This thread is opened to investigate the methods available to measure the Q of an LCR circuit.
Some information is already available in another thread, but to not further interrupt that thread this new one is opened.
That old thread information can be found starting from here: https://www.overunityresearch.com/index.php?topic=4927.msg118776#msg118776
Itsu
The last information on that old thread was about measuring a T200-2 toroid which according this datasheet: https://datasheets.micrometals.com/T200-2-DataSheet.pdf should be able to reach a (theoretical?) Q of 332:
Q test winding N=40, #20 AWG
Q frequency 2 MHz
Q min on HP4342A 332
I guess we could set that as a goal and try to come close to that figure.
Itsu
A T200-2 toroid with 40 turns / 0.8mm diameter (#20 AWG) coil should have a parallel or series capacitance of 300pF to be resonant at 2MHz.
I will make such a resonator and do some initial (Fres / Bandwidth (-3dB) ) measurements to see where we stand using this simple method.
Itsu
This is the device under test (DUT):
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55896;image)
FG input through a single turn: ~2MHz at 20Vpp.
At resonance (2180kHz), output P6139 probe 10MOhm / 8pF see yellow trace, input blue trace:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55900;image)
Measuring the frequency at -3dB points (Fres x 0.707 = 26.159V) gives High 2260kHz, low 2116kHz = difference of 144.
Q = Fres/difference = 2180 / 144 = 15.1, so a far cry of the theoretical Q of 332
Itsu
Hi Itsu,
Thanks for starting this topic.
I referred to this Amidon data sheet page 15 https://alexsradioshop.de/wp-content/uploads/2023/01/AmidonAMI.pdf (https://alexsradioshop.de/wp-content/uploads/2023/01/AmidonAMI.pdf) in the other thread. It has graphs for the T200-2 core's Q values in the function of frequency and in the function of the number of turns.
From the graph, it is shown that to get Q > 300, the number of turns would be between roughly 25 and 40 (i.e. the coil inductance is between 8uH and 20 uH) in the frequency range between 3 MHz and 4 MHz.
For the number of turns I would suggest using between 20 and 25 if the intended resonant frequency remains around 3.8 MHz, this gives coil inductance between 4.8 uH and 7.5 uH. So the tuning capacitor should be chosen accordingly.
Just seen your first measurement. First try to measure the Q for the same LC (remove only the 1 turn coupling coil) in the test circuit you showed for the nanoVNA here https://www.overunityresearch.com/index.php?topic=4927.msg118906#msg118906 (https://www.overunityresearch.com/index.php?topic=4927.msg118906#msg118906) Try to use less than 10 pF coupling capacitors like 2 x 10 pF at both the input and output or less.
Gyula
Itsu, from the scope shot I assume the blue trace shows the generator input voltage feeding the 1 turn coupling coil, is this so?
If yes, its peak to peak amplitude is about 1200 mV and dividing the 37.77 Vpp by this it gives a Q of around 31, twice of the Q of the 3 dB point method.
Perhaps the coupling capacitor test will bring a better Q value, though probably not yet over 300.
Gyula
Gyula,
QuoteItsu, from the scope shot I assume the blue trace shows the generator input voltage feeding the 1 turn coupling coil, is this so?
that is correct, i added the blue trace pp value (1.17Vpp) in that screenshot, and indeed this method yields a Q of 31.6.
I will use the nanoVNA method using the coupling capacitors.
I tested earlier with self-made dual bonded magnet wire caps trimmed to 1pF, but that gave a too marginal signal to show a reliable measurement, i will start with 2x 10pF in series (5pF) at the in- and outputs.
Itsu
Here i connected the DUT as shown using 3x 10pF (~4pF) caps in series at the in- and output of the nanoVNA:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55904;image)
The results are shown here but are kind of rough due to the low (4pF) coupling at the in and output of the nanoVNA:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55902;image)
But it shows again a Q of 15.3.
I will try to keep "score" from the different method at each measurement post:
Method 1: FG / Scope single loop coupling Fres / bandwidth (-3dB): Q = 15.1
Method 2: FG / Scope single loop coupling Output Voltage / input voltage: Q = 31.6
Method 3: nanoVNA 4pF coupling in and out Fres / bandwidth (-3dB): Q = 15.3
Itsu
Ok Itsu, thanks.
In this paper I have found, in the bottom part of page 9 there is info on using a (nano)VNA for measuring Q. https://w7zoi.net/oldtech/zoi-nano.pdf (https://w7zoi.net/oldtech/zoi-nano.pdf) No coupling caps are used, the series LC circuit shunts the TX-RX ports to ground that are connected together via an SMA Tee, the series LC makes a notch in the frequency response.
Then the author refers to a formula 7.4 in page 7.34 in this book, EMRFD which is Experimental Methodes in RF Design, see here https://www.scribd.com/document/336323413/Experimental-Methods-in-RF-Design (https://www.scribd.com/document/336323413/Experimental-Methods-in-RF-Design)
We could try this and see what Q it gives.
Another method for Q measurement is the use of the 50:1 turns ratio transformer I already referred to in the other thread, https://www.overunityresearch.com/index.php?topic=4927.msg118917#msg118917 (https://www.overunityresearch.com/index.php?topic=4927.msg118917#msg118917) using the T520-2 with the 50 turns for primary and a single turn for the secondary coils.
Here is another very good paper based (also) on measurements which includes the Q of several capacitor types https://w7zoi.net/twofaces.pdf (https://w7zoi.net/twofaces.pdf)
Of course, these need time to digest and decide on the test circuit, no need for any hurry.
Gyula
Quote from: Itsu on 2026.04.22, 09:17:45
This thread is opened to investigate the methods available to measure the Q of an LCR circuit.
My 2 cents so you don't feel abandoned.
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=55923)
Measure the current in the yellow wire using your contactless current probe.
Keep the yellow wires thick and short.
Make the R1 resistance such that the maximum current flowing through the MOSFET and D1 and D2 approaches their current ratings (i
max) but does not exceed them ( R1≈V
CC/i
max )
Keep the R2 gate resistor high and R3 gate resistor low resistance (as low as it is safe).
Once the Q2 MOSFET turns OFF, the LC tank becomes isolated and can ring down with minimal loading (which is the resistance of R1 combined with the reactance of MOSFTET's output capacitance C
OSS in series with the reverse capacitances of D1 and D2, C
TOTAL=C
D1C
D2C
OSS/[C
D1C
D2+C
D1C
OSS+C
2C
OSS] ).
The MOSFET's V
D-S_max and the D1 and D2 blocking voltage should be at least 2*V
CC. The D1 and D2 should have a small junction capacitance under reverse biased conditions. ...the MOSFET's C
oss should be as small as possible, too (this capacitance decreases as the drain voltage increases). Ferrite beads on D1 and D2 might decrease the loading further (worth trying with and without them).
CALCULATIONS:The time to ring down to 50% of the initial amplitude and multiplied by π/ln2 (4.53236...) and divided by the ringing period is the Quality Factor (Q).
The time to ring down to 37% of the initial amplitude and multiplied by π (3.14159...) and divided by the ringing period is the Quality Factor (Q), too.
The time to ring down to 4.32% of the initial amplitude divided by the ringing period is the Quality Factor (Q), too.
Verpies, would you tell Itsu whether he use the series aiding or bucking connection for the two coils? (The two dots show series aiding, of course.)
Gyula
Quote from: gyula on 2026.04.22, 20:13:10
Verpies, would you tell Itsu whether he use the series aiding or bucking connection for the two coils? (The two dots show series aiding, of course.)
He sometimes uses aiding and sometimes bucking and compares them.
Quote from: gyula on 2026.04.22, 20:02:12
Ok Itsu, thanks.
In this paper I have found, in the bottom part of page 9 there is info on using a (nano)VNA for measuring Q. https://w7zoi.net/oldtech/zoi-nano.pdf (https://w7zoi.net/oldtech/zoi-nano.pdf) No coupling caps are used, the series LC circuit shunts the TX-RX ports to ground that are connected together via an SMA Tee, the series LC makes a notch in the frequency response.
Then the author refers to a formula 7.4 in page 7.34 in this book, EMRFD which is Experimental Methodes in RF Design, see here https://www.scribd.com/document/336323413/Experimental-Methods-in-RF-Design (https://www.scribd.com/document/336323413/Experimental-Methods-in-RF-Design)
We could try this and see what Q it gives.
Another method for Q measurement is the use of the 50:1 turns ratio transformer I already referred to in the other thread, https://www.overunityresearch.com/index.php?topic=4927.msg118917#msg118917 (https://www.overunityresearch.com/index.php?topic=4927.msg118917#msg118917) using the T520-2 with the 50 turns for primary and a single turn for the secondary coils.
Here is another very good paper based (also) on measurements which includes the Q of several capacitor types https://w7zoi.net/twofaces.pdf (https://w7zoi.net/twofaces.pdf)
Of course, these need time to digest and decide on the test circuit, no need for any hurry.
Gyula
Thanks Gyula, good finds, i will study the first one first (nanoVNA).
Itsu
Quote from: Verpies on 2026.04.22, 20:04:59
My 2 cents so you don't feel abandoned.
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=55923)
Measure the current in the yellow wire using your contactless current probe.
Keep the yellow wires thick and short.
Make the R1 resistance such that the maximum current flowing through the MOSFET and D1 and D2 approaches their current ratings (imax) but does not exceed them ( R1≈VCC/imax )
Keep the R2 gate resistor high and R3 gate resistor low resistance (as low as it is safe).
Once the Q2 MOSFET turns OFF, the LC tank becomes isolated and can ring down with minimal loading (which is the resistance of R1 combined with the reactance of MOSFTET's output capacitance COSS in series with the reverse capacitances of D1 and D2, CTOTAL=CD1CD2COSS/[CD1CD2+CD1COSS+C2COSS] ).
The MOSFET's VD-S_max and the D1 and D2 blocking voltage should be at least 2*VCC. The D1 and D2 should have a small junction capacitance under reverse biased conditions. ...the MOSFET's Coss should be as small as possible, too (this capacitance decreases as the drain voltage increases). Ferrite beads on D1 and D2 might decrease the loading further (worth trying with and without them).
CALCULATIONS:
The time to ring down to 50% of the initial amplitude and multiplied by π/ln2 (4.53236...) and divided by the ringing period is the Quality Factor (Q).
The time to ring down to 37% of the initial amplitude and multiplied by π (3.14159...) and divided by the ringing period is the Quality Factor (Q), too.
The time to ring down to 4.32% of the initial amplitude divided by the ringing period is the Quality Factor (Q), too.
Verpies, Thanks, I feel better now ;)
I did use the ringing method on Smudge's device, but very simple, like pulsing with a short pulse from the FG via a single loop, scope the ringing output and count the cycles to half the amplitude etc.
It turned out to match the Fres / bandwidth method Q value.
But using the contactless current probe to measure the ringing omits the scope probe loading O0
So i will try that too.
In this DUT (T200-2) we only have 1 coil in the LCR circuit, so no aiding or bucking modes available.
Itsu
Quote from: gyula on 2026.04.22, 20:02:12
In this paper I have found, in the bottom part of page 9 there is info on using a (nano)VNA for measuring Q. https://w7zoi.net/oldtech/zoi-nano.pdf (https://w7zoi.net/oldtech/zoi-nano.pdf) No coupling caps are used, the series LC circuit shunts the TX-RX ports to ground that are connected together via an SMA Tee, the series LC makes a notch in the frequency response.
That would be the measurement #
2.
Quote from: Verpies on 2026.04.13, 20:18:39
Let's keep this organized - there are four S21 measurement possible:
1) Shunt-through S21 measurement of a parallel LC DUT
2) Shunt-through S21 measurement of a series LC DUT
3) Series S21 measurement of a parallel LC DUT
4) Series S21 measurement of a series LC DUT
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3924.0;attach=35778)
Quote from: Itsu on 2026.04.13, 20:48:10
I did use the 2) Shunt-through S21 measurement of a series LC DUT in the measurement shown.
Measurement #
3 would make sense, too.
Quote from: Verpies on 2026.04.22, 20:04:59
My 2 cents so you don't feel abandoned.
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=55923)
Measure the current in the yellow wire using your contactless current probe.
Keep the yellow wires thick and short.
Make the R1 resistance such that the maximum current flowing through the MOSFET and D1 and D2 approaches their current ratings (imax) but does not exceed them ( R1≈VCC/imax )
Keep the R2 gate resistor high and R3 gate resistor low resistance (as low as it is safe).
Once the Q2 MOSFET turns OFF, the LC tank becomes isolated and can ring down with minimal loading (which is the resistance of R1 combined with the reactance of MOSFTET's output capacitance COSS in series with the reverse capacitances of D1 and D2, CTOTAL=CD1CD2COSS/[CD1CD2+CD1COSS+C2COSS] ).
