Impeadence measurement with scope?

[rwgast_lowlevellogicdesin]

Can i just use channel 1 to probe a point on a circuit with a standard 10x probe, then use a current probe on channel 2 and set the scope up to divide chan1/chan2 and that will give me impeadence? Then you can take this a little farther by inputting a signal gen in to the circuit, sweep it and then plot it?

This just sounds so simple i figure im missing something.

[mawyatt]

Might find these notes helpful.

https://www.eevblog.com/forum/testgear/capacitive-impedance-plots-with-sds2104x-plus-bode-function/msg4335745/#msg4335745

https://www.eevblog.com/forum/testgear/admittance-measurements-with-dso-awg-with-bode-function/msg4491952/#msg4491952

Best,

[radiolistener]

Can i just use channel 1 to probe a point on a circuit with a standard 10x probe, then use a current probe on channel 2 and set the scope up to divide chan1/chan2 and that will give me impeadence? Then you can take this a little farther by inputting a signal gen in to the circuit, sweep it and then plot it?

This just sounds so simple i figure im missing something.

It can work if there will be zero phase difference between both channels - for voltage and for current. But it may be tricky to get exactly the same phase delay, especially for a high frequency, where even small cable length difference leads to a significant phase difference.

More easy and reliable way to measure impedance is to measure amplitude for different load resistance and then calculate impedance. There are at least two most usable ways:
- with two known resistors which value is around expected value (but not the same)
- for one known resistor close to expected value and for no load (infinite resistance)

When you know resistances and their amplitude, you can calculate impedance.

There is also third way - find 50% amplitude drop with using variable load and then measure variable load resistance. But this way is hard to use on high frequency, because variable resistors have pretty high inductance which will affect the result.

[rwgast_lowlevellogicdesin]

Im using a siglent sds1104x e, it can do bode plots reporting phase using channels one and two, so i would assume if poper precuation is taken with the leads phase shouldnt be an issue? I will definately look through the links posted after work.

Btw im only intrested in like 0 to maybe 10mhz more like a mhz probably. I would use my nanovna for rf focused stuff. Im more intrested in this technique for extended LCR measurements, maybe audio stuff. A vna is definatly a better solution but vnas that work from dc are really exspensive. I would love a bode100, it would be one of my most used tools if i owned one but since im not currentlty making money off electronics i mess with at home 6 grand is a pretty far reach.

Why is it a box like the bode 100 gives impeadence values, but a scopes bode plot does not? Like i said im on a siglent, but ive watched demos of keysight and tek scopes do bodies and they never calculate impeadence. ʻ

[radiolistener]

Im using a siglent sds1104x e, it can do bode plots reporting phase using channels one and two, so i would assume if poper precuation is taken with the leads phase shouldnt be an issue?

If you're planning to use frequency sweep to get impedance vs frequency chart, I think there may be an issue, because there is needs voltage and current probe with exactly the same phase delay over wide bandwidth. So, you may get wrong value on the frequency where phase delay difference will far from zero.

Such issue can happens for example if your current probe has some filter with non linear phase response.

High impedance voltage probe is also a kind of low pass filter which also may have phase delay non linearity. So it's hard to predict if both your probes will have exactly the same phase delay over frequency sweep range. In order to avoid unexpected issues its better to use exactly the same probes for both channels. But since you're needs one probe for voltage and one for current, it's impossible. So it's hard to predict how it will works.

[MrAl]

Can i just use channel 1 to probe a point on a circuit with a standard 10x probe, then use a current probe on channel 2 and set the scope up to divide chan1/chan2 and that will give me impeadence? Then you can take this a little farther by inputting a signal gen in to the circuit, sweep it and then plot it?

This just sounds so simple i figure im missing something.

Hi,

With a regular scope you have to be able to measure the amplitude and phase angle after you excite the circuit with a sinusoidal source.
Impedance is a two-valued quantity, while resistance is only one.  In the general case, it's a two-valued quantity though so you need to measure two things with the scope in order to calculate the actual impedance.
It's not entirely difficult to do with mains line circuits where the frequency is maybe 50 or 60 Hertz.

If you know the load is just R and L, or just R and C, you can calculate the R and the L (or C) as well as the impedance.

[jonpaul]

Will not work over 0..10 MHz range.

Read and learn:

 HP/Agilent app note

"impedance measurement"

https://pearl-hifi.com/06_Lit_Archive/15_Mfrs_Publications/20_HP_Agilent/HP_App_Notes-Handbooks/Z_Measurement_H-book_3rd_Ed.pdf

j

[rwgast_lowlevellogicdesin]

That hp paper is pretty informative, but not to helpful reguarding errors that may be in my set up.

