"The Natural Philosopher" wrote in message
...
the only reliable scales are beam balance. ALL the standonm pressiure ones
are in the end utter ****.

The problem with beam balance scales is that they take a long time to get a
reading because you have to add measuring weights in a binary chop until the
beam is just balanced. And they don't give a reading on an analogue or
digital scale: you have to manually add up all the balancing weights that
you have added.


Spring or strain gauge scales gives an instant reading - though they *may*
be less consistent. Is that the usual way in which they are "utter ****" -
that reading the same object on multiple occasions gives different results?
Or is it that there is a systematic error - eg they under-read at the low
end of the scale and over-read at the high end, so a 1 kg weight added to a
1 kg weight and a 1 kg weight added to 100 kg will register as different
extra weights?

I've seen scientific "balances" for laboratory use which can read to the
nearest milligramme for overall weights up to maybe 10 grammes. Are you
saying that this degree of accuracy is spurious? If so, a lot of lab
readings would be found to be incorrect and people would soon start to
complain. So strain gauges must be reliable.

I suppose one could argue that *any* device (not just weighing scales) that
uses a spring to balance out the position of a needle on a gauge against an
analogue of the quantity that you are measuring (eg voltage proportional to
speed, weight, oil pressure, engine revs, temperature) is equally "****"
because you may not get a linear reading across the scale or may get
different readings for several measurements of the same quantity, because of
non-linearities in the rate of the spring (constant of deflection versus
force) or because the needle sticks slightly on its bearings.