¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io
  1. QRPLabs
  2. Topics

L C Meter

 

With respect to several posts here on the discussion group;

I would like to wake up an old (1950s) L C measuring technique where calibration and absolute accuracy was not an issue.

You will need at least two (three is better) reactive components of simular value and quality. We will define one component as our 'standard' and the other as 'unit under test'.
?
Set up...
To measure C ... connect two (2) capacitors in PARALLEL
?> OR <
To measure L ... connect two (2) inductors in SERIES

... if meter allows set to a realistic operating frequency ...

Measure and record the value of two (2) components.
Remove the component 'under test' leaving only the quote 'standard' value connected to the meter.
Record the 'standard' value.
The value of the component under test is the difference between these two measures. Substract the small number from the large number.

I feel the accuracy of this technique is independent of absolute? meter accuracy and as far as I know, there is no way to define the accuracy of the value of the component under test but to say 'empirically' it may be spot on. Remember too component lead length and or stray capacitance of test fixture will play a role in the final 'measured' value.

We used this method long ago when commercial test equipment was not an option for a struggling teenage ham but only had a homebrew grid dip oscillator and general coverage HF receiver. Guess who was king of the ham club when I got a second hand Heathkit GDO?

John
N3AAZ
Arv Evans
 

A bit more on accuracy and tolerance.? Your commercial or homebrewed LC meter
will only be as accurate as its calibration.? Even commercial meters can benefit from
a custom calibration session.? Then it is an issue of how long it will hold this accuracy.
Commercial calibration lab services used to track individual instruments by serial
number and logged the change in adjustments required to bring them back into
calibration.? This provided indication of calibration drift over unit of time so that calibration
offsets could be predicted between calibration dates.

Inductors and capacitors have specific impedances at specific frequencies.? You can
use this to accurately determine the actual value of a capacitor or inductor.? Measurement
can be done with a precision resistive load and RF detector or oscilloscope.?

Another way is to use a bridge network with one leg containing known value precision
resistors and the other leg containing the D.U.T and a precision resistor.? Using the
bridge method does not require a precise frequency...a simple crystal oscillator will
work for the RF source.?

Over the past few years I have slowly converted from leaded components to SMD
types.? Reason is that most SMD components are bulk produced and then laser
trimmed to specific values.? This seems to imply that most, if not all, SMD components
are 1% tolerance, or better.? If the idea of non-leaded components bothers you it is
possible to add leads to SMD devices by holding one end in a metallic (heat absorbing)
vise while soldering a short piece of wire to the opposite end.? SMD components can
also be installed "tombstone style" (standing up on one end) on existing PCB pads
or component leads.? And best of all, SMD components are usually very inexpensive.
Surplus stores sometimes have plastic bags of them for a dollar or so per pound (there
is a lot of SMD components in a pound).? Tweezers type connection probes help to
quickly sort through a bag of random value (floor sweepings) SMD components.














_._



On Sun, Nov 12, 2017 at 11:49 AM, John Kirby <n3aaz_qrp_1@...> wrote:
With respect to several posts here on the discussion group;

I would like to wake up an old (1950s) L C measuring technique where calibration and absolute accuracy was not an issue.

You will need at least two (three is better) reactive components of simular value and quality. We will define one component as our 'standard' and the other as 'unit under test'.
?
Set up...
To measure C ... connect two (2) capacitors in PARALLEL
?> OR <
To measure L ... connect two (2) inductors in SERIES

... if meter allows set to a realistic operating frequency ...

Measure and record the value of two (2) components.
Remove the component 'under test' leaving only the quote 'standard' value connected to the meter.
Record the 'standard' value.
The value of the component under test is the difference between these two measures. Substract the small number from the large number.

I feel the accuracy of this technique is independent of absolute? meter accuracy and as far as I know, there is no way to define the accuracy of the value of the component under test but to say 'empirically' it may be spot on. Remember too component lead length and or stray capacitance of test fixture will play a role in the final 'measured' value.

We used this method long ago when commercial test equipment was not an option for a struggling teenage ham but only had a homebrew grid dip oscillator and general coverage HF receiver. Guess who was king of the ham club when I got a second hand Heathkit GDO?

