whiz115

Somebody told me that capacitors with leakage tend to give higher reading than their actual capacitance, is that true?

thanks!

mvs sarma

Try to connect a high value resistor across (creating leakage deliberately)and see you would have first hand knowledge.

__________________
Regards,
Sarma.

Nigel Goodwin

Depends how you measure it, if you're applying a current to the capacitor and timing how long it takes to charge, then a leaky capacitor would take longer to charge - so give an artificially high reading.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

Hero999

Some multimeter measure the impedance of the capacitor at a certain frequency. The lower the impedance the larger the capaciticance. Therefore, if the capacitor is leaky, the impedance will be lower than expected which will give a correspondingly high reading.

__________________
What's so bad about Microsoft?

Get Opera it's simply a superb browser.

whiz115

mvs sarma thanks man!

herro999 there are other methods other than measuring the impedance?!

Nigel Goodwin

I've just bought one of these at work

http://www.peakelec.co.uk/acatalog/jz_esr60.html

Measuring capacitance isn't much use, it's ESR you need to test, and this does both.

A great little device, and PIC based as well

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

HS3

Come on! That's a silly thing to say! ESR can be an important spec just like admissible voltage and others but if there is one spec you need to choose to define a capacitor it is its capacitance.

That is a neat instrument there but a bit too expensive for amateurs and students. I have built myself a very simple ESR meter which is good enough for everyday testing in that it gives you a rough idea. When repairing power supplies (the electrolytics tend to go like crazy) I can run a quick check with the capacitors in-circuit and this is a great first test.

blueroomelectronics

Care to share the schematic?

__________________
Bill
Smart Kits build Smart People
http://www.blueroomelectronics.com

Nigel Goodwin

That's quite simple, you read the value off the side of the capacitor, why would you want to measure it? - a faulty electrolytic may well read the same capacitance, but it's ESR will have increased making it useless. Measuring capacitance of an electrolytic doesn't test it in any worthwhile way.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

Dean Huster

If we take that to the nth degree, the National Institute of Standards and Technology (NIST, formerly the National Bureau of Standards) should shut down their division on capacitance standards. Somewhere in the chain, SOMEONE has to measure the capacitance, whether it's the QC department of the cap manufacturer or their design division.

One excellent excuse for testing for capacitance that I do is sorting caps for the range capacitors on a function, pulse or signal generator. They are sorted according to the significant digits of the value for each decade. You install caps that have the same significant digits regardless of their actual value. That way, the single frequency-calibration adjustment will be good for all ranges and not just the one range that's used for the adjustment. If you're using an analog dial (or better yet, restoring an older generator), the overall accuracy of the instrument is improved by a full magnitude that way. Yes, you can use a counter to measure the output frequency in actual use, but for many applications (alignment of an antique radio comes to mind), you don't need counter accuracy but would be content with better than off-the-design-shelf accuracy for an older instrument.

Dean

__________________
Dean Huster, Electronics Curmudgeon
Contributing Editor emeritus, "Q & A", of the former "Poptronics" magazine (formerly "Popular Electronics" and "Electronics Now" magazines).

R.I.P.

Nigel Goodwin

Manufacturing is obviously a completely different application

The exception that proves the rule perhaps?

But notice that in either of those two cases, measuring the capacitance doesn't tell if the capacitor is any good or not - although during manufacture they test ESR as well anyway (which is where the ESR spec for the capacitor comes from).

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

HS3

This is symplifying things a tad too much for my taste. Why would I want to measure the capacitance of a capacitor? Well, for the same reasons I might measure the actual value of a resistor or an inductance or the performance of an active component or the length of a piece of lumber even though all that information is given by the manufacturer. Maybe because the value given has a wide tolerance and is only nominal and I want to have more precission. Maybe because I want to confirm that the component actually meets the specs and is not bad. It can be mislabeled or the label may not be readable. If I have a box of recycled electrolytics I am more interested in knowing their capacitance than their ESR. A capacitor known to have 10 uF and of unknown ESR can still be used in applications which do not require low ESR. A capacitor of unknown capacitance is useless.

