- The patent in suit,
EP(UK) 0309596, is entitled 'Pumping apparatus for delivering liquid at
high pressure'. The proprietor is the first claimant, Hewlett-Packard GmbH.
The second claimant, Agilent Technologies Deutschland GmbH, is an exclusive
licensee under the patent. I do not need to distinguish them and I shall
refer to them as 'Agilent'. The defendants, Waters Corporation and their
English subsidiary Waters Ltd, petitioned to revoke the patent in suit on
the 17th September, 1997. I shall refer to them as 'Waters'. The patentee
complained of infringement by Writ served on 4 August 1998, and the two
sets of proceeding were dealt with as a single action for infringement of
patent with a counterclaim for revocation.
- Agilent applied to
amend the patent in suit. This amendment was not opposed by Waters at the
trial, and was not opposed by any other opponent. No argument was addressed
to me on it. The only amendments are (1) to claim 10, and (2) the removal
of the words 'a preferred embodiment of' in the general description. The
former amendment was obviously unobjectionable (although it would not have
saved the patent if claim 1 had been invalid) and the latter is intended
to remove any suggestion that the invention contemplates anything other
than a particular adjustment of the stroke of the pump with flow rate. This
amendment brings the specification into conformity with the characterising
portion of claim 1, and is equally unobjectionable. I shall accordingly
allow the amendments.
The
patent in suit
- Notwithstanding the
generality of its title, the patent in suit is concerned with a pump suitable
for use in a particular analytical technique used in chemistry. This technique
is called high-performance liquid chromatography or HPLC. In HPLC, a sample
of material to be analysed is dissolved in a solvent. The solution is then
injected into a stream of solvent which is passing over the contents of
a column called a 'separation column'. The idea is that the different compounds
present in the sample adhere to the material with which the column is packed
with different degrees of enthusiasm, and are washed along the column at
different rates. The result is that different components of the sample appear
at the far end of the column at different times, and, by using well characterised
mixtures the column can be calibrated to permit the analysis of the unknown
sample. The technique is described in more detail in the primer which the
parties have made available, and which has proved to be of some assistance,
although in this case it was clear from its comparatively narrow scope that
the parties considered it difficult to agree on, for example, the common
general knowledge at the priority date concerning HPLC pumps.
- HPLC requires pumps
which can provide a smooth flow of solvent through the column. The pressure
required to achieve the desired solvent flow can be as high as 400 bar.
The desired flow rate can be very small, ranging from 20 microlitres (µl)
per minute to 5 millilitres (ml or cc) per minute. The flow rate must be
steady because variations between different compounds in retention times
in the column themselves vary with the flow rate. Accordingly, a very uniform
flow rate is desirable, and the evidence was that this was recognised throughout
the art as one of the aspects of existing HPLC pumps in need of improvement
at the priority date.
- At column 1 line 14,
the specification says this:
'In
the chromatographic analysis, it is desirable that the flow rate of the
liquid delivered to the column is adjustable over a wide range of flow rates.
It is furthermore desirable that the solvent delivery system permits the
generation of mixtures of solvents and to change the mixing ratio of the
various solvents of the mixture in the course of time (gradient operation).
Such versatility of the solvent delivery system permits to optimise the
analysis conditions for the specific sample to be chromatographically separated.
'Although
the flow rate should be adjustable, it is very important that a once adjusted
flow rate is kept as constant as possible. If the flow rate through the
separation column would fluctuate, variations in the retention time of the
examined sample would occur so that the areas of the chromatographic peaks
produced by a detector connected to the outlet of the column, e.g., an absorption
detector, a fluorescence detector, or a refractive index detector, would
vary. Since the peak areas are representative of the concentration of the
chromatographically separated substances, fluctuations in the flow rate
would impair the accuracy and the reproducibility of quantitative measurements.'
- It may be convenient
to note here that the language of the specification reveals that it has
been translated rather clumsily from an original text, perhaps in German.
The specification proceeds to examine the difficulties associated with existing
pump types.
'Some
pumping systems like reciprocating pumps with a single piston have inherent
flow variations because the piston delivers only during a portion of a pump
cycle. To reduce such pump pulsations, it is known to use a dual piston
pump having two interconnected pumps each with a reciprocating piston.'
- Herr Riggenmann, the
expert who gave evidence on behalf of Waters, said that at the priority
date of the patent in suit there were four types of HPLC pumps commercially
available, of which dual and triple headed reciprocating piston pumps were
one type. He referred to a book entitled 'Practice of high performance liquid
chromatography' by Engelhart, and in particular the chapter by Colin entitled
'Liquid Chromatographic Equipment'. Colin is particularly clear about the
need for a stable flow rate. Herr Riggenmann puts Colin forward as representing
the common general knowledge in the art. I do not believe that this was
really challenged. Colin observes that flow instabilities are created by
a pump action and he deals in particular with the periodic refill of the
pump head. He says
'Other
instabilities are associated with the periodic refill of the pump head.
