Measuring Self-Capacitance and Self-Resonant Frequency SRF of Inductors
ฝัง
- เผยแพร่เมื่อ 9 ก.ย. 2024
- The parasitic capacitance of every imductor limits it´s usability as a filter. It even behaves like a resonant tank-circuit at the so-called self-resonant frequency.
Support Roger on Patreon: / kainkalabs
Roger explains an easy way of measuring the parasitic properties of an inductor with just a square-wave generator and an oscilloscope, even an "old" analog one.
No special test equipmet required :-)
The best and easiest method to do this I have ever seen. I did this on two oil-filled ignition coils and it works!
Thank you for the video I have enjoyed watching it. I also found a website that derives the formula for calculating the Q of the resonant system, (number of cycles for envelope to diminish to 50%) X (4.53) = Q. The frequency of oscillation and the inductance can be used to calculate the capacitance in the resonant system. The Q and the inductive reactance can be used to calculate the effective resistance in the resonant system. I use the term resonant system because there is more capacitance in the system than just the stray capacitance of the inductor. I used this method to try to determine how stray capacitance varied as a function of the number of turns of an air core inductor. What I have learned is the oscilloscope probe can add between 15pF and 20pF to the system. 15pF to 20pF can be many times greater than the stray capacitance of the coil itself and what initially appears to be the self resonant frequency of the inductor is actually significantly less than the self resonant frequency of the inductor. I have experimented with coupling the signal from the inductor to the oscilloscope probe through a 1pF capacitor and this has its own set of difficulties. In my case I have at least verified that the stray capacitances of the coils I am experimenting with are small enough that they can be ignored in the preliminary design of a crystal radio. It is almost impossible to determine the stray capacitance of an inductor precisely when the stray capacitance is in the range of 2pF to 10pF. Minor adjustments can be made in the design using trimming capacitors to improve performance of the initial design. I really enjoyed this video and it has led me to several interesting experiments and observations over several months. This method is theoretically sound but the capacitance added be oscilloscope probes makes it extremely difficult to measure the very small stray capacitances associated with some inductors such as air core inductors used in crystal radios.
It is of utmost importance that the square-wave source and the oscilloscope-probe are isolated from the inductor-under-test each by a resistor of at least 10x the value of the calculated (or initially guessed) of the resonance-resistance of the inductor/self-capacitance tank-circuit.
If you connect the oscilloscope probe directly to the inductor you of course get the capacitance of the oscilloscope probe added to the self-capacitance of the inductor.
10 Megohm should be the minimum, 100 Megohm is better for High-Q inductors.
Of course it´s not easy to get reasonable results when the self-resonant frequency goes into the many mega-Hertz or even 100 MHz and above.
There it even becomes important to have the inductor-under-test isolated with a wooden holder from any metal, ground-planes, earth-connected cables etc.
Otherwise you get an extra capacitive coupling that skews your results.
You can notice these effects even when touching the inductor (as can be seen in the video).
Super job teaching the subject . Best I have seen on TH-cam.
What a wonderful explanation. Well illustrated. Thank you sir. I liked it.
Most excellent! I was able to immediately go down to the KY8D lair and verify that the basketweave loading coil for the 472kHz TX antenna would, after all, be good for its intended purpose. This after my LCR meter gave alarming mis-results, when used by itself. Ditto the Rx loop. But that neither may do double duty for 160 meters. Good to know. Thanks!
Very informative video about inductors.
Thanks! This is a very clear explanation.
After some thought, what you presented has caused me to reevaluate one of my designs to determine why it wont oscillate. I overlooked stray and parasitic capacitance. I built the circuit on Protoboard....a bad place for stray capacitance.
In this video he used "Thompson's formula" f = 1/(2PI*sqrt(LC)) based on the self resonant frequency and L measured at a much lower frequency to calculate the self capacitance (which he called "parasitic capacitance" - a term we don't use in english).
