How to Use LCR Meter to Check if PSU Needs Recapping

[alligatorblues]

How does one use a LCR meter to check if a PSU requires recapping?

[beatman]

unsolder caps and measure with lcr

[MathWizard]

What type of PSU, linear or switch-mode ? The main caps to check are the electrolytic's, soup can shaped ones. A lot of times that means they need to be de-soldered, or if you are lucky, you can just de-solder and raise 1 leg free.

The ones that get the most punishment will usually be on the AC to DC power-in area, and the output filtering caps. But any of them can get overheated and age in some PSU's, near the wrong part.

Usually people check the ESR at 100kHz, and measure tan delta, at 1kHz, and datasheets have that info, but not the ESR. The tan delta is ratio of the ESR / the reactance tho.

Then there's DC leakage testers, and high voltage testers. Depending on how old the caps are and the type, some people would throw all of them out.

[TERRA Operative]

In switchmode power supplies, the output filter capacitors are the ones that cop a good hiding with all the ripple.
I usually check those first and replace with low ESR caps with high ripple current rating.

To test, I check capacitance and ESR, preferably at working voltage if your tester supports applying a DC bias.
But be aware that bad caps can sometimes read high capacitance values...
Capacitors are tested by charging the capacitor then discharging through a known resistance and the time to charge and discharge is used to calculate the capacitance.
If the capacitor is leaky, the altered charge/discharge times can result in an incorrect reading.

The easiest way to test leakage is to hook the capacitor to a current limited bench power supply set to the working volage of the capacitor and a couple mA of current.
Connect your leads through a multimeter set to mA DC, and measure the current as the capacitor charges.
It should quickly drop to a very low value as the capacitor charges. Good capacitors will have a leakage of tens of uA, even single digits for high quality caps.

If it's more that a few mA of leakage after a minute or two, replace the cap.

If one cap in a group of caps is bad, replace the lot. If one is bad, the rest won't be far behind...

[inse]

Not all types of LCR meter are suitable to check the electrolytic caps.
It needs to be able to measure the ESR as the capacity alone is not sufficient to decide on the quality of an electrolytic cap.

[Wallace Gasiewicz]

I think Terra has a really good point.
Big PS caps often leak in failure mode if they do not self destruct
It is important to check for leakage more so than ESR
His setup is nice suggestion
I use an old Heathkit cap tester for old caps that are rated in hundreds of volts
It can be surprising to see that a nice looking cap with good ESR is leaking badly.

[David Hess]

How does one use a LCR meter to check if a PSU requires recapping?

Usually I already know that the aluminum electrolytic capacitors need to be replaced, and I will replace all of them except for the bulk input capacitors unless I suspect they are also worn out.

But for testing, I use an LCR meter to measure the capacitance and dissipation, and sometimes ESR if that is specified, to evaluate the health of the capacitor.  Capacitance and dissipation are usually enough to decide.

[alligatorblues]

These PSUs are in HP branded 34401As. I think it would be best to replace the lot with Panasonic low ESR units in both DMMs. Much easier than testing each cap. And I usually purchase at least 10 caps of anything I don't have in stock, so the price isn't bad. Both DMMs function normally, so there's no rush.

I keep unused electrolytics for longer than shelf life. Before I use one, I form it with a DC supply. Anyone else do that? Also, when I purchase used DC supplies, I load them, and bring them up in cycles with a Variac (i.e. 0-10V, 0-15v, 0-20V ...., for five minutes per cycle), for the sake of the electrolytics inside, to build the dielectric if the unit might have been sitting for five years. I've got a shelf full of large resistors for the job, up to 1000W. Any wisdom to this practice?

One other question, I talked to a guy in Florida, and he ups the capacitance by one size, because newer caps have different plus/minus tolerances that can make a difference in older equipment. Any thoughts?

[David Hess]

I keep unused electrolytics for longer than shelf life. Before I use one, I form it with a DC supply. Anyone else do that? Also, when I purchase used DC supplies, I load them, and bring them up in cycles with a Variac (i.e. 0-10V, 0-15v, 0-20V ...., for five minutes per cycle), for the sake of the electrolytics inside, to build the dielectric if the unit might have been sitting for five years. I've got a shelf full of large resistors for the job, up to 1000W. Any wisdom to this practice?

I do not think that matters except for high voltage capacitors, like you would find in a tube design, so I do not bother with "modern" equipment.

One other question, I talked to a guy in Florida, and he ups the capacitance by one size, because newer caps have different plus/minus tolerances that can make a difference in older equipment. Any thoughts?

For old capacitors the plus side of the tolerance is often 100%, so going to the next higher 20% tolerance is still within the specifications and should make no difference.  I usually select the capacitor based on physical size and lead spacing first, and then use a higher voltage part to get something physically larger.