The MOSFET's VD-S_max and the D1 and D2 blocking voltage should be at least 2*VCC. The D1 and D2 should have a small junction capacitance under reverse biased conditions. ...the MOSFET's Coss should be as small as possible, too (this capacitance decreases as the drain voltage increases). Ferrite beads on D1 and D2 might decrease the loading further (worth trying with and without them).
CALCULATIONS:
The time to ring down to 50% of the initial amplitude and multiplied by π/ln2 (4.53236...) and divided by the ringing period is the Quality Factor (Q).
The time to ring down to 37% of the initial amplitude and multiplied by π (3.14159...) and divided by the ringing period is the Quality Factor (Q), too.
The time to ring down to 4.32% of the initial amplitude divided by the ringing period is the Quality Factor (Q), too.
Initial test run using the above shown method and this calculation: "The time to ring down to 4.32% of the initial amplitude divided by the ringing period is the Quality Factor (Q), too".
These are the components i used and the results in green:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55939;image)
Here the screenshot from the current ringdown in green (blue is input pulse from FG):
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55941;image)
Q calculates as 19 if i understand the calculations correctly.
Edit: D1 is also a 1N4148 (100V / 300mA and junction capacitance of 4pF).
Coss of the MOSFET is 185pF
Itsu
Quote from: Itsu on 2026.04.24, 10:03:03
Q calculates as 19 if i understand the calculations correctly.
Edit: D1 is also a 1N4148 (100V / 300mA and junction capacitance of 4pF).
Coss of the MOSFET is 185pF
Looks correct. I count a little more than 19 ringing cycles to your cursor, though.
The major improvement would be to install low-capacitance low-voltage TVS diodes (like SP0402B-ULC) in place of D1 and D2 and increasing V
CC by their combined voltage drop.
Another and smaller improvement would be to use a MOSFET with a lower C
OSS. Most likely a higher voltage MOSFET or GaN like EPC2037.
i
max of present D1 and D2 limit you to 300mA anyway (a big mismatch with the 14A I
Dmax of the present Q2).
Cosmetically, I'd increase R2 a lot to minimize ringing and maximize stability when Q2 turns ON.
QUESTION: What is the capacitance of your low-voltage bidir-TVS diodes ?
With low-capacitance TVS diodes like SP0402B-ULC you could decrease the LC tank loading 100x
My 1.5KE18CA bidi TVS (18V) seem to have a capacitance between 2 and 4.3nF, so not really an option and compared to your SP0402B-ULC a nono.
The RFD14N05L was the MOSFET with the lowest Coss i had available, other MOSFETs including higher voltage ones like IRF740 / IRF840 and IRFP460 etc. all had higher (330pF) values.
Would changing those components you suggested make such a difference that it would near the (theoretical?) Q 0f 300 as specified in the T200-2 toroid datasheet?
Itsu
Quote from: Itsu on 2026.04.24, 14:25:17
Would changing those components you suggested make such a difference that it would near the (theoretical?) Q 0f 300 as specified in the T200-2 toroid datasheet?
I don't know.
The SP0402B-ULC diodes would decrease the LC loading not only by their lower capacitance but also by blocking conduction in both directions below their combined breakdown voltage (2*9V).
With these diodes, the LC tank could ring freely below ±18V
p-p amplitude while being loaded down by a reactance of less than 0.13pF
Ok, good enough for me, i will order those TVS's and some EPC2037's.
Meanwhile i will proceed with the other mentioned methods.
Next i used the method mentioned by Gyula in an earlier post:
QuoteIn this paper I have found, in the bottom part of page 9 there is info on using a (nano)VNA for measuring Q. https://w7zoi.net/oldtech/zoi-nano.pdf No coupling caps are used, the series LC circuit shunts the TX-RX ports to ground that are connected together via an SMA Tee, the series LC makes a notch in the frequency response.
Then the author refers to a formula 7.4 in page 7.34 in this book, EMRFD which is Experimental Methods in RF Design, see here https://www.scribd.com/document/336323413/Experimental-Methods-in-RF-Design
We could try this and see what Q it gives.
The suggested nanoVNA and DUT setup looks like this:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55943;image)
The formula mentioned in the book 7.4 in page 7.34 look like this:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55945;image)
So after setting it up the nanoVNA produced this result:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55947;image)
Using the shown formula (i used the series LC setup Qs) where f = 2.239MHz, A = -7.1dB, Lu = 19uH and Z =50 Ohm, i calculated Q to be 13.46 (please check my result).
Again very close to the earlier used methods results.
"score" from the different method at each measurement post:
Method 1: FG / Scope single loop coupling Output Voltage / input voltage: Q = 31.6
Method 2: FG / Scope single loop coupling Fres / bandwidth (-3dB): Q = 15.1
Method 3: nanoVNA 4pF coupling in and out Fres / bandwidth (-3dB): Q = 15.3
Method 4: verpies his current probe ringdown setup Q = 19
Method 5: FG / Scope 3pF coupling in and out Fres / bandwidth (-3dB) Q = 15.25 (i also tried with ONLY 1pF capacitor as input, result was Q = 15.8 )
Method 6: nanoVNA Tee setup / formula used Q = 13.46
Itsu
Quote from: Itsu on 2026.04.24, 20:05:34
So after setting it up the nanoVNA produced this result:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55947;image)
That's a very handsome notch.
Quote from: Itsu on 2026.04.24, 20:05:34
Method 1: FG / Scope single loop coupling Output Voltage / input voltage: Q = 31.6
Method 2: FG / Scope single loop coupling Fres / bandwidth (-3dB): Q = 15.1
Method 3: nanoVNA 4pF coupling in and out Fres / bandwidth (-3dB): Q = 15.3
Method 4: verpies his current probe ringdown setup Q = 19
Method 5: FG / Scope 3pF coupling in and out Fres / bandwidth (-3dB) Q = 15.25 (i also tried with ONLY 1pF capacitor as input, result was Q = 15.8 )
Method 6: nanoVNA Tee setup / formula used Q = 13.46
The average of these 6 measurements is
18.285
Quote from: Itsu on 2026.04.24, 19:08:15
Ok, good enough for me, i will order those TVS's and some EPC2037's.
EPC2038 is the lowest C
OSS (1.6pF) MOSFET but it can handle only 500mA of pulsed current.
EPC2037 has a 6.5pF C
OSS but it can handle 2.4A, which is similar to the 2A i
max of the SP0402B-ULC TVS diode.
EPC8010 can handle 7.5A at the expense of 25pF C
OSS.
EPC2036 can handle 18A at the expense of 50pF C
OSS.
All of the above are 100V MOSFETs.
A noteworthy 200V MOSFET is the EPC2012C which can handle 22A at 64pF C
OSS.
Thanks, i ordered some of these parts.
Meanwhile Gyula was still wondering why the Q results are so low (around 15), and in a simulation he showed that a high ESR of the resonance capacitor (287pF) could cause this.
Using my Agilent 1733C LCR meter which has buttons for ESR and DCR measurements showed that the 287pF resonance capacitor had an ESR of 83 Ohm.
Looking at the datasheet of that TDK 100pF ceramic capacitor (i use 3 parallel) it turns out that that is according to its specs:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55954;image)
I don't think this causes a problem to see which method is the better one as they all measure the same thing, but it could explain why we see such a low Q.
Itsu
Quote from: Itsu on 2026.04.25, 21:01:03
Using my Agilent 1733C LCR meter which has buttons for ESR and DCR measurements showed that the 287pF resonance capacitor had an ESR of 83 Ohm.
So add a low-ESR cap to your parts order before it ships on Monday.
I have been pretty happy with Murata ERB series (https://www.digikey.com/en/product-highlight/m/murata-electronics-north-america/high-frequency-high-q-capacitors) capacitors that have 40mΩ ESR at 3MHz.
Quote from: Verpies on 2026.04.25, 21:35:25
So add a low-ESR cap to your parts order before it ships on Monday.
I have been pretty happy with Murata ERB series (https://www.digikey.com/en/product-highlight/m/murata-electronics-north-america/high-frequency-high-q-capacitors) capacitors that have 40mΩ ESR at 3MHz.
I had some 270pF SMD caps: kemet c1206h271jggact250 which suppose to be low ESR, so I measured one to be 269.8pF and an ESR of 7.5 Ohm at 100kHz (my max).
I tested it with the T200-2 core with the 40 turns 1mm diameter wire measured L= 18.82uH and DCR of 0.065 Ohm (100kHz) and a single turn input loop from the FG.
Result is:
270pF SMD cap, 500mVpp input voltage, Fres 2204kHz at 19.9Vpp output voltage with -3dB points (19.9V x 0.707 = 14Vpp) at High 2210.2kHz, low 2196.6kHz = 13.6kHz difference with a Q of 2204 / 13.6 = 162
I also used a big variable air capacitor set to 270.2pF which measured an ESR of 2K at 1kHz, but that was lowered to 3 Ohm at 100kHz.
Result is:
270pF Air cap, 400mVpp input voltage, Fres 2174kHz at 15.6Vpp output voltage with -3dB points (15.6V x 0.707 = 11Vpp) at High 2182kHz, low 2165.6kHz = 16.4kHz difference with a Q of 2174 / 16.4 =132.5
So the ESR of the capacitor is a major factor for a good Q result.
Those Murata ERB series capacitors look even better, so worth a test with them.
I don't think comparing this measurement result with the ones done before is valid, so i won't list them here.
Itsu
Indeed, we cannot compare the previous Q measurement results with this new result because both the coil and the tuning capacitor affects the resulting Q value. So in this case the capacitor was to blame for the miserably low Q values.
We can trust in the Amidon toroidal core specifics, namely for the T200-2 type, it has Q=425 at and around 2 MHz (N=40 turns, 20uH, awg #18), this has been tested during the decades and proved correct data.
So a good quality tuning capacitor is needed, with a Q > 1000 but preferably even higher Q like 2000-3000. These are expensive, unfortunately.
I attached the graph on the Q values of the T200-2 core from its data sheet.
At 2 MHz the 20 uH coil on this core has an inductive reactance of XL=251.3 Ohm and has a Q=425. Its total loss resistance (core + wire) is r = XL / Q i.e. r = 251.3 / 425 = 0.59 Ohm
So the ESR of a tuning capacitor should be way below than this, preferably around 0.1 - 0.2 Ohm at 2 MHz so that it should not ruin significantly the coil Q. Reverse calculating the Q for such a capacitor would give a Q = 251.3 Ohm / 0.1 = 2513 or 251.3 / 0.2 = 1256
Then the Q value of the LC parallel circuit at resonance will be 425 x 2513 / (425 + 2513) = 363 or 425 x 1256 / (425 + 1256) = 317 (The resulting Q can be calculated like the resulting resistance of paralleled resistors.)
Gyula
Gyula,
thanks for the clear explanation, so we should aim for a Q of >300 at least.
The problem is to find a good enough (low ESR) capacitor, preferable one from which we can measure the ESR at the resonance frequency of the LCR circuit, in our case around 2240kHz (40 turn coil on a T200-2 core with a parallel capacitor).
There are procedures to measure the ESR of a capacitor, but only at its self resonance frequency which often is much higher than our LCR resonance frequency.
Like my variable air capacitor set to 270pF which measures an ESR of 0.7 Ohm at 15.7MHz self resonance frequency.
That is nice to know, but does not say anything about the ESR of it at 2240kHz.
I have some low ESR SMD capacitors on order, so it probably will be a trail and error method to find the correct one.
Itsu
Well, you can use a series LC circuit too, either using the nanoVNA which has 50 Ohm terminations at its RX-TX terminals (and no coupling caps), measuring S21 at the bottom of the notch and look the ESR in the table a youtuber guy links to.
or you could use the the 50:1 step down transformer as referred to in the HP Q meter, also in series LC connection, scope probe would be hooked up across the tuning capacitor.
Lets hope the ordered capacitors will have an ESR of < say 0.2 around 2-3 MHz. 8)
EDIT: You wrote:
Quote from: Itsu on 2026.04.28, 20:46:32
... so we should aim for a Q of >300 at least.
Well, in fact we should aim for as high a Q value as we can achieve with practical components (considering design goals for our LC circuit, of course).
Gyula
Gyula,
QuoteWell, you can use a series LC circuit too, either using the nanoVNA which has 50 Ohm terminations at its RX-TX terminals (and no coupling caps), measuring S21 at the bottom of the notch and look the ESR in the table a youtuber guy links to.
Good idea, you are talking about this video: https://www.youtube.com/watch?v=lO97K8th_uU and this chart at the 11-minute mark:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55990;image)
So using my nanoVNA on the SERIES LCR circuit of the 40 turn coil on the T200-2 core and the 270pF variable air capacitor, i get these graphs:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=55992;image)
The upper graph is the S21 Log Mag sweep pointing to a series LCR resonance of 2300kHz at -23.02dB which according to the above chart points to around 2 Ohm.
The bottom chart shows in red the real resistance R at series resonance as 1.83 Ohm.