John
N3AAZ
 

May have already been seen, but here are a few very cool RLC cheapo meter... using a laptop's sound card and FFT software:






73, Joe W2JEJ

On Sun, Nov 12, 2017 at 3:03 PM, Arv Evans <arvid.evans@...> wrote:
A bit more on accuracy and tolerance.? Your commercial or homebrewed LC meter
will only be as accurate as its calibration.? Even commercial meters can benefit from
a custom calibration session.? Then it is an issue of how long it will hold this accuracy.
Commercial calibration lab services used to track individual instruments by serial
number and logged the change in adjustments required to bring them back into
calibration.? This provided indication of calibration drift over unit of time so that calibration
offsets could be predicted between calibration dates.

Inductors and capacitors have specific impedances at specific frequencies.? You can
use this to accurately determine the actual value of a capacitor or inductor.? Measurement
can be done with a precision resistive load and RF detector or oscilloscope.?

Another way is to use a bridge network with one leg containing known value precision
resistors and the other leg containing the D.U.T and a precision resistor.? Using the
bridge method does not require a precise frequency...a simple crystal oscillator will
work for the RF source.?

Over the past few years I have slowly converted from leaded components to SMD
types.? Reason is that most SMD components are bulk produced and then laser
trimmed to specific values.? This seems to imply that most, if not all, SMD components
are 1% tolerance, or better.? If the idea of non-leaded components bothers you it is
possible to add leads to SMD devices by holding one end in a metallic (heat absorbing)
vise while soldering a short piece of wire to the opposite end.? SMD components can
also be installed "tombstone style" (standing up on one end) on existing PCB pads
or component leads.? And best of all, SMD components are usually very inexpensive.
Surplus stores sometimes have plastic bags of them for a dollar or so per pound (there
is a lot of SMD components in a pound).? Tweezers type connection probes help to
quickly sort through a bag of random value (floor sweepings) SMD components.














_._



On Sun, Nov 12, 2017 at 11:49 AM, John Kirby <n3aaz_qrp_1@...> wrote:
With respect to several posts here on the discussion group;

I would like to wake up an old (1950s) L C measuring technique where calibration and absolute accuracy was not an issue.

You will need at least two (three is better) reactive components of simular value and quality. We will define one component as our 'standard' and the other as 'unit under test'.
?
Set up...
To measure C ... connect two (2) capacitors in PARALLEL
?> OR <
To measure L ... connect two (2) inductors in SERIES

... if meter allows set to a realistic operating frequency ...

Measure and record the value of two (2) components.
Remove the component 'under test' leaving only the quote 'standard' value connected to the meter.
Record the 'standard' value.
The value of the component under test is the difference between these two measures. Substract the small number from the large number.

I feel the accuracy of this technique is independent of absolute? meter accuracy and as far as I know, there is no way to define the accuracy of the value of the component under test but to say 'empirically' it may be spot on. Remember too component lead length and or stray capacitance of test fixture will play a role in the final 'measured' value.

We used this method long ago when commercial test equipment was not an option for a struggling teenage ham but only had a homebrew grid dip oscillator and general coverage HF receiver. Guess who was king of the ham club when I got a second hand Heathkit GDO?

John
N3AAZ
 


?" "A bit more on accuracy and tolerance.? Your commercial or homebrewed LC meter will only be as accurate as its calibration" "

I respectively disagree with the above quote comment. The point I tried to make is that an extremely accurate capacitor value can be determined regardless of test equipment calibration status or "standard" component tolerance.

I failed to make my point above, maybe an example...

One analogy of the old (1950s) GDO empirical measuring technique I describe above.

?If you put 73 marbles into a bag, then I remove a number from the bag, how would you determine how many marbles I took out of the bag? ...
...Count the marbles left in the bag and substract from 73...
That was an empirical measurement ...

If you parallel two capacitors and measure the total capacitance, then I remove one capacitor, how would you determine the value of the capacitor I removed?

Again ... meter accuracy and or 'standard' component tollerance play no roll in this type of L/C measurement.
The value of the capacitor I removed is 'spot on' assuming a clean and proper test fixture.

John
N3AAZ
 

Well said.
?Even homebrew LC meters can be as accurate , if not more accurat. Of course i ahave read that while using the microcontroller, the test results drift twice the erroe in the clock it is using,? It has relevance to the code that is being used.
?Suppose we are using VK3BHR design , we can pre calibrate the clock comparing it with known stadard reference.
?There after the code reflected errors also would be minimized.
?Thanks for the thoughts Jhn.

regards
sarma
?vu3zmv

regards
Sarma
?

