08/02/2026
From Diyaudio.com, explaining reasons for my now future use of polymer electrolytic capacitors:
“Across my many projects I have finally concluded electrolytics have no business in a signal path. Not "audio grade" ones, not fresh ones, not expensive ones. The only place they still earn a seat is bulk duty behind the rectifiers where hands are tied needing thousands of microfarads. Everywhere else — coupling, decoupling, feedback shunts — it's polymer or film, and in digital I bypass the bulk cans too. I'm done, electrolytics - fired.
Took me three projects to actually believe it, for what it's worth. I resisted for a while because everyone says polymer is wrong in signal.
Here's the thinking. In a series coupling position the whole signal charges and discharges the dielectric on every note. Aluminum electrolytic dielectric absorption runs 10-15%, polymer is roughly an order of magnitude better, film better again. Dielectric absorption is basically the cap holding onto a bit of the signal and handing it back late — thats the mechanism that eats low level detail, reverb tails, room cues, texture on a cymbal. I pulled a 10µF Rubycon out of a phono input recently that measured 4.69µF with a dissipation factor near 0.8 at 1kHz. Half a capacitor, half a resistor, sitting in front of a 5mV cartridge signal for thirty years. The polymer that replaced it reads 10.2µF and about 20 milliohms.
The leakage objection needs to stop. Yes, polymer datasheets spec higher DC leakage. Read the test condition — that's measured at full rated voltage. In a coupling seat at near-zero bias, which is where most of them actually live, leakage collapses to nearly nothing and drops further as the part settles. The objection could be real for a polymer coupling a 40V tube stage but irrelevant for a buffer output at millivolts into 20k. Most people repeating the rule have never run the numbers for their own operating point. Second thing nobody says out loud -- there's no money in polymer. The audio grade electrolytic market has decades of mythology and a boutique tier at three figures a cap. A polymer is a two dollar commodity part from a Digikey reel with a datasheet instead of a story. Nobody reviews it.
In the signal path the thing that goes away is a kind of congestion you stop noticing until it's gone — the sound opens up and low level stuff that was being smeared into a haze comes back with its own texture. Bass gets tighter and paradoxically bigger, because the leading edge of the note arrives intact instead of soft. Anyone who's described electrolytics as stuffy or metallic is hearing ESR and dielectric absorption. ESR is just the cap's internal resistance, and on old electrolytics it climbs with frequency, right where the air lives. Both problems just leave.
The digital side surprised me more. Polymer across the supply rails on a DDC, DAC and clock board did more than I expected — flat milliohm ESR across the band is exactly what a regulator feeding a converter wants, and the result is a blacker background and better spatial cues. Same story on decoupling around the digital sections. You don't get "more detail" so much as less noise sitting on top of it, and the difference in how solid an image sits is not subtle.
Three components done this way now: a DAC, a preamp, a pair of power amps. Somewhere north of a hundred caps. Zero aluminum electrolytics anywhere between the DAC chips and the voice coils.”
The diyAudio Community