As the used method is for measuring / calculating the ESR of a capacitor ALONE, i am not sure if the measured 2 / 1.83 Ohm values are the ESR of the capacitor alone or that it could include also some other resistances.
But it clearly shows that some improvement is possible, so we now have something to compare with O0
Quoteor you could use the the 50:1 step down transformer as referred to in the HP Q meter, also in series LC connection, scope probe would be hooked up across the tuning capacitor.
I am working on this 50:1 step down transformer using my big T520-2 core, so we would have another means to compare results.
Itsu
Hi Itsu,
I think the 2 Ohm or so ESR you got includes that of the coil losses too.
Because the coil's total ESR gives about 0.59 Ohm (core + wire loss) as I calculated from the data sheet in Reply #25, I think if we substract this from the 2 Ohm, we can
receive the approximate ESR of the capacitor, i.e. 2 Ohm - 0.59 Ohm = 1.41 Ohm. So it looks like the variable cap has around 1.4 Ohm ESR at this frequency.
Thanks for your efforts.
Gyula
Quote from: Itsu on 2026.04.29, 09:04:58
So using my nanoVNA on the SERIES LCR circuit ...
Measurement #2 (https://www.overunityresearch.com/index.php?topic=4927.msg118809#msg118809) in shunt-through mode ?
Quote from: Verpies on 2026.04.29, 11:48:55
Measurement #2 (https://www.overunityresearch.com/index.php?topic=4927.msg118809#msg118809) in shunt-through mode ?
Yes, correct
Here i tested the earlier in post #25 measured (Q = 162) 270pF SMD capacitor kemet c1206h271jggact250 which suppose to be low ESR:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56000;image)
As the higher Q predicts, it has a lower ESR than the variable air capacitor (its Q was measured to be 132.5).
-29.57dB at series resonance (2222kHz) with an ESR according to the chart of 0.85 Ohm.
Itsu
I finished my 50:1 T520-2 toroid with 0.8mm diameter magnet wire and which measures 49uH : 0.66uH and a DCR of 0.198 Ohm : 0.031 Ohm.
I hooked up my FG to the 50 turn primary and the 1 turn secondary to the series T200-2 40 turn coil and my 270pF variable air capacitor.
I hooked up the scope across the variable air capacitor.
The frequency was set to resonance which was 2223kHz.
The result can be seen here:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56004;image)
Blue is the FG input signal acros the 50 turn primary
yellow is the signal across the 270pF variable air capacitor.
Using the Fres / bandwidth (-3dB) method for calculating the Q i get 2223 / 10.4 = 213.7
Itsu
Itsu, did you have a 50 Ohm resistor across the primary coil? I assume you had. 8)
Thanks!
Gyula
Quote from: gyula on 2026.04.29, 21:00:54
Itsu, did you have a 50 Ohm resistor across the primary coil? I assume you had. 8)
If the cable connecting the FG to the primary winding was 10cm long then the 50Ω resistor was 10cm away from the primary winding, because the FG's output is terminated with an internal 50Ω resistor.
Quote from: Verpies on 2026.04.29, 22:02:25
If the cable connecting the FG to the primary winding was 10cm long then the 50Ω resistor was 10cm away from the primary winding, because the FG's output is terminated with an internal 50Ω resistor.
Hi Verpies, yes but both in the HP4342A and in the W7ZOI paper the 50 turn primary coil has a 50 Ohm termination, establishing a 25 Ohm impedance across the primary coil. The turns ratio is 50:1 and the square of 50 is 2500, so the 25 Ohm impedance created across the primary coil (by the 50 Ohm resistor + the generator 50 Ohm output impedance in parallel) create 25 Ohm / 2500 = 0.01 Ohm secondary output impedance. They mentioned 0.001 Ohm output i.e. 1 milliOhm for the secondary coil output impedance but IMHO it is a typo.
What is your understanding on this? I attached the Figure from the W7ZOI paper.
Gyula
Quote from: gyula on 2026.04.29, 21:00:54
Itsu, did you have a 50 Ohm resistor across the primary coil? I assume you had. 8)
Thanks!
Gyula
Gyula,
no i did not have that, as i also expected this to be the FG input resistor.
But i added one across the primary which changed some things in the created voltages, see screenshot (FG input was set to 5V sine wave p2p in both cases):
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56019;image)
Blue trace again the input signal across the primary (and 50 Ohm), which is now higher.
Yellow trace is across the variable air capacitor set to 270pF still, at resonance (almost the same) we here now have a lower voltage.
But the Q seems almost the same with 215.9 using Fres / bandwidth (-3dB) method (2224 / 10.3).
Itsu
Hi Itsu,
Thanks. As I see this, the additional 50 Ohm across the primary coil made the impedance across the primary coil to be 25 Ohm, the nearly 2.5 Vpp you measured across it means that because the FG output set to 5 Vpp is halved.
Without the additional 50 Ohm across the primary coil, the stepped down FG 50 Ohm output impedance should appear as 50 Ohm / 2500 = 0.02 Ohm in series with the series LC circuit, while with the added 50 Ohm the output impedance of the transformer
should appear as 0.01 Ohm.
From the present ESR values of the L and C components points of view this additional 0.02 or 0.01 Ohm means a quasi negligible additional loss this measurement method represents, this explains why there is small difference between the Q values 'without' and 'with' the additional 50 Ohm. Now it seems the ESR of the capacitor ruins the operating Q which would be over 400 with this toroid core.
Would you mind checking the stepped down voltage amplitude across the secondary coil? Because the turns ratio is 50 to 1, the nearly 2.5 Vpp from the primary coil should appear across the secondary coil as nearly 50 mVpp.
The secondary coil voltage is labelled as "e" in W7ZOI's schematic. The ratio of this "e" voltage to the voltage across the capacitor labelled as "E", so (E / e) should give the same or very close Q value to that of the 3 dB method.
You can increase the FG output to a higher value so that the stepped down secondary voltage should be higher than 50 mVpp.
Gyula
Gyula,
thanks for the information.
Without changing anything else since the above screenshot, i added a 3rd probe across the 1 turn secondary, see purple trace:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56021;image)
It is around 200mVpp, so much higher than expected.
If i increase the input voltage from the FG to 20Vpp i get the following result:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56023;image)
Not much better.
Perhaps i need to put things on a ground plane PCB, so all grounds are at the same point (now grounds are mostly via the scope probes).
Itsu
Thanks. Well, the total circuit inside the HP4342A Q meter includes additional structural solutions which are understandably not present in your setup.
We can see photos from the original HP4342A meter in this HP Journal, see it from page 10 https://hparchive.com/Journals/HPJ-1970-09.pdf (https://hparchive.com/Journals/HPJ-1970-09.pdf) In page 14, Figure 5 shows the tuning capacitor with the connection posts on top to which the unknown coils are to be hooked up.
Anyway, it is good the 3 dB method with this setup i.e. feeding the series LC via a very low source impedance reveals much better LC circuit Q than the previous ones you have tested.
Gyula
Quote from: Itsu on 2026.04.27, 08:31:46
I had some 270pF SMD caps: kemet c1206h271jggact250 which suppose to be low ESR,
Itsu
you can get a better ESR by putting low ESR caps in parallel.
Soundstream amplifiers claimed an ultra low ESR cap bank at the input of their amps. the caps were already labeled low ESR.
mags
Thanks Mags, i have some "47pF Murata ERB series capacitors" on order that suppose to "have 40mΩ ESR at 3MHz" and which i want to parallel 6 of them to get to around 300pF.
Itsu
Instead of the variable air capacitor set to 270pF, i used some very small (0402) :D smd capacitors of 47pF and put 6 of them parallel measuring 285pF, and they suppose to be very low ESR.
So i tune for resonance frequency using the FG which turns out to be at 2126.3kHz.
FG input is again 5Vpp.
The results are here:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56029;image)
So we have in:
blue; input across the 50 Ohm resistor
purple; the signal across the 1 turn loop
yellow the signal across the fixed 285pF capacitor.
Using the Fres / bandwidth method i have for a Q: 2126.3 / 7.3 = 291.2
Here a picture of the cleaned up (ground plane used) setup used:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56033;image)
Itsu
Itsu, very good result.
Thanks,
Gyula
Quote from: Itsu on 2026.04.30, 20:23:51
Here a picture of the cleaned up (ground plane used) setup used:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56033;image)
I can't see the annotated single loop on that photo.
If I can't see it then it is too small.
Consider using a very thick wire for it. Like the pencil-thick speaker wire... or a slice of copper pipe. You will cross 300 then.
The single loop is of the same 0.8mm diameter magnet wire as the primary 50 turns coil.
But i see it could be better to use a thicker wire there as we are dealing with very low impedances, so i will see what i can do.
Itsu
Yes, a thicker secondary wire used for the secondary single turn coil of the matching transformer will reduce the present loss in the 0.8 mm piece of wire. The improvement will be small though, I think.
In the present setup the ESR of the tuning capacitor represents the highest loss, after the coil. When the new capacitors arrive, hopefully they will have lower ESR than the present cap assembly that establishes the Q=291 now.
Gyula
Gyula,
to avoid any confusion, the 6 parallel 47pF SMD capacitors i am using now instead of the variable air capacitor are the newly ordered ones.
But they are not the " Murata ERB series capacitors" verpies had recommended as they seem to be obsolete, and instead i was directed to this substitute: GJM1555C1H470FB01D which are also 47pF smd (very small 0402 type) capacitors with LOW ESR.
No indication on how low the ESR is however, but better than the variable air capacitor.
Itsu
Did the low-capacitance low-voltage TVS diodes (like SP0402B-ULC) and low COSS MOSFET like the EPC2037 arrive, too ?
Yes, they did, but they are unhumanly small :D especially for the elderly :o
I am working on a small!! PCB in Kicad to be able to have the correct footprints to put those fleas on:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56046;image)
Itsu
Hi Itsu,
Thanks for the clarification on the ordered capacitors, I missed that you got it.
Unfortunately, the GJM1555C1H470FB01D type has no specs under 100 MHz, only for higher than that frequencies.
Gyula
Quote from: Itsu on 2026.05.01, 20:44:18
Gyula,
to avoid any confusion, the 6 parallel 47pF SMD capacitors i am using now instead of the variable air capacitor are the newly ordered ones.
But they are not the " Murata ERB series capacitors" verpies had recommended as they seem to be obsolete, and instead i was directed to this substitute: GJM1555C1H470FB01D which are also 47pF smd (very small 0402 type) capacitors with LOW ESR.
No indication on how low the ESR is however, but better than the variable air capacitor.
Itsu
Quote from: Itsu on 2026.05.02, 07:51:22
Yes, they did, but they are unhumanly small :D especially for the elderly :o
Of course. Just like with any capacitor construction - the smaller the surface area, the smaller the capacitance.
The permittivity of the material between these surfaces matters, too. Vacuum/Air is the smallest, FR4 is larger.
Quote from: Itsu on 2026.05.02, 07:51:22
I am working on a small!! PCB in Kicad to be able to have the correct footprints to put those fleas on:
The placement of components wrt to the LC tank connector matters, ...the inter-trace capacitance and trace lengths, matter too.
P.S.
I just thought that if you are going to manufacture a PCB, you might as well add a comparator/zero-crossing detector like the TLV3202 (https://www.ti.com/lit/ds/symlink/tlv3201.pdf) and some jelly-bean ICs to make a nifty Q-meter with a digital readout.
Once the ringdown waveform (green trace) is converted to a rectangular waveform (blue trace), the digital pulses can be simply counted by BCD counters like the 74HC4518 (https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=56053) and displayed by latched 7-segment LED drivers like the NE589 (https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=56050) or 74HC4511 (https://www.ti.com/lit/ds/symlink/cd74hc4511.pdf).
At 4% hysteresis, the pulse count equals the Q.
Quote from: Verpies on 2026.04.30, 22:16:38
I can't see the annotated single loop on that photo.
If I can't see it then it is too small.
Consider using a very thick wire for it. Like the pencil-thick speaker wire... or a slice of copper pipe. You will cross 300 then.
Instead of the 0.8mm diameter single loop secondary, i now used as this single loop secondary a piece of 5mm diameter copper tubing, see picture:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56056;image)
The signals using the voltage probes are like this:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56054;image)
Yellow is the output signal across the 284pF capacitor (6x 47pF smd parallel)
Blue is the input signal across the 50 Ohm input resistor (FG set to 5Vpp)
purple is the signal across the single loop secondary.
Fres was 2141kHz at 15.3Vpp, -3dB points (10.8V) are at 2144.1 and 2137.6 difference is 6.5, so Q is Fres / difference = 2141 / 6.5 = 329.3
Itsu
Quote from: Verpies on 2026.05.02, 11:12:38
Of course. Just like with any capacitor construction - the smaller the surface area, the smaller the capacitance.
The permittivity of the material between these surfaces matters, too. Vacuum/Air is the smallest, FR4 is larger.
The placement of components wrt to the LC tank connector matters, ...the inter-trace capacitance and trace lengths, matter too.
