The Case: Bell Cups and Rotary Atomizers

Dürr and EFC both manufacture rotary atomizer equipment used in automotive spray painting. At the center of the dispute is a component called the bell cup — a spinning disc whose surface geometry controls how paint is distributed across a substrate. The shape of the bell cup's overflow surface directly determines whether paint atomizes evenly enough to produce consistent color matching. Minor geometric differences can produce significant coating non-uniformity.

Dürr held patents covering bell cup designs. EFC's products, Dürr alleged, fell within the scope of those claims. The dispute came down to the meaning of three terms: generally conical, substantially conical, and attached. The district court ruled for EFC on all three. The Federal Circuit reversed on all three.

Dispute One: "Generally Conical" and "Substantially Conical"

What the Claims Said

Dürr's claims described a bell cup overflow surface that was generally conical or substantially conical. These terms are textbook approximation modifiers — phrases added to claim language to provide a buffer against design-arounds based on minor geometric deviations from the idealized shape described in the claim.

The District Court's Error

 District Court — Reversed
The district court construed "generally conical" to require a geometrically clean conical slope — a surface with no undulations, curvatures, or deviations from a true cone. Because EFC's overflow surface had a slight curvature, it did not satisfy the claim term. The court further invoked prosecution history to support this narrow construction, reasoning that the applicant had, during prosecution, distinguished its design from prior art that featured an uneven overflow surface.

Two specific errors compounded the problem. First, the district court pointed to a preferred embodiment showing a component labeled element 144 — described as a "generally conical rear surface" — alongside element 142, described as a "rounded central rear surface." The contrast between these two labeled elements appeared to support the court's view that a conical surface, by definition, could not have any rounding.

Critical Component Misidentification

The district court's cited embodiment — elements 110, 142, and 144 — describes the deflector plate, not the overflow surface (element 126). These are different functional components of the bell cup assembly. The geometric constraints applicable to the deflector plate say nothing about the permitted geometry of the overflow surface. The district court's entire evidentiary foundation for the narrow construction rested on the wrong part of the specification.

Second, the prosecution history argument was overbroad. During prosecution, the applicant had argued that its overflow surface was different from the prior art — specifically, a design covered by the Schneider patent, which featured an overflow surface with pronounced, irregular undulations and a sharp raised lip at the spray edge. The district court read this argument as a blanket disclaimer of any curvature in an overflow surface.

The Federal Circuit's Correction

 Federal Circuit — Reversed
Approximation modifiers like generally and substantially exist specifically to prevent a claim from being construed to require mathematical or geometric exactitude. Reading them to demand absolute precision — a geometrically perfect cone — renders the modifier meaningless. A construction that strips a claim term of its operative function is presumptively wrong.

On the prosecution history issue, the Federal Circuit applied a familiar but frequently misapplied principle: the scope of a disclaimer extends only as far as the statements made during prosecution actually go. The applicant had argued that its design was different from the Schneider prior art's dramatically undulating overflow surface — one with large-amplitude surface variations and a protruding lip edge. That argument disclaims the extreme case (severe undulations, sharp lips). It does not disclaim all curvature across the board.

The scope of prosecution disclaimer is measured by what the applicant actually said, not by the broadest inference an adversary can draw from it. A distinction from extreme undulation does not amount to a covenant that the protected surface must be geometrically perfect.

This is a clean articulation of a principle that trial courts consistently underweight. Prosecution history estoppel and disclaimer doctrine are tools for enforcing the bargain the applicant made with the examiner — no more, no less. Stretching a targeted argument into a global disclaimer converts a shield into a trap that swallows the claim.

Dispute Two: "Attached to the Bell Cup"

What the Claim Said

A second claim element described a rear cover attached to the bell cup. EFC's accused product had a rear cover that was integrally formed with the bell cup body — manufactured as a single piece rather than as two separate parts assembled together.