During this refill and the time needed to sufficiently compress the solvent
for opening the outlet valve, no flow is delivered from this head. This
results in flow and pressure pulsations if no means are taken to compensate
for the flow variations from this head. The resulting effect of these pulsations
will depend on their amplitude (which is related to the number of
pump heads, to the piston displacement profile, to the possible use of flow
feedback loops, to the ratio of the volume of the pump head to the volume
displaced by the piston, to the solvent compressibility and seal elasticity),
on the efficiency of their damping (which is related to the ratio
of the system volume to the stroke volume, to the column pressure drop,
the solvent compressibility and the possible use of damping devices) and
on their time (or volume) period compared to the peak retention
time (or volume) and peak width.
- Colin points out that
if several cylinders are used in the pump, the pulsation frequency is a
small multiple of the individual piston frequency, usually 2 for duplex
and 6 for triplex pumps. He identifies two types of double-header pump.
These are reciprocating pumps with two parallel heads and reciprocating
pumps with two heads in series. As the claim makes clear in features a),
b) and c), it is concerned with pumps which possess two heads arranged in
series.
- After an acknowledgement
of cited prior art, the specification continues:
'At
the high pressures encountered in high performance liquid chromatography,
compressibility of the solvents becomes noticeable resulting in an additional
source for flow pulsations. The reason is that during each compression cycle
of the pump, the first piston has to move a certain path to compress the
liquid to its final delivery pressure before actual delivery of liquid starts.
As a consequence thereof, pulsations in the outflow occur at the pump frequency.
These flow pulsations are particularly disturbing at low flow rates. The
reason is that the percent magnitude of pulsations remains substantially
constant over a wide range of flow rates but that the amplitudes of the
peaks in the chromatogram become smaller when the flow rate is reduced,
in particular when smaller separation columns are used, so that the influence
of the flow pulsations on the chromatographic results is more pronounced
at lower flow rates.
…
Relative
to this prior art, it is an object of the invention to provide a pumping
apparatus for delivering liquid at high pressure according to the preamble
of claim 1 which has a simpler mechanical design and which substantially
avoids over a wide range of flow rates the problems caused by interferences
of pulsations of the flow of the delivered liquid with the chromatographic
measuring results.
- The specification indicates
that the problems identified in the passage which I have quoted above are
solved by the characterizing features of Claim of 1. Claim 1 is as follows,
omitting the reference numerals:
'A pumping
apparatus for delivering liquid at a high pressure at which compressibility
of the liquid becomes noticeable, and at a selectable flow rate, comprising
a) a
first piston for reciprocation in a first pump chamber, the first pump chamber
having an inlet port and an outlet port,
b) a
second piston for reciprocation in a second pump chamber, the second pump
chamber having an inlet port and an outlet port,
c) a
conduit connection between the outlet port of the first pump chamber and
the inlet port of the second pump chamber,
d) an
inlet valve connected to the inlet port of the first pump chamber for allowing
flow of liquid into the first pump chamber and for inhibiting flow in the
opposite direction,
e) an
outlet valve connected to the outlet of the first chamber for allowing flow
of liquid into the second pump chamber and inhibiting flow in the opposite
direction,
f) drive
means for reciprocating the first and second piston,
g) wherein
the liquid in the first pump chamber is compressed to a high pressure before
delivery of the compressed fluid into the second pump chamber,
Characterised
by
control
means coupled to the drive means for adjusting the stroke lengths of the
pistons between their top dead centre and their bottom dead centre, respectively,
in response to the desired flow rate of the liquid delivered at the outlet
of the pumping apparatus, with the stroke volume (i.e., the amount of liquid
displaced during a pump cycle) being decreased when the flow rate is decreased
and vice-versa, such that pulsations in the flow of the liquid delivered
to the output of the pumping apparatus are reduced.'
- The specification continues
with a commentary upon the characterizing feature of the claim.
'In
order to see how the provision of an adjustable stroke volume leads to a
reduction of the flow pulsations, the following is to be considered:…'
- This introduction suggests
that the principal feature of this invention is thought by the inventor
to lie in the provision of an adjustable stroke volume. The particular apparatus
used to drive the piston so as to facilitate an adjustable stroke volume
also forms part of the invention, but one that seems to me to be of secondary
importance. The specification continues:
'In
known solvent delivery systems, the flow rate is changed by changing the
frequency of reciprocation of the pistons so that the pistons move at a
higher frequency when a higher flow rate is selected, whereas the stroke
volume remains the same when the flow rate is altered. According to the
present invention, however, the flow rate is changed by changing both the
frequency of reciprocation of the pistons and the stroke volume. In a preferred
embodiment of the invention, the stroke volume is decreased with the flow
rate. Thus, when the stroke volume becomes smaller, the volume which has
to be compressed to the final pressure before delivery starts also becomes
smaller. Since the volume to be compressed is smaller, the compression phase
becomes shorter resulting in smaller pulsations in the outflow of the pump.
