Measuring capacitance of leaky capacitors?

[Agent24]

I've recently built a capacitance meter, and am wondering what would happen when measuring a leaky capacitor?

By that mean a capacitor which does not hold charge well, not one that is literally leaking electrolyte.

I read someplace that on most testers the test voltage (3v, as mine is) is too low to show up the problem and so the capacitor reads with the right value but may still be bad.

If this was not the case (capacitor was very leaky even at low voltage) what would the outcome be?

Reduced capacitance reading or just nothing?

[Longbow]

Re: Measuring capacitance of leaky capacitors?

I've recently built a capacitance meter, and am wondering what would happen when measuring a leaky capacitor?

It is not possible to predict the readings. If you have already identified a leaky capacitor, then there is no point in taking a capacitance reading, since it has to be replaced. Your ESR meter will give you the information you need when troubleshooting electrolytics. If electrolytic caps are used as coupling, then your DVM is the best tool, used in-circuit and powered-up. (Obviously, it is of great help to know what type of circuit you are working on.) Once in a while the value of a non-electrolytic may be in question - that is where your capacitance meter may be of assistance. Otherwise, remember that all electrolytics lose capacitance over time depending upon the original uf value, so an actual capacitance reading in itself will not give you much information. And finally, a capacitance meter must be used on a part out-of-circuit (or at least one leg desoldered).

[Agent24]

Re: Measuring capacitance of leaky capacitors?

I don't actually have any known leaky capacitors, I was just wondering if there was a predictable outcome on the meter which might give a heads up if one was leaky.

I guess other methods must be used, as you say.

[PCBONEZ]

Re: Measuring capacitance of leaky capacitors?

To check leakage current apply rated DC voltage and measure current through cap after it's charged fully.
[With the voltage left applied.]

.

[PlainBill]

Re: Measuring capacitance of leaky capacitors?

I don't actually have any known leaky capacitors, I was just wondering if there was a predictable outcome on the meter which might give a heads up if one was leaky.

I guess other methods must be used, as you say.

It depends on the meter, the method used to determine capacitance, and how leaky it is. With the old capacitance - inductance bridges leakage would not affect the reading. Most DMMs charge the capacitor with a fixed current and determine how long it takes to rise to a set voltage. In all cases the meter will indicate a capacitance higher than the actual value - how much higher will depend on how bad the leakage actually is. As a practical matter, that error may be insignificant. While I haven't determined this with a scope, my meter seems to charge the cap up to less than 10 mV.

PlainBill

[Longbow]

Re: Measuring capacitance of leaky capacitors?

To check leakage current apply rated DC voltage and measure current through cap after it's charged fully.
[With the voltage left applied.]

This is a bit misleading in my opinion. It should be noted that perfectly good electrolytic capacitors do have a certain amount of leakage current. That's just the physics of these parts. In most circuit applications, the small leakage current can be ignored, but it is there. The measurement of leakage current on any kind of cap is rarely if ever done except by the manufacturer. Troubleshooting capacitors that actually "leak" voltage, on the other hand, is done with a voltmeter with the part in-circuit.

Here is an example: take a vintage audio amplifier that uses tubes. There will be mylar or polyester capacitors in the range of .047 - .1 ufd used as coupling caps between the audio stages. These caps might operate with a working voltage of 50v up to 400v. The plate (input) end of the cap might be at a dc potential of 350 volts, while the output side of the cap will usually be around 0 volts because it will be feeding signal to the grid of the next stage. Once in a while, you might see 50 or 100 volts on the output side of the coupling cap. That is wrong. By lifting up the output leg of the cap, you will then see the 50 to 100 volts or more on the unsoldered leg. That voltage is due to the faulty cap, which should act as an insulator at dc. This voltage leak is of course measured between the open leg of the cap to ground, not to the opposite leg of the cap. At this point, knowing the actual capacitance value of that faulty capacitor is irrelevant.

In that case, you have located a leaky capacitor. This is not the same as an electrolytic that is physically leaking electrolyte onto the pcb. To summarize, good electrolytics by their nature have a finite leakage current. That current is rarely a problem if the cap is used as a power supply filter. In other circuits, the normal leakage current might upset the operation - in which case a low leakage cap such as a tantalum might be used.

Some typical uses of a capacitance meter: 1) a circuit works, but not exactly right - could be a change in value of a capacitor. As a last resort you start checking values. 2) you have a box of miscellaneous parts from scrapped pc boards with no markings - measure those unknown part values with your capacitance meter and put them in the correct drawer.

[PCBONEZ]

Re: Measuring capacitance of leaky capacitors?

This is a bit misleading in my opinion. It should be noted that perfectly good electrolytic capacitors do have a certain amount of leakage current.

How exactly is that misleading?
And from exactly what vacuum did you shit the idea that there will be no current when you measure a good cap?
.

The measurement of leakage current on any kind of cap is rarely if ever done except by the manufacturer.

Not even close to the truth.
Hobbyists may not do it, but -real- Electronics Technicians do.
In fact there are testers made specifically for that purpose and most -real- Electronics shops have one.
.
TC does. ... His Sencor Z-Meter has the function.
.

[Toasty]

Re: Measuring capacitance of leaky capacitors?

[severach]

Re: Measuring capacitance of leaky capacitors?

Capacitor leakage is equivalent to a parallel resistor. If the resistance is low enough the Atlas ESR60 will say "In Circuit" and not show a uF value. If the cap isn't in circuit that's a clue that the cap is too leaky. A few seconds on a bench power supply will prove it.

[PCBONEZ]

Re: Measuring capacitance of leaky capacitors?

Yes, and in a lytic if you don't have at least a little leakage the cap can't rebuild the oxide layer after the ripple blasts pits in it.
Leakage is what makes lytics 'self healing' caps.
.

[Tube_Dude]

Re: Measuring capacitance of leaky capacitors?

Capacitor leakage is equivalent to a parallel resistor. If the resistance is low enough the Atlas ESR60 will say "In Circuit" and not show a uF value. If the cap isn't in circuit that's a clue that the cap is too leaky. A few seconds on a bench power supply will prove it.

Yes. Thats why I like the Atlas ESR60 so much...