A discussion on Polyphon tuning scales and comb material

I note Mark Singleton's contribution to the question of different
scales for the 24-1/2-inch Polyphon [161005 MMDigest].  Kevin McElhone
lists just the one example starting at C but wrote there were other
scales for the same model.  I presume that there were indeed other
scales.  If so, the question has to remain: Why?  I postulated two
possibilities and Mark postulates others.  In fact neither of us has
any firm evidence to show who is right.  I made my speculations in
order to see if there is any known evidence to justify or reject them.
Speculation is a valid part of an investigative process.  Measurements
and records play an important part and much of that is still missing.

As to points of agreement, Mark is right to assume Graham Webb's scales
are close to what we accept as normal, and also for stating that scales
do not always conform exactly to a standard pattern.  Thus we agree
there are definite variations in the precision of a scale and that comb
materials and treatment may not be entirely uniform or consistent.
However, I doubt if this non-uniformity is due to the carbon content
and presence of other alloys in the comb's structure.

The Bessemer process was invented in 1856 and became the basic process
for cheap high quality steel.  However, English steel had a reputation
for high quality before that date based upon Huntsman's crucible steel.
As with all processes, the Bessemer-Gilchrist open-hearth process
superseded the Bessemer, also called the basic Bessemer process.
Gilchrist changed the lining of the furnace to dolomite or limestone
instead of clay.  Clay had acidic properties but Gilchrist's lining
helped to remove unwanted contaminants such as phosphorous.  The
material was of excellent quality.

Mark then considers the use of steel for discs and considers how
brittle they can be.  I agree that it was not likely to be the result
of the tempering process but I presume that material quality and
consistency was not as critical as that for a comb.  But it begs the
question about the effect of carbon content on comb tuning.  I find
that very difficult to justify on the basis that comb steel must have
been of a consistently high specification for the purpose of machining,
hardening, tempering and then tuning such a critical component.

His comments on tuning were interesting but not entirely relevant to
the possible reasons for different scales.  We do not disagree on the
quality of tuning where Polyphons were known to 'ring out whilst others
were as dull as ditchwater'.  But that was equally true for much
earlier cylinder musical boxes.

There are several factors that can give rise to different musical
performances.  A change of scale is, I believe, one factor but only
to the discerning ear.  More important is the state of the comb after
years of use as well as the effectiveness of the soundboard and overall
condition of the casework to which it is attached.  A soundboard with
splits that is also not adequately secured to its case can almost
destroy the potential performance of a well-tuned comb.

I said that the [Polyphon] machines were expensive.  I do not think
that statement conflicts with Mark's comment that machines were
designed to sell -- and yes they did, in their thousands.  Pricing
policy has always been balanced with maximising the market, reducing
the cost whilst retaining an acceptable quality.  Many machines were
coin-operated and hence commercial investments.  Was that a reason for
scale change in one of the most expensive parts of the machine, its
comb?  It is a fair proposition but is it true?  I believe 'the jury
is out'.

Mark Singleton also confirms that he came across a 24-1/2-inch Polyphon
in a different key, so we appear to have another agreement that there
were different keys.  Was that due to a personal taste?  Most people
would not care about a transposition of scale for whatever reason
unless it made the instrument more noticeable, for example, in a public
place.  All we now can say that is that there were scalar changes and
the reason has eluded us.

I would still argue that, supposing there was a managerial decision
to change permanently from one scale to another, there might have been
a reduction in manufacturing costs as a result.  How many times have
small changes in a mass-market manufacturing environment saved money
and put the business on a competitive footing with other manufacturers.
It is called marginal costing.  I find it difficult to accept that it
was the grade of the steel that affected the tuning scale.  Should that
be the case it would have to be justified by a saving in the marginal
cost whilst sustaining the market's competitive advantage of musical
quality and performance reliability.

Exhibition quality: One cannot disagree with the statement that models
prepared for exhibition would be of the highest quality both to an
individual wishing to purchase or the agent wishing to sell for a good
profit.  Were these of a different scale though?  That brings us back
to the question: Why?

In a sense, the starting point of any scale is arbitrary unless all
combs were tuned to a recognised standard.  It seems reasonable to
accept that there was such a standard for a particular maker and the
range of instruments produced.  That was standard practice for cylinder
musical box combs so it seems reasonable to assume that it would apply
to disc musical box combs.

Did the quality or composition of the steel make a difference to the
tuning scale for a chosen pitch?  Did the hardening and tempering
process make any difference?  Did the process of manufacture make any
difference?  To answer those questions is not easy but there are some
metallurgical factors that indicate the answer is 'No.'

As a graduate engineer in the 1950s I studied ferrous and non-ferrous
metallurgy, the dynamic performance of materials subject to natural and
induced forces and machine design.  That does not make me an expert,
though.  Steel manufacture during the 1800s and well into the last
century was largely down to the skill of the master steel founder.
Even in the 1950s when I worked at GEC [General Electric Company, UK]
in the manufacture of steam and water turbines and gained some foundry
experience, steel production was a formulated art practiced by the
Master steel founder.

