Hello Richard,

I have seen this method one day in an QST and, as you, found the idea
very fine. Then keep the small diagram in my favorite notebook. ;-)
It was published in QST 1982/05.

73,
Francis
F6AWN

Hello Richard,

I have seen this method one day in an QST and, as you, found the idea
very fine. Then keep the small diagram in my favorite notebook. ;-)
It was published in QST 1982/05.

73,
Francis
F6AWN

I saw the subject line, and was ready to warn against trusting a
return loss bridge at high reflection coefficients (assuming you'd
just hooked the L directly across the RL bridge unknown terminals),
but of course the method of putting an R and C in series gets around
that problem. If you have a swept system, you don't have to tweak the
cap, just look at the freq where the return loss is minimum...or where
the phase changes the fastest, or is zero. Then perhaps you don't
even need the series R (see next paragraph...).

In the "Measurements and Analysis" chapter of Reference Data for
Engineers, there are several bridge circuits suitable for measuring
inductance. One is a "resonance bridge" which really is doing just
what Richard wrote about. One arm of the bridge is a series RLC. But
in that case, the R is the equivalent series resistance of the LC, and
is balanced by the other bridge arms at resonance, just as a simple
Wheatstone bridge. Note that if your return loss bridge has been
calibrated to be accurate at small resistances, you can use this
without the series R in your LC to get a very quick estimate of Q, but
you can also do that by sweeping the frequency. -- See also circuits
for HP and Boonton (and GR?) Q meters.

Cheers,
Tom

"Richard Hosking" wrote in message .au...
My Friend Rod VK6KRG uses a snazzy method of calculating inductance using a
return loss bridge
I have tried it to measure a 500nH inductance and the result agrees closely
with the calculated result for an inductor of that physical size and number
of turns.

Basically he uses the unknown inductor, in series with a known capacitor, in
series with a 50 ohm load across the "unknown" arm of the bridge. When
XC=XL, then the two reactances cancel (assuming they approximate to an ideal
reactance) and load presented to the bridge by the RLC circuit = 50 ohms.
Thus you tune the system to best return loss which is a null and calculate
XL at this frequency You can even do this at the end of a section of 50 ohm
coax, assuming the loss isnt too great

No doubt all you RF experts out there have known about this for a long
time - has it been published as an idea before?

Richard

I saw the subject line, and was ready to warn against trusting a
return loss bridge at high reflection coefficients (assuming you'd
just hooked the L directly across the RL bridge unknown terminals),
but of course the method of putting an R and C in series gets around
that problem. If you have a swept system, you don't have to tweak the
cap, just look at the freq where the return loss is minimum...or where
the phase changes the fastest, or is zero. Then perhaps you don't
even need the series R (see next paragraph...).

In the "Measurements and Analysis" chapter of Reference Data for
Engineers, there are several bridge circuits suitable for measuring
inductance. One is a "resonance bridge" which really is doing just
what Richard wrote about. One arm of the bridge is a series RLC. But
in that case, the R is the equivalent series resistance of the LC, and
is balanced by the other bridge arms at resonance, just as a simple
Wheatstone bridge. Note that if your return loss bridge has been
calibrated to be accurate at small resistances, you can use this
without the series R in your LC to get a very quick estimate of Q, but
you can also do that by sweeping the frequency. -- See also circuits
for HP and Boonton (and GR?) Q meters.

Cheers,
Tom

"Richard Hosking" wrote in message .au...
My Friend Rod VK6KRG uses a snazzy method of calculating inductance using a
return loss bridge
I have tried it to measure a 500nH inductance and the result agrees closely
with the calculated result for an inductor of that physical size and number
of turns.

Basically he uses the unknown inductor, in series with a known capacitor, in
series with a 50 ohm load across the "unknown" arm of the bridge. When
XC=XL, then the two reactances cancel (assuming they approximate to an ideal
reactance) and load presented to the bridge by the RLC circuit = 50 ohms.
Thus you tune the system to best return loss which is a null and calculate
XL at this frequency You can even do this at the end of a section of 50 ohm
coax, assuming the loss isnt too great

No doubt all you RF experts out there have known about this for a long
time - has it been published as an idea before?

Richard

You don't even need a bridge. Just put a BNC "tee" in line between
any signal generator and any detector. Now put your LC series
resonant tank on the third port of the tee. At resonance, the LC
will suck out the signal and you will see a dip in the response.
This method is more accurate than the bridge method you describe.

Regarding small inductances: the capacitor has to have negligible
inductance compared to the inductor you want to measure.

Rick Karlquist N6RK


"Richard Hosking" wrote in message
. au...
My Friend Rod VK6KRG uses a snazzy method of calculating inductance using
a
return loss bridge
I have tried it to measure a 500nH inductance and the result agrees
closely
with the calculated result for an inductor of that physical size and
number
of turns.

Basically he uses the unknown inductor, in series with a known capacitor,
in
series with a 50 ohm load across the "unknown" arm of the bridge. When
XC=XL, then the two reactances cancel (assuming they approximate to an
ideal
reactance) and load presented to the bridge by the RLC circuit = 50 ohms.
Thus you tune the system to best return loss which is a null and calculate
XL at this frequency You can even do this at the end of a section of 50
ohm
coax, assuming the loss isnt too great

No doubt all you RF experts out there have known about this for a long
time - has it been published as an idea before?

