Is there a functional difference between a "dc link" cap and an "ac filter" cap?

[Deke609]

I think they must be the same in terms of basic function and operation: block DC and allow AC to pass (with an impedance varying with frequency and capacitance), all subject to voltage handling ratings based on dialectric strength. But the differences in nomenclature gives me pause.  I kind of like the block form factor of some Wima polypropylene caps for B+ smoothing, but they're called "DC Link" capacitors. Any cause for concern?

MTIA, Derek

[Paul Joppa]

Yes there is a difference. They are designed for different applications, specifically safety margins and types of fault conditions. You canprobably6 find information from major manufacturerssich as Cornell-Dubilier. I have not studied this, so I don't know what the difference is - but I do know there are some!

[Deke609]

Many thanks PJ. Rats. Time for a deep dive into the datasheets then.

[Paul Joppa]

I mentioned CDE partly because they publish a book-size catalog. My copy is some 15 years old, so no modern DC link capacitors. Presumably there's an update, or something similar from another manufacturer. But it's nearly an inch thick. It's chock full of in-depth descriptions, application guides, and selection guides - much more info than just a data sheet. Recommended, even though there's no commentary on subjective audio quality ...  :^)

[Deke609]

Thanks for the tip PJ. Datasheets revealed nothing to me (other than suggesting that they do the same thing in power supplies).  I'll poke around the Cornell-Dubilier website and see if I can find that catalog or some white papers.

cheers, Derek

[Paul Birkeland]

From the Wima website:

"Due to their typical circuit position AC filter capacitors have to exhibit good high-frequency characteristics and at the same time high AC voltage capabilities. They in general fulfil two requirements:

Low AC impedance to dissipate high-frequency interference signals
Attenuation of transient voltage pulses caused e.g. by switching."

and..

"WIMA DC-Link capacitors are especially designed for applications in high power converter technology where due to increasing electrical requirements they are more and more substituting electrolytic capacitors. Manufactured with a low loss polypropylene dielectric they show a higher current carrying capability as well as lower dissipation/self-heating at high frequencies compared to electrolytic capacitors. Further outstanding features are, e.g.:

Very high capacitance/volume ratio
High voltage rating per component
Very low dissipation factor (ESR)
Very high insulation resistance
Excellent self-healing properties
Long life expectancy
Non-polar construction
Particularly reliable contact configuration
High shock and vibration resistance
Outstanding mechanical stability"

[Deke609]

Further to PB's post, I've discovered the following from reading some of CDE's technical papers (which are remarkably reader friendly and clear even to a non-engineer):

AC filter caps are designed to take a hell of a beating and keep performing - and if they fail, to fail reliably open-circuit: e.g., to withstand power surges, including lightning strikes! They need to be able to handle both huge voltage spikes and huge current spikes. CDE's AC filter caps (like Kemet's and others) have a pressure triggered "fail open" design that is supposed to work (at least in the case of CDE) up to 10,000A. My understanding is that high AC current through the equivalent series resistance of the cap produces heat-> expansion of cap fluid -> increased pressure in the cap -> blows the pressure-sensitive circuit interrupter.

CDE also mentions some sophisticated self healing tech in their AC filter caps. The metalized film is deposited in discrete, tiny squares that are connected to one another by minute traces. These traces act as "fuses" when a fault occurs within a square: the sudden increase in current through the faulty square vaporizes the "fuse", resulting in the electrical isolation of the failed square -- i.e., the cap "self-heals" and continues to operate normally with the failed square now removed from the circuit.

Cool stuff. But my take-away for B+ filtering where one is dealing with well less than 1A and relatively stable AC voltages: either one should work fine.  Unless one plans to install the amp in, say, a tank or the control cabin of a pile driver, in which case the extra vibration damping typical of ac filter caps (although also available with some DC link caps) is the right choice.

cheers, Derek