MEASURING CAPACITORS

[MrsR]

Hi! Guys,
I need to read the value of some Electrolytic Capacitors. (There showing their age) actually There on a RPM meter that reads the shadow of a propeller blade. I have checked every other component so that leaves the Caps.
So can someone recommend a method to read the Capacitors.
I have a DSO  a DMM and a variable Voltage supply.

THANKS GUYS
Rachael

[dannyf]

Drive the capacitor via a large resistor (10k or 100k) from the 1KHz source on your scope. The voltage changes on the capacitor is proportional to its capacitance.

[kfitch42]

w2aew


and mjlorton


have made videos about this.

[Mr. Coffee]

I'm not sure if you can fully test the capacitors (that is, ESR) with just a DC power supply...usually also need a signal generator for that (gets complicated), or else a dedicated tester (easy).

You can test the 'lytics capacitance if you have an assortment of resistors along with the power supply.
The time constant for charging a capacitor is its capacitance in farads, times the series resistance in ohms, that you are charging it through. That is, R x C. It takes 5 time constants to full charge a capacitor though a resistor.

The easiest format is to use a high impedance analog meter if you have one, but a DMM can do it. Discharge the capacitor fully, and connect the meter across the capacitor. The capacitor is then hooked to a DC supply through a high value resistor. (Mind polarity of the capacitor and the analog meter, and don't exceed the voltage ratings of either.)

You turn on the power count the seconds until the voltage stops rising across the capacitor indicating that the cap. is fully charged, and divide that by 5 times the ohms of the resistor. That is the capacitance in farads.

For testing leakage, you hook a 10M ohm DMM in series with the discharged capacitor and the power supply. Read the voltage and use ohm's law to determine the series resistance of the capacitor's dielectric. The power supply should be set to close the working voltage of the capacitor, but not more. (Mind polarity as always.)

[G0HZU]

W2AEW and MJL are both excellent presenters but I think you can go a stage further in the testing if you have access to a function generator and a typical modern scope.

I.e. you can measure both the capacitance and ESR across frequency by using a simple sense resistor and a 2 channel scope and a clean sine wave generator (plus some basic trig sums in an excel spreadsheet)

If I had a large/dodgy electrolytic cap then I would want to know the ESR vs frequency and also the capacitance vs frequency.
This method I am suggesting is decades old. You basically measure the ac voltage each side of the series sense resistor and measure the phase angle between them with the scope.

Then do some sums and you get the complex impedance of the cap at that test frequency. You can also find the self resonance of the capacitor. It takes about the same amount of effort to do and makes much better use of the test equipment as you get to measure the impedance of the capacitor vs frequency

Most caps fail with a trend of increasing ESR so this is a much more thorough test method

[Mr. Coffee]

Most caps fail with a trend of increasing ESR so this is a much more thorough test method

This is quite true.

[dannyf]

I'm not sure if you can fully test the capacitors (that is, ESR) with just a DC power supply

Obviously you cannot. And I would be willing expand that statement further, to say that if you cannot fully test anything with a finite number of equipment: there are simply infinite ways a device can fail, or an infinite ways to quantify a device.

But, capacitors have to act like capacitors (ie they have to have capacitance) first, before any further measurements are necessary to deem them failures. In that sense, it is perfectly ok to establish if a capacitor has (sufficient) capacitance before attempting any other tests.

[MrsR]

Thanks Guys!
I have a little test board using a 555 to generate a 1KHz square wave. It works well with small Ceramic and Monolithic. but I can't get an accurate reading above 1uf. I would need a slower time base. I do have a good analog Multi Meter so will use that method, at least I understand it. lol........

Thank you all for the suggestions
Rachael

[dannyf]

I can't get an accurate reading above 1uf.

It can be too fast or the current too small -> the voltage rise is too small to be easily measured.

[DmitryL]

I'm not sure if you can fully test the capacitors (that is, ESR) with just a DC power supply

But, capacitors have to act like capacitors (ie they have to have capacitance) first, before any further measurements are necessary to deem them failures. In that sense, it is perfectly ok to establish if a capacitor has (sufficient) capacitance before attempting any other tests.

BTW, you can easily measure capacitor's capacitance and ESR using a constant current source. And there are some simple microcontroller - based projects that do this.

Another thing that electrolytic capacitor's ESR may depend on how you measure it (on AC, DC of various frequencies etc.) On the other hand, who needs the exact value of ESR if it is used to sort caps on fail/pass basis...

[Conrad Hoffman]

Here's another trick, though it's a bit tough to get quantitative data. It's more of a signature method. Use your 555 square wave generator and drive the capacitor through a 1kohm resistor. With a sensitive scope, look at the voltage across the capacitor. Adjust the frequency (depends on the capacitance) until you see a triangle wave. For a given setup, the amplitude of the triangle will indicate the capacitance, and the sharp offset seen at the peaks will indicate the esr. The easiest thing to do is set up with a known good capacitor of similar value, then sub in the questionable ones. This method is very revealing with caps in the tens to hundreds of uF.

The underlying principle is simple. You're seeing how good an integrator the cap makes. A perfect cap would convert the square wave to a triangle wave. Ideally this would take a constant current source, but the 1kohm resistor is close enough for the voltage swings involved. If the cap has some non-linearities, the triangle will have curved sides (try a tantalum). As the esr increases, the reversal (which is a high frequency component) becomes lossy, and you'll see a noticeable offset.