Richard

You don't even need a bridge. Just put a BNC "tee" in line between
any signal generator and any detector. Now put your LC series
resonant tank on the third port of the tee. At resonance, the LC
will suck out the signal and you will see a dip in the response.
This method is more accurate than the bridge method you describe.

Regarding small inductances: the capacitor has to have negligible
inductance compared to the inductor you want to measure.

Rick Karlquist N6RK


"Richard Hosking" wrote in message
. au...
My Friend Rod VK6KRG uses a snazzy method of calculating inductance using
a
return loss bridge
I have tried it to measure a 500nH inductance and the result agrees
closely
with the calculated result for an inductor of that physical size and
number
of turns.

Basically he uses the unknown inductor, in series with a known capacitor,
in
series with a 50 ohm load across the "unknown" arm of the bridge. When
XC=XL, then the two reactances cancel (assuming they approximate to an
ideal
reactance) and load presented to the bridge by the RLC circuit = 50 ohms.
Thus you tune the system to best return loss which is a null and calculate
XL at this frequency You can even do this at the end of a section of 50
ohm
coax, assuming the loss isnt too great

No doubt all you RF experts out there have known about this for a long
time - has it been published as an idea before?

Richard

Can this be used to check antenna resonance?

MikeN ZL1BNB


On Wed, 10 Sep 2003 03:38:12 GMT, "Rick Karlquist N6RK"
wrote:

You don't even need a bridge. Just put a BNC "tee" in line between
any signal generator and any detector. Now put your LC series
resonant tank on the third port of the tee. At resonance, the LC
will suck out the signal and you will see a dip in the response.
This method is more accurate than the bridge method you describe.

Regarding small inductances: the capacitor has to have negligible
inductance compared to the inductor you want to measure.

Rick Karlquist N6RK


"Richard Hosking" wrote in message
.au...
My Friend Rod VK6KRG uses a snazzy method of calculating inductance using
a
return loss bridge
I have tried it to measure a 500nH inductance and the result agrees
closely
with the calculated result for an inductor of that physical size and
number
of turns.

Basically he uses the unknown inductor, in series with a known capacitor,
in
series with a 50 ohm load across the "unknown" arm of the bridge. When
XC=XL, then the two reactances cancel (assuming they approximate to an
ideal
reactance) and load presented to the bridge by the RLC circuit = 50 ohms.
Thus you tune the system to best return loss which is a null and calculate
XL at this frequency You can even do this at the end of a section of 50
ohm
coax, assuming the loss isnt too great

No doubt all you RF experts out there have known about this for a long
time - has it been published as an idea before?

Richard

Can this be used to check antenna resonance?

MikeN ZL1BNB


On Wed, 10 Sep 2003 03:38:12 GMT, "Rick Karlquist N6RK"
wrote:

You don't even need a bridge. Just put a BNC "tee" in line between
any signal generator and any detector. Now put your LC series
resonant tank on the third port of the tee. At resonance, the LC
will suck out the signal and you will see a dip in the response.
This method is more accurate than the bridge method you describe.

Regarding small inductances: the capacitor has to have negligible
inductance compared to the inductor you want to measure.

Rick Karlquist N6RK


"Richard Hosking" wrote in message
.au...
My Friend Rod VK6KRG uses a snazzy method of calculating inductance using
a
return loss bridge
I have tried it to measure a 500nH inductance and the result agrees
closely
with the calculated result for an inductor of that physical size and
number
of turns.

Basically he uses the unknown inductor, in series with a known capacitor,
in
series with a 50 ohm load across the "unknown" arm of the bridge. When
XC=XL, then the two reactances cancel (assuming they approximate to an
ideal
reactance) and load presented to the bridge by the RLC circuit = 50 ohms.
Thus you tune the system to best return loss which is a null and calculate
XL at this frequency You can even do this at the end of a section of 50
ohm
coax, assuming the loss isnt too great

No doubt all you RF experts out there have known about this for a long
time - has it been published as an idea before?

Richard

"Rick Karlquist N6RK" wrote in message news:E0x7b.407103$Ho3.62141@sccrnsc03...
You don't even need a bridge. Just put a BNC "tee" in line between
any signal generator and any detector. Now put your LC series
resonant tank on the third port of the tee. At resonance, the LC
will suck out the signal and you will see a dip in the response.
This method is more accurate than the bridge method you describe.

Regarding small inductances: the capacitor has to have negligible
inductance compared to the inductor you want to measure.