It seems like you should be able to get useful measurements up to 5 maybe 10mhz if you ditch the probes, use coax cut to length and mesure current using a resistor instead of a current probe.

Could you also not run a sweep through the current probe while monitoring the original signal to find the point at which the the current probes phase starts to shift and then skew the signal gen output to match the phase shift?

[rwgast_lowlevellogicdesin]

I just read through the linked posts and they were very insightful, it also reminded me that the analog discovery has an impeadence analyzer.. they are pulling it off with way less hardware than a scope and sig gen that are programmable with scpi.

The last link seemed very promising and then just stoped.

[jonpaul]

Hello again:

1/ Depending on the impedance levels and accuracy desired, yes a sense R for current can work, but the resistor must be carefully chosen and non-inductive.

The impedance of the scope input (eg 1 M ohm, 22 pF) and any cable ( eg 100 pF/foot of cable) or probes (eg 10 M ohm/10 pF will have a substantail effect above a few kHz, let alone 1..10 MHz.

Use the formulars for Xc, Xl, Z at the freqencies mentioned to find the residual effect of the probes, scope etc on the supposed test setup.

2/ Professional LCR/Bridges for high frequeny impedance measuements are often rackmount and cost many 100s..1000s $. We used Wayne Kerr 3245, HP 4262A, 4277A, 4275A, GenRad 1689 and HP 4195A Network/Impedance analyzer with impedance test fixtures...(see pix) .

3/ For  hobbiest use, a handheld digital LCR meter is  only $40..100. Test at a few kHz. Some are surprising accurate over the limited ranges.


4/ RE the Pearl Audio link ....  third party, may cause issues.

 original Voila! The latest version of this classic HP reference.

https://www.keysight.com/us/en/assets/7018-06840/application-notes/5950-3000.pdf

Bon Chance,

Jon

[MrAl]

Hello again,

I would imagine that at 10MHz this would be hard to do.  At 50 or 60Hz it should be sort of easy.

The key to a lot of this is to be able to calibrate your setup.  You use equipment known to be good and compare those measurements with your own setup measurements.  You would have to borrow or rent good equipment to do this.

I would also think you might get some idea how your setup is working by measuring some known impedances.  Start with a noninductive resistor, then maybe that resistor with a known inductance, then capacitance, then all three RLC.  You can use calculations to compare to your actual measurements.

For a few examples:
R L C all in series: Z=R+j*w*L-j/(w*C)
R L C all in parallel: Z=(j*w*L*R)/(-w^2*C*L*R+R+j*w*L)
R in series with L, and that in parallel with C: Z=(R+j*w*L)/(j*w*C*R-w^2*C*L+1)

There are not that many combinations of connecting R L and C so you could try other cases too.

Of course it should go without saying that you have to figure in the specifications of any probes used, and also some theory from transmission lines if you use just cable.  It could be rather difficult, so the calibration routine is probably the only way to be sure as you can get away with a lot that way but you have to be willing to do some careful testing with good equipment to be able to specify your own setup so you get good results once you give the expensive stuff back.

[Grandchuck]

Might find these notes helpful.

https://www.eevblog.com/forum/testgear/device-impedance-using-dso/msg3677731/#msg3677731

https://www.eevblog.com/forum/testgear/capacitive-impedance-plots-with-sds2104x-plus-bode-function/msg4335745/#msg4335745

https://www.eevblog.com/forum/testgear/admittance-measurements-with-dso-awg-with-bode-function/msg4491952/#msg4491952

Best,

Those are interesting links.  I decided to make a few plots and was puzzled by excessive error and then realized the ceramic capacitor I was testing was quite sensitive to the signal levels used to create the Bode  plots.  Below, the plot with 1 volt drive is about right and with 4 volt drive the cutoff frequency is too high.  The only difference is the ac source voltage.

[jonpaul]

Hi D dielectrics like X5R have a high TC and VC. See mfg specs.

Use an NPO or mica for a stable accurate cap.

