From: Paul
Date: Wed, 20 Dec 2000 21:30:50 +0000
Subject: [TSSP] Measurement Techniques
The other day Marc Metlicka wrote to remind me that I promised to put
up a web page describing how to go about making the kind of precision
measurements needed by this project. Well I began ages ago but ground
to a halt when I needed to check out a few things but haven't been
able to find the opportunity.
Marc, sorry to keep you in suspense, and I hope you don't mind me
replying to you on-list, but it will allow the others to chip in
with info and suggestions and to correct my mistakes, as I'm probably
not the best person to consult on measurement techniques.
Marc wrote:
> did you ever get those pages on making measurements for your inputs
> finished? with all this talk of base input impedance and lcr meters
> and good equipment, i'm stupefied.
> i have an excellent lcr bridge, a set of fluke true rms. 83 and 87
> III certified meters, fluke counter/tracer, tek 466 storage scope
> and actually another nice lcr meter, nice signal gen. and two other
> scopes to use
> i just don't know how to do the measurements, it is very
> frustrating.
> i have a 10-1 secondary that performed very well, 21g wire on a
> 4" by 42" wound form with both a flat and an inverse conical pri.
> and a series of other coils including the 3k coil system, 3000
> turns of 28g on an 8.75" by 40" wound form.
Sounds like Marc has all the right kit and I'm sure that with a little
guidance from this group he'll be able to put it all to good use.
I'm particularly interested in the h/d=10 secondary - we've not had
any results for such a large h/d before, so this is an opportunity to
stretch the envelope over which our tssp secondary model is validated.
We also have not had results for a 3k turn coil before either, so
both of Marc's coils promise to be interesting and reliable results
would be most valuable.
Perhaps if between us on this list we can steer Marc towards carrying
out some sound measurement procedures, we can collate the postings as
a basis for a how-to web page on measurements.
I'll start with my recommendations!
The most important thing from my point of view is to obtain a
repeatable coil setup. This means finding a suitable place - a test
environment - in which to set up the test coil in such a way that it
can be either left set up permanently, or can be replaced exactly the
same each time it is used.
It's quite difficult for me if someone sends measurements in, and then
a fortnight later supplies some more, but from a slightly different
test setup. It makes for lots of extra work to 'align' the data for
comparison purposes, and sometimes unequivocal comparisons become
impossible because of the 'fiddles' necessary.
So try to find a place where the coil can be repeatably set up for
test. The location should be as far as possible from anything else
other than the floor or table on which the coil sits. You must be able
to operate the instruments without having to enter the field of the
coil, so you may need some long-ish test leads and things.
It helps considerably if the coil has a well defined ground plane
beneath it. This is so that the coil's capacitance can be accurately
calculated and the ground plane will help to localise the E-field so
that it is less disturbed by other objects further away in the room.
Most of our tests require that the coil be driven from a signal
generator by applying the signal directly to the base of the coil,
rather than through a primary winding. Also it is often important that
the impedance of the driver (as seen by the coil) is very low. The
following connection to the base is suggested:
| |
| |
| |
| | coil
| |
| | ____________
| |----------------[50 ohm]---O____________O-----*
base | | coax | sig gen
terminal [1 ohm] | -----*
| |
------------------------------------
ground plane
My ascii art attempts to show the signal generator connected
through a long coax leading to the coil base. The center of the
coax connects to the coil base via a 50 ohm resistor (assuming
your sig gen requires a 50 ohm load) and there is also a 1 ohm
resistor connected between the coil base and the ground plane.
The sig gen should be operated near its maximum output voltage and
the 50 ohm resistor will need to dissipate most of the power,
so if your sig gen delivers 20v RMS into a 50 ohm load, thats
20*20/50 = 8 watts to dissipate.
Neither the 50 nor the 1 ohm values are at all critical.
OK, that takes care of putting a signal into the coil. Now we need
some means of detecting the resonance of the coil. The idea is
to sample either the E or the B field with some kind of probe in
such a way that you can watch what the coil's doing on the scope,
but without disturbing the E or B field that you are coupling to.
There are lots of ways to do this. I've used the following method and
I'm sure the others will suggest alternatives.
Make up a small flat search coil, a few turns, 5 or so, of
hookup wire, say an inch or so in diameter. The search coil is
placed on the floor or table, above the ground plane, and near to
the coil base. Connect the search coil to another long coax leading
back to your scope.
Run the generator into the coil, tuning it slowly across the resonant
frequency, and you should see some RF pickup from the search coil
appearing on the scope. Use max gain on the scope. If the scope signal
is too large, take advantage of that to move the search coil away from
the secondary.
You should now be able to tune the generator across its frequency
range and locate not only the quarter wave resonance, but several
other higher frequency resonances too.
So the first step is to organise a test setup and get the hang of
locating the resonant frequencies. You'll then need to send me details
of the coil dimensions, height above groundplane, and a few other
things about the test setup, so that we can then calculate the
resonant frequencies and compare them with your measured values.
You can also measure the low frequency or 'DC' inductance of the coil,
and the 'DC' capacitance between the coil and the ground plane.
Once we've got to the stage where the Ldc and Cdc agree with our
calculations, and also the lowest three resonances agree with the model
to around 1% accuracy, we can move on to the more exciting and delicate
stuff.
Marc, I hope these notes help you to get started, and let me say that
the enthusiasm which comes through with your postings is most welcome.
Regards,
--
Paul Nicholson,
Manchester, UK.
--
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.