How good is the crossover in your loudspeakers?

Yes, this thread started (and has largely remained) a discussion about loudspeaker crossovers. But now that the subject of open baffle subs has come up (I accept the blame for that ), here’s something that hasn’t yet been mentioned regarding them.

All the open baffle speakers on the market (at least all those I am aware of) have the speaker’s drivers---including the woofer---mounted on a flat baffle, the baffle then mounted on a base. It is common knowledge that one of the penalties of eliminating a sealed or ported enclosure for a woofer is the resulting dipole cancellation phenomenon: without an enclosure to separate the front and rear waves of a woofer, those waves "wrap" around the open baffle on it’s two sides (and top, for that matter), the positive (forward) and negative (rearward) waves meeting on the baffle’s sides (and top), that meeting resulting in a drop off in bass response (+1 plus -1 = 0).

The frequency at which drop off begins is determined by the dimensions of the baffle. The bigger the baffle, the lower the frequency.at which drop off begins (if the baffle is big enough---like the woofer mounted in a wall---it becomes an infinite baffle woofer. A subject for a different thread).

What the W/M and H open baffle "frames" do is increase the distance between the front and rear of the woofer(s), thereby lowering the frequency at which drop off begins. However, if the enclosed space created by adding side and top panels to a flat baffle is deep enough, a "cavity resonance" is created, which we don’t want. The open baffle frames used by Rythmik/GR Research and Linkwitz are all about 14 to 16 inches in depth, which provides a healthy amount of front-to-rear woofer separation (thereby lowering the frequency at which dipole cancellation commences) while avoiding a cavity resonance within the passband of the woofer system.

Ergo, no penalty to the open baffle sound, with only positive benefits. Another benefit is that the added side and top panels provide increased baffle structural stiffness, hence decreased resonance from an unsupported front baffle. In addition, Brian Ding installs a dipole cancellation compensation circuit in the plate amp (Rythmik model A370) that comes with the OB/Dipole Woofer kit (I believe it is a simple 1st-order/100Hz filter, the filter boosting frequencies below 100Hz at a rate of 6dB per octave). The only people who poo-poo the OB/Dipole Woofer system are those who have not heard it. A Rythmik Audio/GR Research OB/Dipole Servo-Feedback Subwoofer fanboy? Oh, you betcha!

My experience is based on LS3/5as.

I had an original Rogers 11ohm pair with it's fairly complicated crossover.

When the Cicable external crossovers came out for them I bought them and removed the internal crossovers - I also replaced the cabinets with thinner walled ones (9mm instead of 12mm like in the original Kingswood Warren built prototypes) so that might muddy the picture slightly.

The Cicables were designed by Derek Hughes, who used to run Spendor and learnt his craft from his father (designer of the BC-1) and at the BBC, and has since designed all of Graham Audios and Stirling Broadcasts speakers and a passive speaker system in the Royal Opera House in London, which is very unusual as such systems are normally active.

The crossovers use expensive inductors and caps from Mundorf and at points reduce the THD of the LS3/5a by about 40%.

The difference is stark - much more transparent with more detail and much smoother.

Derek apparently doesn't just implement the standard type of crossover curves, he considers the driver characteristics more - there are interviews with him on YT where he mentions this more.

Derek designed the Stiring Broadcast V2 LS3/5a when Kef stopped making the LS3/5a drive units, which use modern drivers from SEAS and Scanspeak.

The crossovers for these are less complicated than the originals. They sound good but not as good as the Cicable crossovers which is not surprising considering the price point they were designed for.

Later they produced the V3 version which is basically the same crossover but with the iron-cored inductors replaced with large air-cored ones. These are more transparent and detailed, closer to the Cicables, but there is a 'shizzle' to the upper end, maybe exposing some of the THD in the original V2 crossover and it's more basic components.

Now there is a V3.2 version which has a more boutique crossover which is handbuilt by Derek, or his daughter who works at Audionote). Components are Jantzen air-cored inductors and Crosscap capacitors. Again another step up from the V3.

I also have some Xtracable external crossovers, also by Derek H, for my V2s and they are again slightly better than the v3.2s - also using expensive Mundorf components.

This progression in sound quality is whilst using the same cabinets and drivers, so clearly isolated.

Lastly they had a bass extender called the AB-2 which used a band pass conenction to the LS3/5as with a sinlge 15mH iron cored inductor with a 220uF electrolytic and 6.8uF bypass polyprop capacitor per channel.

