There are inumerable uses for solenoidal wound coils.

Over the years there have been been hundreds of discussions and
contributions to newsgroups about the Q of single-layer solenoid
coils. But what is ENTIRELY missing is the measured data for
particular coil dimensions and frequency.

Does nobody have a Q meter? It would appear nobody has any confidence
in Q meters in the HF range.

QUESTION:

What is the measuring accuracy of the best commercial Q meters in the
ranges of 1 to 100, 100 to 500, 500 to 2000 and above?

Or do not manufacturers state measuring accuracy? Are they themselves
uncertain of what its all about?

A subsidiary question is what use is made of Q values after a
measurement has been made? Does an inacurate Q value matter very much
anyway?

Please give numbers in your reply. For once I confess to a minor
troll. But I hope I get a few sensible answers to sensible questions.

Nil answers will be considered to be of equal consequence to the
others.
----
Reg, G4FGQ

Reg Edwards wrote:
There are inumerable uses for solenoidal wound coils.

Over the years there have been been hundreds of discussions and
contributions to newsgroups about the Q of single-layer solenoid
coils. But what is ENTIRELY missing is the measured data for
particular coil dimensions and frequency.

Does nobody have a Q meter? It would appear nobody has any confidence
in Q meters in the HF range.

I just purchased an old Boonton 160A Q meter that seems to still be
working, correctly. It supplys test waves from 50kHz to 75 MHz.

It was made in 1949 (the same year I was), is a 3 tube design,
including the supply rectifier, and probably last calibrated when I
was wearing diapers, but everything on it still seems okay. I have
been using it to compare different coil and ferrite rod designs to
improve my ability to build optimal rod antennas.

QUESTION:

What is the measuring accuracy of the best commercial Q meters in the
ranges of 1 to 100, 100 to 500, 500 to 2000 and above?

Absolute accuracy is hard to quantify, because the connection losses,
radiation losses and near field losses vary a lot, depending of the
setup. The 160A has made a valiant attempt to have a low loss tuning
capacitor available to resonate the coil at various frequencies, but
if I pad this with an external capacitor, I never achieve as high a Q
reading as I do with just the internal cap. The best I can do is
compare variations with a single setup. In other words, if i can make
two tests with a single setup, I can clearly tell which variation has
the higher Q. But if I sit closer or further away from a large coil,
it changes both results. At high Q, very little things count a lot.
Measuring a large coil that is 6 inchs from any metal surface, and
then putting it in a metal box, an inch from the side, kills the Q,
anyway.

The scale on the 160A ranges from 20 to 200, but you can double that
with another setting that cuts the excitation. However, at Qs above
100, the tuning is so touchy that I often can't find the exact peak.

Or do not manufacturers state measuring accuracy? Are they themselves
uncertain of what its all about?

A subsidiary question is what use is made of Q values after a
measurement has been made? Does an inacurate Q value matter very much
anyway?

For tuning and filter purposes, it predicts the bandwidth. For rod
antenna purposes, Q gets into the region of space the antenna couples
to. The higher the Q, the larger volume of space the antenna pulls
energy from.

I have also had some success at low frequency (below 1 MHz) measuring
tank Q by driving the tank through a 10X scope probe, and measuring
the resonant voltage with an AC volt meter through a second 10X scope
probe. I find the frequency that produces the peak voltage, then tune
up and down to find the two frequencies that produce .707 of that
peak. The Q is the square root of the product of those two
frequencies divided by their difference. I don't know if it is the
loading effect of the probes, or lack of calibration for the Boonton,
but I consistently get lower Q by this method than the Boonton shows.
The Boonton may be 1.2 to 1.5 times higher. But, at least both
methods allow comparison of variations, so I can have a "getting
warmer" indication of which variations are better.

John, thanks for describing your detailed and recent interesting
experience and your thoughts on the subject.

However, once again the question is raised - how does one calibrate a
Q meter? Rhetorically, is calibration traceable to National or
International Standards?

So much depends on the Q quality of the meter itself. Meter
manufacturers are unable to state degrees of accuracy at various
frequencies and actual values of Q. Nobody knows what the actual
value actually is! Least of all the user!