P.S.
I just thought that if you are going to manufacture a PCB, you might as well add a comparator/zero-crossing detector like the TLV3202 (https://www.ti.com/lit/ds/symlink/tlv3201.pdf) and some jelly-bean ICs to make a nifty Q-meter with a digital readout.
Once the ringdown waveform (green trace) is converted to a rectangular waveform (blue trace), the digital pulses can be simply counted by BCD counters like the 74HC4518 (https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=56053) and displayed by latched 7-segment LED drivers like the NE589 (https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=56050) or 74HC4511 (https://www.ti.com/lit/ds/symlink/cd74hc4511.pdf).
At 4% hysteresis, the pulse count equals the Q.
Well, such a nifty Q-meter with a digital readout would be a nice thing to have if one is in the habit of measuring the Q of coils often (which i do lately, but i can't remember doing it earlier).
But even then, it will make a nice project to work on, so let me see what jelly-bean parts i have available here.
You said: 'At 4% hysteresis, the pulse count equals the Q.", so does that mean that in my present configuration, the coil will produce a ringing signal consisting of 329 pulses?
Itsu
Hi Itsu, very good Q result, the single turn copper tubing together with the shorter wiring and the direct soldering of one end of the tubing to the PCB increased the Q from 291 to 329.
The T200-2 core still have inherent reverse Q of up to 425 at this frequency but to utilize it a tuning cap of even less ESR would be needed than the one now. Of course, only in case if a Q > 400 is needed.
Gyula
Quote from: Itsu on 2026.05.02, 19:43:21
Instead of the 0.8mm diameter single loop secondary, i now used as this single loop secondary a piece of 5mm diameter copper tubing, see picture:
The signals using the voltage probes are like this:
Yellow is the output signal across the 284pF capacitor (6x 47pF smd parallel)
Blue is the input signal across the 50 Ohm input resistor (FG set to 5Vpp)
purple is the signal across the single loop secondary.
Fres was 2141kHz at 15.3Vpp, -3dB points (10.8V) are at 2144.1 and 2137.6 difference is 6.5, so Q is Fres / difference = 2141 / 6.5 = 329.3
Itsu
Quote from: Itsu on 2026.05.02, 19:57:25
You said: 'At 4% hysteresis, the pulse count equals the Q.", so does that mean that in my present configuration, the coil will produce a ringing signal consisting of 329 pulses?
Yes, down to 4% of the initial amplitude.
The only thing that can dampen that parallel LC ringdown is poor connections, thin wires and the active feedback circuit inside of your contactless current probe (or the ferrite of the probe itself but this ferrite should be very high Q because your probe is sensitive down to mA and has 70MHz bandwidth, AFAIR) .
Quote from: Itsu on 2026.05.02, 19:57:25
But even then, it will make a nice project to work on, so let me see what jelly-bean parts i have available here.
It's a rewarding project that doesn't cost much. Good for times when you find yourself idle, e.g.: waiting for the center post of that exotic core.
Quote from: Itsu on 2026.05.02, 19:43:21
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56056;image)
That's more like it. I can see the secondary now.
Do you get higher Q with the secondary connected to the ground plane ...or without that connection ?
Quote from: Itsu on 2026.05.02, 19:43:21
Fres was 2141kHz at 15.3Vpp, -3dB points (10.8V) are at 2144.1 and 2137.6 difference is 6.5, so Q is Fres / difference = 2141 / 6.5 = 329.3
You crossed 300 !
Quote from: Verpies on 2026.05.02, 21:51:46
That's more like it. I can see the secondary now.
Do you get higher Q with the secondary connected to the ground plane ...or without that connection ?
You crossed 300 !
I did not measure the Q without the secondary connected to the ground plane.
It was quite hard to solder that 5mm copper tubing to the ground plane so i would rather not try to remove it again.
Itsu
I went back to verpies his ringdown circuit here (https://www.overunityresearch.com/index.php?topic=4949.msg118990#msg118990) and left everything the same except for the parallel capacitor of the DUT.
Instead of the High ESR ceramic cap i now used the 6x 47pF parallel low ESR caps (284pF).
I expected that the Q will increase from the initially measured 19 to the around 300 which i reached with Gyula his method.
This is the result:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56074;image)
The time between pulses was measured to be 464ns, so with 300 pulses we would have a ringdown time of 139.2us which is about the time between the 2 vertical cursors.
So now we can see how that looks like and i wonder if the pulses still could be measured by using a comparator.
Itsu
Quote from: Itsu on 2026.05.03, 12:56:03
So now we can see how that looks like and i wonder if the pulses still could be measured by using a comparator.
Zoom around the right cursor and see.
Quote from: Itsu on 2026.05.03, 12:56:03
I expected that the Q will increase from the initially measured 19 to the around 300 which i reached with Gyula his method.
Without the upgraded diodes and low C
OSS MOSFET I don't expect it to reach 300.
When you upgrade them - do not pulse them for a long time. Terminate the high pulse before the drain current gets up to 2A. Keep the PRF low too, e.g. 10Hz.
Quote from: Itsu on 2026.05.03, 10:06:05
I did not measure the Q without the secondary connected to the ground plane.
Too bad.
Are you familiar with the performance of symmetric vs. asymmetric cables for e.g.: microphones ?
Quote from: Verpies on 2026.05.03, 13:39:26
Zoom around the right cursor and see.
Without the upgraded diodes and low COSS MOSFET I don't expect it to reach 300.
When you upgrade them - do not pulse them for a long time. Terminate the high pulse before the drain current gets up to 2A.
Zoomed in around the right cursor 139.2us:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56076;image)
The horizontal A cursor is on the middle line, the B cursor at the top peak pointing at 3.4ma amplitude which is about 4% of the starting amplitude of 84mA.
Pulses look OK to me to be measured by a comparator.
OK about the pulse length, so start with a low duty cycle.
Quote from: Verpies on 2026.05.03, 13:53:22
Too bad.
Are you familiar with the performance of symmetric vs. asymmetric cables for e.g.: microphones ?
Sorry, edited your post instead of answering :D
No not really, is there a relationship with the secondary being grounded versus floating?
Quote from: Itsu on 2026.05.03, 14:13:13
Sorry, edited your post instead of answering :D
I know, it can sometimes happen accidentally. Just pay attention to the yellow window outline - its presence means that you are editing someone else's post.
Quote from: Itsu on 2026.05.03, 14:13:13
No not really, is there a relationship with the secondary being grounded versus floating?
Yes, just like with symmetric vs. asymmetric microphone cables.
It has to do with flapping the ground plane wrt Earth ground.
Because of its large surface area, the ground plane is one plate of a large capacitor.
Ideal capacitors don't dissipate energy but ESR, radiation and lossy dielectrics, do.
i agree with verpies. that ground plain probably introduces capacitance issues, considering the small value of intended caps involved. i see the spacer between the large core and the copper pcb, but again, how ever small the capacitance there is between the 2 could alter calculations product.
mags
OK, but as i followed this design of the HP4342A, which has the single turn firmly attached to ground, i also solidly grounded it:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56082;image)
Itsu
Quote from: Itsu on 2026.05.04, 15:05:47
OK, but as i followed this design of the HP4342A, which has the single turn firmly attached to ground, i also solidly grounded it:
This schematic already has one error, namely the 0.001Ω vs 0.01Ω impedance miscalculation.
It is conceivable that the grounding is erroneously or ambiguously marked on that schematic, too ...and in reality is as depicted on the attached schematic.
Also, note that an autotransformer would have be sufficient for the impedance transformation if galvanic isolation between the primary and secondary winding was not planned.
Hi Itsu and all,
In the meantime I have found further useful information on the issues mentioned.
The 0.001 Ohm versus the 0.01 Ohm umbiguity came from my reverse calculation, I did not think the source impedance labeled as Z0 could be much less than 50 Ohm or even much less than 25 Ohm,
the latter came logically from the 50:1 turns ratio that insures an impedance transformation of 2500 from the supposed 25 Ohm but not to 1/1000, only to 1/100.
So it is not 25 Ohm but around in the range from 0.5 Ohm to 1.5 Ohm, this now comes from
first: the schematic diagram found in the online user manual of HP4342A, see the 1st attached picture.
and second: it comes from a very thorough series of measurements on the HP4342A Q meter described in this paper:
http://www.ve2azx.net/technical/HP4342A_Q%20Meter_Tests1.pdf (http://www.ve2azx.net/technical/HP4342A_Q%20Meter_Tests1.pdf) This paper needs careful reading because it is full of data and includes measurements also on replicated 50:1 transformers.
So the impedance converter circuit seems to be an emitter follower having an output impedance in the range from 0.5 Ohm to 1.5 Ohm impedance within 20 kHz and 70 MHz,
this drives the primary coil of the 50:1 transformer which steps it down to around 1 milliOhm.
Yes, Math wise the 1 Ohm divided by 2500 gives 0.4 milliOhm, of course but as it turns out from VE2AZX measurements, this small value is little higher in the practical implementation (mainly due the estimated K=0.8 coupling factor between the primary and the secondary of the 50:1 transformer) and gradually increases in the function of frequency.
Taken from his text, it is 1.8 milliOhms at 100 kHz, 14 milliOhms at 10 MHz, 130 milliOhms at 50 Mhz and 228 milliOhms at 70 MHz.
I have also found a youtube video which shows how the 50:1 transformer is built, see it here: https://www.youtube.com/watch?v=tqjJgGce05Y (https://www.youtube.com/watch?v=tqjJgGce05Y) and from video time around 8 minutes he starts disassembling the structure which directly encloses the toroid core.
I attached a screenshot of the 50:1 turns ratio transformer.
Probably there remains some further questions that are not clear yet. For instance in the tuning capacitor assembly circuit there is a 75 Ohm resistor shown across the primary coil of the transformer: most likely it trimms, compensates the impedance favorably on the primary side.
Gyula
I like the 1 turn secondary that spans the ENTIRE circumference of the toroidal core.
I was going to suggest something like this but decided not to so as to avoid nitpicking too much.
Gyula,
thanks for this, it shows how an amazing instrument this HP4342A really is, nothing to compare with my setup.
Itsu
Itsu, no worry, your setup is able to measure Q values in the range of 300 in the 2-3 MHz range and this Q included the higher than desirable ESR of the tuning cap, And with a HP4342A you could measure a Q of around 410 or so the T200-2 core is capable of. 8)
And with the inverter circuit (emitter follower) driving the primary coil the performance can be improved (in case the need arises, now it is not really needed).
Gyula
Of the three schematics above one shows 2 separate types of ground for the left and the right of the total circuit. usually meaning they are not grounded together. where one side may be grounded to earth ground and the other to maybe an isolated chassis ground, or just negative reference, etc. so the top original circuit may not be incorrect.
for me the copper PCB under the large toroid could cause issues, again, the tiny SMD caps value needs to be THEE capacitance values for the test to give an accurate result as possible, and that copper plane may skew results by adding capacitance (small but maybe large enough to be a problem in reference to the SMD caps value) without realizing it.
mags
Quote from: gyula on 2026.05.04, 20:04:00
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4949.0;attach=56090)
I think I see feed-through caps next to the toroid.
Mags:
from the HP4342A manual it turns out the main internal circuits are fed from a PS which gives +25V and -25V DC with respect to the common ground.
using split DC supplies for RF and HF circuits is common, helps unwanted interactions between circuit stages.
regarding the copper PCB area under the large toroid: I see what you mean but consider the HP toroid core enclosure revealed in the teardown video,
they used teflon insulation spacers directly under and above the core and the dielectric constant of teflon is 2, i.e. twice that of the air, meaning
the stray capacitance added to the measuring circuit was quasi doubled, yet they used it. The spacer Itsu applied between the large core and the PCB surface
should contribute much less unwanted coupling capacitance I think.
see pages 25 and 26 for replicated transformers and measurements in this paper: http://www.ve2azx.net/technical/HP4342A_Q%20Meter_Tests1.pdf (http://www.ve2azx.net/technical/HP4342A_Q%20Meter_Tests1.pdf)
he used 65 mil plastic spacers at bottom and top
Verpies: Yes, the feed-through cap is included in the schematic too, labelled as C4, 56 pF, see the attached schema in Reply 70, previous page.
Gyula
Quote from: Verpies on 2026.05.03, 13:39:26
Zoom around the right cursor and see.
Without the upgraded diodes and low COSS MOSFET I don't expect it to reach 300.
When you upgrade them - do not pulse them for a long time. Terminate the high pulse before the drain current gets up to 2A. Keep the PRF low too, e.g. 10Hz.
OK, i have the recommended parts put together on a PCB and the schematic looks like this:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56166;image)
The first result with a 5V DC 1kHz 10% duty cycle input pulse looks like this:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56168;image)
At the top the pulse with ring down, and at the bottom the zoomed in part of the start of the ring down.
