Q measurement methods

[Verpies]

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 ?

[Itsu]

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).





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

[Verpies]

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...

[Itsu]

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:



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

[gyula]

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   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

[Itsu]

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) 

 
Here a schematic of the present circuit i work with:

[Verpies]

Maybe I am mistaken but probably verpies meant to protect the jfet cascode amplifier input, ...

You are not mistaken.

[gyula]

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

[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

[Itsu]

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:



With the tapered cap on (soldered later on) it looks like this:



The nanoVNA shows the following results for inductance and resistance (100kHz to 10mHz) on the primary:



So at 2.3MHz we have an inductance of 12.8uH

Itsu

[Verpies]

Looks good.
What is the capacitance between the primary and secondary ?

[Itsu]

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.

[Itsu]

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).



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

[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:



Itsu

[Verpies]

I also build a smaller x-former,

Did you use a one-piece copper pipe cap this time ?

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.

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.

[Itsu]

Did you use a one-piece copper pipe cap this time ?

Yes, i used a 22mm copper pipe cap.

You 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.

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.

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

[Itsu]

The 2 impedance transformers (38:1 and 50:1) finished, together with their measured data;

[Verpies]

...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.

[gyula]

...
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
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

[Verpies]

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.

Considering such low A_L it makes sense now.

[Itsu]

...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 core

L 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 core

L 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