Expect inductors to be much more difficult to characterize

j

[mawyatt]

Might find these notes helpful.

https://www.eevblog.com/forum/testgear/device-impedance-using-dso/msg3677731/#msg3677731

https://www.eevblog.com/forum/testgear/capacitive-impedance-plots-with-sds2104x-plus-bode-function/msg4335745/#msg4335745

https://www.eevblog.com/forum/testgear/admittance-measurements-with-dso-awg-with-bode-function/msg4491952/#msg4491952

Best,

Those are interesting links.  I decided to make a few plots and was puzzled by excessive error and then realized the ceramic capacitor I was testing was quite sensitive to the signal levels used to create the Bode  plots.  Below, the plot with 1 volt drive is about right and with 4 volt drive the cutoff frequency is too high.  The only difference is the ac source voltage.

Gald you found some interest in those previous posts, your Bode plots look nice!! Note the series resonance due the DUT & setup ESL and ESR, well done

As mentioned some ceramic types (high K) have a high voltage sensitivity, NPO/COG are really good but usually limited to under 100nF.

The film types like Polystyrene are excellent (except <80C), Polypropylene is good and Mylar (Polyester) are OK.

You can also have some fun with inductors and even LC combinations for series and parallel resonates.

Edit: You can also evaluate the Capacitor (+ Setup) ESR & ESL from your nice plots, details if you wish.

Best,

[Grandchuck]

Edit: You can also evaluate the Capacitor (+ Setup) ESR & ESL from your nice plots, details if you wish.

Best,

Sure, details please

[mawyatt]

At resonance (Phase is zero) the Inductive reactance has equaled the Capacitive reactance and you are left with the Real part Impedance (Resistance) of the shunt Capacitor and any setup Resistance. So the ESR can be estimated as a voltage division with whatever Resistance R is "seen" suppling the signal to the shunt Capacitance.

The Bode plot displays voltage ratio in dB, so ESR can be estimated as:

ESR ~ R*10^(BP/20), where BP is the Bode Plot value in dB at Phase = 0 and R is the series Resistance.

Likewise, ESL can be estimated as:

ESL ~ 1/[C*(W^2)], where W is 2*pi*F, and F is Frequency where Phase = 0, and C is the shunt Capacitance value.

Anyway, hope this helps.

Best,

[Berni]

You should be able to use the bode plot function to pretty much the bandwidth limit of your scope. With some extra software over SCPI you could even turn this into a proper network analyzer Smith chart.

Yes things get a fair bit trickier once you go past around 10MHz. The phase shift of current probes or cables definitely starts coming into play. However you can make a 'current probe' by simply sticking a 1 Ohm resistor across it and take that as 50 Ohm to the scope input. Will have heaps of bandwidth and a predictable group delay since it is basically just the time of flight over the coax to the scope.

You can easily test your setup by measuring the impedance of a known value resistor (non inductive types). If you have delay or phase issues you will see it on the graph since a resistor should have 0 degree phase delay at any frequency (otherwise might point to a group delay mismatch problem). The impedance should also be a flat line across all frequencies (otherwise might point to a parasitics problem)

If you go into really high frequency it will become impossible to keep your parasitics in check (Tho if you have an active scope probe with low loading you can push it a lot further). At that point you will have to start doing a open/short/load calibration and then subtracting the distortion out of your measurements (like the real network analyzers do it)

EDIT: Oh and i also remembered that i used a scope for measuring very low capacitance values before. I was using a low loading 0.7pF active probe and fed it a sine wave signal to the probe tip using a high value resistor (like 10kOhm). Then when i touch the probe to my circuit i can observe the amount of shift and attenuation this causes, letting me calculate capacitance down into the femtofarads. Useful for measuring parasitics of connections or PCB tracks. Better tool for this is a network analyzer tho(but i don't have a proper one because even the old clunkers are hella expensive)

Those are interesting links.  I decided to make a few plots and was puzzled by excessive error and then realized the ceramic capacitor I was testing was quite sensitive to the signal levels used to create the Bode  plots.  Below, the plot with 1 volt drive is about right and with 4 volt drive the cutoff frequency is too high.  The only difference is the ac source voltage.

Those are nice looking plots. You can clearly see at what frequency the capacitor stops being a capacitor and how good of a capacitor it is.

[mawyatt]

@ Grandchuck,

Be sure and check out these posts here, especially TopQuarks posts starting at #4. The simple fixture he created allows use to ~100MHz with DSO and AWG.

https://www.eevblog.com/forum/testgear/capacitive-impedance-plots-with-sds2104x-plus-bode-function/msg4335745/#msg4335745

Also note the results we achieved (#1) when compared to a Lab Grade LCR Meter (TH2830), impressive for a DSO based measurement IMO.

Of course as always YMMV,

Best,