Hams can not design good Q circuits. I always hear them complaining on the
air. I hear "a lousy Q, a lousy Q" followed by their call signs. :-)
Bob F. W8IL

"Reg Edwards" wrote in message
...
There is a coil.
It is 1" in diameter.
It is 2" long.
It has 20 turns.

How accurately can coil Q be determined at 30 MHz?

(1) Using an Autec antenna analyser?
(2) Using the best commercially available instrument.
----
Reg.

Bob Furtaw wrote:
Hams can not design good Q circuits. I always hear them complaining on the
air. I hear "a lousy Q, a lousy Q" followed by their call signs. :-)
Love it!!
Actually, I measure Q by the following Method (for ferrite or iron
powder cores):
First, wind about 10 turns on the toroid. Then put a capacitor across it
- preferrably a 1% mica or similar. Now, connect to a signal generator
via a non-inductive resistor, put an oscilloscope (preferrably with a
X10 probe) across the parallel tuned circuit and find the resonant
frequency Fr. Add a smaller amount of capacitance across the circuit and
re-measure the Fr. Measure the signal generator voltage and the voltage
across the tuned circuit at resonance.
Now you have all the data necessary to calculate the inductance, stray
capacitance and Q. Vary the resistor so that the voltage drop across it
at resonance is about the same as the voltage across the tuned circuit
to maximise the measuement of Q.
If the Q is lower than expected, try a range of Fr by changing the
number of turns and/or the resonating capacitance.
How accurate is this? Good enough for my purposes
73 de Alan
VK2ADB

Reg Edwards wrote:
There is a coil.
It is 1" in diameter.
It is 2" long.
It has 20 turns.

How accurately can coil Q be determined at 30 MHz?

(1) Using an Autec antenna analyser?
(2) Using the best commercially available instrument.

Those are good questions, although not particularly germane to the
current discussion. A toroid is much easier to measure than an air core
inductor with fair accuracy, since the field is largely confined. What
makes solenoidal coils relatively difficult is the problem of avoiding
coupling to the measuring device and nearby objects.

I'll mention again that ferrites are most commonly used at RF for
wideband transformers, baluns, and EMI suppression. In those
applications, Q is typically very low (1 or less) and generally immaterial.

To answer the question, though, I first note that your program predicts
a Q of about 500 for this coil, with a Z of about 960 ohms and an ESR of
about 1.4 ohms at 30 MHz. If it's correct, an antenna analyzer would be
poor choice for measuring it for several reasons -- poor accuracy at
that high an impedance, poor resolution of the ESR, and residual
resistance in the measuring device. So probably +/- 50% would be wishful
thinking. On the other hand, a good Q meter might make 20% if you could
get the coil far enough away from the fixture, and I could probably do
around 30% with my GR impedance bridge.

But what's the point you're trying to make?

Roy Lewallen, W7EL

By the way, I found another surplus toroid with somewhat better
characteristics....higher permeability and higher Q, as measured around
200 kHz. That has now been incorporated in a small switching power
supply operating at about 30 kHz. It works, though not as well as is
predicted by a SPICE model. Must be some of those vices that Reg
mentions ! My SPICE model does not take into account the variation of
the permeability/inductance with DC current, so this may be at least
part of the difference. One of these years I'll break down and get an
oscilloscope so I can figure out what non-microwave circuits are really
doing, and maybe a signal generator that works below 150 kHz.

It is adequate (barely) for what I need, however, so it has now been
incorporated as a bias supply in my 3.4 GHz transverter. I'm crossing
my fingers that it keeps working over the temperature range it will
encounter in portable operation.

The Q-measurement technique I have been using involves connecting a
signal generator through a 50 ohm attenuator (to set the output
impedance) to a 50-ohm input microwattmeter. The inductor and a
capacitor are connected as a series-tuned resonant circuit and inserted
either in series between the pad and the meter or shunted across the
meter input. The inductance is obtained by finding the resonant
frequency and working backward through the formula, given the known
capacitance. The tuned circuit is then replaced with a resistor which
is adjusted (by substitution) to give the same power on the meter as
the tuned circuit at resonance. This resistance is equal to the
equivalent series resistance of the tuned circuit, from which the Q can
be determined.

As yet the Q results I obtain with the series and shunt connections
tend to be somewhat different, so my techniques certainly have room for
improvement (there are quite obvious stray-coupling issues, even at
LF), but it gives me a rough idea, anyway.

73,
Steve VE3SMA

"Roy Lewallen" wrote
But what's the point you're trying to make?
======================================

No need to be suspicious, Roy.

Nobody can accuse you of suffering from delusions of accuracy.

Actually, I'm waiting for a reply perhaps from somebody who has a
"best commercially available instrument".
----
Reg.

A couple of people have mentioned how they do inductor Q measurement.
Here's how I do it:

I make a parallel tuned circuit with the inductor and an air variable
capacitor. I've found that even mica capacitors often have a low enough
Q to affect the measurement of reasonable Q inductors. The variable C
also lets me do the measurement at the frequency of interest. I couple
into and out of the parallel circuit with 1 pF capacitors, connecting
one to a signal generator and the other to a 50 ohm termination and a
scope. (If you calculate the parallel equivalent of the coupling cap and
terminating resistor, you'll find that you need either a low or very
high value of termination to avoid affecting the measurement.)

I've now got a signal generator with a digital frequency readout, but I
used to use an old high level generator which I tapped into in order to
hook up a frequency counter. I peak the scope signal at the frequency of
interest. Then I vary the frequency slightly and find the precise center
frequency and the -3 dB frequencies. The Q is simply the center
frequency divided by the 3 dB bandwidth. For ease in making
measurements, I built a simple 3 dB switchable attenuator and put it in
line with the signal generator, terminating the output in 50 ohms at the
Q meter so the attenuator would work properly. I measure the center
frequency with the attenuator in, then switch it out and find the -3 dB
frequencies by adjusting the frequency for the same output level as
before. If you use the attenuator, the detector doesn't have to be
linear, so you could do away with the scope and use just about any kind
of detector like a diode and DVM.

Using this method I get within about 10% of an HP Q meter at HF, at
least up to a Q of 300 or so, which is about the best I usually get with
a powdered iron toroid core.

Roy Lewallen, W7EL

Roy Lewallen wrote:
A couple of people have mentioned how they do inductor Q measurement.
Here's how I do it:

.........

Good method. I have done that before but my current RF sig gen doesn't
have very good output level regulation, nor is the output adjustable -
it's only a one-transistor sig gen after all
They wouldn't let me take my H-P with me when I retired ]
Alan

Alan Peake wrote:

Good method. I have done that before but my current RF sig gen doesn't
have very good output level regulation, nor is the output adjustable -
it's only a one-transistor sig gen after all
They wouldn't let me take my H-P with me when I retired ]

Neither of those should be a problem. The absolute level isn't
important, and the amount you have to adjust the frequency for a single
measurement is small -- for inductors of reasonable Q, anyway -- so the
level probably won't change much over that small range unless the level
stability is extraordinarily bad.

I've got a one-FET homebrew oscillator I made a long time ago with an
extra transistor or two for a crude ALC that keeps it reasonably flat
over the whole HF range. But you shouldn't need an ALC for most Q
measurements.

Roy Lewallen, W7EL