'It
is a further consequence of the variation of the stroke volume as a function
of the flow rate that, particularly at low flow rates, the frequency of
reciprocation of the pistons is higher than in a prior art pump having a
fixed stroke for all flow rates. This increase in the frequency of reciprocation
leads to a corresponding increase in the frequency of any remaining pulsations
of the pump output which has advantageous effects on the reproducibility
of quantitative chromatographic measurements. In contrast to low-frequency
pulsations which may affect the retention times and areas of different peaks
in the chromatogram in different ways, high-frequency pulsations are more
like a uniform background signal which affects the whole chromatogram in
substantially the same manner. The increase of the frequency of the pulsations
is particularly advantageous when a detector is used which is very sensitive
to flow pulsations, e.g., a refractive index detector.
- This effect is illustrated
also in Figure 6 of the specification, in which a contrast is drawn between
curve 'c', which is said to be the curve of percentage pulsation against
flow rate for a pump according to the invention, and curve 'd' which is
said to be the percentage pulsation of a pump with a constant volume piston.
- Agilent identify five
issues of construction which arise in relation to the characterising feature
of the claim. These are as follows:
1. Does
the claim require the stroke lengths of both pistons to be adjusted?
2. What
is the adjustment that must be made?
3. Does
the claim require that for any decrease in the flow rate there must be a
corresponding decrease in the stroke volume?
4 How
are 'pulsations in the flow' of the liquid to be assessed.?
5. What
is the significance of the words 'such that … reduced'?
- This is a summary of
the questions which arise when one comes to consider both the prior art
and the alleged infringement. It is most convenient to deal with the allegation
of infringement first, and then consider anticipation and obviousness in
the light of the pleaded items of prior art. Insofar as the questions arise
from a consideration of the alleged infringement it is not in general possible
finally to come to a conclusion as to the proper construction of the claim
until the nature of the alleged infringement has been considered. This is
a consequence of the need, identifiable from the decision of the House of
Lords in Catnic v. Hill and Smith [1982] RPC 183 and that of Hoffmann
J in Improver v. Remington [1990] FSR 181 to identify variants from
the literal meaning of the claim which are represented by the alleged infringement.
The
alleged infringement
- There is a product
and process description relating to the Waters pump in issue in the action.
This is accepted to be accurate. The discussion at trial centred on the
claimants' and defendants' notices of experiments. Both sets of experiments
were concerned with the operation of the pump at a range of different flow
rates.
- In general terms, the
point can be stated quite simply. In the Waters pumps, the user may select
one of a number of ranges of flow rate. The control systems which control
the pump adjust the volume in response to the range selected. Accordingly,
the pump volume does not vary with the rate in a linear manner, but in a
stepwise manner. Furthermore, it appears that while the swept volume of
the secondary or accumulator cylinder increases from range to range, within
each range it either decreases as flow increases, or remains constant. The
swept volume of the primary cylinder remains the same throughout any given
range. The operation of the pump is therefore somewhat more sophisticated
than that contemplated by the claims. Figure 7 annexed to Mr James's report,
which was not challenged on this point, shows the volume delivered during
a pump cycle in the various operating ranges provided for automatically
in the Waters pump and its controlling software. This diagram shows that
in the three lowest ranges the accumulator piston stroke decreases with
increasing flow rate. In the range between 3 and 5 ml per minute, it is
more or less constant, falling again in the range between 5 and 7.5 ml per
minute. It will be observed from this diagram that the substantial increases
in stroke lengths in the accumulator piston take place only on range changes:
within each range the length of the stroke of the accumulator piston either
remains constant or decreases. A substantially similar diagram is produced
by Herr Riggenmann (note that the horizontal axis of this diagram is distorted
below 1 ml per minute). Thus, between range changes, the Waters pump increases
flow rate by increasing frequency of pumping and without also increasing
the swept volume.
- It is convenient to
deal with the point which arises on the concluding words of the claim first.
In its description of the invention, the specification describes by reference
to Figures 5 and 6 a comparison of a pump according to the invention with
a prior art pump. The pump according to the invention is shown to have a
primary cylinder volume adjustable between 20 and 100 µl, and the comparison
pump has a primary stroke volume of 100 µl. Figure 4 is used to demonstrate
that in the prior art, the only way of increasing flow rate is to increase
pumping rate. Essentially, the passage in column 12 between lines 7 and
51 describes in some detail the way in which a pump according to the invention
runs with a higher pumping frequency than a prior art pump. In the result,
as the specification points out, the pulsations in the outflow are also
of higher frequency thereby reducing the 'ripple' in the flow. A graphical
representation of this effect is shown in Figure 6, which is a plot of pressure
ripple. Herr Riggenmann was concerned that it was incorrect to equate flow
variations with variations in pressure. He said that what mattered was the
variations in the flow. However, I did not understand him to be seriously
in dispute with the proposition that a pump pulsation defined in terms of
pressure nonetheless, on the assumption of laminar flow, was strictly equivalent
to a flow variation. It may well be that the assumption of laminar flow
in an apparatus of this nature is unrealistic. Pressure fluctuation is nonetheless
a useful indication of the flow variations which are taking place albeit
that there will be no simple relationship between pressure and flow in the
general case.
