Re: Dimensions of my flat spiral coil

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Jim Lux by way of Terry Fritz <twftesla-at-qwest-dot-net>" <jimlux-at-earthlink-dot-net>

>
> >Can you measure changes of <0.5% reliably?  I am reminded of the problem
> with inexpensive DMMs... It may have 3.5 digits on the display, but only
be
> accurate to 5%.
>
> I'm using a BK Precision model 875A LCR Meter.  According to the spec
sheet
> at the 20mH setting that I used for the coil, the test condition is 1KHz,
> 100uA and the accuracy is +/-(2%rdg + 1dgt).

OK... at 1kHz, the parasitic C isn't going to be a big deal, compared to the
10's of mH
Two percent of the reading is, given a reading of 6.xx mH, around 0.12 mH..
so, your three coils all measure the same, within the uncertainty of the
instrument.    Does the spec sheet give any data on whether that is a 2
sigma uncertainty (often the case when not otherwise specified...  if error
is gaussian, then it's roughly a 95% confidence interval)?


>
> >Off hand, the relative values between windings are probably more accurate
> than the absolute value, especially since they are all about the same
value.
>
> I believe the difference is due to the condition I stated earlier.  The
> three outer wires terminate at 120 degrees from each other so this means
> their lengths are slightly different.  Had I terminated all three wires at
> the same point, I'm confident I would have three identical inductances.
> This is why I brought the point up, btw.  Just in case someone else builds
a
> three wire coil and needs three identical inductances, then they should
> terminate all three wires together.

Naahhh.. the difference in length is so small, compared to the overall coil.
You have hundreds of turns.  A third turn isn't going to make all that much
difference. A bounding case would have L proportional to N^2 (and in
reality, the exponent is lower).  So, if you have 400 turns, adding a whole
turn is only going to change the inductance by (400/401)^2.. about 0.5%,
which is a lot lower than your measurement uncertainty.


>
> >What frequency are the measurements being made at?
>
> Noted above.
>
> >What sort of surroundings and what was the test lead configuration (the
> test leads themselves could easily have 10 uH of inductance, or more
> important, the uncertainty in the test lead inductance is on that order.
>
> The surroundings were clear of any obstruction for a four feet radius.
Then
> there are some electronics equipment such as TV and VCR, 10 feet away is
my
> computer.  The walls and ceiling in my work area are covered with space
> blanket, more for the insulation value than for the Faraday value.  But
> there is an aluminum coating in the space blanket.  The insulated test
leads
> are 10" long and came with the unit.  This said, none of these factors
would
> account for consistent differences in readings.  I have also measured the
> exact same readings while the coil was in different areas of the room.
I'm
> sure the difference are due to wire length differences.

The conductive surface would help for electrostatic shielding, but do
nothing for magnetic shielding.  For "good" metrology, you want to think in
terms of factors of 10... If the dimensions of the unit under test is 1
foot, you want a clear space 10 times that.  Failing that (since it rapidly
becomes unreasonable), you can deliberately perturb the surroundings, and
see how much of a difference it makes.  For instance, if you put a steel
plate 2 feet away, and repeat the measurement, and it doesn't vary more than
the measurement uncertainty, then you can kind of figure that you're
insensitive to disturbances of that type.


>
> >A measurement uncertainty of 10-20 uH would be on the order of the
> differences you are seeing.
>
> Only if the measurements produced random results among the wires.  But a
> specific wire has a specific inductance and it is the same regardless of
> where the reading is taken.

The test leads have noticeable inductance (a free wire in space is around 1
microhenry per meter), particularly if it is anywhere near the other lead.
Ideally, you'd have twisted leads all the way from the meter to the device
under test.  In inductance measurement, minimizing the area of the loops is
the name of the game.  The hot stuff guys use things like braided two
conductor litz wire (i.e. imagine two bundles of fine insulated wire all
carefully intermixed).

>
> >Not to be nit picky here, but I've gotten "beat up" in reviews for
> seemingly simple measurements where all these effects weren't at least
> quantified, if not measured.  Precision metrology is much an art, and
> remarkably painstaking, even for simple measurements.
>
> In this case, it is clear about the accuracy.  And I believe the wire
length
> differences is an adequate explanation for the different inductance
> readings.
>
> >Interestingly, if you hook the windings in parallel, since they are very
> tightly coupled, I'd ballpark the resulting inductance at 9 times that of
> the single winding, or around 50-60 mH, just on the N^2 principle.
>
> Well, let's try that and see what we get.  I hadn't thought of that.

As pointed out other places, I was wrong here... In parallel, it should
measure pretty much the same.  The Q might be better (less resistance).
>
>
>