May 25, 1917.

THE COLLIERY GUARDIAN.

987

to the diameter of the spindle. For this reason, the
waste water from a balancer should not be connected
to the suction pipe of a pump, or means should be
taken in the pump to equalise the areas subject to the
static hydraulic pressure.

The Rotor.

The rotor revolves within the guide chamber and
housing, each impeller running at some points in the
closest proximity to the partitions dividing the pres-
sure stages and return guides.

These so-called running joints or neck rings must be
so maintained by design and construction as to secure
in continued service the smallest possible hydraulic
leakage back from stage to stage, or mechanical loss
by rubbing friction, both of which affect the efficiency
and wear and tear of the pump. As a question of
practical importance, the design of this combination
probably follows next to that of the successful auto-
matic hydraulic balancing of the axial end thrust. We
have in it the question of the deflection of the pump
shaft; the reinforcing effect of the surrounding
impellers; the support afforded by the bushes, if any,
between each stage, and the loss by bush friction
(wear); the form of neck ring to reduce leakage; and
the effect of key ways and keys, etc., all independent of
each other, and affecting the whole.

The ideal condition is that of a rotor supported in
lubricated bearings, with a shaft of such sufficient
stiffness between supports that the deflection, under all
possible running conditions, is less than the clearance
allowed at the neck rings and intermediate bushes, so
that no contact takes place between the rotating and
the fixed members, this clearance at the neck rings
and intermediate bushes being kept down to the
smallest possible limits. To lock up the impellers
together shoulder to shoulder against each other by
the end nuts in order to get the advantage of the rein-
forcing effect is, however, advisable, because, owing to
the slight inaccuracies of manufacture in squaring the
shoulders of the impellers, sleeves, etc., the result
would be to throw the combination out of truth, and
distort the shaft. Moreover, provision must be made
for dismantling in a mine or similar place, with its
obvious attendant disadvantages, and the possibility of
rough treatment taking place; furthermore, as a safe-
guard against heating up and consequent expansion of
the outside combination (bronze) against the internal
(steel) shaft due to accidental contact with the inter-
mediate bushes, running, perhaps, when the pump is
empty, suitable expansion should be allowed for, and
the impellers, to ensure this, must not be locked tight
against each other.

The supporting effect of the bushes on the shaft, in
passing through the diaphragm intermediate between
the impellers, is very difficult to allow for exactly.
Intermediate bushes can only, in some cases, be con-
sidered as water-lubricated supports which will act as
such so long as a certain low surface pressure on them
is not exceeded; and the author believes that too much
use is made of such supports in turbine pump design.
It is found in practice .that an internal bearing to be
successful must have the same water pressure at both
ends, and must be properly lubricated with good grease
—it then gives excellent results. Hence, one is prac-
tically dependent on the shaft itself for the necessary
strength and stiffness to allow of fine internal clear-
ance, and the importance of a good design which will
economically give the minimum deflection of spindle at
once becomes apparent.

Consideration of the factors affecting the deflection
of a turbine pump spindle shows that the following
arrangements of bearings are possible: Two outside
bearings; one outside and one inside bearing; or one
outside and two inside bearings; but experience has
shown the great importance of keeping the distance
between bearings in a multistage turbine pump as
short as possible, for the more or less customary type,
with outside bearings at each end, gives a deflection
of 6J times that of a pump with two internal bearings.

The direct result of this extra deflection is that the
clearances between the outside of the eye of the
impellers and the neck ring must be increased to allow
of the impellers revolving without touching, and this
increased clearance necessarily means a correspond-
ingly increased leakage and loss of efficiency. More-
over, the passage of a large volume of water through
the running clearance means increased wear, and a
tendency to push the walls apart, and therefore force
the impeller and shaft still farther out of the straight,
with consequent friction against the adjacent neck
ring surface.

Securing the stiffness of a shaft by increasing its
diameter is an inefficient method, for, besides the
increased area for leakage through the neck rings,
there are much worse conditions of entrance in larger
and more rapidly rotating eye and inlet edge of the
impeller vanes. The desirable condition, therefore,
from all points of view, is to use the smallest diameter
of shaft with the shortest distance between the
bearings.

Leakage Between Rotor and Stator.

The leakage through the neck rings may now be con-
sidered under different conditions and clearances. The
material of these rings is of first importance, and
should be of such a type that it will not tend to drag
or tear if accidentally touched or rubbed by the
running impeller. Material of a hard “ short ” nature
like cast iron will meet this condition, and in practice
the smallest clearance can be successfully run with
such neck rings; they are, however, subject to corro-
sion, particularly with acid waters as are frequently
met with in mines.

The impeller, for considerations of susceptibility to
corrosion and strength to resist the centrifugal effects
of high speed, is usually made from phosphor-bronze.
Hence, phosphor-bronze impellers running in phos-
phor-bronze neck rings are frequently used, though a
phosphor-bronze impeller with a cast iron neck ring
would be a better combination, and run with smaller
clearances; but to get over the corrodibility of the cast
iron it should be specially treated, or should be similar
to that for renewal balancer rings.

