Measuring capacitance with an oscilloscope

Thread Starter

daanmicro

Joined Nov 15, 2017
50
Hi there,

currently i'm trying to determine the capacitance of capacitor in the most accurate way, with my material availabe. Right now i'm using a method where you can determine the capacity (determine Z --> alfa --> Z = Zcos(alfa) + jZsin(alfa) etc.) of a capacitor in an RC circuit. I'm measuring the phase shift, frequency and voltage magnitudes.

upload_2018-6-27_11-15-44.png
The thing is i'm using a Tektronix TDS 210 and i want to determine the phase shift more accurate. This oscillosocpe is not capable of measuring this by itself. I can use the delay time and period, using the cursor option, but i get the idea this is very inaccurate. Another restriction is that i can't use cursors in the XY-mode of the oscilloscope.

Any of you have any ideas for other options to determine the phase shift (or even another way to measure capacity with an function generator, oscilloscope, breadboard and alll kinds of resistors/capacitors)??

Thanks in advance!
 

Thread Starter

daanmicro

Joined Nov 15, 2017
50
Thanks! But because i'm on a pretty short time frame; Does one of you know a better way to measure the phase shift with my available instruments?

Kind regards.
 

dl324

Joined Mar 30, 2015
10,072
currently i'm trying to determine the capacitance of capacitor in the most accurate way, with my material availabe. Right now i'm using a method where you can determine the capacity (determine Z --> alfa --> Z = Zcos(alfa) + jZsin(alfa) etc.) of a capacitor in an RC circuit. I'm measuring the phase shift, frequency and voltage magnitudes.
Why not just charge the capacitor with a current source and measure the voltage after a known time?
 

PeteHL

Joined Dec 17, 2014
304
Are integrated circuits allowed? If so, configure a 555 timer as a square wave oscillator including the capacitor of unknown capacitance, and measure frequency of oscillation. The frequency of oscillation depends inversely on capacitance of the capacitor.

If ICs are not allowed, you can use a capacitance bridge made up of two resistors and two capacitors.

"If two arms of the bridge are capacitances and two are resistances, balance is achieved at any frequency if the capacitances have the same ratio as the resistances" -Norman Crowhurst, Basic Electronics Course, p.198.
 
Last edited:

Thread Starter

daanmicro

Joined Nov 15, 2017
50
Thank you guys for all thinking along with me!

Sorry, i should indeed be more specific.

My guess is the range will be from 100 nF to 100 uF. (i know this is everything but specific:rolleyes:)

So the purpose of this small measurement setup, is to measure (inkjet printed) capacitors. Considering the non-linearity at higher frequencies, i'll have to be able to measure at different frequencies. The desired accuracy is <5%.

At first instance i wanted to apply a sine wave on a series RC-circuit (and i still kind of want to do this). In this way, when comparing I (or the voltage across the resistance) to the capacitor's voltage (knowing the phase shift accurately), an accurate measurement of the Capacitance can be made.

So that's why i was wondering if there are any other ways to measure the phase shift or capacitance within this 5% accuracy. I'm kinda on short time, so that's why i prefer using my currently available setup.

Thanks in advance.
 

MrChips

Joined Oct 2, 2009
20,335
A simple technique is to incorporate the unknown capacitor in an oscillator circuit such as a 555-timer circuit. Measure the period or frequency of the oscillator's output waveform in order to determine the capacitance in the circuit.
 

danadak

Joined Mar 10, 2018
3,911
The problem with timer approach is meeting the 5% error budget
when taking into account all errors, R tolerance, threshold error,
PSRR, T, Vcc.....

Then there is issue how do you interpret C vs F measured using this
approach. Given the parasitics in a C.

You would have to cal it to get meaningful accuracy.

Bridge probably best approach.



Regards, Dana.
 
Last edited:

MrChips

Joined Oct 2, 2009
20,335
The problem with timer approach is meeting the 5% error budget
when taking into account all errors, R tolerance, threshold error,
PSRR, T, Vcc.....

Then there is issue how do you interpret C vs F measured using this
approach. Given the parasitics in a C.

You would have to cal it to get meaningful accuracy.

Bridge probably best approach.