- But the distinction
between the flow pulsations on the one hand and pressure pulsations on the
other becomes important when one considers the reduction in pulsations which
is said to be achieved by the use of the lower volume stroke for low flow
rates in the Waters apparatus. Herr Riggenmann suggests that the flow pulsation
at a stroke volume of 100 µl amounts to 0.605 per cent of the flow rate,
but at a stroke volume of 25 µl amounts to 0.778 per cent of the flow rate.
From this, he draws the conclusion that the flow pulsations at the pump
outlet of the Waters pump is greater at lower stroke volumes than at higher
stroke volumes and accordingly the last phrase of the claim is not satisfied.
- The nature of the comparison
referred to by the concluding words of the claim is, in my judgment, a comparison
of the kind described by the specification in the passage at column 12 lines
26ff. I have no doubt that like other such comparisons it is difficult to
define with any more precision than this. Indeed, it has some resemblance
to comparisons of the kind called for by the words 'thin and flexible' (Cleveland
Graphite v. Glacier (1950) 67 RPC 149 (HL)) or 'large' (BTH v. Corona
Lamp Works (1922) 39 RPC 161 (HL)). It is a comparison with the prior
art.
- Mr Wyand suggests that
continuous adjustment can be viewed as a 'paradigm case', and the discontinuous
adjustment of stroke volume in the alleged infringement can properly be
considered as an attempt to achieve the benefit referred to in the specification
that 'the pulsations in the flow are reduced' in the manner generally described
in the specification. It seems to me that there is raised here a second
point on construction. The claim does not distinguish between the two pistons
of the pump. As I have indicated, the two pistons fulfil a pumping function,
but it must be remembered that their respective duties are somewhat different.
The function of the second or accumulating piston is to maintain pressure
in the circuit while the first piston is drawing in solvent and the function
of the first piston is both to maintain pressure in and pass solvent to
the circuit while filling the second cylinder.
- On the face of it the
claim is calling for a reduction in stroke volume of both pistons as the
flow rate is reduced. I reject Agilent's contention that there is no requirement
in the claims as literally construed that the accumulator piston stroke
length must increase with increasing flow rate. In my judgment, precisely
the opposite is the case. The behaviour of the pistons of the Waters pump
therefore represents a variant on the manner of operation called for by
the claim in at least two respects: stepwise variation between ranges, constant
volume (primary piston) and constant volume or volume reducing with increasing
flow rate (accumulator piston) within each range.
- I shall follow the
principles of construction set out by Hoffmann J. in Improver v. Remington.
The two questions of fact which must be answered are whether the variant
which I have identified has any effect upon the way the 'invention' works
and, if so, would that have been obvious to the skilled man at the date.
No doubt stepwise variation between ranges accompanied by continuous variation
dependent upon flow rate within each range would be encompassed by the wording
of the claim, but the variant is more subtle than this. It is the combination
of stepwise variation between ranges and absence of variation (or a variation
contrary to that called for by the claim) within each range.
- Herr Riggenmann gave
evidence that the flow rate is selected by Waters and corresponds to particular
classes of columns available commercially. He suggests, therefore, that
once a particular column has been selected, the pump operates to vary flow
rate within the ranges acceptable for that particular column by changing
frequency. Indeed, it would be surprising if Waters had not selected their
pumping ranges to coincide with commonly used columns. In the end, it seems
to me that a rational development of the invention of the patent in suit
to deal with different ranges of flow rate is to provide stepwise jumps
between different flow rates, and to provide for increase in the stroke
volume of both pistons within the pre-selected ranges.
- The evidence as to
the materiality of the effect was not satisfactory. It was complicated by
Herr Riggenmann's observation that the flow variations in the Waters
pump were actually greater, as a percentage, at low flows than at higher
flows. But this is not the comparison called for by the claim. The contrast
is between the pulsations in the output of a fixed-volume pump at a low
flow rate and the pulsations in the output of the variable volume pump at
the low flow rate. Herr Riggenmann looks at the percentage pulsation at
high flow, high volume and low flow, low volume, and observes that the pulsations
in the latter are greater. This is not particularly surprising. The correct
comparison, as I construe the claim, was between low flow, high volume and
low flow, low volume.
