[R-390] The saga cont pt 7

[2002tii]

Barry wrote:

>Can you explain this?  What's the disadvantage of metalized film 
>caps in these applications?

The primary problem is that they use much, much thinner dielectric 
sheets, so (1) the chance for pinholes and other imperfections is 
higher and (2) the field gradient across the dielectric is much, much 
higher.  Both of these make metallized film caps more prone to fail 
shorted by arcing through the dielectric.  To complicate matters, 
manufacturers use a "sputtering" (vapor deposition) process to apply 
the metallization, which itself is hard on the dielectric sheet.  I 
have seen countless failures of metallized film caps in tube 
equipment of all kinds, and almost never a failure of a 
film-and-foil.  Finally, the metallization is much thinner than the 
foil layer in a fim-and-foil cap, so the equivalent series resistance 
("ESR") of the cap is substantially higher (this last is a potential 
performance issue, not a reliability issue).

Here is an excerpt from one of my previous posts:

>Anyway, as I have said before -- but it is worth repeating -- there 
>is a world of difference between film-and-foil caps and 
>metallized-film caps.  Besides the performance differences 
>(metallized film caps generally have significantly greater ESR, as 
>well as exhibiting greater nonlinearity due to the higher field 
>gradient across the thinner dielectric), film-and-foil types are 
>many, many times more reliable than metallized film types.  Further, 
>metallized film caps are designed to be "self-healing" in the event 
>of a dielectric breakdown.  Sounds good, right?  Unfortunately, it's 
>only good marketing.  What this means is that if the dielectric 
>breaks down, the capacitor will arc until it vaporizes enough of the 
>metallization to clear the short.  That means whatever the cap was 
>intended to keep DC off of -- this could be your mechanical filter, 
>or the grid of the next stage -- has to to tolerate the "healing" 
>current while this vaporization happens.  One prominent manufacturer 
>of tube-type hi-fi equipment in the '70s and '80s used big, honking 
>metallized film coupling capacitors from the drivers to the output 
>tube grids of audio amplifiers.  You could actually see flashes of 
>light from inside the caps, often several times a minute, as they 
>drew their "healing" current through the grids of the output 
>tubes.  Needless to say, output tube life was seriously compromised.

So, I would never put more than 100 V or so across a metallized film 
cap, regardless of the manufacturer's voltage rating.  Below 100 V, 
they're great as long as you don't need the lowest possible ESR.  I 
use them all the time in SS designs with great results.

Best regards,

Don