Interesting theorie Menahem.

You think that the capacitor has leaked because of the bad bend solder lead instead of the work-stress where it is under? Does this powersupply has a traditional powerhandling or does it have any form of PWM ?

With the latter, almost every capacitor in even nowadays designed switching powersupply's doesn't seem to have a lifetime lifespan

This is a PWM, but the actual circuit voltage is only 12.7V.

You can see that this cap is rated @ 63V.

So, this cap should have been just relaxing.

I can only guess that the original cap was a 22uF 16 or 25V, but the tech obviously keeps a limited stock on hand, applying the stupid (for techs anyway) maxim "One size fits all".

A year ago, 7 caps were replaced in the PSU, all of them rated @ 63V. None of the circuit voltages on the secondary side approach anywhere near that.

Menahem Yachad wrote:

This is a PWM, but the actual circuit voltage is only 12.7V.

You can see that this cap is rated @ 63V.

So, this cap should have been just relaxing.

Forgive my ignorance but isn't ripple current what's killing the cap? The high frequency content of the PWM signal?

I'm asking because I'm curious.

/  Johan

You are entitled to believe and think, what you like but claiming the leaks has something
to do with the components pins being bent, I think, must be on your own account, Menahem.

It's quite common to see caps leaking like this, also factory
mounted ones that never had their pins bent in any way.
The Beocenter 2100/2200/4000 series springs to mind, let alone
B&Os VX-series of VHS machines. You will find practically nothing but
leaking capacitors in these machines, particularly in the VXs where heat is
a major problem and all caps have blackened pins from the boric acid inside running down.
Often also staining the circuit board surface around the capacitor.
Heat is not an issue in the Beocenters but the caps often leak the same way.

I've seen caps in TVs do the same and also in Penta amplifiers (and displays) and they are quite good quality builds.

Using a 63V rating capacitor in a circuit where the voltage across it never
gets higher than, say, 10 Volts or so is not always a good idea.
The capacitor will not get "excited", it will relax and become "lazy", to put it in
reader-friendly terms. You can reform it (take it out and slowly bring up its voltage) and
it will often be able to live up again but it's not a lasting solution.
Some of the early production individuals of the Beocord 6000/8000 series suffered from
this, where the signal slowly faded out and became distorted in one or both channels
because capacitors with a too high voltage rating were used as coupling capacitors in
low-voltage amplifier circuits.

Martin

OK - all excellent comments and, one at a time:

Martin - once again, this proves that even without taking into account the stress (and there is) on the rubber seals, that there is a VERY CLEAR DIFFERENCE between the different types of capacitors, and one type will hold up very nicely for thousands of hours under conditions of high-stress (be it heat or ripple), and another will give up the ghost, after only a few hundred hours.

So what could cause a cap which on the face of it (105C, 2000hours according to the data-sheet) should last a few years, but didn't?

Answer - another uncounted stress factor, in this case, stress by the legs.

Many manufacturers state express warning to NOT bend the legs like this, so on this issue, I know that I do not, at all, stand alone.

One size does NOT fit all, be it speaker crossovers, signal or power circuits.

Johan - you are correct that ripple is the major cause of failure in PWM circuits, but here we see catastrophic failure after not 10+ years (like in the BeoMachines which Martin mentioned, and on which many of us have much experience), but after less than 1 1/2 years!

Anyway, just for the record, here is the comparison 22uF @120Hz between the Samwha RD (63V) and the Panasonic FM (50V), which is my first choice.

RD 105C 2000hrs - ripple capability 83mA

FM 105C 2000hrs - ripple capability 250mA x 0.55 = 137mA

So, IMHO, the Samwha RD is a piss-poor choice for this PWM circuit. But the tech chose it of his own free will, and charged the customer a not insignificant sum for his repair. The machine worked after the repair - no argument. But does that mean that the tech did his professional best? Not by a long shot.

Samwha, even though it's B&O's choice, will NEVER have a place in my workshop.

So, at the end of the day, it is the tech's absolute freedom to choose which components he wants to use, but blindly selecting for convenience, and not making a professional effort to select the best component for the circuit requirement, will only shortchange the customer.

Regardless of the discussion. I didn't know the brand Samwha. I only use the respected brands like Panasonic, Nichicon, Vishay etc. So I'm on the safe side then 😁

Mostly using the 105 dgr with at least 2000 hrs...

Damages to capacitor seals are often done by to high soldering temps. or to long soldering time, all caps has a recommended soldering temp. and time. The most important part of soldering or desoldering is cleanliness, a clean soldering area lowers your soldering time. One failure that will be common for most of us is: Cleaning the area, desoldering the object, ordering spares, then waiting for delivery, after a couple of weeks the new parts arrive, and as one cleaned it all before desoldering, you would think its still clean, NOT so, copper (and other metals) starts oxidation as soon as oxygen gets to the surface, flush accelerate oxidation.

I always clean my soldering area with IPA, just before soldering or desoldering, this helps me to keep temp and time down.

I have done a lot of copper tube soldering lately, there one can see what happens with the solder tin used, leaving tubes cleaned just one day before soldering, made soldering slower or needed higher temp to get the solder to flow around the tube. 

So: CLEAN CLEAN CLEAN😁 

Menahem,

you are right what you say about bending the legs of the caps. They must go straight out of the cap, there must be no angle, there must be 2x90 degrees for fitting to the board. The seals are stressed additionally and there will be sooner a problem. The cap is already damaged without using.

There is also an additional effect. Bending the wires in 2x90 degree angles to fit the board, makes less stress to the solder points. The cap can move and still does not crack the solders, it does not pull sideways.

There is no guarantee that someone will bend later a new soldered cap. The technician is working on the board, he must handle it, touch it. A cap soldered in without angling the legs will crack the solders, break off the copper wiring. There is a high risk damaging the board. There is also the risk, to pull the wires out of the cap. 

Confirmed, ich have made exaclty the same experiences... and that the reason why i do it myself and i dont use Samwha. Crap caps...

There is also a reason not to use 63V caps if you dont need it: they are bigger and ventilation in a power supply is very important. Bigger caps make a problem here. That's also the reason, why i use miniature caps in the VX recorders... they last longer and stay coller, because better air flow.

Martin

 Menahem, this is quite an interesting theory and usually while soldering components on a PCB, most technicians won’t care about this. I also believe that most capacitors might not have this issue anyways because of a better sealing though and I’m saying this because I have seen electrolytic deposits on the legs of capacitors which I had soldered straight into the PCB without any bend. This could be a consideration along with the quality of the rubber seal by the OEM.