With my manual lathe, a beautiful 1970 Colchester
Student model, I find it hard to turn parts to fit each other accurately.
Too tight and they wonít come apart again and too loose they wobble -
particularly when a mechanism has worn in. The difference between a tight
fit and a sloppy fit is only .01mm, much too small to see. My lathe tools
donít always cut smoothly and sometimes they donít cut parallel,
precision is certainly not as easy as it looks. My inaccuracies are partly
the fault of my old lathe but mainly because I never did an apprenticeship
in how to use it. I bought it and gradually learnt by trial and error,
occasionally picking up tips watching skilled machinists. After years of
struggle, I can just about make things fit, but I have concentrate Ė
itís never something I take for granted.
Until the late 19th century parts didnít
usually fit together. They needed a skilled Ďfitterí to file and
adjust them to make them fit. This changed with the ĎAmerican systemí
of interchangeable parts. Making parts more accurately made them cheaper
because they didnít need this skilled Ďfittingí.
Today we take interchangeable parts for granted and
the idea that precision saves money has spread. Builders use laser levels
to make walls and floors accurately vertical and horizontal because it
saves time Ė thereís less measuring to do. This is partly why itís
often cheaper to knock down an old house and build a new one, rather than
to restore the old one.
Training to be an engineer, I was taught to design
parts only as precise as they needed to be. Every dimension on a drawing
had a Ďtoleranceí, the maximum permitted error from the dimension.
Although I now never do formal drawings of the parts I make, I still,
almost subconsciously, decide how precise each dimension needs to be.
Historically the more precise a part, the longer it
took to make and the more it cost. But now machined parts are made with
CNC tools this is less true. The precision I still struggle to achieve
with my Colchester student lathe is effortless with todayís CNC tools.
My friend Will Jackson, who makes robots, has a Japanese Mazac machine
that is bigger than a small house (it also cost more than one). It works
through the night unattended producing perfect parts. An extra smooth
finish can take longer and really high precision needs a more expensive
machine but generally the cost of precision is now negligible.
But the quest for precision isnít always logical.
Will could have managed with his existing machines, he bought his
Mazac at least partly because he was fascinated by it. Machinists and
other craftsmen who are taught to work to the highest precision possible
find it hard to lower their standards and I sympathise. As my craft skills
have gradually improved, I also now find it hard to go backwards. I have a
sense of pride in my work, even if the precision isnít necessary.
When electric lighting arrived electrical engineers
measured the electrical resistance of their circuits with a device called
a Wheatstone Bridge. I learnt how to use one at school. It was quite an
elaborate procedure involving several measurements and calculations to get
the answer. When young engineers started using multimeters in the 1920s,
which gave an immediate result, the older engineers were outraged and
complained that the meters werenít accurate enough. It is true they
werenít as accurate as Wheatstone Bridges, but perfectly good enough for
most things. Despite this, multimeters were only fully accepted once the
older engineers had died off.
I recently helped with a joint project between
Londonís Royal College of Art and Imperial College (a respected London
art school and engineering college). The students were making arcade
machines. At the end the students complained that the college should stock
a few more essential components like bearings. But a lecturer said there
were over 6 million different bearings and they couldnít stock them all.
He didnít want to stock a few because it might teach the students bad
habits, not specifying the perfect bearing for the job. He turned out to
be a bearing expert. With his specialist knowledge, it was hard to accept
lower standards. For their projects, the students just needed to reduce
friction a bit and their machines only had to work for a day Ė almost
any bearing would have done. I only stock about 12 different bearings in
my workshop and very rarely need any other ones.
In many situations, precision is an illusion. A few
years ago I made some sliding gates for the Tiger enclosure at London zoo.
The architectís drawings were so precise I assumed the dimensions of the
cages, to which my gates fitted, would be too. But I arrived to find the
cage installers cursing and hacking all their galvanised parts with an
angle grinder to make them fit. The resulting spaces left for my gates
were up to 50mm too big or too small. I followed suit and hacked my gates
with my angle grinder too. I donít have a lot of experience of building
sites, but Iím sure this modern day Ďfittingí is still common with
any work on old buildings. Certainly in the case of my gates, they worked
fine after being hacked Ė the precision of architectsí drawings
Precision in science is also often an illusion.