The form of the neck ring is, in practice, confined to
four forms: (a) Internal, (b) external, (c) vertical, (d)
labyrinth. The first is the usual form in turbine
pumps, the second is more rare, the third is practically
restricted to single impeller low-lift centrifugals, while
the fourth is a refined form after steam turbine prac-
tice, and is used in some high-lift pumps.

Between (a) and (b) there is little to choose, but for
manufacturing reasons (u) is usually preferred. Clear-
ances of a few thousandths of an inch per side can be
run with a suitable shaft, and the leakage is very
slight, less than 1 per cent. The vertical form (c) is
used in pumps for low lifts subject to dirty water and
inferior attention. It allows of downward wear of the
impeller without any rubbing surface coming in con-
tact, which (a) and (b) would not; (d) is an attempt
to reduce leakage further by a device of the kind used
on the steam turbine.

The labyrinth type, having'to deal with a non-expan-
sible fluid like water, and not steam, is far less effec-
tive than in the steam turbine, and in fact can only
restrict leakage by the formation of eddies in the
channels or grooves; shock losses at the sudden changes
in area and direction, and the centrifugal effect of the
rotating film of water in the narrow aperture. Its pos-
sible advantages in this way are more than negatived
by its fragile nature if it comes in contact with the
impeller, such a contingency usually involving the
eventual renewal of both parts. Types (a) and (b)
only require a new neck ring after wear. All types,
however, are readily made with loose rubbing rings,
which may be easily changed.

In practice it has been found that when grooves were
cut on the impeller shoulders to secure a labyrinth
effect, and at the same time the neck rings were left
plain, that after funning for some time the marks of
the grooves were always found in the neck rings; the
difference between the worn surface and the original
surface often being as much as ten thousandths of an
inch, showing in these cases a deflection of the shaft
greater than the calculated value. In these particular
cases the fact that the grooves appeared all round the
ring to the same depth indicated that they were due to
actual deflection and not faulty alignment, the reasons
for which may be found in the causes already
enumerated and in mechanical inaccuracies.

When inter-locking grooves are used, the difficulties
of erection and dismantling are increased -unless the
inter-locking is in the plane at right angles to the axis
of the pump ; any other arrangement is, of course, only
possible with horizontally split pumps or with split
or divided neck rings.

In conclusion, the author said that the efficiency of
the turbine pump was considerably less than that of
the water turbine, and the object to be aimed at was
the raising of that efficiency to as near as possible that
of the water turbine. Even now, the turbine pump
was driving practically all other pumps out of the field.
Its greatest enemy was probably prejudice. At the
present time there were various committees for con-
ducting and increasing the amount of research, but he
did not know whether the turbine pump fell within
the scope of any of these committees or of the institu-
tion itself. If it were agreed that the turbine pump
was going to assume, or was assuming, the very impor-
tant position he had outlined, then it seemed to him
that it ought to be the subject of a special research,
just as much as the internal combustion engine. If
the efficiency were raised by even 5 to 8 per cent.,
there would be an enormous value represented in
£ s. d.

DISCUSSION.

The President (Mr. Michael Longridge) said the
paper was one of the type to which he wished to see
more attention given by the institution, as it described
the results of years of work overcoming practical diffi-
culties. It contained information hardly ever given
in text-books, but which yet was quite as necessary to
design and practical working as that found in the
literature. The paper formed a fitting supplement to
Dr. Unwin’s classic work on the subject. It also con-
tained a warning to those coming fresh from college
and feeling in their own minds the strength and ability
to solve all sorts of problems from the knowledge they
had gained there. If they relied solely upon that
knowledge, they would come to unutterable grief, not
because their knowledge was wrong, but because it was
impossible with our present mathematical tools to take
account of all the conditions arising in practical pro-
blems. Adopting the attitude of a “ candid friend ”
towards the author, he regretted working drawings
had not been given in the paper instead of diagrams.
Whether that was from want of time, or a feeling of
secrecy, which undoubtedly did prevail among engi-
neers at the present time—a fear of communicating to
outsiders what they had attained themselves with a
great deal of trouble—he did not know. That was a
natural feeling, but he did not thifik it was one tend-
ing to the benefit of British engineers; and he hoped
the time would come when that secrecy would cease to
be considered necessary.

Mr. T. G. Sherwell supported the author’s plea for
more research work to improve the efficiency of the
turbine pump. Efficiencies had been obtained as high
as 85 per cent., but there was no reason why, with the
larger sizes and greater quantities of water, an effici-
ency of 90 per cent, should not be obtained. The diffi-
culty at present with manufacturers was that their
time was fully occupied in producing for national pur-
poses, and they could not possibly spare time for
research work. Whilst that only applied to war time,
and would not apply to peace, yet he felt there was
room for independent research. With regard to the
advantages of the two types, the ring casing and the
divided casing, he thought a great deal too much was
made of the difference. In his experience, there had
never been any trouble with either.

Mr. H. Munro, speaking on the question of effici-
ency, asked how long the figures mentioned were main-
tained. In reciprocating pumps, he understood that
the efficiency was maintained almost indefinitely; but
in centrifugal pumps, especially with acid waters, the

tips of the impellers were worn away rapidly, and the
efficiency rapidly fell off.

