The long-awaited HUMAN 81 DK is now available.
Best viewed in "landscape" mode.
Capacitors
(Note: this essay is still, and may always be, a work in progress.)
Capacitors are used in virtually all multi-driver speaker systems.
The reason is that their impedance rises in proportion to frequency, and thus they are very useful in restricting the frequencies that reach the various drivers. Along with other components (inductors and resistors) these parts are used to build the crossover network.
At their simplest, as in HUMAN Speakers two-way designs, crossovers can consist of as little as one capacitor to prevent low frequencies from damaging the tweeter. (This only really works if the parts were designed to work together at a fundamental level to start with.)
At their more complex, crossovers can contain dozens of capacitors, inductors, and resistors. While this can look very impressive, it can also have drawbacks, and can mean that the parts used don't really "go together" very well. But it can make very good sense in a well-designed system - for instance, a three-way speaker with fourth-order filters is going to use a dozen or so components to achieve the linear-phase and sharp cutoffs provided by thsi design.
The point of this essay is to discuss the reasons, both real and imagined, for replacing the capacitors in an older speaker crossover.
It is almost a mantra nowadays when advising someone who just scored some old speakers for a pittance at a thrift store or yard sale (or rescued them from the trash) to "refoam and recap". The reason for the refoaming is obvious - the woofers and perhaps midranges on older speakers almost always have foam suspensions, which do not last forever and have often turned to dust. The reasons for replacing the capacitors, typically with ones that are perceived to be of better quality, are not so obvious, if indeed there are valid reasons.
There are two reasons people do this. One is the perception that it will bring better sound. The other is the idea that the old capacitors will have deteriorated somehow, and actually need to be replaced for the speaker to function correctly.
I will address the latter first, since it is the main point of this essay.
If, indeed, a typical crossover capacitor degraded such that its value drifted to 10 or 20 percent away from when it was new over a few decades, replacement every ten years or so would be obvious routine maintenance. We would want to replace them in our active stereo systems before any degradation in sound was apparent - so, perhaps every five years? If any other factor besides its capacitance changed meaningfully, I would say the same rule would apply. This could mean its voltage "rating" has fallen, its deviation from the ideal capacitor has become worse, or - or what?
All a capacitor consists of is two conductive plates separated by a non-conducting dielectric. The plates are usually rolled up around the dielectric, which is why most of them are cylindrical. The distance between the plates and the nature of the dielectric determine its voltage rating, and a formula combining the plate area, distance, and dielectric specification determine its capacitance.
So, do they go bad?
No. There is virtually no reason at all to replace them. (Unless, of course, they have been damaged due to misuse, abuse, or accident. This does happen once in a while.)
So, why does "everybody" do it as a routine matter of course?
Because it is easy, cheap, and feels good. It lets the hobbyist play with a soldering iron, and "improve" the guts of a speaker system. Even a bad soldering job will work in most cases, so it isn't even very demanding to do right.
It is easy, cheap, and feels good. But it does virtually nothing.
I have measured hundreds of capacitors from old speakers, out of curiosity. Perhaps thousands. From many, many speakers that are forty to fifty years old, and quite a few that were pushing sixty. They measure correctly - that is, within their original marked tolerance of their original marked value. Interestingly, I have encountered people "on the internet" who claim this alleged degradation has been proved with solid data. For some reason they do not have access to this data, in any form, when asked for it.
So, that's that. Or would be, if there weren't tens or even hundreds of thousands of people on the internet automatically recommending replacing them. It's easy. It's cheap. It makes you feel like a speaker engineer.
There are a couple of partial justifications for this almost religious doctrine of crossover capacitor replacement, although neither originates in the field of speakers. One is that ancient tube amplifiers were made with electrolytic capacitors in the power supply - which runs several hundred volts - that have often aged poorly due to sitting around unused for decades. Upon powering up, when hit by that surge of DC, they would explode. Two routes are typically taken to avoid this. One, if they "look" ok (no leakage, etc.), is to use a variac to gradually bring the power supply up and avoid shocking the capacitors. The other is just to pre-emptively replace them.
Another occasion where it seems some replacement of capacitors (and, often, resistors) is justified is in old electronics - amplifiers, tuners, etc. - where the original components are seen as noisy or unstable compared to the performance of newer (or more expensive) types of components.
But neither of these reasons really applies to speaker crossovers.
There is, however, a potential, fairly miniscule, reason to replace some capacitors with "better" ones.
Capacitors exhibit a measureable deviation from perfection. A perfect capacitor's impedance is exactly in inverse proportion to the frequency being applied (that is, when we double the frequency, its impedance exactly halves, or at least it should). Nothing in this world is perfect, and this particular deviation from perfection is called the dissipation factor, or DF. When necessary, this is specified when an order for capacitors is placed. This deviation is aklso sometimes referred to as "ESR", or "equivalent series resistance". It amounts to the same thing.
The typical aluminum electrolytic capacitors in consumer-grade speakers were specified to have a DF of under 5% (and usually measure a bit better). This is not bad at all, and only means that their impedance may not drop quite as smoothly as frequency rises as it should, or some similar imperfection. The more expensive film-type (among others) capacitors favored by audiophiles, and the press and industry that caters to them, usually perform much better in this regard - for example, the capacitors I use in my speakers have a DF in the 0.01% or lower range. This is far closer to "perfection".
Do they actually sound better? Because that is all that really matters. I can claim to think I have heard a difference - but it is damn subtle, certainly less than the difference in how music feels when one is in a good mood versus feeling a little cranky. The way I usually put it is "I feel like I have heard a difference, but I am also almost certain I could not pass a double-blind test". So much of the experience of music is emotional that sometimes (often!) simply believing something makes for better sound will result in a higher level of enjoyment. Unfortunately, the flip side of this mental effect is that worrying about the quality of one's gear can take over and interfere in any pleasure to be had from recorded music at all.
In the end, as a small (tiny) manufacturer, I think the difference - which is at least backed up by a solid, measureable aspect of the parts - can be audible, especially in capacitors used in series with tweeters (and to a lesser extent, midranges). So I use them.
* * *
Top - Contact
New Speakers -
EPI and Epicure - Genesis Physics
© Copyright - All Rights Reserved
7 Kelsey Road, Lee, New Hampshire 03861
Prices can change and specifications will improve without notice