This is generally a very inaccurate method and should not be used.
The reason is that at and near self resonance, the circulating capacitance current seriously distorts the distribution of coil turns current. Obviously, if it is resonating, the current in the start and end turns must be very small, but the current in the middle turns very high. In fact the distribution of current in the turns is parabolic - quite different to what happens when resonated with an external capacitor, wherein the current is the same in every turn.
Thus the magnetic field is a different shape, the energy stored different and that means the inductance is different - changing the self resonant frequency away from what Thompson would predict.
The traditional way of measuring self capacitance is to find two external capacitors that resonate the coil at a frequency f1 and 2f1, both much lower than the self resonant frequency, say 1/10th. The self capacitance is then (C1 - 4C2)/3. This will for most inductors typically be about 10 to 15 times more accurate than KainkaLab's method. You don't need to know L to get the self capacitance.
I had to laugh every time he pronounced "series" as "serious". He did it every time.
Thank you!
SG, test load, Osc in paralel like in the circuit doesn't work with my Fnirsi 1014D Osciloscope.
Dont know why.
By connecting all in series(without resistor) then i can find the resonant frequency.
thank you so much. great work
your lesson is great thank you very much
Can you tell me the name of the device you use at minute 28:05 ?
Thanks a lot
You probably mean the "Component Tester" which is an open-source project from a German ham-radio amateur. It sells on Ebay under different names and in a number of models. Search for "Component Tester" or "Transistor Tester". David L. Jones also given it a "thumps up". We have it in our shop for many years:
www.ak-modul-bus.de/stat/laborzubehoer.html
And I made a 4-part video series (kind of shootout) of 4 different transistor/component testers...
@@KainkaLabs Thank you very much for this complete answer. See you soon, a cordial greeting.
excellent lecture thank you very much
great lesson! thank you!
Thank you for an excellent lesson!
yay really enjoyed this :)
Thank you very much!
He said you can find Q by counting the number of cycles to get to half the first cycle amplitude and multiplying by 5. That's a bit rough and ready. For a Q of 50, 10 cycles would drop to 0.522. not 0.5. Logarithmic decrement (the drop from one cycle to the next) is pi/Q, so the amplitude of the 10th cycle is (1 - pi/50)^10 = 0.522. He added further error by starting at the first cycle, whose amplitude and waveform is affected by the residual wave from the previous driving cycle - you can clearly see that in the video. He should have taken say the 5th cycle as his reference or, better, use a much lower frequency triggering square wave.
His real Q is closer to 45, not 50.
If you don't have a fancy scope with Cursors can you determine the SRF by applying a sine wave and adjusting the frequency to get the highest signal amplitude on the scope?
You don´t need a fancy-pancy scope :-)
Every analog scope (with enough bandwidth) can do it.
You just have to trigger on the exciting square wave to get a stable view and just count the cycles where the amplitude drops to half its peak value.
I think in this video I explained 2 other methods to measure Q-factor: th-cam.com/video/gJlkWt3VYxM/w-d-xo.html
Basically it should be possible with what you suggested.
But be careful to couple in the sine-signal with at least a 1Mohm resistor (10 MOhm is better) and you probe the amplitude with a 10:1 scope-probe with 10 MOhm input-impedance.
Just count a full cycle horizontal divisions and multiply by hor sec/div time base.
@@KainkaLabs is there any way to measure srf if I have no scope at all?
gracias!
You said it is the self-resonant frequency measured by the component tester, but that was 546kHz, different than the 680kHz the scope measured.
You said it's not very accurate, but that is a big difference, is it still the same measurement, or did I miss something?
In the calculation, he is using 680kHz. Ignore 546kHz. 546kHz was used by the component tester to measure the inductance of the coil.
