On Fri, 27 Sep 2013 16:31:09 +0100, N_Cook wrote:

On 27/09/2013 15:53, Adrian Tuddenham wrote:
wrote:

On Tuesday, August 4, 2009 10:38:19 AM UTC-4, N_Cook wrote:
[...]
How about a longish length , folded 6 times until 64 wires. Maybe
longer/more bulk. Hand twist together until it will not sensibly tighten any
more. Take average diameter, use packing factor allowance, and infer for 1
wire diameter,..

Make the bundle by winding the wire 32 times around two spaced pegs so
as to be certain that all the 64 wires this produces between the pegs
are parallel and not intertwined. Slip the wire off the pegs and do not
twist it, but squeeze the parallel section so that it takes up a
cylindrical shape.

Wrap another length of the same wire tightly around the outside of the
cylindrical section for a known number of turns (20 at least). Unwind
the wire and measure its length and divide by 20 to calculate the mean
circumference of one turn.

Do exactly the same thing with a length of wire whose diameter you do
know (probably something much larger, so that you can measure it
easily). You may not be able to wrap as many as 20 turns, so adjust
the divisor accordingly.

The ratio of the lengths of the one-turn circumferences will be the
square of the ratio of the wire diameters.




Friday afternoon after a long tiring day yesterday, I cannot
thought-experiment my way into this. I'll have to have a go with some
thicker wire to start with, to work out the method you've outlined

Why not measure the resistance and calculate the diameter from that?
(Assumes the wire is just copper and is round).

On 27/09/2013 16:40, Pat wrote:
On Fri, 27 Sep 2013 16:31:09 +0100, N_Cook wrote:

On 27/09/2013 15:53, Adrian Tuddenham wrote:
wrote:

On Tuesday, August 4, 2009 10:38:19 AM UTC-4, N_Cook wrote:
[...]
How about a longish length , folded 6 times until 64 wires. Maybe
longer/more bulk. Hand twist together until it will not sensibly tighten any
more. Take average diameter, use packing factor allowance, and infer for 1
wire diameter,..

Make the bundle by winding the wire 32 times around two spaced pegs so
as to be certain that all the 64 wires this produces between the pegs
are parallel and not intertwined. Slip the wire off the pegs and do not
twist it, but squeeze the parallel section so that it takes up a
cylindrical shape.

Wrap another length of the same wire tightly around the outside of the
cylindrical section for a known number of turns (20 at least). Unwind
the wire and measure its length and divide by 20 to calculate the mean
circumference of one turn.

Do exactly the same thing with a length of wire whose diameter you do
know (probably something much larger, so that you can measure it
easily). You may not be able to wrap as many as 20 turns, so adjust
the divisor accordingly.

The ratio of the lengths of the one-turn circumferences will be the
square of the ratio of the wire diameters.




Friday afternoon after a long tiring day yesterday, I cannot
thought-experiment my way into this. I'll have to have a go with some
thicker wire to start with, to work out the method you've outlined

Why not measure the resistance and calculate the diameter from that?
(Assumes the wire is just copper and is round).


For this very fine wire it is only ever enamelled/lacquered wire I ever
deal with, being used as magnet wire, pick-up coils and the like

N_Cook wrote:

On 27/09/2013 15:53, Adrian Tuddenham wrote:
wrote:

On Tuesday, August 4, 2009 10:38:19 AM UTC-4, N_Cook wrote:
[...]
How about a longish length , folded 6 times until 64 wires. Maybe
longer/more bulk. Hand twist together until it will not sensibly
tighten any more. Take average diameter, use packing factor allowance,
and infer for 1 wire diameter,..

Make the bundle by winding the wire 32 times around two spaced pegs so
as to be certain that all the 64 wires this produces between the pegs
are parallel and not intertwined. Slip the wire off the pegs and do not
twist it, but squeeze the parallel section so that it takes up a
cylindrical shape.

Wrap another length of the same wire tightly around the outside of the
cylindrical section for a known number of turns (20 at least). Unwind
the wire and measure its length and divide by 20 to calculate the mean
circumference of one turn.

Do exactly the same thing with a length of wire whose diameter you do
know (probably something much larger, so that you can measure it
easily). You may not be able to wrap as many as 20 turns, so adjust
the divisor accordingly.

The ratio of the lengths of the one-turn circumferences will be the
square of the ratio of the wire diameters.




