[andyr]

The recent fascinating article on Zobel networks which Kyrill posted got me playing around with lspCAD and various passive XO configurations.

Without fail, every order of passive parallel XO that I modelled had an impedance peak around the XO region - even a 1st order XO!    Interestingly enough, though, a 1st order series XO does not exhibit this sharp peak!

Anyway, my Q is this ...

* a 2-way passive parallel XO consists of a LP & a HP filter.  If we are using 2nd order filters we have a coil in series with the woofer plus a cap across it ... and a cap in series with the tweeter with a coil across it.

As mentioned above, this produces a sharp rise in impedance around the XO point and, as this is an inherent "feature" of the parallel XO - and is between the amp and the drivers, it negatively affects the amp's ability to drive the speaker.

* in an active 2-way setup, the XO is situated before the amps ... and it is easy to make the active XO match the HP & LP filter slopes and elbow frequencies of the passive setup.

So my Q is ... is there or is there not, a rise in impedance across the XO region in an active situation? 

I personally cannot see how there can be - which means that an unexpected benefit of active XOs is that they do not generate this impedance peak!!      

What say you, gennlemens?

Regards,

Andy

[Davey]

A first-order parallel crossover using the same components as a series crossover should yield the same impedance curve.
Edit  It will, but only under certain circumstances.  Thanks for correcting me Andy.  As I said, I'm not a crossover guru.

Other setups will indeed create impedance peaks/valleys in the overall system.

However, I don't think I would conclude that it "negatively affects the amps ability to the drive the speaker."    A good amplifier will act as a voltage source and provide whatever current is required to the load.

It can, however, cause the response to skew when driven by an amplifier with less than ideal output characteristics.  Highish output impedance, etc.  You can model that easily with your LspCAD setup Andy.

An active configuration is a totally different setup.  The raw driver impedance is what's seen by the amplifier and that's it.  Generally you'll see a peak created by the driver resonance lower in its operating range and a rising impedance as you go up in frequency caused by driver inductance.
There are no complex impedance interactions created by crossover networks and multiple drivers connected together in parallel or series configuration.

I don't believe Zobel networks are necessary for active drive configuration systems.  (Especially speakers like yours.)  Assuming you are using a decent amplifier and there aren't RFI problems that would be solved by terminating the speaker lines at the far end.

You probably should move this thread to The LAB or another more appropriate circle.

I'm not a crossover guru.    But neither is the moderator of this circle. 

Dave.

[AKSA]

Davey,

That was a cheap shot.  Not necessary, and gratuitous.  I never purported to be a crossover guru, and whatever people say about me, the fact is I know precisely what I do not know and signal it accordingly.

You are welcome to my circle, but please be civil and stop trying to show people up to prove some obscure point.  I know you are a clever engineer, and I do respect you for that, but I suspect you feel that non-engineers have no place in audio.  That is patently absurd....

However, no offense taken,

Hugh

[Davey]

Oh for crying out loud!  If I was going to take a shot at you it would be a lot better than that.  You must be kidding.

Andy, you've got my email.

I'm out.

Dave.

[AKSA]

Davey,

I think you mean '...it would be a lot worse than that'.

Nough said, your choice Sir,

Hugh

[rabbitz]

Without fail, every order of passive parallel XO that I modelled had an impedance peak around the XO region - even a 1st order XO!    Interestingly enough, though, a 1st order series XO does not exhibit this sharp peak!

Davey is right that the series 1st order should have the same impedance curve as the 1st order parallel. A lot of the series camp, myself included, have moved from the standard series crossover to the AR series crossover or variations of.

http://www.acoustic-reality.com/ar-sxo2.htm

This crossover has great advantages for the passive designer and one being impedance friendly for amps. An example is Vifa P13 + D27TG35 when using a 1st order series or parallel, you get about 14R at the crossover frequency (using series resistor padding) but with an AR series you get around 7R7 which is not a bad place to be for amp loading.

Another example is a 2nd order AR series using Peerless 830875 and DX25TG... impedance of 6R5 to 9R1 (excluding woofer resonant peak) which is very sweet and that's using a mild zobel of 4.7uF + 6R8. Zobels when used help with the impedance correction but also help to tailor the shape of driver roll off... so it becomes part zobel and part higher order filter.

I only mentioned this as there is a passive crossover that can give an amp a smooth impedance ride and sound exceptionally good.

So my Q is ... is there or is there not, a rise in impedance across the XO region in an active situation? 

I personally cannot see how there can be - which means that an unexpected benefit of active XOs is that they do not generate this impedance peak!!      

You are right... 2 amps, 2 speakers on different planets and their paths do not cross so individual driver impedance only.