So if I have a coil with a capacitance of 290pf that I measured. I want a resonant tank for the natural frequency of the coil and have derived that capacitance needed for the resonant frequency. The problem I see is the coil capacitance is greater than the derived capacitive value. Am I correct in stating the I have to take into account the coil capacitance in creating the LC tank? If this is the case then I can not get the capacitance as low as I need unless I add another capacitor in series to try to get the value low enough.
You can only change the resonant frequency to lower frequencies (below the SRF of the coil). The only way to change/tune the resonant frequency of an existing coil is to add a capacitor in parallel. And that in turn lowers the freqeuncy. Adding a capacitor in series does not change the resonant frequency!
So you have to find or wind another coil to get the desired resonant frequency.
@@KainkaLabs Thank you. I Want the tank to resonate at its S)elf R)esonating F)requency. That is the value of capacitance that I am trying to achieve.
Does the frequancy change with a changing voltage? So if we measured it at 2 volts will it change at 5 volts and 10 volts respectively?
I didn´t try, but why should it change (as long as there are no saturation effects)
How do you select the output frequency of your function generator for this? Many thanks.
It doesn´t matter so much because it only serves to excite the inductor. Your osciloscope will automatically trigger (only) on the start of the oscillation. For the video I just set the frequency so that you can see 2 of the damped oscillations on the screen. For measuring it is better to have only 1 damped oscillation stretched fully on the screen.
15:00 Probe also have some capacity..you should not add 1pF capacitor with probe?
Yes, somebody else remarked this in another post. I forgot to isolate the probe with a 10 Megohm resistor from capacitive influence.
Hello Roger, Great Video! I am trying to do the same with a cylindrical oil filled ignition coil. I seem to be having some issues with this on this sort of coil. I took my 10 mOHM resistor from the inside of a microwave oven capacitor. Can you explain how to do the same experiment on both the primary and secondary of one of these sorts of coils? Perhaps I need less resistance? Remember the ignition coil is an autotransformer with the primary and secondary attached (usually at the negative pole of the primary). I know the inductance of the primary because I have a UNI-T LCR meter that caps out at 20 Henries (my primary is 5.27 mH, but the secondary (which should be around 49 H is too large to measure with this unit).
10 miliohm? Did you mean 10Mohm?
@@AlienRelics Of course, thanks!
@@MisterBonetti Ironically, I mis-spelled "milli"! 😅
Nice. How can I determine SRF of a 15H choke?
Same principle. Where´s the problem?
I believe in general the SRF stands for Self Resonance Frequency.
Correct. Self Resonant/Resonance Frequency.
Why your measured Q factor differs so much from calculated one? According to your parameters Q factor can be calculated as: Q=1/R*sqrt(L/C).... It gives 281..? But anyway perfect video.! :D
You can´t take the ohmic (DC) resistance for calculating Q-factor. You have to take into account skin-effect and proximity effect which raises the DC-resitance sometimes by orders of magnitude.
@@KainkaLabs Thank you very much for your answer.... You really think that skin effect and proximity effect would have such an impact on the resistance in frequency 700kHz? .. and moreover u used a Q factor formula for calculating AC resistance... I am just so struggling with my LC circuit, because my measured value of Q factor is significantly lower than calculated one... Do you think u can use the same way to measure quality factor of LC circuit (where value of capacitance is in tens or hundreds of nF)?? Thank you
I think you used the wrong term 7~ min
- so called 'impedance' wouldn't peak at the resonance frequency, that would be so called 'anti-resonance' so this tells me the peak on the graph is not the resonant frequency. (i may be wrong there) but i think the manufacture is showing just the impedance peak. which does not necessarily align with the frequency of resonance.
they are showing the peak for the use of the device. (as intended)
No. I think you´re wrong.
In a *parallel* LC tank-circuit the impedance is in fact at maximum at (very near) the resonant frequency.
And as is explained a few minutes later the parasitic elements of an inductor always make it a *parallel* LC tank-circuit.
