purely resistive crossovers?

[googler]

From sonicweld.com

Unique Crossover Design
Would you buy an amplifier with a large capacitor in the signal path of its output stage? Of course not. Why then accept one in the crossover of your speaker, which is nearly the same thing?

You don't have to anymore. Rarely does High End audio witness a true crossover innovation such as that featured in Sonicweld products. The same tired formulas, parts, and construction methods have been used for decades. Every conventional loudspeaker crossover uses reactive components (capacitors and inductors) in the signal path. Unfortunately, these filtering networks produce measurable and audible anomalies. Sonicweld circuits neatly sidestep these heretofore insurmountable problems by the use of a unique circuit with the following characteristics:

1. There are NO reactive parts in the signal path. The only element between the input terminals and the drivers is carefully-selected premium-grade wire, yet the driver array benefits from the electrical attenuation provided by conventional filter networks. The benefit to you: transparency, inner detail, speed, and the overall lack of noise and grain like you've never heard before. After extended listening to a Sonicweld speaker with this superior crossover, you may find it difficult and fatiguing to resume listening to the sonic atrocities introduced by conventional crossovers.

2. The Sonicweld crossover acts as almost a purely resistive load. Sonicweld speakers typically measure flat within ± .5 ohm to beyond their operating bandwidth. Your amplifier will perform much more effectively when driving an almost purely resistive load. This approach to crossover design eliminates reflections in the amplifier-to-speaker interface caused by reactivity, which is of particular benefit to tube amplifier owners.

3. All Sonicweld crossovers exhibit virtually no electrical phase shift. A variation of ± 5 degrees for their full bandwidth is typical.

I know absolutely nothing about crossover design....
Has anyone tried a x-over network like this? How would it be implemented? Were the results as claimed above?

[DonoMan]

Sounds like BS to me.

[mac]

For $40,000USD they had better be very special.

http://www.sonicweld.com/main.html

[DigitalJunkie]

Nothing between the input terminals,and the speaker except for high-grade-super-spiffy-wire...

In other words,there is no crossover?

Maybe it's an active x-over?

[kelticwizard]

By sonicweld There are NO reactive parts in the signal path. The only element between the input terminals and the drivers is carefully-selected premium-grade wire, yet the driver array benefits from the electrical attenuation provided by conventional filter networks.

Some linguistic trickery, plus perhaps some good engineering as well.

"Not in the signal path" has come to mean, in crossover circles, as "not in series with the driver". That still means the reactive component(s) can be hooked up in parallel to the drivers,and no doubt are.

If they can keep their speaker resistive or nearly resistive throughout the whole range, though, they have done a great job, even if they do play some games with the language.

[Stocker]

Exactly. Note that inductors are reactive and made of wire only...

[kelticwizard]

Originally posted by Stocker
Exactly. Note that inductors are reactive and made of wire only...

LOL . But callling an inductor "a piece of premium grade wire" would really be cheating though. Although they may be doing it.

Using reactive components only in parallel is just sort of taking advantage of a linguistic convention presently being used. Although that, too, is intended to mislead. Not as bad as calling an inductor a piece of premium wire, though.

Below is a 6 dB low pass crossover. Things are not quite so simple as that, since at high frequencies this becomes a short circuit. However, other reactive components can be added in series with the capacitor to take care of that.

As long as none of the reactive components are in series with the driver, they theoretically are "not in the signal path".

[5th element]

HOw have they done this? Lots of notch filters lol all set up to provide the desired roll off characteristic.

OK i tried this and its a no go, put one single capacitor in series and you can notch to your hearts content until you arrive at the desired filter characteristic. But you cannot get a parrallel trap to do anything unless there is a cap or coil in the way first.

It took 5 notches and a single cap to get 4th order linkwitz Riley on a scan speak d2905/95 at 3000hz. Anyone know how to get parallel components to work without a series?

[5th element]

OK had another play, yeah its possible but its very complicated and uses a HUGE amount of components. But it can be done. Seems a bit of a stupid thing to do, might aswell just use an active xover.

[kelticwizard]

Originally posted by 5th element
.....yeah its possible but its very complicated and uses a HUGE amount of components. But it can be done. Seems a bit of a stupid thing to do, might aswell just use an active xover.

For $40,000 a pair, you might wish consider it, yes.