Although scientific apparatus can measure ever more precisely, the results
of most experiments are now shades of statistical significance. The New
Scientist magazine recently admitted that up to 80% of todayís
experimental results are un-reproduceable. They claim this is due to a
combination of poor statistical methods and the pressure to come up with
Science and engineering are usually lumped together
but they are not the same. Science is the pursuit of truth whereas
engineering is finding the best compromise for a situation Ė minimising
cost and making something work. Although the pursuit of truth is
more profound itís never really achievable whereas an engineering design
that works well is difficult but possible and is to me more satisfying.
Sadly most engineers have to be intensely
conservative about their designs. Itís
expensive to try anything new and even more expensive to rectify something
that doesnít work. Also because they rarely make their own parts, any
detail thatís imperfect is alarming to them even though the part might still
work perfectly well.
My mentality is more inventor than engineer. The
parts I make often end up a bit of a mess, particularly if Iíve had to
modify them. As long as a part works, I rarely go back and remake it. Itís more important to me to get on and finish a machine
than to fuss about details like a messy weld or a bad paint finish. When a
machine is finished there are sometimes a few parts that break or wear and
have to be remade, but this is usually because of the design was bad, not
because of the part's lack of craftsmanship.
Of course precision is partly aesthetic and nothing
to do with function. When one of my machines is nearly finished, I
sometimes remake a part just because it looks wrong and annoys me. This is
like a bad film edit. Itís usually possible to do ridiculous cuts and
they look fine but just occasionally one Ďjumps outí and distracts
from the story. Messy details on a machine can be the same - but I tend to
be over sensitive. Its often good to have another person around to
bring me back to earth.
Aesthetically, precision varies with the situation.
To look precise, things never need to be more precise than the eye can
see. Tradesmen decorators have a saying Ďhigh work isnít eye workí
(so high ceilings are less well finished). Sets and props for TV programs
used to be amazingly rough (one of my first jobs was copying a rough BBC
Dalek to make it presentable to be seen live in an exhibition). When TV
cameras only captured 625 horizontal lines, the prop makers knew exactly
what they could get away with. Most of the colours on todayís TV screens
and computer monitors still arenít precise. Lots of compromises are made
to ensure flesh tones look real, because our eyes and brains are much more
sensitive to this than any other colour.
A table I made using only a jigsaw
There is often charm in the visible lack of
precision, particularly as itís a change from all the precisely made
manufactured things around us. Iím no fan of letting it all hang out Ė
ugly craft objects like heavy mugs and chainsaw furniture. But things like
vernacular architecture, usually without a single right angle, have a
richness and charm that no architect can match. My taste is often a middle
path, things that are almost precise, but not quite. When I made furniture
in my twenties, I cut almost everything with a jigsaw. This was mainly
because I didnít have a workshop, but it was satisfying that although a
final piece looked precise, you could feel it was hand made by running
your finger along an edge. If something is hand made, thereís no point
in trying to perfectly imitate a manufactured process.
In todayís art, anything goes. Iím a fan of Clive
Headís paintings which are ultra realistic and precise. But I also often
like cartoons that are freely drawn. The craftsman mentality can get in the way. I
was once asked to do a workshop for the National Theatre prop makers and
they proudly told me they made everything to look realistic from the
nearest seat. Its an odd rule as part of the fun of theatre is the
suspension of reality, we enjoy being fooled. My wife is a puppeteer and
when watching one of her shows I enjoy getting absorbed and quickly forget
she is manipulating the characters.
A few years ago at the Kinetica Art fair I found that
most of todayís Ďkinetic artistsí look down on anything precise and
reliable. Their taste is for machines on the edge Ė that are almost
self-destructing. I enjoy watching their work even though itís not what
Itís the confusions and contradictions about
precision that make it interesting. At one end of the spectrum fine art
students aren't remotely interested. They are taught that ideas are
everything so thereís no need to have practical skills because Ďthe
workí can be fabricated by Ďothersí. Occasionally this approach can
produce amazing results, but Iím sure it could never result in a good
arcade machine. At the other
end of the spectrum skilled tradesmen find it hard to make anything that
lacks precision.They usually have no control of the design of the parts
they make, so their pride in their work is expressed by making everything
Iíve become a jack of all trades but master of
none. Woodwork, metalwork, electronics, software, graphics and sculpture
all contribute to my arcade machines. I couldnít design them without
this intimate, if also imperfect, knowledge of all the details Ė the
devil is always in the detail. Increasingly I think Iím lucky not to be
constrained by the history of art (by going to art school) or by craft
skill conventions (by doing an apprenticeship).
Iíve almost come to accept that never completely mastering my
lathe is a compromise thatís worth it.