- It seems to me that
the omission of a continuously variable stroke volume within the flow ranges
which might be selected for operation for a particular column is, in fact,
a variant which does have a material effect upon the way the invention works.
There is no suggestion in the specification that the dependence of stroke
volume upon flow rate is in some way optional, quite the contrary. Thus,
had I reached the third question in Improver, I would have come to
the conclusion that had this variant had no effect upon the way the invention
worked, nonetheless on a true construction of the specification it was excluded.
Validity
- The principal item
of prior art upon which reliance was placed was US specification 4,681,513
(Saito). Saito describes a two-stage pump for use in HPLC. He describes
the pump of the prior art as a two-stage parallel pump rather than the series
pump with which I am concerned in this case. At column 4 line 17, he describes
the main object of his invention as follows:
'It
is a main object of the present invention to provide a pump assembly equipped
with a mechanism for preventing the average flow rate from dropping, which
would otherwise be caused by higher delivery pressure of the pump assembly.
It is
another object of the invention to provide a pump assembly producing only
a small amount of flow rate pulsations.
It is
a further object of the invention to provide a pump assembly whose flow
rate can be varied without making the period of flow rate pulsations very
long.
It is
a still further object of the invention to provide a pump assembly capable
of delivering fluid over a quite wide range of flow rates.
It is
a yet further object of the invention to provide a pump capable of delivering
a trace of fluid having a large compressibility, such as liquefied gas,
with the same accuracy as ordinary liquids.
These
objects are achieved in accordance with the teachings of the invention by
a two-stage pump assembly comprising plunger pumps A and B, the suction
port of the pump A being connected to the discharge port of the pump B.
Thus the discharge port of the pump A and the suction port of the pump B
act as the discharge port and the suction port, respectively, of the whole
pump assembly. When the pump B is delivering fluid from the whole pump assembly,
the pump A fills and fully elevates the pressure inside the pump A. When
the pump B is in the filling stage, the pump A is in the displacement stage
and pumps fluid out from the whole pump assembly.'
- It is apparent from
this introductory passage that Saito considered that he had invented the
series pump. It is not this aspect of his invention with which I am concerned,
but rather the preferred embodiment which has additional features which,
Waters say, results in an anticipation of the patent in suit. The crucial
figure is Figure 8, which shows how the velocity of the pistons in the pump
varies with time. It is necessary to bear in mind that Saito has a pump
both of whose pistons are spring-loaded against cams. The cams are mounted
on a shaft which oscillates in rotation, and the pistons move so as to follow
the profile of the cam. Each of the cams is operated by a separate stepper
motor, which is driven by a control circuit. The cams do not have a single
gradient, but three working regions each of different slope, called M, S
and N. The apparatus itself is shown in Figure 7:
- There are two passages
in the specification which gave rise to a great deal of discussion, and
it may be helpful to set them out here. At column 6 line 36, Saito says
this.
'The
operation of the novel pump assembly constructed as described above is now
described. It is to be understood that the cams of the conventional system
rotate in one direction. In contrast, in the novel assembly, the cams are
rotated back and forth within the 360°. The novel assembly is similar to
the conventional system shown in Figure 1 except in these respects. Data
indicating a desired flow rate is first entered into the central control
circuit 63. Then, the circuit 63 determines the rotational velocity of the
cam 45 and the angular range within which the cam swings back and forth,
according to the flow rate, the cam 45 acting on the pump 40. The central
control circuit 63 then delivers a control signal to the driver circuit
44 to control the operation of the cam 45 in such a way that if the flow
rate assumes a large value, the angular range within the cam 45 swings back
and forth, hence the period Ta1 of the displacement stroke and
the period Ta3 of the filling stroke, is extended as the rotational
velocity is increased. In this way, the cam is swung right and left repeatedly
within the selected angular range.
…
(column
7 line 3) 'The central control circuit 63 to which the output from the variation
detector circuit 65 is applied monitors variations in the pressure inside
the pipe 62 in synchronism with the output from the encoder 66. These variations
are presented on the display means 67, for example, as shown in Figure 9.
When the pressure characteristic is inclined upwardly to the right as shown
in Figure 9(a), the pre-compression made by the pump 50 is [not sufficient].
When it is inclined downwardly to the right as shown in Figure 9(c), the
precompression is [excessive]. When it is horizontal shown in figure 9(b),
the precompression is adequate. When the central control circuit 63 detects
the condition shown in Figure 9 (a), it causes the motor driver circuit
54 to increase the velocity of the plunger 53 during the periods Tb1
and Tb2 for augmenting the precompression. In this way, velocities
indicated by the broken lines in Figure 8 change to velocities indicated
by dot and dashed lines 68 and 69. Inversely, when the condition shown in
Figure 9(c) is detected, the control circuit 63 instructs the driver circuit
54 to decrease the velocity of the plunger 53 during the periods Tb1
and Tb2 for obtaining the condition shown in Figure 9(b). Although
control operations are carried out in this way, the following conditions
are invariably satisfied regarding Figure 8:
(1)
the area bounded by both the solid line and the time axis to the period
Ta1 is equal to the area bounded by both the solid line of the
time axis during the period Ta2.
