Accurately measuring diameter of very fine copper wire ?

[N_Cook]

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?

 

[alchazz]

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?

In my analytical work on failed electronics, I would take a wire like
that and mount it vertically in a mounting medium that is used for cross-
sections. I would mount several samples near to each other. Then I would
cross-section the mount and measure the diameters using a microscope with
a calibrated filer eyepiece. The diameters with/and without the coatings
would be provided to the customer along with an average value.

At the time I was doing it, my lab would charge about $100 for that.
Depends how important it is to you for spending that much. Since I am
retired, I'm sure it's more expensive now.

 

[William Sommerwerck]

Here's a thought...

If you could find a table of resistance-per-unit-length for various wire
gauges, and had an ohmeter that could accurately read low resistances...

 

[Smitty Two]

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?

You only need two things:

1. A toolroom micrometer (not to be confused with a homeowner mic, or a
machinist's shop mic.)

2. The skill to use it.

I can't think in fractions of a millimeter very well, but a good
toolroom mic will read directly to 0.0001", and inferentially to
0.00001" within +/- 0.00002" or so.

The trick to not "crushing" the wire is to use the wire like a feeler
gauge, between the jaws of the mic., closing the mic slowly while
feeling for friction.

We buy 42 AWG single poly for guitar pickups, and order it specifically
as "min to nom." (minimum to nominal diameter.) The supplier checks his
stock with a toolroom mic, and ships only those spools on the low end of
the tolerance range. We verify it before putting it to use.

 

[GregS]

Here's a thought...

If you could find a table of resistance-per-unit-length for various wire
gauges, and had an ohmeter that could accurately read low resistances...

I thought getting good contact would be a problem. Soldering might work.

If you could not get an optical scale, perhaps a printer could be used to make a scale.
600 dpi would be less than .04 mm.


greg

 

[N_Cook]

You only need two things:

1. A toolroom micrometer (not to be confused with a homeowner mic, or a
machinist's shop mic.)

2. The skill to use it.

I can't think in fractions of a millimeter very well, but a good
toolroom mic will read directly to 0.0001", and inferentially to
0.00001" within +/- 0.00002" or so.

The trick to not "crushing" the wire is to use the wire like a feeler
gauge, between the jaws of the mic., closing the mic slowly while
feeling for friction.

We buy 42 AWG single poly for guitar pickups, and order it specifically
as "min to nom." (minimum to nominal diameter.) The supplier checks his
stock with a toolroom mic, and ships only those spools on the low end of
the tolerance range. We verify it before putting it to use.

Yes for the Landola pickup, the number of turns from measuring wire diameter
then weight/weighted circumferences and also by the 7.7Kohm ,assuming both
pickups are the same then this gives the same number , but does not agree
with counting a sample 1000 turns and ratioing , they were in fact laquered
together to defeat counting-off

For 2 reels of enamelled wire labelled as 2.4 thou/mil and 2.8 thou/mil and
my micrometer that has a 2Kg closure force (just checked via spring and
kitchen scales) before the torque clutch disengages.
I have to DIVIDE the reading by 1.15 for the 2.4 thou wire and 1.25 for the
2.8 thou wire. I suppose this is to do with the enamel thickness and the
wire is specified as the metallic diameter. I assume my wire gauge v
wight/resistance tables are for bare copper, not actually specified.

 

[N_Cook]

We buy 42 AWG single poly for guitar pickups, and order it specifically
as "min to nom." (minimum to nominal diameter.) The supplier checks his
stock with a toolroom mic, and ships only those spools on the low end of
the tolerance range. We verify it before putting it to use.

I would like to email directly and discuss winding guitar pickups.
Please email:
robert DOT a DOT macy AT gmail DOT com

Where do you buy 42 Awg wire?

We buy small quantities of 30-36 Awg at exceptionally competitive
pricing from
Fay Electric Wire Corp. (800) 245-9473
752 North Larch Avenue
Elmhurst, IL 60126
[not associated with them, or gain by posting this.]

Robert


I have emailed to you

 

[William Sommerwerck]

I wind 100 turns of fine gauge wire on a smooth rod ...

Tightly, of course.

... and measure the length with good calipers,
then divide by 100.

There's a story that Edison asked a young man (presumably an apprentice) to
measure the volume of several light bulbs. The apprentice stated with a
ruler, a pair of calibers, trying to get precise measurements so he could
calculate the volume. When Edison saw him fussing around, he grabbed one of
the bulbs, filled it with water, and dumped into a graduate. I suspect the
whole exercise was intended to make a point, rather than measure the bulbs.

Unfortunately, I can't think of an equivalent "clever" way to indirectly
measure the wire's diameter.

 

[Smitty Two]

they were in fact laquered
together to defeat counting-off

They were in fact laquered together, hopefully in a vacuum chamber, to
keep them from buzzing.

 

[Smitty Two]

There's always a way. Measure out a long length of wire such as 100
meters. Carefully weigh the 100m of wire. The density of copper is
8.92 gm/cm^3. The rest is math, which I won't attempt without my
morning coffee blast.

