Q of ferrite inductor versus air-core inductor

[Roy Lewallen]

As Rocci said, it depends on the equipment you have available. I have
a good method using a signal generator, frequency counter, and scope
if you have them.

Roy Lewallen, W7EL

Charlie T. wrote:
>
> I'm trying to measure the Q of a ferrite core inductor
> and an equivalent value air-core inductor. I know the
> Q of the ferrite may be less but so is the size!
>
> Any suggestions on Q measurement techniques
> appreciated.
>
> -Thanks,
> W5CDT
> --
> >>>Remove the X's from my email address before using<<<
>
> Xc...@onr.com

[Charlie T.]

[Rocci]

There are numerous techniques that can be used, but they depend on the
instrumentation available to you and the frequency range you're working
in. Do you have a signal generator? How about a spectrum analyzer or a
calibrated signal level meter?

There

[Jim Bixby]

"Charlie T." <xc...@onr.com> wrote:

>I'm trying to measure the Q of a ferrite core inductor
>and an equivalent value air-core inductor. I know the
>Q of the ferrite may be less but so is the size!
>

I'm assuming you are talking about an inductor wound on a slug, not a
pot core or toroid. If so, you can think of the inductor as a
loopstick antenna.

A reasonably quick way requires, as noted in the other replies, a
signal generator and a scope. The trick is to not load down the coil
with either the signal generator or the scope. What I do is take the
signal generator, and drive it into a loop of a couple of turns of
hookup wire, maybe 2 " in diameter, to make a 'transmitter', I put
that a distance away (inches to yards, depending on the geometry and
inductance -- it doesn't matter very much) from the ferrite inductor,
with the axes arranged parallel, so that the loop will transmit to the
inductor. With an oscilloscope and a high-z 10:1 probe, add
capacitance across the inductor until the inductor-capacitor resonate
at the desired frequency. Then, just vary the signal generator
frequency to either side of resonance to find the points that the
output signal drops 3db from the peak. The center frequency divided
by the difference in the -3db points is the Q.

After doing that, you really need to check that the scope probe did
not load down the coil. To do that, you should have an idea of the
inductance. If you don't know it already, you can find it
(approximately) by adding another capacitor across the resonating
capacitor and then finding the new resonant frequency. Armed with the
values of the resonating capacitor and the additional capacitor, and
the two frequencies, you should be able to solve for both the
inductor's inductance and its distributed capacitance.

From the inductance, calculate the reactance at the resonant
frequency, and then multiply by the Q you measured. That result
should be much much less than the scope probe impedance (like a factor
or five-ten or more). If it is not, then you will need to go to the
trouble of making a little jfet or mosfet preamp to use in place of
the scope probe.

Depending on the ferrite and frequency, don't count on the ferrite Q
being lower. Q's of more than 200 for low frequency ferrite loop
antennas are not uncommon, and these are, after all, just ferrite slug
inductors.

If you are in fact using a magnetically closed structure, like a pot
core or a toroid, you can do the above, but you will need to couple
the signal generator to the inductor in a way which does not load the
inductor. For a toroid, its pretty easy -- put one turn through the
toroid and then use a very high value resistor in series, to the
signal generator. This won't work for the pot core, so you have to
connect directly to the coil winding, with a resistor much larger than
the reactance times the Q you are looking for.
----------------------------------------------------------------------------
Jim Bixby
b...@san.rr.com
http://home.san.rr.com/bix

[Reg Edwards]

Yes Jim, it is a very simple and accurate method.

If you havn't a digital signal generator then use your
transmitter on low power connected to its antenna. The
coil under test with its tuning capacitor on the bench will
pick up sufficient signal. If the coil core is a closed
ferrite ring it may need a foot or so of wire to act as an
antenna.

To make sure the scope or other RF voltmeter does not load
the coil, use a 1 or 2-turn link coupling to the scope.