(2)
the area bounded by both the broken line and a time axis during the period
Ta1 is equal to the sum of the area bounded by both the broken
line and the time axis during the period Tb2 and the area bounded
by both the broken line and the time axis during the period Tb3.
(3)
the delivery of the pump assembly during the period Ta3 is equal
to the difference between the area bounded by both the broken line and the
time axis during the period Tb3 and the area bounded by both
the solid line at the time axis during the periodTa3.
- These passages are
best understood by reference to Figure 8 itself:
Finally,
it is necessary to refer to a short passage at column 8 line 17:
'The
central control circuit 63 shown in Figure 7 selects one of three angular
ranges according to the flow rate entered, and drives the pump within that
angular range. The inlet pump 50 is required to deliver fluid at a larger
flow rate than the outlet pump 40. Where the cam 55 is identical in profile
of the cam 45, it is necessary to rotate the cam 55 over two or more of
the angular ranges. In this manner, the novel pump assembly alone is capable
of covering a very wide range of flow rates.
- The teaching of Saito
is far from simple to understand. The first question is whether Saito teaches
the use of oscillations of different amplitudes within a given region of
the cam so as to provide a range of different stroke volumes, the amplitude
of oscillation of the cam being selected according to the desired flow rate:
a lower flow rate meaning a lower amplitude of oscillation within a fixed
period. I think it was agreed that if this was the teaching of Saito then
it was directly relevant to the patent in suit.
- Mr Wyand submitted
that it is the task of the court to determine what Saito clearly and distinctly
taught the skilled person at the priority date, not what can be read out
of Saito by the application of hermeneutical stress. This admirable phrase
concisely describes the process of squeezing a document to extract every
last drop of meaning. The submission is correct: to anticipate, a document
must contain a clear description of, or clear and unmistakable directions
to do or make, something within the claim: see General Tire v Firestone
[1972] RPC 457. When considering obviousness, on the other hand, ambiguities
in the disclosure of the document may be obviously capable of resolution
in a particular way without the exercise of ingenuity: but it is not legitimate
to try to resolve obscurity by an exercise in imaginative reconstruction
to ascertain what it was that the patentee must have been trying to describe.
- Mr Wyand pointed out
that before one approaches the teaching of Saito it is necessary to keep
a few basic principles in mind. The first is the dependence of the flow
rate on piston velocity. Ignoring for a moment the compressibility of the
liquid in the system, the actual flow rate is the product of the cross-sectional
area of the piston and the piston's velocity. If the piston is driven by
a cam, its velocity can be changed by changing the speed of rotation of
the cam, regardless of the angle through which the cam moves. The stroke
is determined by the angle through which the cam rotates and the gradient
of the cam. If one considers a change in the rotational speed of the cam
without a change in the angle through which it rotates, it will be apparent
that the speed at which the cam moves has gone up and accordingly so has
the velocity of the piston: so the flow rate increases. If on the other
hand the angular velocity is not changed, but the cam moves through a greater
angle, the piston will travel through a longer stroke, but at the same speed.
In this case, the flow rate will not increase. Accordingly, if the period
of rotation of the cam is fixed, an increase in angular velocity will mean
an increased stroke length, and an increase in rotational angle will also
mean an increase in stroke length. The difference between these two methods
of increasing the stroke length is that in the former case, the flow rate
is increased whereas in the latter the flow rate remains the same.
- Mr Wyand's submission
depends upon a proper analysis of figures 8 and 9 of Saito. He says that
Figure 8, examined carefully, makes it clear that it is only ever the primary
piston stroke that is adjusted in Saito. Mr James, Agilent's expert, says
this in paragraph 5.2.11 of his report:
'Both
pistons are moved via cams by stepper motors. The cams swing back and forth
(i.e., oscillate) within three discrete angular ranges each of which has
a different slope. Saito's oscillating multi-range cam is therefore designed
to cater for a wider range of flow rates than could be achieved using a
single slope cam rotating through 360°. Both of the motors are controlled
by drive circuits which are in turn controlled by control signals from central
control unit.
'5.2.12
Liquid in the reservoir is drawn into the primary piston while the accumulator
piston delivers fluid. During the filling stroke of the accumulator piston,
the compressed fluid is drawn from the primary piston. In the only embodiment
described in Saito … the diameters of both pistons are the same and the
primary piston therefore needs to move over a longer stroke length than
the accumulator piston in order to displace a greater volume of liquid than
the accumulator piston.'
