boom: https://www.youtube.com/watch?v=7Br3L1B80ow

As Dave says in the video, that should be a good option for pass/fail testing and rough component identification. When caps go bad, their ESR values tend to explode, so while you wouldn't be able to differentiate between, say, Panny FR and FM series based on readings alone, you could probably tell if the cap has failed or not.

Quality modern LCR meters tend to be pricey, but you can find good prices on eBay for older analog gear. Using a double dial bridge meter should be a rite of passage for everyone.
https://www.ebay.ca/itm/Genrad-1650-LCR-Bridg … XAAAOSwRbtaEj-x

Ozzuneoj wrote:

Being able to tell if they need replaced or not

I rest my case.

But seriously, once they are out of the motherboard, you might as well replace them with new.

luckybob wrote:

But seriously, once they are out of the motherboard, you might as well replace them with new.

Exactly.

As the capacitors will be connected to other components as well, you cannot measure a single capacitor when it is in the system.
For example in CPU VRMs there are many output capacitors in parallel.
If there is a problem in one of them, you don't know which one if you just try to measure them in-system.
So that is why they need to be removed from the motherboard before measuring them individually.
So if the old ones are bad then you need to put new caps in for sure.
If the old ones are good, then half of the job is already done and you can still put new ones in.

But at least you know that if old caps were good and it does not work, then it's something else that is broken.

luckybob wrote:

Ozzuneoj wrote:

Being able to tell if they need replaced or not

I rest my case.

But seriously, once they are out of the motherboard, you might as well replace them with new.

and testing then on the board / in system is useless, so remove + replace anyway!

It can be useful short of a full recap to pull sample caps and check them.

just recap the whole damn board. don't do a half assed job by only foing some caps.

Le sigh.

Checking sample caps allows you to tell if they've actually failed (as some fall out of spec without any visual evidence). There's no sense in replacing a whole board of perfectly good caps, nor is there sense in not replacing caps that have actually failed just because they look okay.

Furthermore, in terms of reliability just in time maintenance is actually better than preventative maintenance -- it avoids excess early failure. Replacing caps in high heat areas or that have a higher stress function at a higher rate than others is ultimately more reliable than performing a full recap every time.

Consider getting a real LCR meter like the DER EE DE-5000. Yeah, it's a bit more expensive, around 100 bucks from Japan. But unlike those cheap Chinese DIY projects, a proper LCR meter performs an actual vector impedance measurement at various selectable frequencies, which is important if you want to compare your measurements to the capacitor's datasheet (which is really the only way to accurately tell the health of a capacitor). Also, because it measures both the phase shift and the voltage drop at the same time, it can distinguish between a shorted cap and a good cap (or between an inductor and a capacitor and everything in between). And of course it has a Kelvin connection which allows you to make very accurate low resistance measurements accurate to a milliohm. You can even use it to measure the impedance of a battery.

luckybob wrote:

boom: https://www.youtube.com/watch?v=7Br3L1B80ow

i can't stand this guy

gdjacobs wrote:

Le sigh.

Checking sample caps allows you to tell if they've actually failed (as some fall out of spec without any visual evidence). There's no sense in replacing a whole board of perfectly good caps, nor is there sense in not replacing caps that have actually failed just because they look okay.

Furthermore, in terms of reliability just in time maintenance is actually better than preventative maintenance -- it avoids excess early failure. Replacing caps in high heat areas or that have a higher stress function at a higher rate than others is ultimately more reliable than performing a full recap every time.

This is exactly my thinking. If I have a board that "looks" fine and has 10 identical name brand caps on it, I can pull one or two for testing. If its still well within spec, then why recap the entire board? Maybe people who have replaced thousands of caps can do a whole board in a few minutes with little effort, but it takes me a fair amount of time to do something like this and it can often be tedious. Buying hundreds of replacement caps is also not a trivial outlay of money (especially figuring in all the hours involved in using them all before they too are "old").

I'm also interested in testing caps to see how much they truly degrade over time.

omg you cheap bastards...

If the board is unstable, just recap it from stem to stern AND FUCKING BE DONE WITH IT. I bet you wankers are the kind of people that go to the tire store and buy ONE TIRE AT A TIME and then bitch about how you are at the tire store every other week.

I just recapped a socket 462 board last week. took all of 30 minutes and maybe $10 worth of caps.

Just get the thing I linked to. I just bought one. (With a box, because I'm worth it.) It is perfect for us shade-tree mechanics for a "go/no-go" setup.

Living wrote:

luckybob wrote:

boom: https://www.youtube.com/watch?v=7Br3L1B80ow

i can't stand this guy

Meh. When he drags his crotch goblin on for a video I can't stand it. Past that I have no issues. As someone who never "technically" finished his EE degree, he was a great source to fill in the gaps in what I know.

luckybob wrote:

I just recapped a socket 462 board last week. took all of 30 minutes and maybe $10 worth of caps.

Did you replace every single electrolytic cap, even the tiny ones?

If so, congratulations. You obviously have a far greater level of expertise than I in this area, so we're not even close to being on the same page. What may work best for you is not what would work best for me.

When I do electronic repairs it tends to be a surprise when it doesn't take a long time and when it goes without a hitch. 🤣

I'm getting better, but with the volume of boards i have around I simply don't have time to replace the caps on all of them. However, if I get a tester and it turns out that many DO have caps that should be replaced, then I can order large quantities knowing that I will actually need to use them. I'm certainly not going to waste my time replacing a thousand caps on dozens of perfectly stable boards without knowing if they're even close to failing or still well within spec.

All the electrolytics, yes. All told, 20 caps?

You will get faster, the more you do it. I did my first recap so many years ago. It looked horrible. I'll post the pic later.

Just get a decent iron and a bit of desoldering braid and with practice, it goes quick. I even ENJOY doing it. I like to work with my hands. I put on some Dimmu Borgir and just spend a leisurely afternoon doing 2-3 boards. I just got 5 socket 775 boards that need the treatment, so my next saturday is planned. ^.^

luckybob wrote:

All the electrolytics, yes. All told, 20 caps?

You will get faster, the more you do it. I did my first recap so many years ago. It looked horrible. I'll post the pic later.

Just get a decent iron and a bit of desoldering braid and with practice, it goes quick. I even ENJOY doing it. I like to work with my hands. I put on some Dimmu Borgir and just spend a leisurely afternoon doing 2-3 boards. I just got 5 socket 775 boards that need the treatment, so my next saturday is planned. ^.^

I have a Hakko digital temp controlled iron and a Zhongdi vacuum desoldering station. My problem is that half the time solder still doesn't want to melt, which makes things frustrating. I use the wet sponge and the brass wool and I tin the tip before setting the iron down... and on some boards the solder just never liquefies. Any tips are welcome.

get a roll of the old lead type solder. Flow a bit onto each joint. remove cap, then remove solder via pump or braid.

luckybob wrote:

get a roll of the old lead type solder. Flow a bit onto each joint. remove cap, then remove solder via pump or braid.

I have lead (60\40 I think?) flux-core solder, and I do try to do this, but my technique is probably lacking. Its like the new solder won't stick to the old joint most of the time.

What do you do to prep a solder joint on an old board?

luckybob wrote:

If the board is unstable, just recap it from stem to stern AND FUCKING BE DONE WITH IT. I bet you wankers are the kind of people that go to the tire store and buy ONE TIRE AT A TIME and then bitch about how you are at the tire store every other week.

Not the sort of maintenance strategy that will fly with anything having expensive downtime or repair cost. Sure, your car tires will tend to wear fairly evenly (unless your front alignment is bad or the weight distribution is uneven). That won't always be the case, and you should never assume so.