The only problem here is that some crossovers are designed with parts in mind.
That is
impossible and also not true. We've had this discussion numerous times, I believe, and no one making these claims ever has any proof, in terms of certain circuit modifications which would even suggest such changes. It's just erroneous audio dogma.
As I previously set forth in this thread,
using plots and data from the capacitor manufacturer, and which you could independently verify, the electrolytic's capacitance varies with frequency, so the crossover cannot, by definition, properly function with electrolytics. Simply cannot and trying to fix this limitation is impossible. Bypassing is an attempt, but it cannot overcome the limitations.
As I previously wrote on that subject:
Once the oats have been through the horse it is generally impossible to reverse the process.
I do not know of any crossover ever made which compensated for electrolytic rolloff. Why would one? It's like the Lady from Philadelpha dealing with Mrs. Peterkin after she'd added salt to her coffee instead of sugar.*
* For those of you whose childhood did not include this story, here it is for your education and entertainment.
"
The Lady Who Put Salt in her Coffee",
The Peterkin Papers by Lucretia P. Hale. (1886; 2nd edition)
http://digital.library.upenn.edu/women/hale/papers/salt.html
This was Mrs. Peterkin. It was a mistake. She had poured out a delicious cup of coffee, and, just as she was helping herself to cream, she found she had put in salt instead of sugar! It tasted bad. What should she do? Of course she couldn't drink the coffee; so she called in the family, for she was sitting at a late breakfast all alone. The family came in; they all tasted, and looked, and wondered what should be done, and all sat down to think.
KEFs 104/2 conjugate loading crossovers come to mind.
Sigh.
I think you've been reading too much propaganda about those, as I see plenty of it. What KEF did was simple enough: add
impedance compensation for the driver to prevent rise in impedance with frequency. The compensation is not, and could not be, to adjust for poor-quality crossover components, including electrolytics or ferrite-core inductors. That is not what the "conjugate impedance compensation" — a five-dollar marketing sales name for the venerable Zobel network combined with an LCR resonance flattener — does.
Really. Think about this. How does taming a rise in driver impedance modify or compensate for the poor properties of electrolytics?
What could it do? It somehow increases capacitance with frequency? But the network does impedance compensation, not capacitance compensation.
I have seen much mythology about these complex crossovers, which is then oft-repeated until it has the force of dogma, but it simply is not correct and cannot be correct for the reasons I above set forth. Every major capacitor manufacturer sets forth the substantial limitations of electrolytics.
The KEF crossover, furthermore, is very complex (I deliberately use that word) being a third-order with piles of impedance compensation, much of which is to fix the filter's ringing and overshoot in the pass-band and stop-band. The analysis is beyond the scope of this rebuttal, and nothing interesting falls out of it. But I will say that adding more stages to the crossover creates distortion via phase shift (delay in time) and group delay (uneven delay by frequency). Is this distortion audible? I would say that depends upon the sophistication of one's listening, and whether or not one cares. Simple first-order crossovers are used for a reason.
Yes, you can definitely replace NPEs with polys or films, but you may disrupt the balance of the speaker which the engineer set, using the parts he was given. Now, this isn't every speaker out there - but a spendy speaker design? Sure, I'd expect the engineer to consider NPEs in the equation if they were mass producing them with NPEs. [...] However, in many cases, it's a better investment to replace NPEs with NPEs.
I have previously explained in this thread, and elsewhere, why electrolytics are unsuitable for filters.
Any such use creates a degraded filter and it is impossible to compensate for that. Bypassing is one approach, but that's lipstick on the pig.
Adding an L-Pad sounds easy, but actually, it is not. You must find a place to mount it or make it accessible while tuning the speaker. Oh yes, if you meant it is easy to wire in.. I agree with that.
I understand your point, but let's take a step back on this. We both, I believe, agree that addition of the L-pad is trivial from a conceptual standpoint, i.e. the need to balance the tweeter against the midrange and woofer as a theoretical concept. The issue is the practical: hooking up wires to the driver.
I would argue that anyone able to replace crossover capacitors surely ought to be able to add an L-pad. But, at the same time, anyone not able to perform such modification should either (a) not be rebuilding crossovers, or (b) should obtain assistance. It really isn't very hard.
We're not talking about building speakers here - we're talking about refurbishing pre-made ones. If from scratch, of course you would use polys, budget allowing. But having to modify a speaker beyond simply replacing capacitors may not be a feasible option for many, and probably doesn't make sense in a lot of cases.
In a situation where hobbyists replace degraded surrounds, upgrade degraded capacitors, and even swap cored inductors for air cores, I don't think any of these upgrades are unreasonable.
If one is going to upgrade the crossover then adding an L-pad is the sensible option. It is easy enough to find the requisite assistance with such a project if the owner lacks it.
Also, it's not like speakers with NPEs sound bad. As flawed a product they are, they do their job. For the average hobbyist out there, many would be perfectly happy using them. A lot of people get hung up on the "grass is always greener" aspect of this hobby and honestly, it's not always worth it, especially when it comes to modifying crossovers.
That is a different argument.
As long as a replacement is being made, the superior decision is to install polypropylene capacitors with excellent properties instead of using electrolytics with terrible ones, plus a short lifespan. This is not being "hung up", it is understanding the limitations of components and the concomitant degradation of sound from using unsuitable ones.
The difference in switching to bypassed film capacitors is so obvious — modify one speaker, play a mono signal, and do an A/B comparison — that the arguments in favor of NPE are obviously seen to be specious and of utterly no value.
Such components were chosen to meet a specific price point. Same reason why lower-quality resistors and capacitors are used in amplifiers. Capitalism and all that. We, as hobbyists, can undo those decisions and build a circuit closer to the ideal performance than what we currently have.
I understand that most commodity speakers sound sort of ok with poor components (electrolytic capacitors, cored inductors), non-chamfered baffles, no time-alignment for drivers, no impedance compensation, poor cabinet lining, cabinets which are too thin and resonate, etc. etc. etc. And that owners simply do not care.
That doesn't mean we shouldn't suggest how to fix those limitations for those who
do care. Such changes may take a commodity speaker and kick it into an entirely new realm of performance. Doing an A/B comparison between speakers reveals the difference.
Particularly when the original reason for those poor component choices was fabricating a speaker for a specific price-point, not because those abysmal choices were purportedly made for sonic bliss, a claim which cannot be verified in practice.
We have better options since we are only building one pair of speakers, not a hundred thousand, or a million. Always good to have options like that.