Regards, Dana.
He wants to measure capacitance with an oscilloscope.
There goes the 5% already.
 

danadak

Joined Mar 10, 2018
3,911
This approach will take you to 3Khz using sound card, author Stebber.

http://wb6dhw.com/RLC_Meter/A Low Cost Automastic Impedance Bridge.pdf

Program attached, run it in compatibility mode right click .exe file, choose XP, SP3.

He wants to measure capacitance with an oscilloscope.
There goes the 5% already.
Depends on scope, some modern DSOs will do very precise timing/phase
measurements, and amplitude. Like 11403A, 1% vertical accuracy, time
interval .002% +/- 100 ps.

Regards, Dana.
 

Attachments

Last edited:

Thread Starter

daanmicro

Joined Nov 15, 2017
50
Yeah i ran into that soundcard prog. already. Looks useful, although i already own a pretty capable function generator.

Unfortunately my scope isn't capable of measuring phase.

At this moment i'm considering to buy a precision inductor and measuring with an tank circuit and freq. sweep. In my idea this should be pretty accurate as well. (depending on the accuracy of the components and function generator ofc.).
 

MrAl

Joined Jun 17, 2014
7,185
Hello,

After reading the posts here it looks like the best way to do this is to create an oscillator and use the cap under test as one of the frequency determining components. You can not use a 555 chip however because that works the cap in its exponential charge/discharge mode, and that's not the same measurement as using a sine wave, which brings us to the next point which is the oscillator must be a sine wave oscillator.

This still presents some problems because you may also want to measure with different AC voltages. You'd have to have a way to vary the DC supply to the oscillator for example so the AC voltage across the cap changes.

The other way this is done is to use a bridge circuit where you use a known capacitance and compare it to the unknown capacitance.

Measuring phase shift with the scope is only so accurate, that's the way it goes. If you want to measure phase shift better you can set up a digital circuit with comparators and a time base and frequency counter that measures pulse widths. It's not that easy to construct but you can do it. You basically detect the zero crossing of the reference sine and the zero crossing of the phase shifted sine and that generates a pulse, and you measure the pulse width with the frequency counter set to measure pulse widths (or perhaps duty cycle).

Also you did not give values for f and R in your example, nor explain the method well enough and why you need two voltages.
 
Last edited:

Thread Starter

daanmicro

Joined Nov 15, 2017
50
Hi,

Thanks! I now understand the phase shift measurements won't get better than this. So i ordered a few inductors (1% tolerance). Because i need to have accurate measurements this week, this looked like the best alternative to me. I chose inductors with a quality of min. 60, which in my idea would be sufficient. (still a noobie)

Now when i sweep the frequency, i can find a pretty solid peak of the amplitude. Unfortunately when i calculate my capacitance values, i continously get values around 10-20 percent lower than the actual value. My guess (which might be obvious) is that this is due to parasitic capacitance/inductance. If i'm right, how could i improve this result? I'm using a BNC cable which is directly connected to my oscilloscope and to a breadboard with a parallel inductor and capacitor. I can't imagine these parasitics would have such a big influence on the calculated values.

For example:
upload_2018-7-3_13-55-36.png
While the actual value is 10uF with 5 % tolerance (multimeter also gives 10.03 uF)


Thanks a lot.
 

MrAl

Joined Jun 17, 2014
7,185
Hi,

What is the tolerance on the inductor, is it really 1 percent?
And you say the cap is 5 percent?

What is the value of the inductor? That would help determine if it was parasitic inductance messing up the measurement.
A 10uf cap is not likely to be influenced by parasitic capacitance.

What is actually driving this combination of components?
 

Thread Starter

daanmicro

Joined Nov 15, 2017
50
Yeah the datasheet says 1 percent. And the capacitor is 10%, sorry my bad.

These are the components:

https://nl.mouser.com/ProductDetail/810-FA20X7S1H106KRU6
https://nl.mouser.com/ProductDetail/807-4470-13F

I indeed couldn't imagine the influence could be that big. I just don't see what's wrong here. Another question, if i use crocodile clips for example and make sure the components are hanging free in air. Would this have influence? Right now i have the components just straigth plugged into the bnc connector.

I'm having a hard time to gain insights in these parasitic properties.

Inductor value is 10 uH 1%
 
Top