Capacitor specifications

[acheld]

I'm in the process of assembling a BOM for a Dumble style amp and looking at a lot of different capacitors.

When ripple current is called out in a capacitor datasheet, what does this mean?  Is lower better, or worse?

[PRR]

Dumble didn't look at ripple current specs.

IMHO it has zero meaning in audio, but people try to make it so.

[pdf64]

Higher ripple current spec is ‘better’. The only place it might conceivably be relevant is for the reservoir cap.
Max operational temperature, expected operational life and maybe ESR are more relevant and apply to all decoupling caps.

[acheld]

Dumble didn't look at ripple current specs.

 
I can see that.

Max operational temperature, expected operational life and maybe ESR are more relevant and apply to all decoupling caps.

That's what I was thinking.  But I'm curious; any idea what it (the spec) is actually saying?

[PRR]

The makers of the kinds of caps we would use for reservoir caps know the usual uses. A tiny physical size would overheat on 100/120Hz ripple, but the capacitance would give 50% ripple which would buzz too bad to use. And high frequencies can run the same ripple current 100 times more per second which would heat much faster, so rip-current IS critical in "switchers" working 10kHz-100kHz.

In audio, if the can is big enough to absorb 90% of the ripple, it is big enough to throw-off the heat of 100/120Hz ripple.

I am aware that a very sharp-pencil proof can find even commercial amplifiers which violate their caps' ripple limits at full power. Mostly these will be class B at lower idle current most of the night. Some just "eat caps"... hey my dog eats a bag of chow a month and I tolerate the expense.

[acheld]

Got it, so it is the ripple current that the cap will "tolerate," so to speak.   TY!  Tried to look this up on the google, but failed miserably.

[PRR]

> what it (the spec) is actually saying?

Internal series resistance vs surface area to throw-off the I^2*R heat.

The math is no fun. But assume 5 Ohms ESR and 0.3A of both-ways ripple current on 0.1A of DC current. That's 1.5 Watts. The case is liable to be >1"x2" or >6 square inches. 0.25Watts per square inch is warm but not hot.

[PRR]

Capacitor testing in 1949:

[acheld]

Wow, now that is great ad!      

Who knew that Sprague had a sense of humor . . .   

And thanks for heads up regarding the relationship between ESR, ripple current and internally generated heat dissipation.  Makes complete sense.