So i need to count the ring down pulses to acquire the Q
Itsu
Quote from: Itsu on 2026.05.11, 10:52:51
So i need to count the ring down pulses to acquire the Q
Yes. If you want to do it automatically, the first step is to convert the ringing to digital pulses with a fast comparatpr like the TLV3202 (https://www.ti.com/lit/ds/symlink/tlv3201.pdf).
Also the exciting pulse can be automatically determined by a second comparator that turns off the MOSFET as soon as it drain voltage or current reaches max value.
Right, so i need to have a current transformer of some sort where now the current probe is to pick up the ringing signal to be processed by the TLV3202.
I have some LTS15-NP (perhaps to sturdy) which i could try for this.
Quote from: Itsu on 2026.05.11, 14:24:52
Right, so i need to have a current transformer of some sort where now the current probe is to pick up the ringing signal to be processed by the TLV3202.
Yes, so you could wind as many non-touching turns with a very thin wire as possible on a toroidal core and thread the parallel LC tank's wire inside it to see if your scope's voltage probe can pick up the ringing current.
Well, the problem with that is that the toroidal core (i tried several small and medium cores) hugely impact the number of rings thus the Q of the parallel LC tank circuit (like a ferrite bead would to avoid oscillations).
There are high-loss ferrite materials such as the ones used in ferrite beads and there are low-loss ferrite materials such as the core used in your current probe. Smudge should chime in here...
If you don't have a low loss core then we can think of more exotic ways of measuring the current. This method does not have to be very linear because the goal is only to count the pulses so you can use such things as Rogowski or Hall sensors or a TMR or GMR read heads (w/ preamplifier) from a broken hard-drive (circa 2006 or later).
There are also non-magnetic Ohmic methods of measuring µV at two different points of the copper wire that connects the p-LC tank.
Finally, there are also capacitive methods that sense the voltage across the LC tank's capacitor, but they worsen the low <1pF isolation capacitances that the TVS diodes provide when the MOSFET turns off.
Ok, i had an defective P6302 current probe which i toke the ferrite core off which sits in the clamp part, so i think it should be the correct low loss core:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56170;image)
I used this core the first, and it gave me the reduced ringing, then some other cores seems worse.
Quote from: Itsu on 2026.05.11, 20:48:51
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56170;image)
I used this core the first, and it gave me the reduced ringing, then some other cores seems worse.
I did not have this core in mind.
I had the one in mind that the DUT wire is threaded through.
Yes, i was meaning to say that i used this core to thread the DUT wire through (next to the current probe).
How does the resistance of this (https://www.allegromicro.com/-/media/files/datasheets/acs37100-datasheet.pdf?sc_lang=en) non-inductive current sensor compare with the DC resistance of your LC tank (ESR of the cap + DCR of the coil) ?
To measure the DC resistance of the capacitor and coil i connected them in series LC configuration so that at resonance they should cancel out their reactances.
I put my nanoVNA in "series measurement" to measure this expected low impedance.
I set the sweep frequencies as close to the resonance point as possible.
This is the result:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56200;image)
It seems to me that the total ESR and DCR resistances at resonance is 0.084 Ohm (84 mOhm) as can be seen in the red circle.
I checked the DCR of the coil to be 0.040 Ohm using an Ohm meter so that would make the ESR of the capacitor in the 0.044 Ohm range.
Itsu
Hi Itsu and all,
I still think that the best for estimating the Q for the coil wound onto the T200-2 toroid core is to consider its data sheet info, see here:
https://www.overunityresearch.com/index.php?topic=4949.msg119037#msg119037 (https://www.overunityresearch.com/index.php?topic=4949.msg119037#msg119037) It says a Q of 425 at and around 2 MHz for a 20 uH coil on this core from #18 (1mm) wire.
The 20 uH has XL = 251 Ohm reactance at 2 MHz, so the measured coil in the data sheet had a total loss resistance of XL / Q i.e. about 251 / 425 = 0.59 Ohm.
Here we have to assume the tuning capacitor in their test (if they resonated their coil for the measurements) should have had a Q = > 10,000 or so.
Your coil has about 19.3 uH, the inductive reactance is 262 Ohm at 2157.35 kHz. So the Q of your coil, assuming the same total loss of 0.59 Ohm should be 262 / 0.59 = 444 maximum, assuming the tuning capacitors have a Q = > 10,000 this involves an ESR of 0.0262 Ohm.
If we use the DC resistance of your coil as 0.04 Ohm, the coil Q appears to be 262 / 0.04 = 6550. This does not seem correct, too high.
Data sheet would have included such a high Q. We need to figure out why you got such a low loss resistance.
Comments are welcome.
Thanks,
Gyula
Gyula,
i have put the T200-2 core / coil to the test by using the nanoVNA, so JUST the core with coil.
Sweeping from 100kHz to 4MHz and looking at:
S11 phase (i understand that for a coil to test we need to check it at its +90 degree point),
S11 series inductance,
S11 series resistance and reactance.
The results are here:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56202;image)
So at 90 degree phase (which is at 429kHz) noted by the red marker 1 we have an induction of 18.483uH and a series R of 1.367 Ohm, see in the green circle on the right.
The reactance of this 18.483uH at 2MHz is 253 Ohm as can be seen by the green circle bottom middle and the line pointing to the data.
So there are some differences with the earlier measured resistances both by the Ohm meter and the nanoVNA when using the LC circuit (coil with capacitor).
Next i will try to measure in this same way the used parallel capacitor (284pF).
Itsu
Okay Itsu, thanks, will comment tomorrow.
The problem i have is with the resistance (1.367 Ohm).
This is just a 2 meter or so 1mm diameter magnet wire, which should not have such a high resistance.
My Fluke Ohmmeter shows 0.1 Ohm with short (5cm) leads, so it must be in that range.
Itsu
I zoomed in on the lower graph (series R) and it shows that the resistance starts low as expected, but then quickly rises to 2.8 or so Ohms around 1MHz and stays there.
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56217;image)
I will run some further tests.
Itsu
Hi Itsu,
The resistance is one issue, the other one is reactance: if we accept the inductance of 18.483 uH or around that value, this is close to the 19.3 uH the L meter showed at 100 kHz,
then its inductive reactance at 2 MHz should be around 232 Ohm (instead of the 253 Ohm the nanoVNA shows).
Gyula
Gyula,
I carefully recalibrated the nanoVNA and made the following measurements (shunt measurement):
Frequency 2MHz
Coil
S11 series L 18.8uH
S11 series X 238 Ohm
Capacitor
S11 series C 287pF
S11 series X -275 Ohm
The DC resistance (DCR) of the coil needs to be measured using the Shunt-thru measurement, but gives nonsensible results (3.3 Ohm).
But the Agilent U1733C LCR meter shows at 100Hz an R of 40mOhm, which agrees with the DC resistance specification of an AWG 18 (1mm diameter) of 21 Ohm/km, so (40 turns of 5cm = 2m) 2m = 42mOhm.
The ESR of the capacitor was measured using the method showed here: https://www.youtube.com/watch?v=eytQh_ucDxw and turns out to be 49.4mOhm (-54.1dB using the table / graph mentioned in the video description).
Itsu
Hi Itsu,
These results sound much better I think, though there is still one thing to discuss.
I think we have to consider the T200-2 total core loss which can be reverse calculated from the Amidon data sheet as about 0.59 Ohm, see my reply #88 above.
This 0.59 Ohm by Amidon measurements should already include the DC resistance too and because you use the same #18 wire, we can say with good certainty the total loss of your coil is 0.59 Ohm at 2 MHz.
If we add to this the ESR of your capacitors, we get 0.59 + 0.0494 = 0.6394 Ohm. So the Q of this series LC (18.8 uH, 287 pF) circuit at 2.1667 MHz where they resonate, would be 255.939 Ohm / 0.6394 Ohm = 400.2
This is very close to the data sheet Q value. (The 255.939 Ohm comes from either the XL or XC reactance at the resonant frequency 2.1667 MHz).
Thanks for your kind efforts and hopefully we can find a measuring setup with which this Q of around or above 400 can be measured for this series LC circuit, the very low, 49.4 mOhm ESR of your capacitors makes this possible too.
Gyula
Gyula,
OK, so besides the DCR of the coil of 42 mOhm, we also have the core losses totaling 590 mOhm.
I removed the ground plane underneath the 50:1 coil, and thus isolated the single secondary loop from it.
I connected the T200-2 core/coil directly to the 287pF capacitor in series and to this single secondary.
Resonance is at 2143.1kHz (3Vpp input from FG) and Vpp across this single loop is 135mVpp with a Vpp across the 287pF capacitor of 20.1Vpp.
Taking this for Q measurement i got 20.1Vpp / 0.135Vpp = 148.8
IF i use the -3dB bandwidth method, i get a Q of 2143.1 / 6.5 = 329.7
But this is with the voltage probes attached, which might have some influence on the measurement.
Itsu
Well, the voltage probe adds its own 5 pF or so self capacitance to the 287 pF capacitors, lowering resonant frequency from 2.1667 MHz to 2.1431 MHz (this comes with about 293.35 pF) and also adds its input resistance in parallel with the 287 pF, which is certainly less than 10 MegaOhm, and a certain loss is also introduced by the 1 turn secondary coil feeding the series LC circuit.
These two losses (introduced) could explain, more or less, the Q of around 330 instead of the around 400 value with the -3 dB measurement method
Why the ratio of the secondary voltage and the capacitor voltage brings a much lower Q of 148.8 value (i.e. the HP Q-meter measuring principle) remains to be solved. We will figure it out. 8)
Gyula
Hi Itsu and all,
I noticed an interesting lossless transformator model among the LTspice Example files and decided to use it for simulating this setup.
The received Q was 349.2 which is close enough to your measured 329.7 value. I used the same component values, loss resistances that you measured with the nanoVNA.
I included the probe impedance across the 287 pF tuning capacitor but to get the 20.1 Vpp across it I had to use a 680 kOhm as the probe's resistor.
This reduces the simulated unloaded Q of 398 (no probe) to 349.2 mentioned above. With no any probe resistor the e2 voltage across the C1 capacitor is 22.9 Vpp (8.14 Vrms).
What still remains to be figured out is why this e1 and e2 voltage measuring method (which now works in the simulation) does not work in your setup?
Because the -3 dB Q measuring method gave to you Q=329.7 but the e1 and e2 voltage ratios gave Q=148.8 in the very same setup.
Comments are welcome.
I blown up the voltage scale in the scope display vertically so that the e1 voltage 57.7 mVpp should be seen correctly, this is why the e2 voltage 20.15 Vpp goes out of the display vertically. (Otherwise the e1 voltage would be a flat line.)
Gyula
Gyula,
nice find this lossless transformer model, now we perhaps can find where the difference between theoretical / simulated Q values and my real life Q value comes from.
It seems strange to me that my single loop output (e1 voltage) is higher than your simulated one while the output voltage (e2) is the same.
Itsu
Hi Itsu,
Yes, it is strange and I suspect if we find the cause of the higher single loop output, the e2 / e1 voltage ratio will bring a Q value very close to the -3 dB Q measuring method.
It has to be caused by the 50:1 transformer itself I believe, might be leakage?
Comments are welcome from anyone. 8)
Gyula
Gyula,
i measured the leakage induction of the both primary and secondary to be (Frequency measured is 2143kHz ):
Normal inductance of the primary coil with the secondary coil OPEN is 55uH
Leakage inductance of the primary coil with the secondary coil SHORTED is 45uH
Normal inductance of the secondary single loop coil with primary coil OPEN is 136nH
Leakage inductance of the secondary single loop coil with primary coil SHORTED is 116nH
I will measure the primary / secondary voltages to see if they are at a 50:1 ratio (which they should) and if not, try to find a way to correct this.
Here is the present circuit i work with:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56239;image)
Itsu
When using the above-mentioned normal and leakage inductance of a coil, we can calculate the coupling factor with this calculator: https://e-magnetica.pl/doku.php/calculator/magnetic_coupling_coefficient
This means for the 50:1 toroid primary (55uH versus 45uH) a coupling factor K of 0.42
For the secondary it means (136nH versus 116nH) a coupling factor K of 0.38
So roughly I think we can say we have a 50:1 toroid coupling factor of 0.4
Itsu
Quote from: Itsu on 2026.05.20, 14:15:18
So roughly I think we can say we have a 50:1 toroid coupling factor of 0.4
At the risk of being boring, I will repeat that a winding, which symmetrically spans the entire circumference of a toroidal core, maximizes flux confinement in this core and consequently minimizes flux leakage, mutual magnetic coupling to other coils/cores, EMI generation and susceptibility to external EMI.
A single-turn winding can also span the entire circumference of the core by using a cuppola-style winding.
A multi-turn winding can symmetrically span the entire circumference of the core by existing in even number of reversing layers as shown below (the terminals of this winding should depart and arrive at the same place).
