This opposition appeal deals with sufficiency of disclosure for a mechanical invention. Claim 1 of the main request in this opposition appeal read as follows:
A gear wheel with a plurality of teeth capable of meshing with the teeth
of another corresponding gear wheel, characterised in that the profile
of each tooth falls within a band of tolerance of ± 1 /20th of the depth
of the tooth (H) with respect to a theoretical profile similar to a
profile
defined by a natural spline function passing through a plurality
of nodal points having pre-established coordinates {X, Y}, expressed in
a system of Cartesian coordinates having their origin at the centre (O)
of the the pitch circle (P) of the gear wheel,
corresponding to tables 1
to 6, also given below, for gear wheels with a number of teeth equal
respectively to five, six, seven, eight, nine and ten:
The board is unhappy with this claim and the corresponding description for two reasons. It appears that the gear defined by a spline through these points would not produce a working gear, and that the claim puts an undue burden on the skilled person to select working gears within the band of tolerance.
Reasons for the Decision
2. Background
2.1
The patent concerns gear pumps having a driving and a driven gear
wheel, specification paragraph [0001] and [0002].
In this context the
gear wheels must mesh without encapsulation, that is without trapping
fluid in cavities between interlocked teeth, (specification paragraphs
[0004] and [0005]).
To this end claim 1 as granted proposes a gear
wheel having a plurality of teeth with particular tooth profiles
falling within a tolerance band about a theoretical tooth profile. The
theoretical tooth profile is similar to a profile defined by a natural
spline function passing through a plurality of nodal points tabulated
for five, six, seven, eight, nine and ten toothed gear wheels
respectively. The term "natural spline function" commonly refers to the
piece-wise polynomial function resulting from spline interpolation of a
set of known data points, where each of intervals defined between
successive data points is interpolated by a respective polynomial
function that passes through the interval's end points such that it is
twice differentiable in those points. For a natural spline function the
polynomials are cubic and together form a set of cubic polynomials, one
for each interval between nodes where the nodes define their end points,
and where adjacent polynomials have the same slope at their common
node.
2.2
The gear wheel claimed is also limited by the functional
feature that it must be capable of meshing with a corresponding gear
wheel. Giving the terms their normal meaning, gears mesh when, viewed
in profile, the flank of a driving gear's tooth profile contacts that of
the driven gear, at a single point which moves as the two flanks roll
over one another until they separate.
(...)
4. Sufficiency of disclosure
4.1
Article 83 EPC requires that the European patent application shall
disclose the invention in a manner sufficiently clear and complete for
it to be carried out by a person skilled in the art.
4.2 According
to established jurisprudence, an invention is sufficiently disclosed if
it can be
performed by a person skilled in the art in the whole area
claimed, using common general knowledge and taking into account further
information given in the description of the patent or patent
application, see the Case Law of the Boards of Appeal of the EPO, 7th
edition, 2013 (CLBA hereinafter) II.C. introductory portion.
Furthermore,
the skilled person must be able to perform the invention without undue
burden. Where the skilled person can only establish by trial and error
whether his choice of numerous parameters will provide a satisfactory
result, this amounts to an undue burden, see CLBA II.C.5.6 and 5.7.
4.3
As explained above, a key element of claim 1 is the theoretical tooth
profile (one each for 5-10 tooth gears) that is similar to the profile
defined by a natural spline function passing through the tabulated data
points.
The decision under appeal found that the resultant
profiles using all data points were so irregular that they would not
practically mesh, as required by the claim, nor prevent encapsulation.
It further held that if, as argued interpolation was only through
selected but not all of the tabulated points, this would put an undue
burden on the skilled person.
4.4 Contrary to what was argued
before the opposition division (cf. decision, page 7, second paragraph)
and until recently in appeal (cf. Meneghello, page 8, 2nd paragraph,
last sentence), the appellant in the oral proceedings before the Board
argues that the spline function profile does pass through all the points
of a given table in the claim, and that the resultant profile will mesh
and prevent encapsulation. Insofar as the profile passing through all
points is concerned the Board can agree with such a reading of claim 1.
