Phase issue- active vs passive xo

[rogermadison]

How's it
When designing a passive x-o one can can vary the phase of the drivers to bring them into phase/time alignment, at x-o freq, using various x-o slopes and x-o freq's. How do you address these issues when designing an active x-o? Would you use the same x-o freq's and slopes that work for the passive designs? Any help would be greatly appreciated. Thanks. Cheers.
Roger

[bill]

...with a passive XO you can do with an active one. Generally, passive XOs don't use inductors because they're large and expensive. Instead, you use alternate topologies that essentially replace the inductors with capacitors.

The same slopes and frequencies that work in a passive design will work in an active one. As a bonus, there's far less interaction between XO sections in an active design than in a passive one, and there's no interaction between the XO and the speaker; you don't need Zobels in an active XO.

HTH,
Bill

> How's it
> When designing a passive x-o one can can
> vary the phase of the drivers to bring them
> into phase/time alignment, at x-o freq,
> using various x-o slopes and x-o freq's. How
> do you address these issues when designing
> an active x-o? Would you use the same x-o
> freq's and slopes that work for the passive
> designs? Any help would be greatly
> appreciated. Thanks. Cheers.
> Roger

[Paul O]

> ...with a passive XO you can do with an
> active one. Generally, passive XOs don't use
> inductors because they're large and
> expensive. Instead, you use alternate
> topologies that essentially replace the
> inductors with capacitors.

Huh???

> The same slopes and frequencies that work in
> a passive design will work in an active one.
> As a bonus, there's far less interaction
> between XO sections in an active design than
> in a passive one, and there's no interaction
> between the XO and the speaker; you don't
> need Zobels in an active XO.

Yes, an active crossovers go before the amplifier channels, so all the problems normally associated with passive crossovers don't apply. All drivers are connected in-phase directly to their particular amp channel, so you have descrete control of the level of each pass band.
If you have an active that operates in the digital domain, then you likely have the added benefit of adjustable delay and phase for time alignment of the drivers, as well as individual parametric and graphic EQs on each channel. This is the ultimate solution with abslute control, but of course requires individual amp channels for each driver.

[rogermadison]

>How's it
I'm quite familar with the benifets of the active XO. What I'm trying to figure out is how to bring the drivers into time/phase alignment. I want to build my own active XO's to Bi-amp my home HiFi. I'm using Speaker Workshop and in designing passive XO's you can play around with the freq and slopes to get your correct time/phase alignment at the XO freq. The question is would the same XO freq and slopes used in a passive XO qive you the same results when used in a active XO? If not is there a way to calculate XO freq and slopes for an active XO to give you time/phase alignment at XO freq? thanks. Cheers.
Roger
Huh???

> Yes, an active crossovers go before the
> amplifier channels, so all the problems
> normally associated with passive crossovers
> don't apply. All drivers are connected
> in-phase directly to their particular amp
> channel, so you have descrete control of the
> level of each pass band.
> If you have an active that operates in the
> digital domain, then you likely have the
> added benefit of adjustable delay and phase
> for time alignment of the drivers, as well
> as individual parametric and graphic EQs on
> each channel. This is the ultimate solution
> with abslute control, but of course requires
> individual amp channels for each driver.

[robertcottiers]

> I'm quite familar with the benifets of the
> active XO. What I'm trying to figure out is
> how to bring the drivers into time/phase
> alignment. I want to build my own active
> XO's to Bi-amp my home HiFi. I'm using
> Speaker Workshop and in designing passive
> XO's you can play around with the freq and
> slopes to get your correct time/phase
> alignment at the XO freq. The question is
> would the same XO freq and slopes used in a
> passive XO qive you the same results when
> used in a active XO? If not is there a way
> to calculate XO freq and slopes for an
> active XO to give you time/phase alignment
> at XO freq? thanks. Cheers.
> Roger
> Huh???

