Xover 3 impedance

[parus]

when i model a crossover in xover3 using a parallel crossover under the impedance window it shows that the net impedance of the speaker system to be the same as the drivers ie if i am using two 8ohm speakers the net impedance is 8ohm with a slight rise at the crossover frequency. I had thought that when you parallel two drivers the net should be half of the ohms 8+8=4ohms. also how much of a rise in impedance over the nominal is okay at the crossover frequency it rises up to over 20ohms how can i level that out
thanks

[cjd]

> when i model a crossover in xover3 using a
> parallel crossover under the impedance
> window it shows that the net impedance of
> the speaker system to be the same as the
> drivers ie if i am using two 8ohm speakers
> the net impedance is 8ohm with a slight rise
> at the crossover frequency.

Yup.

> I had thought
> that when you parallel two drivers the net
> should be half of the ohms 8+8=4ohms.

Yup. If you're paralleling them *without* a crossover.

> also
> how much of a rise in impedance over the
> nominal is okay at the crossover frequency
> it rises up to over 20ohms how can i level
> that out
> thanks

Depends on the amp. Solid state amps generally don't have issues with a rise at crossover to 20ohm. A tube amp very likely would be less happy with it. If it's a single clean peak, that's less an issue than some of the bumpier rides through that region that can occur.

I have found that optimizing impedance while also optimizing frequency response is not particularly simple. You can't *just* flatten impedance though.

C

[jimhunt]

> Yup.

> Yup. If you're paralleling them *without* a
> crossover.

> Depends on the amp. Solid state amps
> generally don't have issues with a rise at
> crossover to 20ohm. A tube amp very likely
> would be less happy with it. If it's a
> single clean peak, that's less an issue than
> some of the bumpier rides through that
> region that can occur.

> I have found that optimizing impedance while
> also optimizing frequency response is not
> particularly simple. You can't *just*
> flatten impedance though.

> C

zobel networks!!!!!!!!! come on people.
jim