This reading corresponds to the skilled person's understanding of
"natural spline function" (see above) in relation to "a plurality" of
nodal points which contextually can only be the plurality given in the
table. This point is thus no longer in dispute between the parties.
4.5
It is also undisputed that, based on the nodal points listed in claim
1, parts of the theoretical profiles claimed undulate, that is when set
out in a graph these data points produce a profile that is irregular,
i.e. that is not smooth but has considerable local variations. See for
example figures 1 to 3 in Meneghello, or the tabulated points in the
various figures of D18 Figure 1.
4.6
The only point in dispute is
thus whether or not the resultant profile and band of profiles as
defined in claim 1 can mesh, as required by the claim, and prevent
encapsulation. The respondent has consistently argued that a profile
passing through all points does not, and indeed from the appellant
proprietor's initial response in opposition and appeal that not all
points were meant it can be inferred that he initially agreed. In appeal
the appellant now however expressly states that the main profile
resulting from interpolation through all points does result in a meshing
and non-encapsulating gear.
4.7
Normally, as explained in CLBA
II.C.8, and in particular T 63/06, after the grant of the patent a legal
presumption exists that the patent meets the requirements of the EPC.
In order to establish insufficiency, the burden of proof is upon an
opponent to establish on the balance of probabilities that a skilled
reader of the patent, using his common general knowledge, would be
unable to carry out the invention.
However, when the patent does
not give any information as to how a feature of the invention can be put
into practice, the presumption of sufficiency of disclosure is weak.
The opponent can discharge his burden by plausibly arguing that common
general knowledge would not enable the skilled person to put this
feature into practice. The patent proprietor then has the burden of
proof for the contrary assertion that common general knowledge would
indeed enable the skilled person to carry out the invention.
4.8
In the present case the patent provides no information, by tests or
otherwise, that would support the claimed meshing and non-encapsulating
of the resultant gears. The patent, at specification paragraph [0011]
mentions experiments that led to the identification of the tooth
profiles, but the relevant data has not been included or provided
subsequently in opposition or the following appeal. The
respondent-opponent on the other hand has presented a plausible,
substantiated case why there would not be meshing and non-encapsulating.
Given further the varying positions of the proprietor-appellant in this
point the Board concludes that there can be only a weak presumption of
sufficiency in the present case, and that the onus now rests on the
proprietor to prove sufficiency. In weighing the parties' submissions
against each other he can therefore no longer enjoy the benefit of the
doubt.
4.8.1 In the respondent's submission of 2 July 2012 in
reply to the appeal, see page 5, reference is made to his submission in
opposition dated 15 December 2010, in particular figures 1 and 1b which
plot the data points of table 3 for a 7 toothed gear wheel (table 3 of
the claim) having a pitch circle of approximately 11 mm and an outer
diameter of approximately 19mm, that is for a gear wheel of typical
dimensions. Further figures 2 to 15 of the same submission show the
contact area between an enlarged section of the flank profile of the
driving gear with the flank and the corresponding identical driven gear
profile at various angles of rotation. It is evident that during
rotation the profiles cross and overlap each other due the irregularity
of the opposing flanks. He argues that since gear teeth cannot overlap,
at worst they would not fit together, at best they would not mesh but
touch at more than one point, thereby jamming and encapsulating fluid.
This
was reiterated in D18, page 18, first paragraph, and further
substantiated in D18 by calculating profiles and their so-called
conjugates using natural spline functions applied to limited sets of the
tabulated data points. A conjugate profile is derived from the main
profile under condition of conjugation, that is that the normal to the
instantaneous point of contact always intersects to centres of the two
gears, D18, page 16, first paragraph. These limited set conjugates
already exhibit unphysical anomalies, see D18, figures 5b, 6b, 7b, 8b/c,
9b/c and 10b/c, with interference as above also for these limited set
profiles, see D18 figures 10c,d,e and 11c, and page 39, 5th and 6th
paragraphs. Thus even for limited set approximations the irregularities
and anomalies persist.