Unfortunately, if you don't know what the insulation is, you don't know
squat.

 

[N_Cook]

If the coil has been covered in lacquer, then repeated soakings in lacquer
thinner should remove the lacquer that was used to bind the coil.

Some careful, light brushing, and agitation could speed up the removal. It
should become apparent when the lacquer has been removed as all of the
windings will be loose, and a final rinse in clean thinner should be enough
to remove any traces.

It would probably be helpful to have a jig/form handy to place the coil
windings on, something like two spaced dowels about the same size as the
inside dimension of the coil, mounted to a piece of wood or other material.
Then with a center point located between the dowels, the coil would turn
fairly easily for counting turns.

The former has to be demountable. I've made but not had the time to try
winding yet. A platic binder spine warmed up and opened out to a flat
bottomed V. Doubled up , end over end, and 2 cuts per side , so can easily
cut the remainder after winding. Some thick PTFE wound around to decrease
the width a bit to fit the trough, some thin elastic laid across , so under
the wiring, to tie together in loops after winding. Well thats my theory.
The original one probably failed because in the process of wrapping with
tissue paper and adjusting to fit into the trough , the layup was seriously
disturbed/ stressed, before lacquering.

 

[Smitty Two]

<snip>

Just a heads up that I emailed you per your request. My email as shown
on usenet is munged.

 

[Smitty Two]

Ummm... the wire gauge is a measure of the copper diameter (or more
correctly, the cross sectional area), without the insulation. The
added weight of the insulation coating is quite small when compared to
the larger density of the copper. With flexible small diameter copper
wire, there are a few additives, but they constitute a tiny percentage
of the wire composition. My guess is the biggest error will be in the
accuracy of the weight, not in the effects of the insulation.

Yes, you're probably right. I studied my supplier's reference chart some:

<http://www.mwswire.com/pdf_files/mws_tech_book/copper_magnet_wire_data.p
df>

and it looks like single poly adds about 4% to the weight of 42 AWG,
while a triple build would add nearly 10%. Not exactly negligible, but
not enough to say that 42 triple could be mistaken for 41 single, by
weight.

Still, when it comes to alternative approaches, I think measuring the
resistance of a substantial length is going to be easier in terms of
instrumentation precision than weighing it would be.

But, I've been happy with a good micrometer; it's a little quicker than
measuring off a football field's worth of frail wire.

 

[Nelson]

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?

I remember a "homebrew science" experiment I once saw where someone was
able to do very precise measurements with a homemade apparatus that
used the principle of a lever. If I recall, basically you put the
thing you want to measure the thickness of under the end nearest the
fulcrum and sheets of paper of a known thickness under the other.
Knowing the length of the lever arms, the thickness of the sheets of
paper, and the difference in the number of sheets which can be inserted
with and without the object to be measured under the other end allows
you to calculate it's thickness. I don't remember the details but
remember the precision was surprising for such a simple apparatus. And
by varying the length of the lever arms you can increase or decrease
the precision.

 

[N_Cook]

Yes, you're probably right. I studied my supplier's reference chart some:

<http://www.mwswire.com/pdf_files/mws_tech_book/copper_magnet_wire_data.p
df>

and it looks like single poly adds about 4% to the weight of 42 AWG,
while a triple build would add nearly 10%. Not exactly negligible, but
not enough to say that 42 triple could be mistaken for 41 single, by
weight.

Still, when it comes to alternative approaches, I think measuring the
resistance of a substantial length is going to be easier in terms of
instrumentation precision than weighing it would be.

But, I've been happy with a good micrometer; it's a little quicker than
measuring off a football field's worth of frail wire.

I think I agree too. Keep with your own personal same micrometer. Check the
release torque (i will keep the same flat faced compression spring in a bag
in the micrometer case and check the separation of faces is the same each
time of thin wire measurement), clean faces and check zero before use . Is
probably just as valid without going to full lab precision. And be aware the
variable thickness of enamel has to be taken into account.

 

[Smitty Two]

I think I agree too. Keep with your own personal same micrometer. Check the
release torque (i will keep the same flat faced compression spring in a bag
in the micrometer case and check the separation of faces is the same each
time of thin wire measurement), clean faces and check zero before use . Is
probably just as valid without going to full lab precision. And be aware the
variable thickness of enamel has to be taken into account.

The old guy that taught me insisted that the clutch on a mic is really
meant as a safety release. If it clicks, you've already overtorqued it
for precision measurement. That's why developing a "feel" for using it
with a bit more finesse is important, especially in the sizes we're
discussing here.

 

[Smitty Two]

It might be possible to disolve off the enamel (more
probably silicon varnish) insulation

There's quite a selection of insulations available.

http://www.mwswire.com/insspec.htm

We use polyurethane for guitar pickups. We used to make coils for
magnetic bearings, out of much thicker wire, and we used the
heat-bondable stuff. An hour in the oven gives a nice solid coil without
need for a vacuum bath. We had to do it that way because the coils were
bobbinless. After winding, they went in the oven, then the four-piece
winding bobbin was dismantled leaving the coil intact.