Having
dealt with the generation of the signals indicating the movement of the
accumulator piston and the manner in which the flow rate is controlled by
measuring pressure, Mr James continues:
'5.2.16
The desired flow rate for a pump is entered into the central control circuit,
which then selects one of the three angular ranges for the cams, S, N or
M depending upon the desired flow rate. The central control circuit then
sends a control signal to the driver's circuit for the accumulator piston.
This circuit controls the operation of the cam associated with the accumulator
piston. … '
- I understand Mr James's
evidence to be that in Saito the stroke length is indeed variable. This
is true both of the primary piston and of the secondary piston and is because
both primary and secondary piston stroke ranges are set to a value which
depends upon the flow rate desired. Mr James is certainly of the view that
in Saito the passage which I have quoted above which refers to the 'angular
range within which the cam 45 swings back and forth … or the rotational
velocity is increased' can be plausibly taken to mean that if for a set
flow rate a particular angular range of the accumulator piston cam has been
selected and the operator then wishes to increase the set flow rate, the
higher set flow rate can be achieved by changing the angular range within
which the cam 45 swings back and forth (by changing the cam region) or by
increasing the angular velocity (the period of the swing remaining unaltered).
In other words, the description is unclear as to what Saito is discussing:
it may be stepwise variation or it may be a variation in angular velocity
within a given range, resulting in increased swept volume if the cam region
will accommodate the increased swing necessary for a fixed period. As I
understand his evidence, Mr James prefers the first view because Saito lacks
any explicit teaching that the period of the oscillation is to remain constant
as the velocity increases, and this is, for the reasons I have given, essential
if the swept volume is to increase with flow rate.
- Mr James identifies
a further difficulty with Saito's description of the movement of the primary
piston. He points out that the cylinders of the inlet pump and the outlet
pump are said to have the same diameter. Saito says that it is possible
to make the maximum flow rate in the inlet pump larger than the maximum
flow rate of the outlet pump. This means that the primary piston has to
move at a higher speed than the accumulator piston, since they must have
the same period. Accordingly, the angle of oscillation of the primary cam
must be greater than that of the accumulator cam, and Saito clearly suggests,
says Mr James, that the primary cam may oscillate so that it travels over
more than one of the angular ranges S, M, and N. This would result in a
very bumpy ride for the output of the pump, because of the sudden change
in rate between the two ranges.
- Herr Riggenmann, on
the other hand, did not find such difficulty in the disclosure of Saito.
He concentrates upon the passage which refers to '… the following conditions
[which] are invariably satisfied regarding Figure 8'. He says (in agreement
with Mr James) that the first piston must be able to compress and then deliver
a greater volume of fluid than the intake of the second piston. Since the
cams are identical, and since the period of operation is the same, he draws
the conclusion that the second piston must operate so that its chosen cam
segment is not fully utilised when the first piston operates over the whole
of a corresponding cam segment. He says this is clear from Figure 8. In
coming to his conclusions, Herr Riggenmann performed certain calculations
on the basis of Figure 8, and it, in particular, set up what he considered
to be a plausible view of the speed of the accumulator piston during the
period Ta1. He produced some of his working papers from the witness
box, and it became clear that he had indeed carried out the calculations
that he said he had carried out in checking the torque requirements of the
motors. It is clear that preparations were being made to attack Herr Riggenmann's
bona fides, and I should say that Mr Wyand withdrew the suggestion as soon
as Herr Riggenmann's working papers had been analysed.
- In my judgment it cannot
be fairly said that Saito contains clear directions to vary the size of
the sector of the cam segments traversed. As Mr Wyand points out, even if
there were such a direction, it does not necessarily follow that the flow
rate is increased by increasing stroke length. Everything would depend upon
the effect of the control system, and in particular how it regulated the
speed of rotation of the cam within a single cam segment. It has to be borne
in mind that part of the cam segment (so far as the primary pistons concerned)
must be used to compensate for the compressibility of the fluid. Herr Riggenmann
accepted that so far as range M was concerned, there was insufficient capacity
to compensate for compression if the fluid had a compressibility greater
than 0.5 per cent, which is a very low figure for the solvents in use in
this analytical technique.
- There is plainly no
express teaching in Saito to arrange the control means 63 so that the stroke
volume alters with the flow rate, other than by way of selecting the different
cam ranges. For the reasons which I have already expressed in considering
infringement of this claim, I do not consider that a stepwise variation
unaccompanied by variation within each step falls within the claim. However,
there this, in my view, a dilemma facing the patentee. However that may
be, I do not consider that this claim is anticipated by Saito.
Obviousness
- The argument of obviousness
based on Saito cannot proceed in the manner summarised by the Court of Appeal
in Windsurfing International v. Tabur Marine [1985] RPC 59. The reason
is that it is not possible to identify a clear disclosure which can be used
as the basis of the assessment of the step represented by the invention.
The question is rather whether the ambiguities and obscurities of the disclosure
would be resolved by the skilled man in such a way that he came up with
a something falling within the claim without the exercise of inventive ingenuity.
