Dimensions from patent drawings

EPO decision T 1488/10 considers a new twist on a hoary old question: can one rely on measurements taken from a prior art drawing to invalidate a patent? The twist in this case was that the drawing in question had some measurements indicated for the device shown, just not the particular ones that were needed to invalidate the claim under attack.

Background
If an inventor claims in a patent application that the length of the secondary doohickey is at least three times the sprocket pitch diameter, it can be hard to prove lack of novelty because there may be considerable difficulty in finding a piece of prior art that expressly discloses the same dimensional constraints.

Which line is longer? Sometimes you really
do need to take a ruler to the drawing.

Frequently therefore, an examiner or opponent who is unable to find the necessary lengths or their ratio described expressis verbis in a prior art reference, has taken a ruler to a prior art diagram and shown that a sufficiently lengthy secondary doohickey was visible alongside an appropriately miniscule sprocket. Should such measurements be regarded as part of the prior art disclosure?

The basic rule at the EPO is that stated in T 204/83 (OJ EPO 1985, 310):

Features shown solely in a drawing form part of the state of the art when a person skilled in that art is able, in the absence of any other description, to derive a technical teaching from them. Dimensions obtained merely by measuring a diagrammatic representation in a document do not form part of the disclosure.

In T 1664/06 the Board followed the same path, refusing to take accurate measurements from a drawing, and distinguished over T 748/91 in which a schematic drawing was relied on simply to infer that one dimension was evidently larger than another.

In decision T 451/88 a distinction was drawn between scaled construction drawings and the schematic drawings conventionally included in patent documents, the latter not being intended to allow manufacture of the product, and it being held (once again) that schematic drawings could not be used to derive a ratio between two dimensions.

So we have a rule that pure schematic drawings cannot be accurately measured, while scaled construction drawings can. But what about a drawing where some of the illustrated components carry information about their dimensions? Does this indicate that the drawing is to scale?

T 1488/10
In this latest decision, the prior art drawing below was at issue. The IPKat has reproduced the drawing and has added some cheery colours, at no extra cost above and beyond your usual monthly IPKat subscription.

There is a connector, shown in yellow, which is threaded top and bottom to the green and blue components. So the yellow connector can be notionally divided into three segments along its length: the upper and lower "clamped" portions, and the middle segment which is "free of connections" (in the words of the patent application under attack).

The prior art drawing: if L1 and L2 are known, what does
this tell us about the yellow connector's dimensions?

The novelty of the claim turned on the feature that "the segment of the connection which is free of connection constitutes between 50% and 75% of an overall length of the connector." A ruler measurement of the prior art drawing showed that the relevant segment constituted about 66% of the overall length, slap bang in the middle of the claimed range. The Examining Division refused for lack of novelty on this basis, despite being fully aware of the general rule that one should not take measurements from schematic drawings.

The reasoning of the Examining Division was that the general rule in T204/83 did not apply, as that decision was only concerned with features "shown solely in a drawing". If you look again at the drawing above you will see two lengths, L1 and L2 in red. Although not directly related to the length of our yellow connector, they are dimensions along the same axial direction, and the description of the document required that they be in a ratio of between 1:2 and 1:3 (in the drawing they are 1:2). The Division reasoned therefore:

Since Fig. 4 shows that L2 = 1/2 * L1 and because it is indicated in [0026] that L1 = 23mm and L2 = 12mm, it can be concluded that, at least in the longitudinal direction, the relative dimensions as shown in Fig. 4 can be considered as conforming to the relative dimensions of the real device.

In other words, the inclusion of actual dimensional constraints on a schematic drawing, especially along the same directional axis, were considered to transport that drawing from the schematic realm into the real world, and the drawing could be regarded then as dimensionally accurate.

The Board disagreed on appeal. After discussing the meaning of L1 and L2 (and also of the other indicated dimension, t) and the fact that these parameters were important to the performance of the apparatus being described, the Board turned to the components whose length was neither indicated in the drawings nor discussed in the description of the prior art document [Reasons for the Decision, 3.5]:

However, the dimensions of the other parts of the ultrasonic transducer shown in Figure 4 have to be distinguished from L1, L2, and t. First of all, they are not referred to by any reference signs in the Figure. Furthermore, they are not described as having any specific significance for the claimed invention. For example, no advantages are mentioned in the description concerning particular choices of the dimensions of these other parts.

Regarding the representation of the dimensions of these other parts of the device, factors like the clear visibility in the Figure of these parts may well have induced the draughtsman to represent their dimensions in a manner that does not correspond to their actual dimensions. Therefore, in relation to these other parts of the ultrasonic transducer, the board is of the opinion that dimensions obtained merely by measuring the schematic drawing in Figure 4 does not form part of the disclosure of document D3 (see T 204/83, point 7 of the Reasons).

Accordingly, the relative dimensions of the length of the stud 85 and the length of the connection-free part thereof, as obtained by carrying out measurements on Figure 4 does not constitute part of the disclosure of document D3. Hence, feature (ii) of claim 1 of the main request has not been disclosed in document D3.

While the IPKat thinks the decision is legally correct (and isn't that what matters?), and that any other decision would have been a poor precedent, he wonders what the inventors behind document D3 might have thought about the contention that they did not disclose a connector free from connections along 66% of its length. Anyone who has drafted a decent number of mechanical cases will get a feel for drawings that are dimensionally more-or-less accurate as opposed to purely schematic illustrations, and the figure illustrated above has the feel of being the former.

It also seems unlikely, given the uncrowded arrangement of components, that requirements of clear visibility would have necessitated the draftsman lengthening or shortening the connector to any major extent. But here we are entering the realm of speculation. The bottom line is that to invalidate a patent for lack of novelty you need a clear and unmistakeable disclosure of all claimed features, and so the Board got it right in terms of the standard of disclosure required.

In the end, the victory for the applicant on this point was Pyrrhic: although D3 did not anticipate, all of the claim sets were held to be obvious over another document entirely.