Fortunately, the exact value of Q of a coil is never required. It is
used only to provide coarse estimates of other quantities. And there
are usually other means of finding the other quantities. They can be
estimated by calculating from values which CAN be measured or
estimated.

So Q meters provide support and back up for experimenters who have
other means of finding the answers they are looking for. By itself a
measured value of Q is inaccurate and of no use. What matters is what
can be derived or guessed from it.

It is merely an intermediate variable in a chain of deductions or
calculations. Above Q equal to a few hundreds it is anybody's guess.

In some ways it is similar to an SWR measurement on a line which isn't
there.
----
Reg.

Reg Edwards wrote:
John, thanks for describing your detailed and recent interesting
experience and your thoughts on the subject.

However, once again the question is raised - how does one calibrate a
Q meter? Rhetorically, is calibration traceable to National or
International Standards?

Q is the ratio of two values (and several different pairs of values
can be used to arrive at the same end result). For a coil, one ratio
that produces a value of Q is the peak energy stored during an AC
cycle to the total energy dissipated during a cycle. If you can
measure those two quantities, separately, you can calculate Q. But it
is often much easier to work with a resonant tank and use a capacitor
that is known to have a much higher Q than the coil being tested, so
that you can assume that all the losses are in the coil. Then the
tank circuit Q is the coil Q. This is the bases of the Boonton 160A
and also the signal generator, voltmeter method I have been using.

So much depends on the Q quality of the meter itself.

The meter quality has to be higher than the Q of the device being
measured, of a compensation has to be made for the meter losses. For
instance, with the signal generator voltmeter method, I have
calculated the losses in the two 10X probes, to prove to myself that
the errors they cause are not significant up to the highest Q values I
measured, this way. I would have had to measure a Q near 1000 before
they would have altered a significant digit of the measurement. But I
did convince myself that energy absorption outside the coil in
surrounding objects is significant, since shifting my position in my
chair did change the measurement.

Meter
manufacturers are unable to state degrees of accuracy at various
frequencies and actual values of Q. Nobody knows what the actual
value actually is! Least of all the user!

Boonton originally sold Q standard coils (inductors with known Q) to
be used to check the accuracy of the Q meters. I don't know how those
coils calibration got back to basic measurements traced to the Bureau
of standards.


Fortunately, the exact value of Q of a coil is never required. It is
used only to provide coarse estimates of other quantities. And there
are usually other means of finding the other quantities. They can be
estimated by calculating from values which CAN be measured or
estimated.

I agree. Usually, proving that a given device has at least a certain
Q is enough, or tests made on various devices by the same method can
show which ones have higher Q than others. This is what I am doing
with the measurements.

So Q meters provide support and back up for experimenters who have
other means of finding the answers they are looking for. By itself a
measured value of Q is inaccurate and of no use. What matters is what
can be derived or guessed from it.

Q is a way to measure losses. If losses are important to the
application, Q is one way to get information that is useful. An
infrared thermal imager may be another.

It is merely an intermediate variable in a chain of deductions or
calculations. Above Q equal to a few hundreds it is anybody's guess.

In some ways it is similar to an SWR measurement on a line which isn't
there.
----
Reg.

On Sun, 19 Feb 2006 10:49:48 +0000 (UTC), "Reg Edwards"
wrote:

So much depends on the Q quality of the meter itself.

Hi Reggie,

The "Q of the meter?" What a hoot. Would that be the impedance of
the handles over the resistance of the cover?

Let's also observe the Madison Avenue flair: "Q quality...." Or are
we to believe you are a proponent of measuring the Q of quality? Think
Lord Kelvinator would want a number put to it?

Nobody knows what the actual value actually is! Least of all the user!

Let's see, if the user picks up an Ohmmeter to measure a resistance,
he doesn't initially know the resistance, he doesn't know the
accuracy, hence the meter is invalidated for existing, the user
suddenly lacks a metaphysical basis for being and all disappear in a
cloud of doubt.

Fortunately, the exact value of Q of a coil is never required.

Now that meters no longer exist, users have evaporated, exact Q is
never required, the coil unwinds itself in existential abnegation.

It is used only to provide coarse estimates of other quantities.

Ah! But if "It" is unknowable, "It" offers nothing - coarse or
vulgar.

And there are usually other means of finding the other quantities.