Another caveat: keep the leads between your LC to be measured and the
coax tap point very small compared with the wavelength. For example,
if you have a quarter wave of line between the "T" and the LC, at LC
resonance it reflects back an open instead of a short. And if it's
even just 0.05 waves, for example, with 50.66nH and 50pF on the end,
the suck-out frequency drops from 100MHz (the LC resonance) to about
83MHz, and you're going to make a big error in your estimate of the
inductance. That 0.05 waves of line is only about 4 inches at 100MHz.
In a typical BNC "T", you probably have 0.01 waves at 100MHz, and
even that shifts the apparent resonance by over 4MHz in the example
given, resulting in 9% or so error in the estimate of L. Looking at
it another way, not only does the capacitance have to have negligible
inductance compared with the inductor, but the connection to the line
must also. -- So though the return-loss-bridge technique where a Zo
resistor is added may not be as accurate as a resonance measurement,
it does have the advantage that it can be done at the end of a line
without regard to the line length, and sometimes that can be very
handy.

Another poster asked about using these techniques to find antenna
resonance. The problem for the typical ham's instrumentation is that
the antenna is generally a low Q and presents a fairly high resistance
(in the neighborhood of 50 ohms, presumably, but not necessarily very
close to that) at resonance. It's easy enough to measure the system
resonance (lack of reactance) at the shack end of the line feeding the
antenna, but AFAIK, most hams don't have an analyzer they can
calibrate at the end of a piece of transmission line like you can with
modern lab network analyzers. So you'd have to measure the impdeance
(R+jX, complex value) you see at the shack end of the line and put
that into a calc that backs the line out so you can know the impedance
at the antenna feedpoint. -- But why bother? It doesn't generally
matter whether an antenna is _resonant_ or not, only that it radiates
efficiently with the desired pattern, and that you have a way to get
power to it efficiently. So at least at a given frequency, you can
trim the antenna and/or put a small matching network at the antenna to
get the line SWR low enough to suit you. If you can stand the
resulting SWR over the range of freqs you want to use, then any
additional matching can be done at the shack end of the line.

Yet another note. ;-) When the measurement is done in a system with
a 50-ohm impedance, and the components (L and C) in the resonant tank
have fairly low reactance, you won't get a very sharp null with the
return-loss bridge and additional series R as the OP suggested. The
additional R makes the tank Q very low. With Rick's suck-out circuit
suggestion, you'll need a detector with pretty wide dynamic range
(like a receiver perhaps) to see where the deepest null occurs,
because even fairly far off resonance, there will be significant
attenuation.

Cheers,
Tom

"Rick Karlquist N6RK" wrote in message news:E0x7b.407103$Ho3.62141@sccrnsc03...
You don't even need a bridge. Just put a BNC "tee" in line between
any signal generator and any detector. Now put your LC series
resonant tank on the third port of the tee. At resonance, the LC
will suck out the signal and you will see a dip in the response.
This method is more accurate than the bridge method you describe.

Regarding small inductances: the capacitor has to have negligible
inductance compared to the inductor you want to measure.

Another caveat: keep the leads between your LC to be measured and the
coax tap point very small compared with the wavelength. For example,
if you have a quarter wave of line between the "T" and the LC, at LC
resonance it reflects back an open instead of a short. And if it's
even just 0.05 waves, for example, with 50.66nH and 50pF on the end,
the suck-out frequency drops from 100MHz (the LC resonance) to about
83MHz, and you're going to make a big error in your estimate of the
inductance. That 0.05 waves of line is only about 4 inches at 100MHz.
In a typical BNC "T", you probably have 0.01 waves at 100MHz, and
even that shifts the apparent resonance by over 4MHz in the example
given, resulting in 9% or so error in the estimate of L. Looking at
it another way, not only does the capacitance have to have negligible
inductance compared with the inductor, but the connection to the line
must also. -- So though the return-loss-bridge technique where a Zo
resistor is added may not be as accurate as a resonance measurement,
it does have the advantage that it can be done at the end of a line
without regard to the line length, and sometimes that can be very
handy.

Another poster asked about using these techniques to find antenna
resonance. The problem for the typical ham's instrumentation is that
the antenna is generally a low Q and presents a fairly high resistance
(in the neighborhood of 50 ohms, presumably, but not necessarily very
close to that) at resonance. It's easy enough to measure the system
resonance (lack of reactance) at the shack end of the line feeding the
antenna, but AFAIK, most hams don't have an analyzer they can
calibrate at the end of a piece of transmission line like you can with
modern lab network analyzers. So you'd have to measure the impdeance
(R+jX, complex value) you see at the shack end of the line and put
that into a calc that backs the line out so you can know the impedance
at the antenna feedpoint. -- But why bother? It doesn't generally
matter whether an antenna is _resonant_ or not, only that it radiates
efficiently with the desired pattern, and that you have a way to get
power to it efficiently. So at least at a given frequency, you can
trim the antenna and/or put a small matching network at the antenna to
get the line SWR low enough to suit you. If you can stand the
resulting SWR over the range of freqs you want to use, then any
additional matching can be done at the shack end of the line.

Yet another note. ;-) When the measurement is done in a system with
a 50-ohm impedance, and the components (L and C) in the resonant tank
have fairly low reactance, you won't get a very sharp null with the
return-loss bridge and additional series R as the OP suggested. The
additional R makes the tank Q very low. With Rick's suck-out circuit
suggestion, you'll need a detector with pretty wide dynamic range
(like a receiver perhaps) to see where the deepest null occurs,
because even fairly far off resonance, there will be significant
attenuation.

Cheers,
Tom