Here's a paragraph from IETLab's web site on how their inductance standards
are made:

"Each standard inductor is a uniformly wound toroid on a ceramic core. It
has a negligible external magnetic field and hence essentially no pickup
from external fields. The inductor is resiliently supported in a mixture of
ground cork and silica gel, after which the whole assembly is cast with a
poƫting compound into a cubical aluminum case."

Sounds like their objective is to isolate the winding from as many external
influences as possible.  Of course, the same couild be said of any physical
or electrical standard.

Cheers,
Dave M

Dr. David Kirkby (Kirkby Microwave Ltd) wrote:

Here is a DIY guide to making some lab standards. It is detailed with some
component values.

http://nopr.niscair.res.in/bitstream/123456789/4848/1/JSIR%2065%286%29%20510-513.pdf

On Wed, Aug 19, 2015 at 8:48 PM, Dave M dgminala@mediacombb.net wrote:

Thats a great document, and certainly is a supperior method than doing some DIY coils, provided very stable resistors and capacitors are used. Resistors is straightforward. As capacitors, mica glass is recommended. There are some NOS russian mica high rel capacitors available for little money. I have used those as standards, however I have no history yet, since I only have the inital cal data, second cal verification is still pending, but since they are hermetic, and provided they are used at a constant temperature, should be very good.

It also annoys me that I would have to send C/L meters to calibration regularly. So having own standards is a solution, provided they are stable. Actually, the HP service manual of the older meters says no need to do L verification, as the measurement principle ensures that when doing C verifcation, L automatically is covered, since no good L standards for higher frequencies are available (which seems not really true as we have learnt). Should be true for the 4284A measurement principle too, but I have not checked.

The next question then is, how do you ensure that the drift of these DIY capacitors is good (it is just an assumption they are stable, and 50ppm per year is a challange!!. The IET 1404 primary standards are <20/<40ppm only!!). Actually 50ppm may be pretty hard with anything else than a invar air capacitor. Now these hermetic invar 1404 air capacitors are availabe from time to time on ebay (also ESI SC1000), and thay have that guaranteed low drift over years. Most often they are quite expensive, but from time to time ony pops up at a reasonable price, got mine for about 500 usd. They could be used as a gold standard for at least one value. Together with an inductive decade transformer bridge other capacitances could be calibrated (the decade transformer principle ensures inherent calibration).
And third, one could build up a quadrature bridge and refer a capacitor to a resitor which can easily be measured. Thats how the national labs work to define capacitance values down to the sub-ppm accuracy (besids using calculable capacitors). Building such a quadrature bridge with relatively high accuracy (10E-5 at 10nF) is not too difficult, provided one has a decavider 7 digit ac voltage divider (e.g. for AC voltage calibration purposes) and a dual channel digital 16 bit DDS in the lab anyway.

On 20 August 2015 at 02:38, Todd Micallef tmicallef@gmail.com wrote:

If this was April the first, I would be convinced that paper was an April
Fools Joke!

I can't imagine how you can make high Q (low loss) inductors, by using a
capacitor and two resistors. It just makes no sense to me, but I will read
the maths later. The mere fact there is resistance makes me think it must
be lossy, so low Q. Also the equations seems to come out with convenient
numbers - R in Ohms, C in pF and H in Henries.

But if the paper is real, it is very close to what I want, although it
still leaves the position of finding stable capacitors. I thought inductors
would be easier than capacitors, but maybe not.

Dave

Hi Dave:

I'd classify this article more like pseudoscience than April's fool.  To get an idea of the quality of the paper just
check out the publisher.

"The Chinese Publisher SCIRP (Scientific Research Publishing): A Publishing Empire Built on Junk Science" what I found
by searching for the publication name on Beall’s List.
This is a list of questionable, scholarly open-access publishers. These publishers are commonly used by "Medical"
doctors to promote the "cure" they sell by referring to a paper in a "scholarly" journal.
http://scholarlyoa.com/publishers/

I have the above link on my Faradic (Quack stuff) web page as the first entry in the Links paragraph:
http://www.prc68.com/I/Faradic.shtml#L

PS Some decades ago I spent over a year full time winding and testing coils on a Boonton 160 Q-meter.  Even at that time
the contemporary electronics college level text books did not cover inductors so bought the classics as used books (this
was prior to internet so took some doing).
http://www.prc68.com/I/Qmeters.html
Using coils built with the knowledge from this testing I home-brewed a Crystal Radio with phenomenal performance.
http://www.prc68.com/I/Xtal1800.shtml#BC1

Mail_Attachment --
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
http://www.prc68.com/I/DietNutrition.html
Dr. David Kirkby (Kirkby Microwave Ltd) wrote:

On 20 August 2015 at 19:11, Brooke Clarke brooke@pacific.net wrote:

Yes, the publisher does seem dubious. If Wikipedia is to be believed, they
were willing to publish a paper written by a random word generator. I did
not try looking very hard, but I can't see anything at indicate the authors
attributed with being at NPL are actually there.