Which then loops this logic back to the impossibilty of knowability
and these quantities suddenly dematerialize from the cosmos.

They can be estimated by calculating from values which CAN be measured or
estimated.

Estimates can be made of estimates - um yas, indeed! Now there's a
authentic statement of clarity and rational self-determination. Would
it be inappropriate to appreciate the irony of your attack on accuracy
where your argument is so conclusively lacking - accuracy? ;-)

By itself a
measured value of Q is inaccurate and of no use. What matters is what
can be derived or guessed from it.

An excellent summary. If it is inaccurate and of no use, we can use
it to derive or guess something from it.

Thanx for the opportunity,
Richard Clark, KB7QHC

On Sun, 19 Feb 2006 12:52:26 -0500, John Popelish
wrote:

Boonton originally sold Q standard coils (inductors with known Q) to
be used to check the accuracy of the Q meters. I don't know how those
coils calibration got back to basic measurements traced to the Bureau
of standards.


Hi John,

And for the sake of trolling, Reggie claims he doesn't either.

Gads this is so simple as to defy the angst that surrounds this. Any
good Impedance Bridge which reports R and X separately will give you
the means to measure the Q (or D) of these standards. All this
folderol of the "Q quality of the meter" is so much hokum distracting
from a simple determination.

Balance the bridge and you will have the resistance that so impacts
the Q. Balance the bridge and you will have the reactance that
establishes the Q in relation to the resistive loss. And what does
the meter have to do with Q? The bridge is adjusted for a zero
reading! What accuracy statement can be said about reading zero when
you return the needle to the position it was in when the unit was
stone cold? THIS is how you qualify the standards Boonton offers. You
then qualify your Bridge against separable quantities of X and R.
For sure, this may relegate us to a tedious cascade of "how do you
know what value those are really?" It is this kind of whining that
leads to warning statements being forced into curriculums by those who
want to teach Untelligent Design.

Reggie has managed to turn the discussion of Q into a mystical,
unknown quantity impossible to determine by his simply ignoring first
principles. You measure the Q of the unknown two ways and compare. By
the Bridge and by the Boonton. I dare say no more than 20%
accumulated error will occur with NONE of it attributed to the "Q
quality of the meter" - whatever that is.

So, let's compare. You can have a determination within 20% of actual,
and continue to design with confidence. OR You can mumble about the
abstract impossibility of ever getting it abso-*&!#ing-lutely right
and find yourself in analysis paralysis.

I have, of course, steeply discounted the accuracy of the Boonton to
include all RSS accumulation of errors in the instrument's
calibration. The manufacturer warrants the device to 5%.

Now, if you strip away all the numbers, you can re-achieve the
distinction of the Qualitative statement that got us here. Lord
Kelvinator would point out that that and $5 will buy you an insolated
cup of Laté.

73's
Richard Clark, KB7QHC

amdx wrote:
The meter quality has to be higher than the Q of the device being
measured, of a compensation has to be made for the meter losses. For
instance, with the signal generator voltmeter method, I have
calculated the losses in the two 10X probes, to prove to myself that
the errors they cause are not significant up to the highest Q values I
measured, this way. I would have had to measure a Q near 1000 before
they would have altered a significant digit of the measurement. But I
did convince myself that energy absorption outside the coil in
surrounding objects is significant, since shifting my position in my
chair did change the measurement.


It has been a while since I measured the 3db points of a inductor, but I
think the probe did load the coil. I recall putting a 1meg resistor in
series
with the probe to help isolate the probe capacitance from the inductor.

Any Q measurement of a tuned circuit must pull less energy from the
tank than the tank consumes each cycle, if the meter's effect can be
neglected. The Boonton 160A injects a voltage into the tank with a
fraction of an ohm source impedance. And a vacuum tube volt meter
measures the resultant voltage across the tank. Both these paths
represent a loss, and limit the highest possible Q reading to about
400. But some tank impedance combinations (very high impedance
affected by volt meter input impedance, or very low impedance affected
by voltage source impedance) may face more severe limits.

I bridge type measurement can be zeroed for its internal losses before
the measurement, so is more self compensating. But any equipment I
have seen will have trouble accurately measuring extremely high Q
tanks, inductors or capacitors.