I have some limited practical experience and have experimented - on various projects in the past. My own personal opinion regarding speaker crossovers…is use the best well behaved drivers you can and the least number of components possible to achieve what is required … everything is a compromise. Years ago, I took apart a Wharfedale 708 - quite a reasonable speaker at the time. It sounded pretty good - I examined the crossover and it was very compact with quite a poor track arrangement. I spaced out the components on 3mm ply and carefully hard wired everything together. The speaker was so much better cleaner, better imaging etc etc. I have also done something similar with some Kans I am building … and currently testing. About 12 months ago I was after new speakers … and briefly listened to the Dynaudio Special 40 … in store…they sounded petty good. I didn’t go any further as the size was an issue. Recently I watched a tear down of this speaker…and given the price, the parts used were not that great…electrolytic capacitors, budget resistors and some iron core inductors. Have others found improvements using better components (to the same value)? I would like to know your thoughts and experience on this…
Also I note that Naim used to do passive crossovers that worked (I think I am right in saying) between the preamp and the power amp. Was this to allow the network to work at lower voltages…with smaller components…??? Was this a good route to take???
My only experience is with my old pair of Mordaunt Short MS310s - a budget bookshelf from the early 90s. I had to replace a tweeter and took the opportunity to hard wire both HF drivers (they used push on connectors beforehand) there was a significant improvement after doing so. I would not be surprised if crossover component quality mattered a lot.
Equally though, if it mattered sufficient to be detrimental to SQ amongst manufacturers’ peer products, you would not think Dynaudio would risk poor reviews for a cost saving on component quality.
I think your second point about crossovers is one of the advantages of the approach. The downside is increased complexity in the amplification requirements, more interconnects etc.
Lots of companies put higher spec components (especially capacitors and inductors) in the crossovers of their upper range speakers… keep an eye out for “Mundorf” in the advertising blurbs. Generally speaking, these better components offer not only tighter tolerances but better performance (particularly lower leakage). So, yes: better components contributes to better sound.
This is one way to “go active”, where you have individual power amplifiers for the high and low frequency drivers. In this way, each amplifier sees only the portion of the frequency spectrum appropriate for its associated speaker driver. Normally, when the crossover lives inside the loudspeaker cabinet, the amplifier sees, and amplifies, all the frequencies and the higher amplitude signal is split just before it is passed along to the high and low frequency drivers.
I’ve heard talk of passive XOs between pre and power amps, but never come across an implementation. I think the challenge is whether the preamp output will be affected by its very different impedance compared to a power amp - active XOs provide buffering and ensure output level is essentially the same as input level.
If Naim did once do a passive AXO, the fact that they abandoned it in favour of active types may give some indication of the relative benefits of the two… But maybe @Richard.Dane is betterplaced to be more definitive info about whatever it is you recall of Naim delving into passive crossovers for line level use between pre and power amps?
As I say, PAXOs were designed for and included with specific speakers; SBL, IBL, DBL etc… Naim made them external so they could be easily removed for active operation with a suitable IXO, NAXO or SNAXO.
You could use an IBL PAXO on a pair of active KANs (there was at least one such pair being used at Naim for piping Spire FM around the factory) but it was a bit ungainly and probably not ideal.
My NBLs came with a passive crossover attached to the speakers. The Snaxo 362 was included since they were released active. I auditioned IBLs once and they had similar passive crossovers as well.
I’ve heard of at least a couple of instances where someone has substituted ‘better’ crossover components and found the result to be worse…
The thought was that the original crossover had been optimised for the actual characteristics of the components, and changing them for ‘better’ components upset the balance.
Your right certain components do have signatures…but generally you can’t tell me that putting an electrolitic capacitor in a passive crossover is a good idea…apart from being cost effective… or using iron core chokes…apart from saving space and cost…
Take apart a magico speaker…and look at the crossover…quality…
As you said yourself, everything is a compromise. I’m just suggesting that simply swapping ‘better’ components of the same value might have unpredictable results. If you want to design a completely new crossover, that’s a different proposition.
The crossover in my PMC EB1i speakers, which I took out to go active, were massive! Something like 25 to 30 cm long and 10 or 12 cm wide, with the air cored inductors and largest capacitors sticking out maybe 6 cm ir so above the board. They appeared to be decent quality components. Compared to that their predecessors, IMF RSPM speakers from about 1970, had very small boards, inductors were ferrite cored, and the capacitors seemed nondescript. I re-capped when they were about 40 years old, using not just new but better capacitors, though choice limited slightly by space. I can’t say there was any obvious difference in the performance and sound. Of course, I couldn’t do A-B comparisons, so I was comparing one day with the next, which is never terribly reliable,
But if you want the best, the answer is very simple: take out or bypass the crossovers, get active crossovers, and multiply up your power amps. The difference is far more evident and distinctly positive than changing components on speakers’ passive crossovers…
I was thinking of experimenting on my experimental Kans…using dsp to control each speaker and allow it to blend…together…that way the crossover is in digital domain…and infinately adjustable…would have to measure and adjust in the garden…to avoid room reflections…
however, if you do it in the room, then you can compensate to some extent for me effects at the same time…
Yes, a digital active crossover is a very flexible tool, obviously depending on what you use. On some, you can select independently the upper and lower cut off frequency of each passband, and the slope of the cut off. You can also adjust relative timing (phase) between the drivers at the crossover point
However, I thought the popularity of the Kans among a sector of the hi-fi community was to do with its sound character, and tweaking the crossover conceivably might change some aspects - or is this just an experimental speaker fir which the Kan was simply astarting point?
Alan Shaw of Harbeth pretty much states that crossover parts quality does not matter (as long as they are durable and last a long time).
Danny Richie of GR Research states that crossover parts quality is a huge difference. The truth might be somewhere in the middle.