In approaching this question in respect of a document like Saito, it is
essential that hindsight is avoided. It is however permissible to consider
the general teaching of Saito to see whether the flavour of the document
points the skilled man in the right direction.
- The difficulty presented
by Saito in this respect is that it lacks any general teaching whatever
as to the association of a variable swept volume with flow rate. Given that
flow rate does not depend upon swept volume at all, the decision to vary
volume with flow rate represents an addition to the teaching of Saito. The
control circuit 63 of Saito responds to pressure variations to adjust the
angular velocity of the accumulator piston during the periods Tb1
and Tb2 , with a view to minimising the pressure of fluctuations.
This alteration to in the velocity of the piston in the accumulator cylinder
is intended to adjust precompression. But it is not an adjustment which
takes place in response to the setting of the desired flow rate. It is,
on the contrary, a feedback mechanism for controlling pressure fluctuations
'on the fly'. Mr James thought that this was a 'Rolls Royce' control method.
- In conclusion, it seems
to me that the objection based on Saito is an objection of the kind often
called 'anticipation or nothing'. There is insufficient general teaching
to find a convincing argument of obviousness, and I have come a to the conclusion
that the objection of obviousness on the grounds of Saito fails.
- It is not therefore
strictly necessary for me to consider the contention that the allegedly
independently valid claims possess patentable subject matter even if claim
1 is invalid. I should however do so. This involves a consideration of claims
2, 3, 5, 10, and 11. .
- Claim 2 calls for ball-screw
spindles as the drive means for the pistons. As Mr Hobbs pointed out, a
ball-screw spindle contains what is to all intents and purposes a helical
cam, but this observation is made at such a level of generality that it
is unhelpful. The pressures with which these pumps are concerned are very
high. There are disclosures of the use of ball-screw spindles in high pressure
pumps for use in HPLC. Three patents were referred to, called Allington,
Nicolson and ICI. They were not pleaded. Mr James accepted, I think, that
that ball-screw spindles are part of the 'toolkit' of the designer of a
pump for use in HPLC. I think that this makes the ball-screw spindles part
of the common general knowledge so far as Mr James is concerned, and this
view was certainly supported by Herr Riggenmann . If such spindles cause
a problem in use, then, of course, the specification may be insufficient
in this respect since it gives no directions whatever as to any special
techniques are which need to be used in order to make the spindles usable
in a pump of this description.
- Claim 3 is a claim
to the detailed construction of the ball-screw spindle. The material feature
of this claim is that the ball should have such a diameter that it projects
above the recess in which it sits but can move freely inside the recess.
This enables the ball-screw spindle to accommodate small errors in alignment
with the piston, without causing the piston to be pushed sideways. This
is no doubt important, having regard to the considerable pressures involved
but I am wholly unpersuaded that it is anything other than the application
of sound engineering principles. The same goes for the separate drive means
it to the spindles called for by claim 5.
- Claim 10 is concerned
with the manner in which the stroke length is to be adjusted. It calls for
the top dead centre of the piston motion to be kept fixed, while the bottom
dead centre is varied. On the assumption that claim 1 is invalid, and the
idea of variation is old or obvious, I do not see how this claim could save
it. Given that one of the problems with which the pump of claim 1 is intended
to deal is the compressibility of the solvent, it seems to me plainly obvious
to minimise the amount of solvent which needs to be pre-compressed. This
is achieved if the top dead centre of the piston motion is kept fixed and
the bottom dead centre is varied. Indeed, claim 10 is a selection of one
out of only three possibilities: either the bottom dead centre is fixed,
or the top dead centre is fixed, or both are variable.
- Claim 11 is concerned
with an aspect of the invention with which I have not previously found it
necessary to deal. This is a matter of the adaptation of the apparatus of
claim 1 so as to accommodate a particular HPLC technique called gradient
operation. In this technique, more than one solvent is drawn into the primary
cylinder, and the proportions of different solvents are changed stroke by
stroke. The result is that in the column a gradient of different solvents
is introduced. Bike42
- The valves of the kind
called for by claim 11 for mixing different proportions of the solvents
on successive pump strokes were on the evidence a conventional part of HPLC
techniques. This claim also does not possess independent validity.
Other
citations
- In the end, the defendants
only relied upon US 4,556,367 ('Schmid') and GB 1,450,400 ('Dawkins') in
support of an attack upon the subsidiary claims with which I have dealt
above. The approach was probably impermissible, since neither Schmid nor
Dawkins was demonstrated to be common general knowledge, and it was not
demonstrated that a skilled man confronted with Saito would obviously turn
to either Schmid or Dawkins. Accordingly, it is not necessary further to
consider these documents.
Conclusions
- The patent in suit
is valid over Saito. The allegation of infringement fails. I will hear counsel
upon the appropriate form of order.