On Mon, Nov 13, 2017 at 8:44 PM, John Kirby <n3aaz_qrp_1@...> wrote:

?" "A bit more on accuracy and tolerance.? Your commercial or homebrewed LC meter will only be as accurate as its calibration" "

I respectively disagree with the above quote comment. The point I tried to make is that an extremely accurate capacitor value can be determined regardless of test equipment calibration status or "standard" component tolerance.

I failed to make my point above, maybe an example...

One analogy of the old (1950s) GDO empirical measuring technique I describe above.

?If you put 73 marbles into a bag, then I remove a number from the bag, how would you determine how many marbles I took out of the bag? ...
...Count the marbles left in the bag and substract from 73...
That was an empirical measurement ...

If you parallel two capacitors and measure the total capacitance, then I remove one capacitor, how would you determine the value of the capacitor I removed?

Again ... meter accuracy and or 'standard' component tollerance play no roll in this type of L/C measurement.
The value of the capacitor I removed is 'spot on' assuming a clean and proper test fixture.

John
N3AAZ


 

The point I tried to make is that an extremely accurate capacitor value can be determined regardless of test equipment calibration status or "standard" component tolerance.
John,

Sorry I do not understand.

I failed to make my point above, maybe an example...
?If you put 73 marbles into a bag, then I remove a number from the bag, how would you determine how many marbles I took out of the bag? ...
...Count the marbles left in the bag and substract from 73...
Yes, but how do you count the total number of pFs in your bag?
Then, how do you count how many remain?
They are not marbles the pFs are hidden.

I use a dip meter, I can get the resonant frequency easily but cannot tell the component values without a standard. Or am I missing something?

73 Alan G4ZFQ
Ken KM4NFQ
 

Disclaimer: I have an Art background, not STEM.

I have a Cheap Chinese LC Meter from eBay of unknown accuracy, not calibrated.
I have two capacitors, that I have scavenged from electronics bought
at Goodwill.
They are of unknown values. I label them A and B.
I hook the capacitors up to the LC Meter in parallel, which gives the
sum of the caps.
My reading is 1376pF. So A+B = 1376pF.
I remove capacitor B. Now the reading is 939pF.
What are the values of capacitors A and B, as measured on that LC Meter?
How accurate are those values?

Later, a friend comes over with an accurate, calibrated LC Meter.
We measure the capacitors, and A=800pF and B=330pF.
How could I have gotten those accurate values from the first LC Meter?

Curious,
Ken, KM4NFQ

On Mon, Nov 13, 2017 at 11:54 AM, Alan G4ZFQ <alan4alan@...> wrote:

The point I tried to make is that an extremely accurate capacitor value
can be determined regardless of test equipment calibration status or
"standard" component tolerance.

John,

Sorry I do not understand.

I failed to make my point above, maybe an example...
If you put 73 marbles into a bag, then I remove a number from the bag,
how would you determine how many marbles I took out of the bag? ...
...Count the marbles left in the bag and substract from 73...

Yes, but how do you count the total number of pFs in your bag?
Then, how do you count how many remain?
They are not marbles the pFs are hidden.

I use a dip meter, I can get the resonant frequency easily but cannot tell
the component values without a standard. Or am I missing something?

73 Alan G4ZFQ
 


Alan,

Thanks for the reply.

Sorry, my marble example mixed (apples and oranges) GDOs and LC meters. I just muddied the water with that example. Also, I stand corrected, yes, a standard reference? component is required to measure reactance using a GDO and balanced bridge.


Ken,

Thanks for the reply.

In your example...
where A + B = 1376 and A = 939
then by arithmetic (component under test) B = 437

In this case 'A' is the 'uncalabrated' 'standard' as such it's absolute value is not an issue.
?
Parallel in a third capacitor defined C, now for example...
(A + C) + B = 1876 then remove B now 1439 is indicated on the meter.?? Therefore? 1876 - 1439 = B = 437

With respect to your friends meter and difference in measured values of same components. Capacitance measurements are frequency dependant. I assume your LC meters were on different frequencies.
......
I will take this topic off line until I better understand probably error at higher frequencies and very small capacitor values.....

72, 73
John
N3AAZ
 

yes, a standard reference? component is required to measure reactance using a GDO and balanced bridge.
Ken,

I spend ages looking at those mathematical puzzles without a clue how to solve them.
It's a relief to know there is no solution to that one:-)

73 Alan G4ZFQ

1 - 9 of 9