cause the blender contents to circulate, rather than settle. Thus, one of ordinary skill would not expect Figure 25 to disclose a settling speed.” J.A. 430-31., When all of “the structural limitations recited in [the *1379claims] are all found in the [prior art] reference[,] ... the absence of a disclosure relating to [intended] function does not defeat ... [a] finding of anticipation.” In re Schreiber, 128 F.3d 1473, 1477 (Fed. Cir. 1997). That Figure 25 contains all of. the structural limitations of claim 1, according to the correct construction, completes the anticipation analysis, regard-less of intended use.
“[A] court should discount any expert testimony that is clearly at odds with ... the written record of the patent.” Phillips v. AWH Corp., 415 F.3d 1303, 1318 (Fed. Cir. 2005) (quotation marks omitted). Here, we conclude that at least these portions of Faerber’s expert testimony are inconsistent with the intrinsic record and therefore should be discounted.
IV
Having concluded that all the claims of the ’688 patent are invalid as anticipated by Wulf, we need not address the Board’s obviousness determination with respect to the Kolar prior art reference, U.S. Patent No. 6,364,522.
REVERSED
Costs
No costs.
. The other independent claim, claim 9, is a method claim for what is claimed in claim 1. The dependent claims cover minor features, including the cycle being repeated (claims 2 and 10), maintaining the operating speed for an unspecified, predetermined time period (claims 3, 4, 5, 11, 12, 13, and 14), decelerating the cutter in a continuous manner (claims 6 and 15), reducing operating speed of the cutter to allow settling (claim 8), and decelerating the cutter by terminating the power to the motor (claims 7 and 16).
. The dissent urges that the Board adopted a construction of "settling speed" in its Institution Decision as "a speed at which the cutter assembly has slowed enough to allow the blender contents to be processed again.” Dissenting Op. 1380. However, in its Final Written Decision, the Board specifically held that it did not adopt any explicit construction of the term "settling speed.” See J.A. 7.
. Whirlpool also took this position that empirical testing is required before the Board. See Appellant Br. 18 (Whirlpool counsel arguing that "[t]he patent does talk about this concept of needing to empirically test and determine what the settling speed is for a specific blender and it's specifically depending in part on the shape of the container, the type of blender it is and the contents, the expected contents as an example”).
. Contrary to the dissent, in holding that Wulf does not anticipate, the Board never made any findings with respect to the length of the settling speed as a part of the claimed invention, nor approved that aspect of the Faerber opinion.