Roger: I think that a fourth order "Linkwitz-Riley" filter will give you what you are looking for with regards to phase integrity at the crossover point,here is a link to Rod Elliot's site for further information.

<A HREF="http://sound.westhost.com/project09.htm">http://sound.westhost.com/project09.htm</A>

You should also check out "Linkwitz Labs" for a more technical discussion of the topic.

Regards Bob C.
You

[dthomas]

> ...with a passive XO you can do with an
> active one. Generally, passive XOs don't use
> inductors because they're large and
> expensive. Instead, you use alternate
> topologies that essentially replace the
> inductors with capacitors.

Generally passives do use inductors.

> The same slopes and frequencies that work in
> a passive design will work in an active one.
> As a bonus, there's far less interaction
> between XO sections in an active design than
> in a passive one, and there's no interaction
> between the XO and the speaker; you don't
> need Zobels in an active XO.

A zobel flattens the impedance of a driver... irregardless of the crossover being used. An active xo will not have eliminate the need of adding a Zobel to smooth a drivers response.

Dave

dave

> HTH,
> Bill

[Paul O]

> A zobel flattens the impedance of a
> driver... irregardless of the crossover
> being used. An active xo will not
> eliminate the need of adding a Zobel to
> smooth a drivers response.

> Dave

Sure it will, otherewise what's the point, you still have pasive components between the amp and the drivers mucking things up. Use an equalizer(parametric or graphic, your choice) ahead of your active crossover for broad strokes, or after the crossover on the individual outputs for detailed contouring.

[dthomas]

> Sure it will, otherewise what's the point,
> you still have pasive components between the
> amp and the drivers mucking things up. Use
> an equalizer(parametric or graphic, your
> choice) ahead of your active crossover for
> broad strokes, or after the crossover on the
> individual outputs for detailed contouring.

You are assuming that your equalizer keeps phase etc... in tact, reread your equalizers manual and notice the inputs of it are rated at a certain impedance... not a pure resistive load. You may very well be mucking things up far worse with an equalizer. The point I was trying to make is that irregardless of the type (active/Passive) crossover used, a zobel will benefit most drivers. In other words you would make a decision on not to use a Zobel based on the use of an active xover. +

Acutally everytime your signal is passed thru a circuit whether it is passive or electronic the signal is mucked up. Especially it's phase. So I am not buying the argument that a passive xo is far inferior to an active. I think all being equal they are just that relatively equal, each with advantages and disadvantages.

I am a proponent of as few components as possible because they all affect the signal but I do not think that adding more electronic equipment prior to the amp is a solution either. It all mucks up the signal..

Dave

[Paul O]

> Acutally everytime your signal is passed
> thru a circuit whether it is passive or
> electronic the signal is mucked up.
> Especially it's phase.

Agreed!!! Yes, each filter will impart some phase and delay artifacts on the signal it passes, active or passive. Correcting for this in the passive domain is cumbersome and tedious at best. So I think the active system has far greater potential to be the lesser of 2 evils. To do active right, you need adjustable delay on at least 1 of 2 pass bands in a 2-way for example. So it gets complicated quickly for a DIY project. I have heard the dramatic improvements made by switching to an active crossover from passives. And I think the relatively small anomolies you may introduce with a little EQ adjustments are far outweighed by the added definition of direct amp-driver connections, and the realtime crossover frequency and level adjustments possible with a good analog active(Rane AC22). You can accomplish in minutes and by ear what would take 100's of hours to match with a passive.

Paul

[dthomas]

> You can accomplish
> in minutes and by ear what would take 100's
> of hours to match with a passive.

