Recommendations for electronic crossover.


I am bi-amping my B&W 804 matrix speakers with a 50 watt per channel tube amp for the top and a 200 watt SS for the woofer. Any suggestions for reasonable priced crossover? I have been told Merchand (?) makes a good one.

Thanks!
jpahere

Showing 5 responses by dhl93449

Well, I finally finished my discrete Class A XO project. I used my own design discrete op amps, based on an early Spectral design modified with Nelson Pass inspired JFet output stages (instead of MOSFET stages as originally used by Spectral). I am using 4rth order Linkwitz Riley filters for both HP and LP filters and the drivers integrate smoothly

Took out the gyrators as I really did not need parametric EQ with the Voxative drivers (switched from the Horn).  

System is now sounding better than it's ever sounded. So that confirms that the active XO system has merit in principle, and the class A discrete designs in practice.
Arguing about whether bridging is good or bad misses the point. Many base designs (such as those from Nelson Pass) are bridged inherently. They allow the use of lower power supply rail voltages and therefor reduce power consumption, particularly with Class A configurations. I don't think that a properly designed bridging configuration increases distortion any more than using fully balanced circuitry does. It is true that each amplifier that is bridged will see half the load impedance, and as such must be designed to drive that lower impedance.

Regarding the subject of active XOs, it is almost impossible to find one that is built without op amps. Bryston is the only company that comes to mind that produced HP and LP versions with discrete Class A circuitry, and I do not recall if these are still in current production. In any case they are not cheap, and if you need more than just a single XO point (two way system), the costs mount up fast.

For the poster who mentioned potential pitfalls of eliminating the factory passive XOs, almost all of these are removed by providing a dedicated amp/driver. Line levels are easy to match with gain controls on the amps, and most other impedance matching gymnastics are eliminated with a dedicated amp. The system is also much more efficient if the drivers have widely different sensitivities, because you do not have volume matching resistors wasting a lot of power. The passive components are a lot cheaper and you can use higher quality. For example, teflon film caps are almost impossible to get (or afford) as large values in a passive XO, but are affordable in an active design where they are a lot smaller in value. For more discussion on the advantages of dedicated amping, see Professor Linkwitz discussion on the matter.
phusis
Its all a matter of degree. Whether an active XO system using IC op amps sounds better than a passive system with the best Mundorf or other high end caps, inductors, and resistors may be arguable. 

But it makes no sense to me to place an EQ/XO having IC opamps between a discrete component preamp and discrete component power amps. I actually tried that with a Phonic i7600 op-amp DSP between my Parasound JC2 and two JC1s when trying too chase down a frequency balance issue (and it was good for that), but the sound quality through that Phonic unit did not sound as clean and detailed and the system with a passive XO and the Phonic removed. 

I am probably in the camp leaning toward an active system, but will not compromise in using IC op amps vs discrete class A circuitry. Every time I have made a comparison of products with IC op amps vs discrete, discrete always wins hands down. I think both Nelson Pass and John Curl would agree with me (along with Bryston). So I am in the process of designing and building an all discrete 4rth order Linkwitz Riley LP/HP XO system for the PAP Horn1 speakers. It will also have discrete op-amp gyrators for a parametric EQ function. Expensive, time consuming, and rather large, but I think it will be worth it.
really provided sonic equivalence, all discrete designs would be gone in a heartbeat.
phusis

" Whether it makes sense or not is a matter decided by the ears, not theory, but of course: if what your ears tell you is in favor of a passive configuration - in a specific, singular component context, that is - then that’s what it is to you in that particular scenario."

Yes but the comparison is not passive vs active w Class A discrete, but active Class A discrete vs digital + IC opamps. No doubt digital processing has the flexibility to provide complex transfer functions, frequency responses (and time delays) that no discrete component system could duplicate, but from a pure sonic purity perspective, they will never match that of discrete Class A designs, assuming you don't need the frequency response complexity. With the Voxativ and PAP bass drivers, their response is reasonably flat enough that just a simple 4rth order HP/LP XO will do, assuming the XO point is well chosen. 

Why are there very little discrete Class A XO products available today?
They are expensive to build, large in size, and use components (ie the transistors) that are getting harder and harder to find, as well as expensive. A single matched pair JFet from Linear Systems now cost more than most IC opamps. The Toshiba parts are long out of production and only available from manufacturers (like Pass Labs) who horded these parts many years ago.  But the fact that companies like Pass Labs, Bryston and Parasound still build their top of the line products with discrete components still attests to the sonic superiority of these designs. If a lower cost IC opamp