4.8.2 The appellant has argued that the
irregularities in the profile are not significant enough to prevent
meshing or cause encapsulation. The profiles would be based on
measurements made on actual meshing and non-encapsulating gear wheels
and fluctuations would merely result from measurement error and/or
surface roughness.
The Board is unconvinced by this argument. Most
importantly, it is in direct contradiction with the Meneghello report
commissioned by the appellant himself, which at page 8, second
paragraph, states that it is "not possible nor permissible to perform an
interpolation passing through each of the digitized points [as] this
.... does not allow to obtain a profile that meets the functional
requirements" (in reference to the geometric-functional characteristics
including meshing and conjugation mentioned on page 2). The tests or
empirical data, or even the actual gear from which the measurement data
is said to derive, and which might have proven his case, are lacking.
Without such evidence the arguments that the irregularities would be
insignificant and the profiles mesh and not encapsulate are nothing more
than unproven assertions that can be of little weight.
4.8.3
Weighing the various submissions against each other - the respondent's
substantiated arguments on the one hand, the appellant's varying
positions and unsubstantiated assertions on the other - the Board
decides that a gear with a profile defined by a natural spline function
passing through all of the tabulated data points (or for that matter
only through a limited set) will not mesh and will encapsulate, and thus
fails to produce the desired, claimed effect of the invention. For this
reason already the claimed invention is insufficiently disclosed.
4.9
In addition, the Board notes that the claim 1 does not define a single
theoretical tooth profile (for each number of teeth), but also other
profiles that lie in a tolerance band of +/-1/20th of tooth depth with
respect to each theoretical profile. The band width corresponds to about
10% of tooth height. This figure, much larger than normal machining
tolerances, represents a range of profiles as explained elsewhere in the
specification and argued by the appellant: experiments on various gears
reveal a "range of tooth profiles" (paragraph [0005]); and "...the
actual tooth profile falls within a band of tolerance T ...", (paragraph
[0011] and figure 1).
As submitted by the appellant, the profiles
within this band are not different scaled versions of the basic profile
defined by the data. The tolerance band is rather meant to encompass an
unlimited number of tooth profiles of varying shape, as long as they
give gears that mesh. Only profiles that meet this important functional
limitation of the claim are to be protected.
The appellant
acknowledges that there are some within the band that do not meet this
requirement and should therefore be excluded from protection. Which
profiles do and which do not meet this requirement would then, so the
argument goes, need to be determined by trial and error.
4.9.1
The
question therefore arises as to whether or not such trial and error
does not constitute an undue burden on the skilled person for him or her
to successfully carry out the claimed invention, see section 4.2 above.
4.9.2
The patent is silent as to how to distinguish a working profile from
one that does not. One way would be to prototype, i.e. manufacture, the
corresponding gears and measure meshing and encapsulation. Another
alternative suggested by the appellant would be to generate the gears
mathematically, or model them, using a variety of analytical techniques
known to the skilled person, see e.g. the Meneghello report, page 8,
last two paragraphs.
4.9.3
Given the myriad of possible profiles,
and the effort involved in prototyping or modelling the corresponding
gear, this is a daunting task. In particular, the Board holds that this
would far exceed routine experimental work. For each possible profile
the skilled person must prototype or model two gear wheels and
investigate whether or not they mesh without encapsulation. This process
of trial and error in the Board's view represents an undue burden for
the skilled person.
4.10 In summary the Board concludes that on
the balance of probabilities the main gear tooth profiles defined in
claim 1, which are central to the invention, do not result in a working
gear wheel. Furthermore, the skilled person would not be able to
determine which of the remaining profiles within the tolerance band
defined in the claim would result in working gears without undue burden.
The invention according to claim 1 of the main request is thus
insufficiently disclosed, Article 100(b) EPC with Article 83 EPC.
This decision has European Case Law Identifier: ECLI:EP:BA:2014:T243111.20140924. The whole decision can be found here. The file wrapper can be found here. Photo by Sergei Golyshev
obtained via Flickr.