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=55528)
To minimize the parasitic capacitances the turns and layers should be kept apart with as much air between them as possible, i.e. they should not touch (use tiny glass beads as spacers held by CA glue at the wire crossovers. These beads are used for jewelry, bracelets, tassels, earrings ....and are very cheap). Yes, this is tedious but the results are worth it.
(https://i.pinimg.com/originals/ec/b0/41/ecb041a5b0c680d93f7eb512bc82006b.jpg)
(https://i.pinimg.com/originals/da/99/77/da99775ad67e3c051242921d09397d2d.jpg)
Nothing boring here, tedious yes, but not boring.
I cannot find any info on this cuppola-style winding, is that the style of the toroid in your picture?
Anyway, i was thinking of putting the toroid in a copper box in such a way it forms a single turn around the toroid, like in the HP Q meter, but i am not sure what is more tedious.
Itsu
Quote from: Itsu on 2026.05.20, 20:49:47
I cannot find any info on this cuppola-style winding, is that the style of the toroid in your picture?
No, that's for multi-turn winding.
Quote from: Itsu on 2026.05.20, 20:49:47
Anyway, i was thinking of putting the toroid in a copper box in such a way it forms a single turn around the toroid, like in the HP Q meter, but i am not sure what is more tedious.
Yes, something like this.
You can use a big copper pipe cap...
(https://cdn.mscdirect.com/global/images/ProductImages/3689185-24.jpg)
...or a short segment of large ID copper pipe with a flat end plate soldered on one side with a center post that are capable of containing the toroidal core with the multi-turn winding, without touching them. Solder a conical concentrator section on the other side.
The conical section is the most important feature because it symmetrically brings the current to the middle where the copper center post is. The finished cuppola looks like a door knob.
You have made conical sections like that in the past:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3670.0;attach=41651)
Thanks for visualizing it, looks great.
First I need to find a suitable smaller toroid, as the present 50:1 is on a T520-2 (big), and in this case smaller is better.
itsu
Remember that if you do not confine all of the magnetic flux to the toroidal core by a symmetrically reversing a multi-turn winding then the flux that leaks out of the core will induce eddy currents in the bulk metal of the cuppola and waste energy.
This phenomenon is similar to the one occurring in this arrangement:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=56154)
Where the circular turn around the perimeter of the core represents the uncancelled circumferential current of the multi-turn winding and the rectangularish turn represents the path of the induced eddy current in some distant bulk metal (without the break/gap).
The flux leaking out of this core is like that of a permanent ring magnet polarized axially as depicted below:
(https://i.sstatic.net/wjaHzHpY.png)
...and this axial flux changes in time under AC drive, inducing EMF and current in distant closed conductors ... and other coils.
This undesirable induction is easily observed in the following experiment, where two one-layer windings are placed on two separate parallel toroidal cores which are placed coaxially above one another:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=56247)
This arrangement creates an unintended 1:1 transformer.
The 1:1 ratio surprises most people because one core evidently has more turns on it than the other.
This happens because the transformer ratio, in this case, is determined by the circumferential MMF that generates the axial leakage flux - not the usual toroidal MMF (toroidal ampturns).
Ok, so even the 50 turn coil need a special winding technic to minimize losses.
I am not sure if it is all worth the (tedious) efforts for building such a Q meter as knowing the Q of a Coil or LC circuit is nice, but no showstopper IMO.
Meanwhile, i was thinking of using a AM503B current probe amplifier (https://w140.com/tekwiki/wiki/AM503B) as pickup for the ringing signals on your Q ringing method.
It needs a 50 Ohm input (into a scope, but could be used in any 50 Ohm input i guess).
So i build this TLV3202 circuit mentioned on page 15 / fig. 8-5 of the TLV3202 datasheet (https://www.ti.com/lit/ds/symlink/tlv3202.pdf?ts=1779347202823&ref_url=https%253A%252F%252Fwww.google.com%252F)
But it needs some AC volts sine wave to get so decent square wave out of it.
So i think it needs some buffer amplifier in front of it to boost the 20mVpp and less ringing signals coming from the AM503B.
Itsu
Quote from: Itsu on 2026.05.21, 14:35:05
I am not sure if it is all worth the (tedious) efforts for building such a Q meter as knowing the Q of a Coil or LC circuit is nice, but no showstopper IMO.
It's just fun to do and such efficient 1:50 impedance transformer has other uses than for Q measurement.
Quote from: Itsu on 2026.05.21, 14:35:05
Meanwhile, i was thinking of using a AM503B current probe amplifier (https://w140.com/tekwiki/wiki/AM503B) as pickup for the ringing signals on your Q ringing method.
It needs a 50 Ohm input (into a scope, but could be used in any 50 Ohm input i guess).
Yes, it can be used for that purpose.
You can also use a GMR or TMR sensor cannibalized from a 2007+ hard drive.
Quote from: Itsu on 2026.05.21, 14:35:05
So i build this TLV3202 circuit mentioned on page 15 / fig. 8-5 of the TLV3202 datasheet (https://www.ti.com/lit/ds/symlink/tlv3202.pdf?ts=1779347202823&ref_url=https%253A%252F%252Fwww.google.com%252F)
But it needs some AC volts sine wave to get so decent square wave out of it.
Even if you delete the feedback resistor (that sets the hysteresis) and set the V
REF voltage at the (-) input to several mV ?
If the (+) input of the comparator is located very close to the output of the AM503B then you can skip all comparator input termination resistors and the feedback resistor.
I removed the feedback 1M and input termination resistors (330K) and put 0V (at +0.3mV it stops working, no output) at the + terminal.
Input is a minimum 50mVpp AC sine wave (less won't work anymore, no output).
I used my FG for now to input the signal using this circuit (used a different symbol for the TVL3202):
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56249;image)
The output looks like this, very noisy and feedback on the input signal:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56251;image)
Itsu
Quote from: Itsu on 2026.05.21, 19:48:51
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56249;image)
Swap (+) and (-) comparator inputs and put a ceramic capacitor between the (-) input and comparator's ground.
Hi Itsu and all,
In the meantime I did some simulations in LTspice. I have found that the needed 50:1 transformation ratio comes out only as being around 22:1, using the measured primary inductance of 55 uH and single turn secondary of 136 nH on the T520-2 core.
If we use this data sheet https://alexsradioshop.de/wp-content/uploads/2023/01/AmidonAMI.pdf (https://alexsradioshop.de/wp-content/uploads/2023/01/AmidonAMI.pdf) and we use the formulas included in page 1, we can reverse calculate the number of turns for the measured primary and secondary inductances
(using AL = 207 uH/100 turns from data sheet for that core).
So the calculated number of turns is 51.5 for the measured 55 uH primary coil (which sounds reasonably close for your 50 turns, see note below).
However, the calculated number of turns comes out as 2.5 for the 136 nH secondary coil. The formula predicts 20.7 nH inductance for a single turn on this core.
The simulations show a 50:1 ratio if I use 28.1 nH inductance for the secondary coil, see attached screenshot. The 28.1 nH is much closer to the predicted 20.7 nH than the
136 nH. With the 28.1 nH secondary the Q of the LC circuit comes out as very nearly 370 at 2166 kHz.
Checking your 40 turns on the T200-2 core for your measured 18.8 uH inductance by reverse calculation, it gives 39.5 turns, very close to 40. (AL for T200-2 is 120 uH/100 turns.)
For consideration only: known rule of thumb for a 1 cm long wire inductance (OD = 1 mm) is about 10 nH and the T520-2 core has a permeability of 10.
I attach a simulation of the transformer with 28.1 nH secondary coil driving the LC circuit at resonance. For the total loss resistance in the LC circuit I used 0.69 Ohm.
Note: The little higher inductance of 28.1 nH I had to use (versus the reverse calculated 20.7 nH) to get the 50:1 ratio may come from the 50 turn primary coil which gave the 55 uH
primary inductance as you measured, probably a little high. If I use 52 uH instead of 55 uH to reverse calculate the number of turns, I get 50.1 i.e. 50 turns you have.
Using 52 uH in the simulation and using the 20.7 nH inductance for the secondary coil, I got V3 / V2 = Q = 45 Vrms / 0.1218 Vrms = 369.4 i.e. the same Q result, just the voltage levels are lower.
Gyula
I'd like to notice that any secondary leakage inductance of L2 will be added to the DUT's inductance L3.
Quote from: Verpies on 2026.05.21, 21:01:39
Swap (+) and (-) comparator inputs and put a ceramic capacitor between the (-) input and comparator's ground.
(+) and (-) comparator inputs swapped and placed a ceramic capacitor between the (-) input and comparator's ground
I also changed the 220k trimmer pot to a 1M one:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56255;image)
Signals seems slightly better, but minimum input went up to 400mVpp sine wave, also the trimmer pot still needs to be very close to 0V on the - terminal.
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56257;image)
Itsu
Quote from: Verpies on 2026.05.22, 17:08:48
I'd like to notice that any secondary leakage inductance of L2 will be added to the DUT's inductance L3.
Yes. Itsu measured leakage inductances, though the value he got seems to be also too high because the normal inductance of the single turn secondary seems to be too high as per
the simulations and reverse calculations show. So some tens of nanoHenry leakage is involved maximum, and this should be considered of course, especially when DUT inductance
happens to be under say 50 nH. Obviously this was the case in the HP Q-meter circuit, calibration was done anyway.
Gyula
Quote from: Itsu on 2026.05.22, 20:40:03
I also changed the 220k trimmer pot to a 1M one:
It would've been better if you'd changed the 200k pot to a resistor connected to the +5V rail and connected a 10k (or lower) pot to that resistor and the other side to ground while its center leg goes to the (-) input of the comparator.
Is the yellow trace the signal from the FG ?
P.S.
For comparator testing, triangle or sawtooth waveforms are the best.
Quote from: Verpies on 2026.05.22, 21:21:48
It would've been better if you'd changed the 200k pot to a resistor connected to the +5V rail and connected a 10k (or lower) pot to that resistor and the other side to ground while its center leg goes to the (-) input of the comparator.
Is the yellow trace the signal from the FG ?
P.S.
For comparator testing, triangle or sawtooth waveforms are the best.
Ok, did that, fixed 220K resistor from +5V to 5K trimmer pot to ground with its center lead to the (-) terminal.
Need about 10mV on the (-) terminal for the same signals as in the above screenshot, still with a minimum of 400mVpp input from the FG.
With a triangle waveform i need a minimum of 600mVpp input.
Yes, yellow is across the 50 Ohm resistor.
Itsu
Quote from: Itsu on 2026.05.23, 08:08:38
Yes, yellow is across the 50 Ohm resistor.
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=56259)
The interference appearing at the INPUT is unusual and unacceptable !
Adjust your FG to add a DC offset to the input signal so it does not go below ground. AFAIR this comparator does not like when its inputs go below its ground pin potential because this activates the comparator's internal protection diodes and the current from these diodes can mess with the inputs. The comparator could be already damaged by these excursions below ground.
By the same token, try lowering the +5V power rail because your scope indicates 8.120V ringing at the output, which is above the V
CC_max and might also activate the comparator's internal protection diodes. A little more load at the output might also dampen this ringing.
Finally, try to decouple the +5V power rail and the scope probe's ground with various techniques.
e.g. use your recently acquired low-ESR wonder-caps between the comparator's supply pins and between the (-) V
REF pin and comparator's ground pin.
Perhaps add a choke and f.bead at the +5V rail (before the cap), too.
Watch for radiative EMI coupling to you scope probe's ground clip (try using the springy type and use an improvised grounded copper foil barrier or mini-shield box).
If this bears fruit, I will think of a discrete JFET solution for the (+) input amplification and decoupling.
I did lower the supply voltage yesterday to about 2V and the output (and input) signals improved a lot.
I will try your other suggestions too, thanks.
Itsu
I used the following settings on the FG (rigol):
Frequency 2145kHz
Sine wave 200mVpp
+100mV DC offset
The supply voltage was set to 2V
The DC voltage on the (-) terminal was set to 50.7mV.
Signals look like this:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56261;image)
Yellow input across 50 Ohm, blue is output across 100K
So input is somewhat lower, but any lower and i lose the output.
Lost a comparator in the process (its output seems grounded).
Still need to improve on the noise filtering.
I have a strong 169MHz intermittent signal being picked up by the (yellow) scope probe producing signal distortion and coming from somewhere outside (license free ISM band) :(
Itsu
Quote from: Itsu on 2026.05.24, 19:37:41
I have a strong 169MHz intermittent signal being picked up by the (yellow) scope probe producing signal distortion and coming from somewhere outside (license free ISM band) :(
If it is 169.65MHz then it could be the old P2000 Dutch emergency services system.
Otherwise it could be the wireless M-Bus (EN 13757-4) and similar standards used for remote reading of utility meters (electricity, gas, water, heat) and other sensors.
Yes, it is 169.65MHz, the (not old) P2000 Dutch emergency services system.
What kind of JFETs do you have ?
