Improved Simple PI - all BJT front end for greater sensistivity

[Beenthereuk]

Well, the front end works, just like SPICE said it would...
The version I've built uses the coil I made, which is larger than the one used in the Russian Simple Pi posts so the biasing resistors needed tweaking, a 5k pot in place of the 3.3K resistor did the job in seconds.

[Waikiki_Sweep]

Good to see it working!
Scope shows coil is ringing. Looks like internal capacitance of coil is too high as well as dumping resistor.
To lower that interwire capacitance you can make better coil from stranded wire (solid wire acts like target) with thick insulation (Teflon) or use spacers (polyethylene) or flat spiral shape of coil or basket shape.
Usually coils for such setups has 200-300uH. I see 777uH and 50 turns. I think 20-22 turns probably will be OK. Also dumping resistor may needs adjustment (lowering) to stop that oscillation.

[Beenthereuk]

Yes, the coil creates problems for the transistors too because it is loaded too easy and the input impedance of the first stage is not that high.

I will make myself another smaller coil soon, just need to make another foam disc to roll it on. It would be nice to find some way of winding a decent spiral pancake.

Have you tried this, BTW?: http://dd6um.darc.de/QucsStudio/about.html

[Waikiki_Sweep]

Coil-how-to.pdf Great Russian coil but it is too much work to made. Also if it made from stranded insulated wire it works just fantastic!

[Beenthereuk]

Clever layout. I'm making one on a round disc of wood using holes instead of notches.

I measured my coil's parasitic capacitance today by using a large poly capacitor in series with it on an LCR meter and was shocked to learn it is 10.5 nF
The right damping resistance for that is 140 ohms !!

I need a better coil...

[Waikiki_Sweep]

Clever layout. I'm making one on a round disc of wood using holes instead of notches.

I measured my coil's parasitic capacitance today by using a large poly capacitor in series with it on an LCR meter and was shocked to learn it is 10.5 nF
The right damping resistance for that is 140 ohms !!

I need a better coil...

Cool!!!! "large poly capacitor in series with it on an LCR meter"

This is a novice method of measuring parasitic capacitance of coil.

Our gurus from the forum measuring it by observing self resonance frequency of the coil.

Interesting to compare results of both methods !!!

[WM6]

I measured my coil's parasitic capacitance today by using a large poly capacitor in series with it on an LCR meter and was shocked to learn it is 10.5 nF

How did you take in account "parasitic" resistance of coil using this method?

[sinclairuser]

you deduct the known capacitance of the cap from the total.

[Beenthereuk]

The resistance does not play a part as the large capacitor blocks all dc. The LCR meter reads two capacitors in series and reports the lowest value (the parasitic capacitance). There is a small error that can be made even smaller by making the difference between the blocking capacitor and the expected parasitic capacitance as large as possible. It is then easy to simulate the coil and compare a resonant ping with what the scope shows. Its accurate enough.

[WM6]

The resistance does not play a part as the large capacitor blocks all dc.

Yes, but didn't block AC. For AC coil resistance still exist and change your measurement depend of your measurement frequency.
Try to measure your large capacitor with resistor equal to coil resistance in series with C to check how pure resistance in series can change measurement results.
10.5nF parasitic capacitance for PI coil cannot be real, I never calculate more than 1000pF out of PI coil resonant frequency.
What is working frequency of your LCR meter?

[Beenthereuk]

My meter will work either at 100 or 1000 Hz. So I measured at 100 Hz as the coil has less reactance at that frequency than at 1000 and the interwire capacitance is easier to measure.

The inductor model is R + jwl - 1/jwc, the DC block eliminates the real part, leaving the imaginary part.
The meter can distinguish between leading and lagging signals (phase + or - compared to input signal) and hence will select either inductance or capacitance which it then divides by omega (2 pi F) to obtain the measured value. In this case, because the intrinsic capacitance << the blocking capacitor, it dominates as the smallest capacitor in series.

There is no ac resistance value - at ac it is all impedance: inductive reactance XL and capacitive reactance XC.

The best way to do this more accurately is to take S parameters of the coil, de-embedd the leads and probes, convert to Z parameters to extract XL, then convert to Y paramters to extract XC all vs frequency, fit a straight line through the values to obtain a mean and divide the mean by omega to get each parameter. This is called direct parameter extraction and it is how RF transistor models are created. Only problem is I had to leave my £300,000 Agilent PNA at my last workplace (my baby... ) and I do not have anything to do S parameter measurements where I work now so I have to use this silly LCR meter.

My coil is unusually large at 50 turns and uses 0.8mm wire which probably explains why its interwire capacitance is so large. I used a variable resistor across it and tuned down to 140 ohms before it got to critical damping - have you ever had to use such a low resistance?

[Beenthereuk]

We can take this one step further, using the formula Rd = 0.5 * sqrt(L/C) and solving for C using my coil inductance (777uH) and measured value of resistor for critical damping (140 ohms), we get C = L / ((2 Rd)^2) = 777e-6 / (2 * 140)^2 = 9.91 nF

So yes, there is a little error, but it is close enough

[WM6]

There is no ac resistance value - at ac it is all impedance: inductive reactance XL and capacitive reactance XC.

Yes, you can get total reactance according formula X(total)=XL-XC but your total impedance is still Z(total)=R+jX. So you can count with R.

I don't remember tuned damping R with my PI coil lower than say 450 Ohms. Most of my PI coils are 25cm (+-10cm) in diameter and of about 20 to 30 turns.
With 50 turns and bigger diameter, as you say, it is possible to get bigger parasitic capacitance, despite 10nF is somewhat hard to imagine to me for useful PI coil.
I think that results of your smart idea are too much frequency dependent to be taken as accurate.
Awaiting your further reports.

[Beenthereuk]

No, the resistance part is elliminated as it is real. When measuring the capacitance, if DC is blocked, there is no resistance part left in the signal as the LCR meter looks at reactance (admitance) only. This is not my smart idea btw, it is a well known method. It does have some error but its close enough to model the inductor and then tweak the model capacitance until the simulate trace matches the measured trace.

[WM6]

This could be true if you have an ideal inductor (coil R=0) but you haven't.

There are different measurement methods applied in (L) C meters.

If I understand you correctly, your (L) coil parasitic C measurement proposal is as this on drawing:



Going to practically speaking:

I done some test. In this test mono coil of 2mH (27cm dia) is used in series with large test capacitor of 390nF(according factory marking).

I tested in mentioned LC configuration accordance with the scheme of drawings and use a variety of measuring instruments from simple c meters to high precise RLC instruments.

Results:
A. Precise LC meter (f=630kHz) - 167pF
B. Precize RLC meter (f=120Hz) - 373nF
B. Precize RLC meter (f=1kHz) - 383nF
B. Precize RLC meter (f=10kHz) - 194nF
C. Couple other multimeters wit C range: results from 371nF to 432nF

Can you say from this what is parasitic capacitance of tested 2mH (27cm dia) coil?