But as Ed Breya points out, the circuit is not actually making an inductor.
Whereas with an LCR meter, or resistance meter one normally injects current
and measured the voltage at approximately the same place, using Kelvin
leads, the authors of that paper (if indeed they are the authors),
purposely inject the current at one place, and measure the voltage across a
quite different place.

I guess whether you consider it good  or pseudoscience depends on ones
perspective. It is certainly "interesting" - if it works at all. Tomorrow I
will get the soldering iron out!!

I got my 4284A LCR meter back from Keysight today, so tomorrow I will knock
one of those up and test the idea out. But I don't believe it will actually
be of any use in the calibration of an LCR meter, or even if verifying if
it drifts. Ultimately it is going to be limited by the drift in the
capacitance.

FWIW, if you did not see my earlier note, I stuck the cal certificate from
the LCR meter that came back from Keysight here.

http://www.kirkbymicrowave.co.uk/cal_certificates/Keysight-standard-calibration-with-uncertainties-for-4284A-precison-LCR-meter-18-08-2015.pdf

No inductors were used to verify the performance - only resistors and
capacitors.

There's a trick one can do with capacitors and a transformer, to "fake"
other values, for want of a better expression. But they are available from
reputable sources like IET labs.

Dave

I did a quick and dirty check, using a standard wire DUT adapter with my (not calibrated) 4275 meter and 2x 100R and 1x 1000pf. According to the guide below, this should reslt in 10uH.
Checked the parts allone with the meter before, all reasonably close to nominal.
One end of the resistors connected to the wire contacts each, other ends hooked up together and connected to the case through the 1nF capacitor.
Not precise parts, ended up at about 9.9uH at 1MHz.
So while this is by no means any quantitive analysis, it seems to work in principle and would be worth building the real ones with shielded case.
Again, calibrating these meters does not require inductors, still it is a pretty interesting approach i think.

On 20 August 2015 at 02:38, Todd Micallef tmicallef@gmail.com wrote:

OK, I tried this.

I was hoping to use a box I had with the connectors on already, but that
shorted all 4 shields, so was no use. I cobbled together 2 x 1000 Ohm
resistors, and a 100 pF cap. I had components mounted on leads, and leads
about 600 mm long, and the return point about the same length, so I could
not expect much accuracy, but I was not going to waste much time on this.

Between 1 kHz and 100 kHz, I was getting values around the 75-100 uH
inductance. I had a self-resonance at a few hundred kHz, so obviously it
went wrong there. Also below a few hundred Hz, the inductance started going
negative too. Given how poor my setup was, I am not going to give the
actual values I measured. I am inclined to believe their basic method works
(of sorts).

But having looked at it in more detail, the paper is exceeding poor, so I
am firmly in agreement with Brooke Clarke that it is pseudoscience.

The first paragraph says that inductors wound on a coil have unwanted
parasitic components, such as coil resistance and distributed capacitance.
OK, we all know that is true, but using the component values they give in
table 1, make for an inductor with an exceeding high series resistance

• 10 uH is 200 Ohms
• 100 uH is 632 Ohms
• 1000 uH is 2000 Ohms

I bought some 33 mH inductors a week or two ago and they had a series
resistance of under 200 Ohms, so I think 200 Ohms for 10 uH is rather
excessive!

I think most engineers, if presented with a component in a black box with
two terminals and asked to guess what it is, they would assume it a wire
wound resistor rather than an inductor.

Also on the first page, the authors mention it uses the 4 terminal pair
(4TP) system, and describes the advantages of the 4TP system. They
conveniently fail to mention their inductors (resistors?) rely on this.

So overall, a totally useless paper IMHO, perhaps befitting of the April
Fools joke I first thought of, but certainly not a credible scientific
paper.

Dave