I am thinking of a low capacitance buffer/amplifier like this to drive the comparator:
(https://i.sstatic.net/SfTSR.jpg)
I have some BF245C on hand
What happens when you simulate this (https://falstad.com/circuit/circuitjs.html?ctz=DwYwlgTgBAZgvAIgIwKgFwM6KQUwLQCcqYI2ADAHQBMZZVA7PQMxkAcVAbAQfaiAEaIArGVQAHQQgAsTVADcIw1AFtMw-EhQA+AFBQowAO5QAHtiqsoBDlCpMqVjqngJRAel36jpxHYfWoJiF-J1hEd08DYzMEIQ4yRyh4y2tncIQPPSifBGTE+jIpRzTXDMjgORykGyZWIuqoKVZRMNKoQxcUWEVkWj6W5QBDEzlwijIhMqzgAHMqmyaEhqEkKhKI6ZB5wLqoJHoauvW+X2oNLrBR13GyLowelrkAE0Q8SgJ9+yoLKTJ6KVWZFkmS80WwBx2RRoCVqUmOIIM0Bi+0O9Qhi2OUB6SH6fSmXjmyPRzT2EJWa1aGy8ACsctDIbZ7JCSvZ5GhXlQKLwoPwcGNaEgmAVcSLbvjEXSmb8En5Gs1Mdi8QjZpKHItGQ5yfDymCELLpbYOEVpdrNnSjY1aIbjfLKSdrrjvsQrpQBah7ukoM8xgRWCiLFQeEhWBwmHDlYTfBb1Zwilq7craTF6TRLLLUyyKXJ2Qg8Jzubz+bchaKRShlbr01QoUymBxWKavEjzCkbKs02wFcJ4rRxcBm8gLLY2HshwwKS4WoqleVI4O0-QHO3AoVGwYtjEAqnEuPMaRkPhuZci3cHvIXsgKFICExuBxqqx6H7WGGGxGclvF4kWHCEzqPzY27Lhmf7TLqwEjimnagU2qo7kySAEBOno9LevaJjk9hLEhgTBHsSGZmyHJcqghYOsWwqln05b-smCE4Vh+HIaUFaYXhARBEuBxrv2bEONunFyi0k6oKhLBUgYc6CeqgnxiJyobogBrblIFogSJUD7rgeBHi6Nynp63qXgcUhUEIj4EBMQi+uwb7lAOqlQiOcQytBGnTn2c4uUJUDeXJ6TvjEeTqnk-kseUgw5HkiEOAU9QEa0RBQEMSjJWAnozIM5C4n2uredueTqQF9k5PlznxEJXa9DO0wDmVljeakdpYtl6ElUFI4okkI5Ne5rUSbxHWWD+3WWBizXTm10wmHSUhCIE-yGjUjCYhg6XIKEaB8ggACKNHTbN-FIFCoa2Psq3rdUqBbYgO1rMqM0xHWTD4TCuyIbIzVrdgm3bTtwLlI92C1LY9BIHs5mg7+GnfRt13bQASgDB3IjwezXmd9AQ8xLSw1d6CI5MD1VFa+wEKOSzsBdP3w4gCP3YDUUvZwL27JwoQw5dv2IAAwsjXhAwgIaNPsex-I0NDU3DBO8-tAtVODvyBIrn2czTMsIAAagzKOIBwBSWkklmNKrnp49zWv8wYgvHS9MhtqdMhJWr0s3QgCNy9bdJfqZzM3hLHNm1ztMIDzOvy8mEyNPrthmY0BRS-jbsI+GjNPcGgQY0wIOcYnFsI04xNPWjdjkwCaamXnIc86nuuxHQjI2ME-G3lXGsI7wRfCAwmfk0IUhY7nX3B+3RCsTETSJMuvXFWaE+tnKTF7tgh7Oie7oPBQXRGcdFDcPvB8H-Wnk5JPMWL980OzxHvjg9U82mXsx-D+rbs84Xs45HF6NFN-1TCdfdcX9Cj5BAcdK+bR9yumomvciBlrjbwvEIG4VFCjViJmnPWHA2xx3rFjdsbc36d0wbkOwuEigcBELhbkLsk6I00LlYB8UHB+jbAleStE9adQhIVOIPEBy8JsN-Mygc2jYmhDVUETDbBxCgH-dhgDgCCyNMzEk+t+IvkIfQ8OXsgrHVHBQmQo4aFB1fvQgagtgj4KEC9fu1jRG4xHsnOy0wEbSICN-VkrR8YdE9EMEYShlRuPnokSeXjOihF8W0fxVwkAYLAlFSCsi8g-h4nOesY01GsAaqsHiilcjZJ8hklcECWhaQ0C0Y8cCN6GQvHmK8rBrAs3+HQOsUh2mMKGiUpIhTUkwV0XrX0OxwZg3JrCLRvN4lSInlaASeEirhWmAAeygDgAAdkpd0Yg9YQJMNKWgoRtmlEiF4MQXozZQPGE+VYUz9AGDOVcT6GBLn9B4JQjgZk+xuCWeUFZ6zEDcgwEco0JQ9n9EOekE59zzltGedgagtzTkwqeS8voby4ifOVN835qyNl6kBcC3ZvxoQQuOVkJFLp3T7iYBQGx-dLI8GsrUKgcTa53OhY8qlRYxRYp+cs3FiBbRAp2aC4lBzxCQvJRyi5SkuS-GyUabJLACDmURdKhAKLuWe2ANinQOrwAQF0EAA) in LTspice ?
Not much if i look at the red and light blue traces, they are around 6mVpp:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56267;image)
Itsu
The bias resistors were calculated for the BF245 JFET model. The B grade preferably.
OK, i found a BF245B LTspice model:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56271;image)
Looks better, having 678mVpp at the light blue probe and comparator output O0
Quote from: Itsu on 2026.05.29, 08:16:05
Looks better, having 678mVpp at the light blue probe and comparator output O0
The gain can still be increased with different biasing. Perhaps someone else could help to calculate it ...?
e.g.: R5→220Ω to 470Ω, R3→2kΩ to 5kΩ or something in-between. C2→1µF to accommodate lower cascode frequencies (kHz). When adjusting these values, have LTspice make a plot of the voltage gain vs. frequency (over the 1kHz - 10MHz span) measured at J1.drain.
Additionally, I think that R4 can be increased to 100MΩ and C4 can be eliminated or substituted with a protective resistor. The C7 should be a high-ESR cap (yes - unusual) and can change R12→10MΩ. Also, it would be beneficial to swap the inputs of the comparator and apply the signal coming from R7 and the R12 hysteresis feedback to the noninverting input so C5 only stabilizes the reference voltage at the inverting input and doesn't shunt the hysteresis feedback to ground (with this change, a low-ESR C7 doesn't hurt anything ...and even hepls)
Some low-C clamping diodes at the junction of R7,R8,R9 might be needed if the voltage at this node gets too high at high input signal levels.
If the comparator's kickback still backpropagates to the input in a real circuit then increase R7 (the one thing LTspice cannot simulate).
Also, to keep the input impedance high, the J1 and J2 and surrounding resistors/caps, should be connected in the air using the "dead bug" style with short legs.
Generally, the amplifier still needs design work.
P.S.
The BF245 transistors have a wide parameter spread due to high tolerances of the old manufacturing process. The C version - the highest spread. The A version - the lowest spread, but it is a lottery with all of them (just the odds differ).
The gain of the JFET amplifier should be judged by the amplitude at the red probe (J1.drain) because J3 is just a source follower for the comparator kickback isolation.
OK, i here first made a basic plot of the voltage gain (drain J1 (node n003) / top R4 (node n010) over a 1kHz to 10MHz range with a highlight on our used frequency of 2143kHz:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56275;image)
So IMO we presently have a 67.95dB gain at 2143kHz.
Quote from: Itsu on 2026.05.29, 19:28:10
OK, i here first made a basic plot of the voltage gain (drain J1 (node n003) / top R4 (node n010) over a 1kHz to 10MHz range with a highlight on our used frequency of 2143kHz:
Yes, keep monitoring this plot at J1.drain as you vary the surrounding component values.
Sanity Checks (without input signal):
Real resistor tolerance: With R7 and R12 removed, the DC voltage measured between the junction of R8/R9 and the junction of R11/R10 should be zero (or very close to zero) - if not, then trim these resistors.
- V(J1.gate) to GND = 2.27V. If your voltmeter reads anything other than 2.25 .. 2.30V then there is something wrong. This is set entirely by the R1/R2 voltage divider: 15V × 10k/(56k+10k) = 2.273 V.
- V(J2.source) to GND: The healthy target window with R5 = 1 kΩ is 0.8V .. 2.0V. Anything outside of this range is suboptimal. This is the most informative single measurement in the circuit. It directly encodes drain current: Id = V(J2.source) / R5. With R5 = 1 kΩ the arithmetic is trivial - voltage in volts equals current in milliamps.
- V(J1.drain) to V(J2.drain) =
8.72V : Healthy. J1 has ample Vds
7.19V : Healthy. Nominal operating point.
4.69V : Still comfortable
1.90V : Barely working. Coming out of saturation. Reduced gain
< 1.0V : Failed cascode.
QuoteYes, keep observing this plot as you change the component values.
OK, so when doing the first part:
Quotee.g.: R5→220Ω to 470Ω, R3→2kΩ to 5kΩ or something in-between. C2→1µF to accommodate lower cascode frequencies (kHz). When adjusting these values, have LTspice make a plot of the voltage gain vs. frequency (over the 1kHz - 10MHz span) measured at J1.drain.
Additionally, I think that R4 can be increased to 100MΩ and C4 can be eliminated or substituted with a protective resistor.
Changing 1 component at a time:
Base voltage gain at 2143kHz was: 67.95dB
R4 from 10 to 100M: 67.95dB
C4 remove: 0dB
R5 from 1K to 220 Ohm: 30.91dB
R5 from 1K to 470 Ohm: 64.90dB
R3 from 3.9K to 2K: 68.56dB
R3 from 3.9K to 5K: 67.56dB
C2 from 100nF to 1uF: 67.95dB (lower frequency 1kHz went from 59.4dB to 68.94dB, changeover point around 11kHz)
Not sure why removing C4 stops everything.
Itsu
The gate of a real BF245 JEFT is very sensitive. Even an ohmmeter can damage it.
Input clamping is advised for higher input signal amplitudes (e.g.: current limiting resistor + sdiodes like the BAT68 with 0.5pF ...or BAT15-03W with 0.25pF ...or better yet: HSMS-2810 with 0.1pF ) ...but we will deal with that later.
The gate of a JFET is normally a reverse-biased P-N junction (like a diode) that has leakage current in the pA range. The BF245B datasheet gives Igss ≤ 1 nA under worst-case test conditions (large reverse voltage, elevated temperature).
if your quiescent (no input signal) DC measurements indicate V(J2.source) to GND that is higher than expected AND V(J2.gate) to GND which is > 50 mV above zero, then the two observations are linked - the damaged gate junction is the root cause of both.
Doing the second part:
Quotechange R12→10MΩ. Also, it would be beneficial to swap the inputs of the comparator and apply the signal coming from R7 and the R12 hysteresis feedback to the noninverting input so C5 only stabilizes the reference voltage at the inverting input and doesn't shunt the hysteresis feedback to ground (with this change, a low-ESR C7 doesn't hurt anything ...and even hepls)
I get:
Voltage gain at 2143kHz at OUTPUT / n004 (R7, R8, R9) is -110,22dB
Change R12 to 10M: 45.56dB
Comparators + and - swapped OUTPUT / n005 (R7, R8, R9) is -103.72dB
Change R12 to 10M: 45.50dB
Changing C7 ESR does not influence these latest measurements.
I don't think these last measurements are useful as there is not really a gain (going from sine to square).
Itsu
Quote from: Itsu on 2026.05.29, 20:47:39
Doing the second part:
Please post the screenshot of the changed schematic.
Quote from: Itsu on 2026.05.29, 20:47:39
I don't think these last measurements are useful as there is not really a gain (going from sine to square).
Yes, measuring the comparator is like comparing apples to oranges because the input is analog and the output is digital. The comparator still has a gain that is listed in its datasheet (AFAIR gain=100000).
The change made to the comparator's circuit (swapping its inputs) might not make much difference in the simulation, but it will make a difference in the real circuit.
Swapped comparators inputs:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56285;image)
We see a change in duty cycle, not 50% anymore
Now also changed R12 to 10M:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56287;image)
Now we see a very small amplitude (476nV) of the output signal.
So swapping the comparators inputs has a negative effect on the output signal i did not notice during the frequency plot yesterday.
Quote from: Itsu on 2026.05.30, 08:47:49
Swapped comparators inputs:
That's why it is important to post schematics after topo changes.
See the corrections in red color:
Thanks, stupid mistake, I surely need to stay off the decaf in the morning.
It does not make any difference at the output indeed, not with R12 as 1.2M nor with 10M.