> Paul

Paul... not trying to argue with you and I think we agree in principal but I do not agree completely that making changes in a passive are truly tedious nor require 100s of hours. I think once the DIYer has determined his driver imedance characteristics and has a fairly good idea of its frequency response making changes is very easy and this tweaking is the art of DIY speaker building. I do not know if it is impatience or a lack of knowledge that has made this not the norm. I do however agree that modeling this in software can be difficult and inaccurate. I think there is not enough emphasis being placed on tweaking the crossover once the modeled design is completed. I think that most DIYers are afraid of this aspect becuase they do not have access to the test equipment necessary to measure real world results. And are afraid to trust their ears. I have built dozens of speakers and only a few of the modeled crossovers wound up in the final design. Tweaking is half the fun, figuring out that the tweeter you chose comes to life when crossed 250 hz higher than originally planned or finding that the Q of that notch filter is just muddying things up too much is what this is all about. Also blending a woofer that needs a 2nd order slope with a tweeter that sounds far better with a 3rd order slope always benefits from a little ear tuning to find the best frequency spread.

Not $100s or $1000s of dollars in active crossover and eq. that is not DIY... at least to me. I think a lot of really good DIYers fall into the trap of not beleiving they can do it on their own. THey need a so-called guru to bless a design. You yourself admit you tune your active xover and equalizers by ear and seem to do so with great satisfaction. Next time try this with a passive xo design here the changes that different slopes, Qs and frequency points make and then use these to finalize your design.

Regards,
Dave

[dlr]

> How's it
> When designing a passive x-o one can can
> vary the phase of the drivers to bring them
> into phase/time alignment, at x-o freq,
> using various x-o slopes and x-o freq's. How
> do you address these issues when designing
> an active x-o? Would you use the same x-o
> freq's and slopes that work for the passive
> designs? Any help would be greatly
> appreciated. Thanks. Cheers.
> Roger

AFA the slopes and phase, the active is the same as the passive. You can't alter phase unless you add some sort of time delay circuit. I've not done any active design, so I'm not familiar with what delay circuits are possible, but I assume that you could do this.

I would expect the active delay circuit to be more useful than a passive version, I've just never done it.

In general you would use the same crossover slopes and frequencies, since these are guided more by the driver raw responses. But the active circuit will allow more possibilities in certain situations. This would most likley be for the woofer/midrange XO, since you can ignore the impdeance bump at Fs for both drivers.

AFA phase alone is concerned, you would do the same thing actively as passively (delays excepted). On a flat baffle, you'll want to do one of three things:

1. Delay electically/electronically (as mentioned)
2. Delay physical (tilt the baffle of mount drivers offset on the z-axis)
3. Use assymetric slopes.

Assymetric slopes often work reasonably well, but it is still situationally dependent. In most cases the larger drivers have the acoustic center farther back (delayed) on the z-axis. By using a lower slope on the delayed driver (e.g. 4th on tweeter, 3rd on midrange) then you reduce the delay compared to 4th on that mid and add more driver overlap. This will often yield a reasonable result.

The effect is the same whether passive or active.

dlr

[Paul O]

Dave... great points. As a matter of comparison, I have a set of speakers that I built way back when I knew nothing, and I have continuously been "editing" them ever since. I've completely lost track of how many times I've changed the crossover, and I've tried different drivers as well, all with varied results. I have found that all the crossover tricks in the world and a lousy driver can yeild decent results, but a better driver and a simple crossover can blow it away. I have also had a chance to stick a mic on these speakers and do some measurements, and I have to say the test results show a very flat response, way better than I would have thought I could do by ear. What that did was boost my confidence in "tuning by ear", and reinforced my decisions to use different slopes on each driver, for example, because the end result sounds good to me. I have only recently, relativly speaking, discovered active crossovers, and the WOW factor would have to be right up there with the top hi-lights of my audio life.
I think knowing what I know now, I could use my active crossover as a tool to design a passive crossover much more quickly, and with better results than at any time in my past. You can quickly discover what works for a driver set, and what doesn't. This is how I would do things, but that's the beauty of DIY, everybody gets to do it their own way. So, I won't discourage anybody from building their own active or passive crossover, and I will reinforce the idea that people should trust their own ears, because that's the ones they have to please.