Advice on in-circuit testing of film and ceramic capacitors

[ultrageranium]

Hello everyone,

Until now I used a MESR-100 to do in-circuit testing of bad/dodgy electrolytic capacitors. It's been working well but in the meantime I've started to read more and more about capacitors, looking up their datasheets, etc, and am now considering upgrading to a small LCR meter which would allow me to test quite a few other things and control the input frequency as well. I don't want to spend too much on it, so I'm looking into the DER EE DE-5000 at the moment.

My question though, is that would such a device allow me to also do in-circuit testing of film and ceramic capacitors? Does not have to be very accurate, but accurate enough to spot components that may have gone far out of specs (granted that they are less likely to fail compared to electrolytic ones). Alternatively, what would be a good way to do this? And finally, is the DER EE DE-5000 still the best option for a €100-€130 budget?

Thanks!

[TimFox]

I am very happy with my DE-5000, but I have not used it for in-circuit checks.
One possibility (with the unit under test powered down and all DC voltages discharged) would be to connect the Guard connector (4 mm banana jack) to several points of the circuit under test that surround the component under investigation, to remove them from the measurement.  This would depend very much on what your circuit is.

[ultrageranium]

Thanks for sharing your experience TimFox. I'm very tempted to buy a DE-5000.

I'm just really puzzled about the in-circuit checks. That's the only thing blocking me to order. Browsing through the forum, I read that it is not designed for this, but I struggle to understand if it's a safety issue, ie would discharging the cap under investigation limit the risk of damage), or if it's a design issue, ie the reading will be wrong. I'm asking because it seems some people are using it anyway for in-circuit testing, so this is super confusing.

I'm also struggling to find info about film and ceramic cap testing. With electrolytic caps accounting for the vast majority of issues with caps, it's quite difficult to find resources for testing such less problematic components.

[TimFox]

There is a definite safety issue, since the DE-5000 is not well protected against external voltages and pre-charged capacitors.  Connecting to a powered circuit is bad.
For film and ceramic capacitors, the Q value is a good test parameter.
NP0/C0G ceramic capacitors and polypropylene or polystyrene film capacitors have typical Q values that may be off-scale for the DE-5000 (not a hazard).  If such a capacitor shows a Q value around, say, 100 or so, then it is either a lesser-quality dielectric (e.g., X7R ceramic or polyester/Mylar film), or it is defective.
Similarly, you should see a reasonable Q value on the "lesser" dielectrics.
Q, ESR, and D are just re-calculations of the same measured "minority" component of the capacitor (the majority being the reactance).

[ultrageranium]

There is a definite safety issue, since the DE-5000 is not well protected against external voltages and pre-charged capacitors.  Connecting to a powered circuit is bad.

I see, but assuming I will discharge the capacitors and of course having the circuit off, will it operate fine or at least good enough to spot defective capacitors?

For film and ceramic capacitors, the Q value is a good test parameter.
NP0/C0G ceramic capacitors and polypropylene or polystyrene film capacitors have typical Q values that may be off-scale for the DE-5000 (not a hazard).  If such a capacitor shows a Q value around, say, 100 or so, then it is either a lesser-quality dielectric (e.g., X7R ceramic or polyester/Mylar film), or it is defective.
Similarly, you should see a reasonable Q value on the "lesser" dielectrics.
Q, ESR, and D are just re-calculations of the same measured "minority" component of the capacitor (the majority being the reactance).

Thanks, this is helpful!

[PA4TIM]

The best way to test caps is not in circuit. That is often useless. Often a lot off them are parallel to each other (over the PSU rails so also parallel to elcos) The only good way is desoldering and then measure C, D (or Q that is 1/D) 

The DE-5000 is a nice meter, I had one for a few years but it died a few months ago (Spontaneous while powering it on) I did not like the testleads/pincet, I lost count the number of times I had to repair those. I now bought a bench LCRZ meter.

[TimFox]

I also don't like testing capacitors in-circuit, but it would be interesting to try a guarded measurement, since the DE-5000 has a guard terminal.
I built my own test leads and fixtures, starting with the stock leads to get the interface, then using better alligator clips or binding posts.
Note that the DE-5000, along with most bench LCR meters, is not designed to work with a DUT that has one side grounded, but the battery-operated DE-5000 can be floated on a plastic box, since the voltages it generates are safe.

[ultrageranium]

The best way to test caps is not in circuit. That is often useless. Often a lot off them are parallel to each other (over the PSU rails so also parallel to elcos) The only good way is desoldering and then measure C, D (or Q that is 1/D) 

I also don't like testing capacitors in-circuit, but it would be interesting to try a guarded measurement, since the DE-5000 has a guard terminal.

But are we talking about precise measurements or just good enough approximations? I'm asking because my use case is as follow: I'm slowly learning electronics so a DE-5000 could fit nicely in the list of handy yet affordable equipment to acquire sooner or later. It would be great to use it for precise measurements on the breadboard as I learn. But next to that I'm also applying my meager new skills for simple fixes/maintenance of old (80s) synths and computers, so I'm often replacing capacitors. Practically speaking what is the best practice to find dried or out of specs capacitors when they already passed the visual test? I would not need very accurate results, but something that would hint already that something is fishy.

Said differently, would I gain anything from upgrading my MESR-100 to a DE-5000 for in-circuit quick scan of bad capacitors?

[bdunham7]

I have a BK 886 that I can use for in-circuit testing.  Its advantages are that you can select frequency (100/120/1k/10k/100k) and level (50/250/1000mV) and it is accurate.  The 50mV level setting allows you to test circuits without meeting the threshold of conductance for most semiconductors.  The disadvantages are that they are expen$ive, a rather old design and AFAIK, are just as susceptible to being blown up by external voltage as anything else.

In-circuit testing can be handy, but frankly I don't use it that way all that often.  I suppose it depends on what you are fixing---if you find that in-circuit testing is frequently helpful, then the BK 885 or 886 would be a good option.  Just remember to discharge!

[TimFox]

The DE-5000 is a portable unit for high-accuracy measurements of components, and is not designed for in-circuit tests.  In-circuit testing is not a good idea if precision measurement is required of a component connected to other stuff, but if you want to test the total impedance at the input terminal, or whatever, then you just have to worry that the DE-5000 terminals (as with other LCR meters) are not grounded, so you need to float the meter (easy, since it is battery powered).
Normally, in-circuit testing is used for troubleshooting, where you are trying to identify a failed component.