With the v3 or v3.2 connected via the band pass there was a loss of audio quality, often seen with a reduction in the soundstage.

Replacing theses AB-2 crossovers with a Jantzen air-cored inductor and Jantzen crosscap or Mundorf Evo Oil caps completely stops this loss of audio quality, most easily seen in a retention of the soundstage. The bass is also very slightly better but limited by the nature of the AB-2 bass extender. 

Commercial speakers have crossovers that are very much price constrained compared to what Danny does. The v3.2s that Stirling are doing are using crossovers that are quite expensive and not commercially sensible, but they are closing down and it is only because of that and that there is an LS3/5a enthusiast building them (me) that they exist. If you were to buy a commercial LS3/5a with a crossover that expensive you would be looking at several times their cost.

Same with those AB-2 crossovers, about £480 in parts alone. Worth it to me and a few LS3/5a nutters that have signed up for them, but not commercially viable.

@texbychoice wrote:

Wins in what specific technical and measurable ways?  Trade offs must be honestly identified and considered.  That is the only damn fact that matters.

Now suddenly measurements are a convenient step (i.e.: "Measurements do not tell the entire story")? Apart from perceived listening impressions (they still count, don’t they?), you have a purer ohm load when an amp channel (a dedicated one, no less) is looking directly into a driver’s terminals, avoiding the likelihood of large impedance swings and steep phase angles through a passive crossover and hereby providing for much better working conditions for an amplifier, with better driver control and lower distortion to follow. 

Buy two more similar amps for 3 way setup, eh?  So say I have a quality 100 Watt amp, so buy two more that would add several hundred dollars of cost.  That is a hypothetical that makes no sense to support the case for all active.  Conflating potential reliability facts with trouble sleeping is an illogical comparison.

There’s no "case for all active" to universally go by. Why would you impose specific terms for active to make sense, other than the potential for better sound quality or to whomever it applies? That’s on you, pal. 

The case for active, from my chair, is sound quality via outboard active configuration and high efficiency speakers where size, by and large, isn’t an issue. Whatever it takes, it takes. For others it may be convenience, simplicity (yes, you heard me right), small size, and even an overall lower price with a bundled, preassembled and -designed package.

The good thing about active is that of being able to make more efficient use of a given amplifiers performance envelope, and thus you can save money per unit and keep yourself from buying overbuilt, hugely expensive amps that would otherwise be needed with passively configured, heavy speaker loads. So, what may seem to be more expensive with the need for more amp channels actively to begin with, can turn out to be much less so than expected or even save you money eventually. 

The sleep remark was hardly to be taken in the literal sense, but merely a play on words to address your claimed issue with reliability.  

The more complexity is added, the more the entire system is at the mercy of the weakest link.  Cheap out on any item and the entire system does not achieve it’s potential.  

You’re making a problem where there needn’t be any. Added amp channels and a DSP in themselves don’t necessarily equate into introducing a weak link. Buying a cheaper, bundled active package on the other hand (not least subs with built-in plate amps) can be an issue where reliability goes, but as I pointed to earlier my advocacy is outboard active config., and this way any quality gear can be in the loop for a purchase decision, where reliability is no bigger issue than it is passively.

The core premise of active being cheaper, easier, better completely fails.  Pick any 2.  You can’t have all 3.

Says who? Why don’t you get your head around the fact that some, if not many of us actually pursue active from a sound quality measure first and foremost, without "cheaper and easier" being part of the primary incentive? 

Do your system as you see fit.  Personal preference extrapolated to claims based on broad generalization does not equate to a clear path for all to duplicate. 

Oh I do make my own system as I see fit. I hope you do too. We’re debating crossovers here, and I added my experience with a way to implement them to make their presence less of an issue in the signal path. Please note that amp to driver interfacing is but one part of many to be considered. I’m not implying it’s a ticket that in itself makes everything else magically fall into place, nor do I mean to impose my views on others. 

@phusis "Practically speaking the only trade-off with active is a higher electrical bill from the multitude of amps. "

active x-over could actually save power and deliver higher SQ than passive one, by enabling mixing D (low/sub) and A (mid/high) class amps in one system. mids/highs don’t need more than 10W in regular listening, thus low power class A amps should work. in addition, active x-overs improve mid/high driver damping factor, which is typically. killed by resistive inductors and actual resistors in passive x-over circuit. 

@phusis +1 

Mike