 

[Smitty Two]

Good luck. You must have a really good micrometer with the requisite
crystal ball accessory.

For single build wire, comparing #41 with #42awg:
min max
#41 .0029 .0033
#42 .0026 .0030
Note the overlap between .0029 to .0030. If your measurements land in
this range, you have no way to determine if it's #41 or #42.

However, that assumes that you know that it's single build insulation.
If I compare the diameters between single build and triple build:
single triple
min max
#41 .0029 .0039
#42 .0026 .0035
we have an even larger overlap between .0029 and .0035. Any
measurement landing in this range requires a decision from the crystal
ball.

True enough. In my case, I'm trusting the label on the spool and merely
verifying that I rec'd "min to nom" as requested. Not quite the same as
reverse engineering an unknown pickup.

OTOH, I knew Seymour Duncan 30 years ago. Never wound pickups for him
but did and still do some of his machining. He started with a couple of
hundred square feet of loft and did the reverse engineering himself,
counting off turns, measuring resistance, checking turns per layer, and
otherwise analyzing the pickups he reproduced. So it's certainly
possible.

Also, based on my limited experience, measuring anything with a
micrometer tends to be rather subjective. A friend that own a local
machine shop once mumbled that he could pass a micrometer around the
shop, ask everyone to make a specific measurment, and get a rather
wide range of numbers depending upon technique. It's especially bad
measuring soft items, such as plastic and possibly thin copper.
Consistency of measurement is probably equally subjective.

Agreed, absolutely. Many people don't realize how much skill and
experience is involved in something that looks so easy. My old boss used
to spend half his time visiting customers' QC departments and teaching
them how to use measuring instruments appropriately.

For
anything in ten-thousanths his 12" Nikon optical comparator is the
correct measuring instrument:
<http://www.nikoninstruments.com/content/view/full/3345>

Unfortunately, my previous suggestion, of using an optical reticule to
measure the diameter, will not work. It can probably resolve
dimensions to within a thousanth, but to distinguish between two
adjacent small wire diameters requires resolution down to the
ten-thousanths. A hand loop and reticule are just not accurate
enough. You might want to have the local machine shop, which
presumably has an optical comparator, do the measuring.

We are a local machine shop, but I never thought a comparator would be
accurate enough for this application. I'm going to give it a try,
though.

 

[Jim Yanik]

True enough. In my case, I'm trusting the label on the spool and merely
verifying that I rec'd "min to nom" as requested. Not quite the same as
reverse engineering an unknown pickup.

OTOH, I knew Seymour Duncan 30 years ago. Never wound pickups for him
but did and still do some of his machining. He started with a couple of
hundred square feet of loft and did the reverse engineering himself,
counting off turns, measuring resistance, checking turns per layer, and
otherwise analyzing the pickups he reproduced. So it's certainly
possible.

It's called "precision" in Metrology.

http://www.westgard.com/isoglossary.htm


precision: closeness of agreement between independent results of
measurement obtained under stipulated conditions (»6)
NOTES
1. Precision is a qualitative concept. Its quantitative counterpart is
imprecision, which is computed as a standard deviation or a coefficient of
variation of the measurement results.
2. Imprecision depends critically on the specified conditions.
3. Standard deviation expressing imprecision may depend on the value of the
measurand; the phenomenon is called heteroscedasticity.


"repeatibility" is another factor

repeatability (of results of measurements): closeness of the agreement
between the results of successive measurements of the same measurand
carried out under the same conditions of measurement (1)
NOTES
1. Repeatability is a qualitative concept. Its quantitative counterpart is
standard deviation of repeatability or coefficient of variation of
repeatability of the measurement results.
2. Repeatability may depend on the value of the measurand.

Agreed, absolutely. Many people don't realize how much skill and
experience is involved in something that looks so easy. My old boss used
to spend half his time visiting customers' QC departments and teaching
them how to use measuring instruments appropriately.



We are a local machine shop, but I never thought a comparator would be
accurate enough for this application. I'm going to give it a try,
though.

I went thru USAF's Precision Measurement Equipment (PME) school back in
1971,9 months in Denver,CO.,then 2.5 yrs at Hanscom Fld's PMEL lab.

 

[N_Cook]

The old guy that taught me insisted that the clutch on a mic is really
meant as a safety release. If it clicks, you've already overtorqued it
for precision measurement. That's why developing a "feel" for using it
with a bit more finesse is important, especially in the sizes we're
discussing here.

But that realm of low forces is where grime/gummy lubricant etc affects the
feel. I would say that , on my Moore & Wright one anyway, someone has
determined that 2Kg clutch release force is about right for consistency.
Obviously fine copper wire is going to compress but it will compress
consistently, gauge for gauge, on a consistent 2 Kg force over the diameter
of the plattens. So I'd disagree with your expert.