ESR has become more of an issue in recent years for two main reasons. One is that old electrolytics were huge by today's standards and had lower ESR. Miniaturization has brought higher ESR and new technologies to lower it are needed. Another is that old power supplies worked at very low frequency and with very low ripple current per capacitance unit. SMPS have brought a huge increase in current per microfarad and this requires lowering the ESR to a minimum. Cheap capacitors with higher ESR may not be suitable for certain applications like high frequency SMPS but they can be perfectly good for other uses and I need to know the capacitance.

I would still say that in the beginner's toolbox it makes more sense to have a capacitance meter than an ESR meter and capcitance meters are cheaper and more plentiful anyway. Note also that capacitance value depends on several factors like bias voltage so it is not like it is a fixed, immutable value. A capacitor which is nominally 1 uF at 200 V will have a very different capacity at 10 V.

Also, for the purpose of repair and general maintenance, the technician does not really need to measure ESR with much precission because just an order of magnitude is enough. As I say, I have a very basic instrument and I have learnt to tell good from bad capacitors pretty quick even though I have no idea of numerical value of ESR. Basically 95% of caps fall into two main categories (a) Needle hardly moves (high ESR) so cap is bad or (2) needle goes to end of scale (low ESR) so cap is good. It is rare that I get a middle of scale reading and then I need to take a closer look. If cap is of very low capacitance then it might be passable, otherwise I discard it. Also, it depends on use. If it will go in a high frequency SMPS then I want to see the needle go to the end of the scale but if it will go in some low-current application then it may be perfectly good. The thing about ESR is that once a cap begins to go bad the use will make it quickly spiral and ESR shoots through the roof.

I have repaired many SMPS which were bad due to bad caps. Generally I will replace them all because if a bad brand was used then I want to replace them before they fail again.

For normal repair work you do not need a precission ESR meter. A basic instrument which gives a general idea is enough.

You can find it here.

Nigel Goodwin

I've never heard of that, and never found it so - certainly a 1uF 200V capacitor will read 1uF (within tolerance) at the low voltages used for measuring.

What does change massively is ESR (sorry to mention it again), high voltage capacitors generally have really rubbish ESR values.

I agree, I've been using a simple home made one for years, and only just bought a cheap professional one the other week.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

HS3

Some people think that replacing a capacitor with another one of the same capacitance but rated for higher voltage and higher temperature is always a good thing and that it will last longer. But higher rated voltage and higher rated temperature nearly always mean higher ESR which leads to higher temperature and shorter life. What is generally true is that it is good if you replace a capacitor with another one of higher capacitance because this means lower ESR. Within limits and space constraints I generally replace SMPS caps with somewhat larger values. And if space permits I will install two caps in parallel rather than just one. This also lowers ESR.

I have read (but not precisely confirmed personally) that caps rated for higher voltages will not keep their capacitance value if continuously biased at much lower voltages because the higher voltage determines the thickness of the oxide dielectric. The idea is that the dielectric degrades over time and needs the voltage to keep in in condition.
The formation voltage used is typically 3-4 times higher than the rated voltage. The idea is that the dielectric degrades over time and needs the voltage to keep it "formed". This is to say that it is better for a capacitor to remain biased than to sit unused. I have also observed that a capacitor which has been unused for a long time has higher leakage current which gradually diminishes as the capacitor is "re-formed" by the voltage applied.

whiz115

HS3 from what i understand nigel speaks about diagnosis of a faulty cap.

what you're explaining are very interesting but for a fast diagnosis of a faulty component i think an esr meter is the way to go, if the capacitance is not always going down but ESR goes, then what's the point playing with a capacitance meter while you only want to know degradation of the components?!

if you tell me that capacitance goes always down while the caps are aging or get misused, then i think i'll forget about ESR meter, because i'm not interested for anything else other than testing if an on circuit capacitor is fine.