Electronic crossovers for loudspeakers
Copyright 2013-17 © Troels Gravesen

A often have the question why I never use electronic crossovers for my speakers and here are my thoughts:

I do not necessarily subscribe to the notion that "less is more". Those who settle with a "coil and a cap" often pay a price on certain performance parametres as no drivers - to my knowledge - are capable of performing well from such simple measures unless we never play very loud.

Any loudspeaker crossover, being active or passive, divides the frequency band between drivers with regard to amplitude and phase and ensures a balanced presentation of bass, midrange and treble. In principle as simple as that.
If you want to use an electronic crossover (analog or digital) for any of my constructions, you need to replicate the transfer function of my passive crossover with regard to amplitude and phase, at least if you want the speaker to sound the same as I made it. Now, what does this mean? Below the transfer function of the passive crossover for the Jenzen-D:

Click image to view large.

As can be seen, the tweeter crossover transfer function looks pretty much like a text-book crossover. The midrange not so much with some equalisation of upper mid and a sharp notch-filter at 5 kHz. The active crossover needs to do the same thing and if not digital, not so easy to do actually.
The bass transfer function display some equalisation around 50-100 Hz as we have an elevated response around 80 Hz due to the impedance peaks created by the TL/vented box. This must be dealt with quite differently by an active crossover as the active crossover sees the linear impedance of the following amplifier and not the one created by the box tuning, thus probably no equalisation is needed. I would tune the active bass crossover section by doing near-field measurements of the bass driver.
The long and the short of it is that in order to implement the transfer function of the passive crossover we need to be able to measure what happens, and most people do not have measuring equipment good enough to do so.


To the left an electronic crossover for Linkwitz-Riley LX521 speaker and to the right my Jenzen-D passive crossover.

Anything between the signal source and the speaker drivers is likely to leave a mark on overall performance and the question is whether a well executed passive crossover is doing more harm compared to the signal passing a complex electronic crossover, being analog or digital. There can be several scenarios:
If using an analog musical source we need to convert to digital after the line stage and have the (digital) crossover divide the signal into bass, mid and treble and further provide the necessary equalisation to mimic the transfer function of the passive crossover. Next the signals must be converted back to analog for our (assumed) analog power amplifiers.
Now, all of this can be done in the digital domain - including listening to digital music - and feed 6-8 digital amps before the final conversion to the analog signal driving the loudspeakers. The all-digital solution indeed appears more simple compared to the first example.
The active filter seen on the left hand photo above is an analog filter including the necessary equalisation of drivers' response. I'm not sure the signal passing quite a number of opamps and less that state of the art capacitors does better than the high-quality crossover components seen on the right hand side photo, but some believe that as long as distortion is low, everything is as good as can be. I have my doubts.

I did my share of analog electronic crossovers before this website started and I'm not going this route again. I like it simple, not for the sake of simple but because I think we can do as well with passive crossovers - and I like my wife to be able to use my system as well. I have one really - really - good amplifier driving my mid-tweeter and having two of these would be crazy. One can do it when only we use a well designed passive crossover and premium components.

I do use a dbx 24 dB LR electronic crossover for experimental purposes when deciding e.g. point of crossover between bass and mid. Very useful.
And there were experiments that turned out very well with this dbx crossover, the AT-SW using the 10C77 bass driver and a point of crossover around 140 Hz.
Next I tried the QUATTRO with a push-pull U-frame, 35(W) x 65(H) x 60(D) cm, made from two Eminence Deltalite 2512 drivers connected in series. Again 140 Hz point of crossover. This system was really smashing and to the full demonstrated the smooth and linear response of the QUATTRO speaker. And it could play insanely loud and be a system for those who think the DTQWT-12 is made for 300 cubic meter rooms, which it is not, rather 60-100 cubic meters.

I think active solutions are a good thing for home entertainment when we deal with seriously inefficient speakers requiring multiple amps - or feeding a subwoofer from a separate amp.
One area where digital management is far superior to passive solutions is equalising the bass response, which by far overshadows most other problems people have when installing sound systems in a less than ideal acoustic living space. Too much bass or too little bass is the most common problem people run into.
My recent experience with DSP/Plate-amps, ATS4-HE and MUN-17, see index, has made me consider more constructions based on these principles. It simply works great!
As in other areas of hifi, technology easily turns into religion and pointing fingers is futile.