When varying transmitter frequency looking for the +/-
70.7 percent voltage points, don't move anything except the
Tx tuning knob and your eyeballs.
--
******************************
Reg, G4FGQ For free software go to:-
http://www.btinternet.com/~g4fgq.regp

[Rocci]

Another way that doesn't require magnetic coupling is to make your signal generator have a 'stiff' known output impedance by, say, putting a 10dB 50 ohm attenuator at the output. Then select a good quality capacitor that resonates the inductor at the frequency of interest. Put the scope or other 'non-loading' measurement device at the output of the attenuator, and hook up the LC circuit in a series configuration across the output.
Tune the generator until a resonant null is observed. The equivelent resistance (R) of the inductor can be calculated from the depth of the null. Q of the inductor then equals inductive reactance divided by equivelent resistance.
This only works if the capacitor Q is high compared to the inductor Q (which is usually true), and if the scope is usable at the frequency of interest.
At higher frequencies, a logarithmic signal level meter, a spectrum analyzer, or possibly a calibrated 'S' meter can be used.

[Charlie T.]

Yep...I have an HP8660C...will that work? It's a synthesized signal
generator w/digital frequency display. The "Q" I'm talking about refers
to the losses in the ferrite core. Not the "bandwith" of the resonant
circuit. In a series LC circuit there should be a Q rise in current
at resonance where the Q is the ratio of the reactance to the real
resistance (i.e. generator impedance + skin effect losses + wire resistance
+ ferrite
core loss). Can this be equivocated to the Q definition: 3dB
bandwidth/frequency?

thanks in advance,

-Charlie W5CDT

Roy Lewallen <w7...@teleport.com> wrote in article
<36F31C5A...@teleport.com>...

[Roy Lewallen]

Here's how I do it. I make a parallel resonant circuit with the
inductor and an air variable capacitor. (Even some mica capacitors
have low enough Q to skew results for good inductors.) I couple in and
out with 1 pF capacitors, to the signal generator and scope. I measure
the resonant frequency (peak amplitude) and -3 dB points. The Q (see
below) is the ratio of center frequency to bandwidth.

A reasonable model of the inductor at one frequency is a parallel RLC
circuit, where the C represents the inter-turn capacitance. The Q
measured by this method is the ratio of the R to XL in this model. If
you define a single equivalent L as a combination of the L and C in
this model, and the Q as the ratio of R to its reactance, the Q
measured by this method will show a bit higher. If you need to find
the shunt C, you can derive it from measurements of the apparent
inductance of the coil at a couple of different frequencies. The Q I
measure with the parallel-circuit method is fine for comparing core
losses, since C should be about the same for the same winding on
various cores.

An advantage of this method over some of the bridge and null methods
is its relative insensitivity to impedances, the test fixture, etc.
The only caution is to keep the inductor reasonably well away from
conductors and lossy dielectrics.

I whipped up a switchable 3 dB pad to put in line for convenience. I
switch it in when measuring the maximum, then swich it out and find
the 3 dB points by adjusting the frequency for the same amplitude as
before. The input and output of the pad have to see the right
impedance, of course.

Roy Lewallen, W7EL

[Jim Bixby]

"Charlie T." <xc...@onr.com> wrote:

>The "Q" I'm talking about refers
>to the losses in the ferrite core. Not the "bandwith" of the resonant
>circuit.

The are the same thing. The coil will have losses from core losses,
dc resistance in the winding, skin effect in the winding, dielectric
losses, eddy current losses for currents induced in the ferrite
(negligible) and currents induced in surrounding metal (usually not
negligible). All of these losses combine into the real part of the
impedance of the inductor. At resonance, the shunt capacitance from
the winding and any external shunt capacitance cancel the inductive
reactance, and you are left looking at a pure resistance which is
solely from the losses. The -3db points are where the magnitude of
the reactive impedance of the (ideal) LC tank equals the resistance.
So when you measure the 3 db bandwidth to get the Q, you are measuring
all the losses. You can measure the DC resistance, and you can
calculate the skin effect loss for the winding, but it is very hard to
separate core losses from eddy current losses. I've seen one paper
which reports that the typical Q degradation from moving a
magnetically "open" inductor from free space to the vicinity of other
electronics and a chassis is typically on the order of 50%.