Mr. Percy Griffith referred to the use of turbine
pumps in boreholes as distinct from wells. He knew of
a case in which such pumps had been very successfully
used in these circumstances. He would like to hear of
recent experience in this connection. The case he had
in mind was that of a relatively small borehole, and
the difficulties arose from the necessity for suspending
the pumps from a vertical shaft driven from the sur-
face, whilst their usefulness for this purpose resided
in convenience of fixing and the possibility of getting
a very large capacity in a relatively small space.

Mr. Druitt Halpin mentioned an interesting appli-
cation of the turbine pump which had come to his
notice in Germany before the war. This was for the
purpose of driving a number of cranes. He believed
they were being tried tentatively, because the pressure
used, viz., 4501b., was only about half that used in
this country for similar work. It seemed, however,
an application capable of considerable extension.

Prof. P. M. Baker gave some figures of efficiency
obtained with German pumps in India. The efficiency
was stated by the manufacturers to be from 70 to 75
per cent., but the best figure obtained on actual tests
was 56 to 57 per cent. He did not know whether that
was the result of the German manufacture or an error
on the part of those who took the tests. Was there
any reason for supposing that in particularly acid
waters there would be trouble from electrolysis due to
the phosphor-bronze impellers being more or less in
contact with the cast iron sleeve rings?

Mr. Mark Robinson asked Mr. Griffith what steps
were taken, in the case of the borehole pump that he
had mentioned, to prevent the casing of the pump
revolving. Did the borehole tube itself act as the
delivery pipe, or was there a separate delivery pipe
inside the borehole?

Mr. Griffith said he had not the information at
hand, but would be able to supply it.

Mr. R. Tomlinson said he knew of an instance, in
the north country, four years ago, in which Mather
and Platt turbine pumps were applied to pumping out
a mine, and it took more than three years to do it.
The pump was suspended by two cables, and an elec-
tric cable was also lowered for driving the pump.

Mr. A. J. Bremner, speaking on the question of
couplings for turbine pumps, said he had fitted pumps
with various kinds of couplings, and the only trouble
that-he had had with the flexible coupling was that it
must be properly balanced. He had known the out-
side bearing of the pump to be ruined in a very short
time because the coupling was out of balance.

Mr. G. W. Thompson expressed the hope that the
author would be able to give the efficiencies of various
makes of pumps, because, from his experience, there
was a great difference between them.

Capt. Sankey asked if any steps were taken in the
works to balance the rotors of the pumps against
dynamic want of balance in the same way that the
rotors of steam turbines were balanced. There were
many ways of doing it.

The Author, replying, said the absence of dimen-
sional drawings was due to lack of time and facilities
for having them made. Trade secrecy did not apply,
because he was not now engaged in the manufacture of
these pumps. As to efficiencies, the variation in effici-
ency was dependent upon the size, but at the present
time a 500-gal. pump would give an efficiency of 76 per
cent.; a water turbine under the same conditions
would give possibly 80 to 84 per cent., or something
even better. It was contended that the reason why
one gave a better result than the other was that the
water turbine gave a convergent flow and the pump a
divergent flow, and that one produced eddies and the
other suppressed them. That, however, must not be
taken as gospel. During his time the efficiency had
been greatly increased, and he was convinced that with
detail investigation it would go much higher. The
maintenance of the efficiency in a turbine pump was a
very variable factor. A splendidly constructed pump
dealing with clear water would run for years without
being touched, but as soon as those conditions were
departed from, there was an increased possibility of
trouble. With gritty water there would be quick
wear, and it had been found in practice, in waterworks,
for instance, that great advantage accrued from having
settling tanks in which the water was allowed to remain
for a while before pumping. Then the wear was very
little. He had tried to bring out in the paper what he
thought should be done to improve the efficiency of the
pump. With even running clearances, due to greater
shaft stiffness, there would be little water passing
through. With no water passing through these clear-
ances, there would be no wear, and that was the thing
to be got over. In many designs of pumps, he
believed that too much was allowed for the inter-
mediate neck rings to carry, and if these could be dis-
pensed with, and the shaft made stiffer, the difficulty
would be got over. That was a point that had had to
be dealt with in steam turbine design. As to turbine
pumps in boreholes, the second installation of turbine
pumps that he had had to deal with was of this nature
in Moscow 18 years ago, and it ran very well. The
bore was about 24 in. diameter, and the pump was
entirely suspended from the rising main. That was an
old Reynolds pump, and the trouble originally was not
with the pump, but with the bevel gear drive at the
surface. That pump ran for years, until there was no
more water to pump, and then it was taken out. One
of the difficulties with borehole pumps was the care
necessary to see that a lot of sand was not pumped up
when the pump was first put to work. A violent suc-
tion was created, and the sand drawn into the pump
wore it out very quickly. The highest pressure used
here was about 1,0001b., and as to the remarks of
Prof. Baker about the German pumps, he himself had
also come across similar instances of actual efficiencies
not being as high as those asserted by the makers. He
could not say whether electrolysis was set up through