This is explained in quite some detail in the German Wikipedia entry about LC-circuits:
de.wikipedia.org/wiki/Schwingkreis#Realer_Parallelschwingkreis
(The English Wikipedia entry does not explicitly mention this but it could be derived by diffentiating the formula for the impedance of a parallel LC-circuit)
The opposite is true for a *series* LC resonant circuit.
There the impedance is minimal at (very near) the resonant frequency.
Perhaps that is what you mean with the "anti-resonance" point.
ah ok thanks for the clarification, ok i see at these frequencies how the effects you mentioned come into play. thanks for the correction.
although i must say i don't trust Wikipedia for any real technical information however i see what you are saying, yes it was negligence on my part here.
although i wouldn't personally use the term 'resonance' to describe that effect however i can see how it is described as such.
thanks gain.
furthermore to this yes you are absolutely correct i apologies in these terms.
the two reactants are in resonance which causes the device to become 'cancellation effect' it's the cancellation effect i was calling 'anti-resonance' so yes it was the terms i was having difficulty with.
thanks again.
I think SRF is Self Resonant Frequency.
There is a fundamental Parallel SRF (peaking) and a fundamental Series SRF (low Z or a short) and harmonics of each; even on the parallel and odds on the series. For instance a parallel SRF of 20 MHz would have a smaller peak response at 40 and 80 MHz and so on. Similarly the Series SRF if at 22 MHz would have lo Z values at 66 and 110 MHz and so on...
er... I got banned from the forum for spam but do not know why can you help?
Register with another email adress and place at least 1 senseful message within the first 3 hours after registering.
SFR = Self Resonant Frequency. Kr, C.
No. SRF = Self Resonant Frequency :-)
@@KainkaLabs :p
scope+probe input capacitance is quite comparable to the value you calculated so, this self capacitance value is quite a bit wrong
No. Because the scope-probe was isolated by the high-Megohm value resistor.
@@KainkaLabs @ 16:00 what resistor? It's a 10Mohm input of a 10:1 probe directly connected to the inductor. there is no extra resistor in between, is there? The lowest capacitance 10:1 passive probe I have ever seen is P6139 of Tektronix which has 8pF total capacitance at its input tip when it is connected to the 1Mohm scope and compensated. Any other passive probe will have more than 8pF at the tip.
@@amirb715 so you're saying the SRF measurement is wrong? I'm not sure if I'm missing something, but you're implying it's impossible to measure anything at MHz.
@@KainkaLabs I agree with Amir B that your scope probe capacitance caused your measured SRF to be lower than it really should have been because you did not have a resistor between your inductor under test and the scope probe. I just did a test using your set up using a x10 probe and my measured SRF on an inductor was 5.1 MHz. When I measure the same inductor using a nanovna the SRF measures 7.3 MHz. If I repeat your measurement system measurement but add a 1 Meg ohm resistor between the inductor under test and the scope probe the SRF measures 8.33 MHz. I suspect the 8.33 MHz measurement is the most accurate out of all 3 of my measurements because the added 1 Meg ohm resistor helps isolates the scope probe capacitance from the inductor under test.
I also calculated what capacitance would make my 44.2 uH inductor resonate at 5.1 MHz and it would be 22.0 pf. To resonate the inductor at 8.33 MHz the capacitance would be 8.3 pf. The difference in capacitance is 13.7 pf which is very close to the 14 pf published capacitance of my x10 probe. Therefore it sure looks like the scope probe capacitance greatly influences the measured SRF using your setup.
Update: I also checked self resonance of my inductor by setting my function generator to output a sine wave and connected it to the inductor using a 10 Meg resistor just like the setup shown in this video, but I did not connect the scope directly to the inductor. I wound a two turn pick up coil and connected it to my scope x10 probe and placed this pickup coil close to one end of the inductor and then adjusted the function generator until the scope waveform peaked, and it peaked at 8.3 MHz which exactly matched the results I obtained when I added the 1 Meg resistor mentioned in my first paragraph.
P.S. great video