I found a TLV3201 model, so i used it in the sim:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56291;image)
R12 = 10M, shown is the output in green and the non-inverted input (+) in blue.
Itsu
Do the present values of R3 and R5 pass the DCV sanity checks and give you maximum voltage gain at J1.drain ?
You mean in the sim?
If so, yes, 1.25mA through R5.
Quote from: Itsu on 2026.05.30, 20:23:21
You mean in the sim?
Yes.
What do you get between V(J1.drain) to V(J2.drain) ?
Voltage between J1.drain to J2.drain is 6.537V
This is still with no input signal and removed R7 as in your post #131 above.
Quote from: Itsu on 2026.05.31, 07:56:36
Voltage between J1.drain to J2.drain is 6.537V
Can you get it above 7V ?
Yes, by lowering R3 from 3.9K to 3.5K it becomes 7.03V (1.253mA) with a voltage gain of 68.08dB at 2143kHz.
Lowering R3 further to 2.1K, the J1.drain to J2.drain voltage becomes 8.77V with a voltage gain at 2143kHz of 68.5dB.
I build the amplifier part on a pcb according to the "dead bug" method:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56294;image)
This circuit schematic is like this (using BF245C's, and the voltages in red are measured without input signal J1 gate and J2 source to ground):
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56296;image)
When applying the input signal (100mVpp at 2143kHz), i get these signals (blue input from FG, yellow signal on J1 drain and purple is at output (R7):
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56302;image)
To compare the signals with the LTspice sim, here the same circuit as in real, but with BF245B's:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56304;image)
Sim: Real:
Blue 100mVpp 109mVpp
Yellow 442.1mVpp 498.8mVpp
Purple 403.7mVpp 455.6mVpp
One thing i do not understand is that the Sim shows a ~12V DC offset on the purple trace like it does with the yellow trace, but the real circuit measurement show this DC offset ONLY on its yellow trace (therefor the yellow probe is set in AC coupling).
Why does the Sim shows this DC offset as there is a DC blocking capacitor (C5) which should block this DC?
Answer, apparently LTspice needs to know that the output has a high (MOhms) reference to ground, so when adding a 10MOhm resistor from OUTPUT to ground it loses the DC offset.
Amplifier gain input to output calculated (https://circuitdigest.com/calculators/decibel-db-calculator)
shows 12.13dB for the sim, and 12.42dB for the real circuit.
Itsu
Updated the above post
When i use verpies his ring down circuit to produce the ring down signal of the LC tank circuit, and use the current probe plus its amplifier to pick up this ring down signal, which then is fed into the Jfet amplifier, i get these Jfet amplifier in- and output signals:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56306;image)
Purple is the input signal coming from the current probe amplifier and blue is the output signal from the Jfet amplifier, taken at R7.
This output signal seems ready to be fed into the TLV3202 comparator for further processing to a square wave.
Itsu
Quote from: Itsu on 2026.06.02, 19:04:44
When i use verpies his ring down circuit to produce the ring down signal of the LC tank circuit, and use the current probe plus its amplifier to pick up this ring down signal, which then is fed into the Jfet amplifier, i get these Jfet amplifier in- and output signals:
You could also use it to pick up the voltage signal from the LC circuit directly. But JFET input protection becomes an issue (especially the one that forward biases the gate-source).
Quote from: Itsu on 2026.06.02, 19:04:44
This output signal seems ready to be fed into the TLV3202 comparator for further processing to a square wave.
So add the output of the TLV3202 to that scopeshot.
So i completed to add the comparator circuit to the dead bug pcb, and added the comparators output to the screenshot in yellow:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56310;image)
Signals are somewhat "dynamic" and glitchy, perhaps because we are ringing at 2.3MHz.
The "end" of the ringing period when signals are becoming small present a problem for the comparator as it starts to "fade out":
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56312;image)
Anyway, we're getting there slowly.
Itsu
It looks like the comparator's kickback is not propagating back to the input anymore.
JFETs clip gracefully so the gain of the amplifier should be increased to cover the 100 - 4.32% (27.29dB) dynamic range ...or a multistage AGC must be added to cover that dynamic range.
Also, keep in mind that when this JFET amplifier is used for current sensing with your P6302 probe then the low capacitance of this amplifier is not important but when sensing the LC voltage directly then the low capacitance of this amplifier is very important.
P.S.
What was your value of R12 when taking that scopeshot ?
R12 is 10MOhm.
Here is the schematic with the note that I am using BF245C's and a dual TLV3202 (not used in and outputs grounded).
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56332;image)
Due to the fact that the comparator needs a certain minimum voltage to work with, and that it needs to cover the ringing range of 100 to 4% of amplitude, it struggles at about half this range to produce the needed square wave signal.
So i was looking to the ringing signal coming directly from the current probe amplifier and counted about 288 pulses in this 100 to 4% amplitude range, which points to a Q of 288 of the toroid coil under test which is in the range value as measured by other means.
So i think that this last method will do fine in the rare case i need to know the Q of a toroid / coil.
Itsu
I have not forgotten about this. In one of my attempted solutions the input protection that extends the dynamic range loads the LC tank.
I could use some help here, guys...
To avoid any confusion that may exist, here i have a detailed video on how my present circuit is put together: https://youtu.be/kg-pN2duzLI
Here the signals from this circuit:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56418;image)
Purple is the signal at the input of the Jfet amplifier
Blue is the signal at the output of the Jfet amplifier (R7)
Yellow is the output of the comparator
Itsu
Hi Itsu,
Maybe I am mistaken but probably verpies meant to protect the jfet cascode amplifier input, I base this on his post here I have just noticed: https://www.overunityresearch.com/index.php?topic=4949.msg119355#msg119355 (https://www.overunityresearch.com/index.php?topic=4949.msg119355#msg119355) Thanks for the video.
If yes, then I think the solution would be to follow the circuit principle applied in the HP Q meter. Will continue this later.
Gyula
Gyula,
you might be right, so not using the current probe pickup, but something else to pick up the tank LC signal (thus potentially disrupt / load it) O0
Here a schematic of the present circuit i work with:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56445;image)
Quote from: gyula on 2026.06.14, 13:04:06
Maybe I am mistaken but probably verpies meant to protect the jfet cascode amplifier input, ...
You are not mistaken.
Hi Itsu and Verpies,
Itsu, thanks for the summary and the schematics on your present setup.
Verpies, please bear with me, I have got to do some other works popping up and will post a possible remedy on the issue you rightly see, in 1 or max 2 days time.
Gyula
I was aware of a composite JFET device which makes it possible to extend the input voltage amplitude into the positive polarity well over the max 0.6V for a normal JFET.
The device can be 'assembled' from a JFET and a bipolar transistor, see the attached files from a paper. The source and emitter resistors influence the transfer and output curves.
The diode connected transistor, Q1 is needed in case the drain-source voltage happens to be lower than the gate-source voltage. It is likely a fast diode could be used instead of Q1.
Verpies, this device could solve the issue of a normal JFET not accepting higher than 0.6V positive input voltage between its gate and source. From the parallel LxCx tank circuit a capacitive divider could reduce the several tens of peak to peak voltages
to as low as say from -4V to +4V or so. Note that the gate voltage of the JFET can be set to a positive bias by a resistive divider so that the operating point can shift upwards to widen the input voltage range.
(Itsu measured the Vpp across the LxCx in his setup and found around 38 Vpp.) Of course the Vpp depends much on the coil Q to be measured but Itsu's present LC tank surely should have a Q of around 380-400
this is what I think as inherent in the T200-2 core and his smd capacitors.
So such M-FET could replace the 1st JFET in the cascode circuit or other variants may also be possible. The JFET in the M-FET could be the BF245C or J310 etc and for the bipolar a 2N3904 etc. I or you can delete the paper if needed.
Gyula
Thanks Gyula,
meanwhile i was working in a 50:1 impedance transformer, using a T94-2 toroid and some litz wire i had.
Unfortunately, this litz wire was too thick for getting 50 turns on this toroid, so i settled for 38 turns (with a 15 degrees opening at the start / finish).
It needs to be encapsulated with copper forming a single turn secondary, so this was the (open cap) result:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56466;image)
With the tapered cap on (soldered later on) it looks like this:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56468;image)
The nanoVNA shows the following results for inductance and resistance (100kHz to 10mHz) on the primary:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56470;image)
So at 2.3MHz we have an inductance of 12.8uH
Itsu
Looks good.
What is the capacitance between the primary and secondary ?
I have not figured out how to do that, the LCR meter shows 12.8uH (of course) when measuring from floating end 38 turn coil to floating end (basically ground) single turn coil when in "auto", and "OL" to -20uF to -200nF when set into "C" at the several frequencies.
Perhaps the nanoVNA can do this when i have connected a BNC to the single turn coil too and use a "thru" measurement.
So i disconnected the primary coil grounded site and measured with the nanoVNA from the primary floating end to ground to see a 9.5pF capacitance (so between the now above ground primary, to ground (which basically the single turn secondary is).
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56472;image)
Bottom trace is the capacitance, while the upper trace is still the inductance which shows zero now).
9.5pF confirmed by the LCR meter.
Itsu
I also build a smaller x-former, but now with 50 turns on a T68-6 core using 0.2mm magnet wire (AWG 30).
It measures 13.3uH and 4.8pF between primary and secondary:
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56476;image)
Itsu
Quote from: Itsu on 2026.06.20, 08:51:12
I also build a smaller x-former,
Did you use a one-piece copper pipe cap this time ?
Quote from: Itsu on 2026.06.20, 08:51:12
but now with 50 turns on a T68-6 core using 0.2mm magnet wire (AWG 30).
You know what I am going to write now: No air gaps between the turns = large inter-turn capacitance.
Quote from: Itsu on 2026.06.20, 08:51:12
It measures 13.3uH and 4.8pF between primary and secondary:
What is the resistance ...since small wire = high resistance.
That 4.8pF is mainly caused by the distance between the primary and the cuppola. The only thing you can do is maximize this distance.
QuoteDid you use a one-piece copper pipe cap this time ?
Yes, i used a 22mm copper pipe cap.
QuoteYou know what I am going to write now: No air gaps between the turns = large inter-turn capacitance.
Yes, with careful alignment of the turns it would be possible to have a small air gap between the 50 turns.
QuoteWhat is the resistance ...since small wire = high resistance.
That 4.8pF is mainly caused by the distance between the primary and the cuppola. The only thing you can do is maximize this distance.
Resistance measures 0.9 Ohm.
The toroid is centered in the middle of the copper pipe cap (cuppola) and in the middle of it height wise.
I am following the design as shown on page 21 of the below PDF
The 2 impedance transformers (38:1 and 50:1) finished, together with their measured data;
(https://www.overunityresearch.com/index.php?action=dlattach;topic=3691.0;attach=56565;image)
...and primary to secondary interwinding capacitance ?
What about leakage inductance (measured with one winding shorted) ?
Also, I am surprised at the low primary inductance. Alas, the primary is wound on a toroidal core with many turns.
Quote from: Verpies on 2026.06.28, 11:10:17
...
Also, I am surprised at the low primary inductance. Alas, the primary is wound on a toroidal core with many turns.
The 50 turns on a T68-6 core should give an inductance of 50 x 50 x 47 / 10000 = 11.75 uH (where 47 is the A
L of the core), calculated from data sheet. https://alexsradioshop.de/wp-content/uploads/2023/01/AmidonAMI.pdf (https://alexsradioshop.de/wp-content/uploads/2023/01/AmidonAMI.pdf)
The measured 12.03 uH is quite close to the calculated.
Likewise, the 38 turns calculates to be 38 x 38 x 84 / 10000 = 12.12 uH on the T94-2 core, the measured 12.72 uH is also close.
Gyula
Quote from: gyula on 2026.06.28, 12:14:39
The 50 turns on a T68-6 core should give an inductance of 50 x 50 x 47 / 10000 = 11.75 uH (where 47 is the AL of the core), calculated from data sheet.
Considering such low A
L it makes sense now.
Quote from: Verpies on 2026.06.28, 11:10:17
...and primary to secondary interwinding capacitance ?
What about leakage inductance (measured with one winding shorted) ?
Also, I am surprised at the low primary inductance. Alas, the primary is wound on a toroidal core with many turns.
The capacitance between the prim. and sec. was given earlier, but for completeness and to add the leakage inductance, see below table:
38:1 toroid T94-2 coreL prim: 12.72uH, R: 230mOhm
L sec: 19.00nH, R: 5mOhm
Capacitance Prim/Sec: 9.5pF
leakage L prim:
LPrim with: sec open 12.72uH
sec shorted 7.60uH
K: 0.63
50:1 toroid T68-6 coreL prim: 12.03uH, R: 900mOhm
L sec: 12.20nH, R: 5mOhm
Capacitance Prim/Sec: 4.8pF
leakage L prim:
LPrim with: sec open 12.03uH
sec shorted 8.40uH
K: 0.55
Thanks Gyula for explaining the "low" primary inductances
Itsu