Friday afternoon after a long tiring day yesterday, I cannot
thought-experiment my way into this. I'll have to have a go with some
thicker wire to start with, to work out the method you've outlined

I believe some variation of this method was the standard way of
determining the gauge of wires many years ago. It had the advantage
that the reading was averaged over a number of wires (some of which may
not have been exactly circular in cross-section) and, once established,
could be used by anyone with a ruler to give a fairly good degree of
accuracy.


--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk

On 9/27/2013 8:46 AM, N_Cook wrote:
On 27/09/2013 16:40, Pat wrote:
On Fri, 27 Sep 2013 16:31:09 +0100, N_Cook wrote:

On 27/09/2013 15:53, Adrian Tuddenham wrote:
wrote:

On Tuesday, August 4, 2009 10:38:19 AM UTC-4, N_Cook wrote:
[...]
How about a longish length , folded 6 times until 64 wires. Maybe
longer/more bulk. Hand twist together until it will not sensibly
tighten any
more. Take average diameter, use packing factor allowance, and
infer for 1
wire diameter,..

Make the bundle by winding the wire 32 times around two spaced pegs so
as to be certain that all the 64 wires this produces between the pegs
are parallel and not intertwined. Slip the wire off the pegs and do
not
twist it, but squeeze the parallel section so that it takes up a
cylindrical shape.

Wrap another length of the same wire tightly around the outside of the
cylindrical section for a known number of turns (20 at least). Unwind
the wire and measure its length and divide by 20 to calculate the mean
circumference of one turn.

Do exactly the same thing with a length of wire whose diameter you do
know (probably something much larger, so that you can measure it
easily). You may not be able to wrap as many as 20 turns, so adjust
the divisor accordingly.

The ratio of the lengths of the one-turn circumferences will be the
square of the ratio of the wire diameters.




Friday afternoon after a long tiring day yesterday, I cannot
thought-experiment my way into this. I'll have to have a go with some
thicker wire to start with, to work out the method you've outlined

Why not measure the resistance and calculate the diameter from that?
(Assumes the wire is just copper and is round).


For this very fine wire it is only ever enamelled/lacquered wire I ever
deal with, being used as magnet wire, pick-up coils and the like

I'll skip the lecture on going to insane lengths for no benefit and ask
one question...

Why do you care?

What is it about your application that requires such precision in the
diameter of the insulation?

The obvious solution, and one that is actually related to winding coils,
is to wind 20, 500, how many you feel are needed, turns on a solenoid.
Measure the length of the solenoid and divide by N. You can get
arbitrary precision for THIS spool of wire.

For manually wound coils, the number of turns you can get in a unit volume
is related more to the skill of the winder than the dimensions of the wire.

On Tuesday, August 4, 2009 7:38:19 AM UTC-7, N_Cook wrote:
Not the first time I've met this problem.
Say nominally about 0.05mm . With a micrometer, how much are you compressing
it? could easily be out by 20 percent out and squaring that if using weight
to length via density or resistance calculation via resistivity, is very
iffy.
If access to a microgram resolution of weighing scales then a few metres of
the wire and density of copper and allowance for enamelling , but no highly
accurate weighing machine. Optically comparing under a microscope needs
known diameter standards.
How about a longish length , folded 6 times until 64 wires. Maybe
longer/more bulk. Hand twist together until it will not sensibly tighten any
more. Take average diameter, use packing factor allowance, and infer for 1
wire diameter, how better accuracy might that be.?
If I start from known good coil of say 46swg enamelled wire and do this 64
wire trick , to work backwards, how accurate/reliable would the manufacture
sizing be ?
Any other ideas?

How about winding the wire onto a piece of 1/16th inch piano wire, do ie; 100 turns and then measure the overall length. Presumed that the wire has no enamel.


My best way.
KW

Adrian Tuddenham wrote:

I believe some variation of this method was the standard way of
determining the gauge of wires many years ago. It had the advantage
that the reading was averaged over a number of wires (some of which may
not have been exactly circular in cross-section) and, once established,
could be used by anyone with a ruler to give a fairly good degree of
accuracy.


I saw a prototype optical measuring system about 30 years ago that
was being refined to measure copper wire on the production line. It was
already accurate to under 1/1000 inch, at production speeds.

--
Anyone wanting to run for any political office in the US should have to
have a DD214, and a honorable discharge.

