Getting to dislike tantalum caps

Randy Dawson rdawson16 at
Sun Feb 14 18:49:30 CST 2010

Further, anybody that does radio restoration knows the first step is to "De Cap it"

Pull all electrolytics and replace.

Ceramic non polar is the way...

> Date: Mon, 15 Feb 2010 00:32:02 +0000
> From: classiccmp at
> To: 
> Subject: Re: Getting to dislike tantalum caps
> Teo Zenios wrote:
> > The aluminum electrolytic were the ones with problems.
> True, that. Google "Capacitor Plague."
> It's been happening since the turn of the century, if not a bit earlier.
> AIUI there are at least three different explanations:
>    - Early SMD electrolytics (the tin-can-on-plastic-base ones) tended 
> to suffer from failures of the sealing plug. Basically, convection or IR 
> soldering made the rubber plug deform, and after a while the cap would 
> just dry out, or leak all over the PCB.
>    - Later SMD capacitors had to be soldered with incredibly exact 
> thermal profiles. Some contract manufacturers upped the peak temperature 
> to get the solder joints to flow better. Catch: this also buggered up 
> the rubber seals on electrolytic caps. Same result as above.
>    - The "industrial espionage" explanation. Some jerkoff stole a 
> chemical formula for a low-ESR capacitor electrolyte from a major 
> Japanese capacitor manufacturer (I've heard the name Rubycon banded 
> about). Of course they only got part of the formula; it was missing a 
> few stabilising chemicals. Result: fzzt-boom after about 500 hours 
> runtime, or electrolyte leaking all over the PCB...
> And then there's "just plain crap parts." Panasonic, Chemi-Con and 
> Rubycon are usually safe bets (though there are a lot of fake and 
> knock-off parts about -- e.g. "Rulycon" and "Fuhjyuu" -- the latter has 
> the audacity to use a very slightly modified version of Hitachi's logo).
> There's a database of capacitor manufacturers and their logos here -- 
> <>. The page is in Japanese, but Google 
> translates it into (mostly) readable English.
> > Tantalums are nice as long as you don't have a voltage spike (they hate 
> > going above their rated voltage and short).
> This is why I usually spec them 50 to 100% higher than the highest 
> voltage I expect to see on the power line the cap is filtering. That is 
> to say, if I'm using a tantalum cap on a 5V line, I'll put a 7.5 or 10V 
> capacitor in there.
>  > At least when the aluminum
> > electrolytic leak/go bad they just quit working (open), hopefully not 
> > eating the circuit board.
> Though more often than not, they *do* eat the PCB when they pack in, 
> especially the SMD ones...
> -- 
> Phil.
> classiccmp at
Hotmail: Powerful Free email with security by Microsoft.

More information about the cctalk mailing list