The District Court's Reasoning

 District Court — Reversed
"Attached to" requires two separately manufactured components that are joined together. Integral one-piece construction does not satisfy this limitation because there is no discrete attachment event — nothing is fastened to anything. Two supporting arguments: (1) a dependent claim explicitly required the rear cover to be welded to the bell cup, implying that the independent claim requires at least some discrete joining; (2) every preferred embodiment in the specification showed the rear cover as a separate component mounted onto the bell cup.

The Federal Circuit's Correction

 Federal Circuit — Reversed
"Attached to" describes a structural relationship between components, not a manufacturing method. Whether two structural regions are joined by welding, bonding, press-fitting, or integral machining from a single billet is a manufacturing choice that does not alter the underlying structural relationship. If the rear cover region and the bell cup body region are structurally connected in the accused product, the "attached to" limitation is satisfied regardless of how that connection was created.

The Federal Circuit also deployed the dependent claim in exactly the opposite way from the district court — and more logically. The district court had used the dependent claim's welding requirement to argue that the independent claim also required some discrete joining step. The Federal Circuit used it to argue the opposite: a dependent claim that adds "welded" to the independent claim's "attached to" would be entirely redundant if "attached to" already required welding. The dependent claim's existence is affirmative evidence that "attached to" in the independent claim carries no welding requirement.

This is the doctrine of claim differentiation — presumed to give each claim independently meaningful scope — applied in its most straightforward form. When a dependent claim narrows the independent claim in a specific way, that narrowing is presumed to have been intentional. Reading the restriction back into the independent claim renders the dependent claim superfluous and defeats the presumption.

What This Decision Means in Practice

The core holdings in Dürr v. EFC are not new law. What makes the decision worth studying is that it illustrates — in a single case — four recurring failure modes in claim construction analysis: modifier stripping, specification-claim conflation, prosecution disclaimer overreach, and inverted claim differentiation. Each is common. Each is avoidable.

01
Use Approximation Modifiers Deliberately
Terms like generally, substantially, approximately, and about are not hedging language — they are structural claim tools that create a geometric buffer against design-around arguments. For shape, angle, and positional parameters, draft with these modifiers when minor deviations from the geometric ideal should not determine infringement.
02
Describe Relationships, Not Processes
For component assembly claims, use state-describing language — attached to, connected to, coupled with — rather than process-limiting language like welded to or bonded to. Manufacturing method is the accused infringer's choice; it should not be allowed to become the claim boundary by default.
03
Scope Prosecution Arguments Precisely
Every argument made during examination is a potential disclaimer. When distinguishing prior art, identify the specific characteristic being distinguished — do not make global statements about category. Saying "the prior art has dramatically undulating surfaces" is narrower than "our surface is smooth." The former forecloses one extreme; the latter may foreclose everything in between.
04
Read Embodiments Against the Right Components
Specification embodiments describe specific embodiments of specific elements. They do not establish geometric requirements for other, structurally distinct elements in the same assembly — even if those elements are nearby in a drawing. Misidentifying which embodiment governs which claim element is a foundational error that invalidates the entire construction analysis.

The Broader Pattern

The pattern across both disputed terms in Dürr v. EFC is the same: the district court confused specificity with precision. In each case, the court took language that was deliberately drafted to be inclusive within a range — a roughly conical shape; components that are structurally joined — and read it to require a specific, narrow limiting condition. The Federal Circuit's reversals on both points are a consistent application of a core principle: ambiguous claim terms should be construed to preserve, not to narrow, the scope the patentee actually claimed.

For practitioners advising clients on claim drafting or litigation position, the takeaway is simple: the words you choose at prosecution are the boundaries you will defend at trial. Approximation modifiers protect against geometric design-arounds. Structural relationship terms protect against manufacturing design-arounds. Scoped prosecution arguments protect against disclaimer overreach. None of these require exotic drafting — they require deliberate drafting.

The bell cup case is a useful anchor because the technology is concrete: the geometry matters, the component identification matters, the assembly method matters. That concreteness makes the doctrinal lessons easier to carry into cases involving more abstract claim language — software architectures, chemical compositions, biological assays — where the same analysis applies but the stakes of imprecision are often higher.