when to recap speaker crossovers?


2010-06-14 8:31 pm
Assuming I have a set of speakers from the 70's that still sound good to my ears, is there some more objective test I can do to see whether the crossover still functions within specification? I assume I could feed a signal sweep to the speaker and digitize back in the feed off each speaker I mean to see if it seems to be rolling off at the expected rate, but I wasn't sure offhand how to do that digitization/probably no easier way to do this with a voltmeter reliably is there, or to do it without even opening up a cabinet which maybe wasn't designed to be user serviceable... (I have multi sets of 70's speakers actually i'd like to test, only some are made to open) Or maybe just a voltmeter test of the components indicating within a specified range even if I don't specifically measure the audio, i'm just curious how others do it. This is relevant to buying future old speakers too/if there's some way to test without opening up the speaker cabinet which makes many people uncomfortable, if I could just plug in signal gen and microphone on a laptop or something and look for something obviously out of whack like an unexpected peak at the crossover points which i'd assume would be a sign it's failing to... cross over.

I'm assuming the main thing that is likely to die with age is the capacitor, or I could even preemptively just replace those anyways just to be safe/extend it's life if i've opened it up anyways... or do resistors and inductors ever likely go out of whack?
Resistors & inductors do not typically show significant change. Capacitors are another matter -electrolytics dry out over time; as an ROT (and it's only an ROT) expect a useful service life of 10 - 20 years depending on specific construction & quality. So if the speakers date from the '70s, there is a reasonable chance they're past the best & the transfer functions have drifted somewhat. They would very likely be worth replacing. Depending on what the speakers are / who made them, the ESR of the caps may have been taken into account in the XO design, so be wary replacing with Mylar or polypropylene etc. types as this too may alter the transfer function away from what was intended. In such cases, replace like-with-like or the nearest practical / possible equivalent unless you want to redesign the filter.

Re testing, you may be able to check; you'd need to pull the boards, measure the impedance load of the drivers, simulate the electrical transfer function of the components into the measured loads, then measure the filter behaviour & see how they compare. An interesting exercise, but I'd probably be inclined to re-cap it as a matter of course in any event given they're in their 4th decade now.
Last edited:


2003-02-11 9:02 am
Simply measure the capacitance. If it is within value (especially not short or open) then it should be okay. You can go overboard of course but I have never had the need to do so.

Axial (NP) capacitor from old British speakers are okay imo. I even prefer using them than my exotic/new ones (for amplifier input).

The chance for capacitor malfunction in a crossover is small. So the chance for the malfunction to happen to both channel is even smaller. So when the left and the right sound the same then stay cool. I once heard different sound from left and right and the troubleshooting led to a capacitor changing value (become less capacitive of course) in one channel.