Failure mode of aluminium electrolytic capacitor....open cct..not short cct?

[treez]

From the below link, it appears that when an alu elec cap fails due to electrolyte drying out…it fails open circuit, which is in some ways harmless…..and so what we can do, is simply put a few solid polymer capacitors in with the capacitor bank, so that when the alu elec's fail open, then the solid polymer ones will be there to give some support….do you agree?

We don’t want to use all solid poly ones, as they are far more expensive.

https://www.illinoiscapacitor.com/pdf/Papers/failure_modes_aluminum_electrolytic.pdf

 

[barry]

No, i dont agree. If they polymer caps are adequate, then why use an electrolytic? If they're not adequate, then when the electrolytic fails youll have an inadequate circuit.

 

[treez]

I am comparing alu poly cap…870055975003 (33uf, 100v)
https://www.mouser.co.uk/datasheet/2/445/870055975003-1726594.pdf
..with wet alu elec cap Rubycon ZLJ (33uf, 100v)
https://www.rubycon.co.jp/en/catalog/e_pdfs/aluminum/e_zlj.pdf


The alu poly one is 2000hrs at 105degc…the wet elec one is 6000hrs at 105degc.
I take it that the wet elec one has a longer lifetime……from searching digikey……….there are no longer life solid poly ones at 33uf, 100v.
This thing about solid poly being longer life than wet elec is looking doubtful?

The caps improve transient response...but the product does not always suffer bad transients.....so to an extent, it is still workable with out the wet elecs.

 

[barry]

I am comparing alu poly cap…870055975003 (33uf, 100v)
https://www.mouser.co.uk/datasheet/2/445/870055975003-1726594.pdf
..with wet alu elec cap Rubycon ZLJ (33uf, 100v)
https://www.rubycon.co.jp/en/catalog/e_pdfs/aluminum/e_zlj.pdf

The alu poly one is 2000hrs at 105degc…the wet elec one is 6000hrs at 105degc.
I take it that the wet elec one has a longer lifetime……from searching digikey……….there are no longer life solid poly ones at 33uf, 100v.
This thing about solid poly being longer life than wet elec is looking doubtful?

The caps improve transient response...but the product does not always suffer bad transients.....so to an extent, it is still workable with out the wet elecs.

Let me be blunt: your design approach is questionable at best, dangerous at worst. I would never buy a product designed by someone whose philosophy is ‘well, transients don’t happen that often, so we’ll cut corners there and hope for the best.’

 

[albbg]

From the link you attached it seems to me that "deterioration of electrolytes" goes to both open circuit failure and increased dissipation factor. I fears than under some circumstancies this could cause severe problems to the cap.
In any case if you put some solid polymer caps with the Al cap how can they do the work since they have a shorter lifetime ?

 

[dick_freebird]

Aluminum electrolytics caps being spiral wound, the
plate current is concentrated at the electrode end and
an open circuit failure at high current (esp. if electrolyte
is depleted and can't help cooling of the inner region
where contact is made from leads to the foil) is a likely
failure mode (electromigration or plain fusing). The
plate will detach from the lead as the aluminum contact
"fuse" blows at the point of highest current*temperature.

Interdigitated plate solid body caps have a more favorable
distribution of plate current. But the mfr may have
(probably has) take advantage of this to shave material
cost and raise capacitance density.