Measured quantities of materials were added to the furnace according
to the desired specification for the steel produced and it its purpose.
There were materials for springs, turbine blades rotating shafts and
a plethora of other uses all to be produced to British Standards for
the target specification.

As samples were drawn off the furnace they were taken to the
laboratory, polished and examined for grain structure and quality; then
various tests were carried out to check whether the material conformed
to its intended specification.  All this was whilst the furnace was
still operating.  The laboratory staff then instructed other measured
quantities of material to be added to the furnace.  The sampling and
testing process continued until the result was within specified limits.
Thus, the result for high quality products was based as much on skill
and experience in the materials mix, furnace temperature, process time,
slag removal, etc.  as it was on analysis and tests.  Not a great deal
had changed since the 1800s.

Nevertheless, the end products in the mass production process were
highly consistent even in the 1800s.  Most comb steel for cylinder
musical boxes was English and of a consistently high quality.  When
machined, hardened and tempered the combs were tuned by hand and ear
against a master comb, often with spectacular results.  There were,
apparently, few rejects.  Even so, most defects were thought to be due
to human error, not faults within the steel.  The fact remains that the
combs of most surviving cylinder musical boxes, that have not rusted
or suffered mechanical damage, have stayed in tune.

The next point to consider is the manufacturing process irrespective
of the quality of the material being processed.  All musical box comb
teeth are cantilever springs (encastré at one end and with elastic
properties.)  Unlike the disc musical box combs, cylinder musical box
combs were tuned to all sorts of different scales (called gammes) to
accommodate all of the notes required.  Those that survive to this day
show little signs of change in pitch of individual teeth other than
that due to wear and tear, rust, lead corrosion and contamination by
grease and dirt.

Cylinder musical box combs were made from standard steel stock.  The
thickness of the steel was chosen to minimise cost of both the steel
and subsequent machining.  The thinner steel lowered the cost and the
weights added to achieve the bass notes.  It made little difference to
the ability to tune a tooth to that of a master comb.  The combs were
machined so that the minimum amount of work was needed to tune teeth
to their target pitch with adjustments to remove metal from the tips
to raise the pitch rather than at the base of the tooth to lower the
pitch.  As Mark Singleton pointed out, it was easy to change a bass
tooth's pitch by removing slivers of lead to raise its pitch.

A common fault is broken teeth, often as a result of metal fatigue
or the metal being overstressed.  Metal fatigue is a result of cyclic
forces acting over a long period of time.  The microscopic granular
structure of the metal deforms to an extent when it can no longer
contain the forces.  The spring effect diminishes and the tooth snaps.

This can occur for a number of reasons.  Assuming satisfactory
hardening and tempering there can still be material and machining
faults.  Inclusions within the material and sharp edges left as a
result of machining (the notch effect) can reduce the strength of
the material and cause fracture at those points.  But this has nothing
to do with tuning scales.

During the life of a comb it may have been slightly repositioned.
The further the engagement between tooth tips and cylinder pins the
higher the lift and the louder the result.  Also, the greater the
stress imposed.  Bulleid's 'rule of thumb' for cylinder boxes was that
the lowest bass tooth lift should not exceed 75 thousandth inches, and
at least half that for the extreme treble.  It would seem reasonable
to assume similar assumptions applied to the manufacture of disc box
combs.  Is there any further information available?

Bulleid stressed (no pun intended) the importance of the stiffness
of a tooth in respect of its musical performance.  Stiffness depends
on the overall dimensions of the tooth and its springiness after
tempering.  Not much historical data exists on where tuning started.
Perhaps it started at the most vulnerable top treble tooth.  Within
limits the actual consistency of spring steel material has little
effect on the ability to tune a tooth of the same dimensions to the
same pitch.

Work carried out by others, such as the late John Powell, confirmed
Mark's other observation that there were irregularities in the actual
pitch of the comb teeth.  To what extent each tooth was exactly tuned
to a master comb is not known.  It would seem sensible to assume that
the master comb was used in the same way that a piano tuner would use
a tuning fork or for an orchestra to tune to the piano.  The majority
of tuning will have been done by ear by different people.

There was a tendency to tune bass teeth a few cents lower than the
designated scalar pitch and the treble notes higher.  The reason is the
way in which the ear and brain processes the vibrations it receives.
The frequency differences between the octaves above or below allows the
ear to distinguish the vibrations it receives.  These 'overtones' can
create the effect of brightness when any instrument is tuned well.  But
does it have anything to do with the reason for choosing a particular
scale?

Finally, to what extent does a variation in material quality and heat
treatment affect the ability to tune a tooth to a predetermined pitch?
The answer is very little but to explain it sensibly would take a long
time.  There is so much more to learn but if only restorers would
record and publish their records we may achieve the answer as to: Why?

The MMD forum is such an excellent way to allow such debates to take
place and I hope their will be much more comment from others.  Mark
Singleton's contribution is a good example.

Paul Bellamy