On 9/27/2013 3:51 PM, wrote:
On Tuesday, August 4, 2009 10:38:19 AM UTC-4, N_Cook wrote:
Not the first time I've met this problem.
Say nominally about 0.05mm . With a micrometer, how much are you compressing
it? could easily be out by 20 percent out and squaring that if using weight
to length via density or resistance calculation via resistivity, is very
iffy.
If access to a microgram resolution of weighing scales then a few metres of
the wire and density of copper and allowance for enamelling , but no highly
accurate weighing machine. Optically comparing under a microscope needs
known diameter standards.
How about a longish length , folded 6 times until 64 wires. Maybe
longer/more bulk. Hand twist together until it will not sensibly tighten any
more. Take average diameter, use packing factor allowance, and infer for 1
wire diameter, how better accuracy might that be.?
If I start from known good coil of say 46swg enamelled wire and do this 64
wire trick , to work backwards, how accurate/reliable would the manufacture
sizing be ?
Any other ideas?


Take 10 meter. Measure the resistance.
resistance is 0.0175 ohm per meter per square millimeter.
From this you can find the area.
Area is 1/4 pi d*d This will give the real copper diameter.

Two prerequisites. Pure copper and the wire is round.

On 30/09/2013 13:02, tuinkabouter wrote:
On 9/27/2013 3:51 PM, wrote:
On Tuesday, August 4, 2009 10:38:19 AM UTC-4, N_Cook wrote:
Not the first time I've met this problem.
Say nominally about 0.05mm . With a micrometer, how much are you
compressing
it? could easily be out by 20 percent out and squaring that if using
weight
to length via density or resistance calculation via resistivity, is
very
iffy.
If access to a microgram resolution of weighing scales then a few
metres of
the wire and density of copper and allowance for enamelling , but no
highly
accurate weighing machine. Optically comparing under a microscope needs
known diameter standards.
How about a longish length , folded 6 times until 64 wires. Maybe
longer/more bulk. Hand twist together until it will not sensibly
tighten any
more. Take average diameter, use packing factor allowance, and infer
for 1
wire diameter, how better accuracy might that be.?
If I start from known good coil of say 46swg enamelled wire and do
this 64
wire trick , to work backwards, how accurate/reliable would the
manufacture
sizing be ?
Any other ideas?


Take 10 meter. Measure the resistance.
resistance is 0.0175 ohm per meter per square millimeter.
From this you can find the area.
Area is 1/4 pi d*d This will give the real copper diameter.

Two prerequisites. Pure copper and the wire is round.



That applies to most copper wire but not these finest dimensions where
that formula breaks down

On Mon, 30 Sep 2013 14:04:04 +0100, N_Cook wrote:

On 30/09/2013 13:02, tuinkabouter wrote:
On 9/27/2013 3:51 PM, wrote:
On Tuesday, August 4, 2009 10:38:19 AM UTC-4, N_Cook wrote:
Not the first time I've met this problem.
Say nominally about 0.05mm . With a micrometer, how much are you
compressing
it? could easily be out by 20 percent out and squaring that if using
weight
to length via density or resistance calculation via resistivity, is
very
iffy.
If access to a microgram resolution of weighing scales then a few
metres of
the wire and density of copper and allowance for enamelling , but no
highly
accurate weighing machine. Optically comparing under a microscope needs
known diameter standards.
How about a longish length , folded 6 times until 64 wires. Maybe
longer/more bulk. Hand twist together until it will not sensibly
tighten any
more. Take average diameter, use packing factor allowance, and infer
for 1
wire diameter, how better accuracy might that be.?
If I start from known good coil of say 46swg enamelled wire and do
this 64
wire trick , to work backwards, how accurate/reliable would the
manufacture
sizing be ?
Any other ideas?


Take 10 meter. Measure the resistance.
resistance is 0.0175 ohm per meter per square millimeter.
From this you can find the area.
Area is 1/4 pi d*d This will give the real copper diameter.

Two prerequisites. Pure copper and the wire is round.



That applies to most copper wire but not these finest dimensions where
that formula breaks down
I'm a machinist and I measure parts that have features smaller than
the wire in question. I use an optical comparator sometimes for these
measurements. The comparator has a large screen that makes these
measurements easy. However, I also have a small hand held comparator
loupe that is used similar to a magnifier. The difference though is
that this comparator has a reticle that touches the item to be
measured. This helps to avoid parallax errors. Mine is made by PEAK
but there are several other brands available. It has a 10x lens in it
and will accurately measure features as small as .025mm.
Link:
http://www.peakoptics.com/index.php?...oducts _id=12
Eric