Crossover Question - Mix passive / active

[Guest Peter the Greek]

Hi all,

 

I have only ever used fully active crossovers with separate amplifier channels for each driver. This makes things relatively simple, albeit expensive as you need more amps and more dsp/active crossover channels….not a huge problem in a 2 channel system

 

Given what is going on with ATMOS, I am facing the concept of another 4 speakers (well 6 actually) and its becoming a bit of a joke in terms of cost. In order to reduce costs, I had thought of doing the following:

 

1. Setting crossovers with a passive circuit

2. Do all subsequent filtering on a single dsp channel

 

Let me give that some more detail. I am talking 2 way speakers here.

 

Choice 1:

- each driver has its own amp and dsp channels (current plan, which costs too much)

 

Choice 2:

- each speaker (two drivers) has one amp and dsp channel

- I make passive crossovers

 

The main problem I see in "Choice 2" scenario is not being able to add (easily) a delay to align the phase at the crossover frequency (have I said that right?)

 

I don’t know the first thing about passive crossovers, so I'll need to have then built/designed for me (cost unknown?).

 

I wouldn't be hard to build the speakers, run one active and get the crossover right and then send off the specs for someone to build?

 

What are people's thoughts on the above?

 

Any help would be most appreciated.

 

Cheers,

Peter

　

[davewantsmoore]

The main problem I see in "Choice 2" scenario is not being able to add (easily) a delay to align the phase at the crossover frequency (have I said that right?)

 

Yes, that is the main problem.    The alignment between the two drivers will need to be physical, and/or compensated for in the design of the passive crossover.

 

 

I don’t know the first thing about passive crossovers, so I'll need to have then built/designed for me (cost unknown?)

 

I strongly recommend you build someone else's well documented speaker design....   or buy a commercial speaker

 

However, if you understand active crossovers well, then all the same things apply for passives  (you are just more limited with the available 'tools' so to speak)

Edited July 25, 2014 by davewantsmoore

[davewantsmoore]

build the speakers, run one active and get the crossover right and then send off the specs for someone to build?

 

 

Yes.... but with the understanding that not everything available from electronic crossovers are possible (or at least practical) for passive circuits.

 

For example.....  high Q filters....  HP and LP filters at low frequencies...  and delay....  are difficult to realise in passive circuits.

[Guest Peter the Greek]

Hi Dave,

 

Thanks.....figured as much. So I need to either (a) man up and spend the coin or ( b ) buy an off the shelf passive (be it DIY or complete)

 

....imagine having 10 surrounds/heights/ceilings all fully active......its a nice idea, not sure its worth the coin though....... > insert groan here <

 

Cheers,

Peter

Edited July 25, 2014 by Peter the Greek

[davewantsmoore]

Not an "off the shelf" passive   ;-)      ....   A passive specifically designed/tuned for your specific drivers and cabinet.

 

 

In an active...  you can put the drivers "where ever you like", and align them with digital delay.    In a passive crossover, the delay (if any) you have available for each driver is not as flexible.   You cannot just consider the acoustic alignment later.  It must be built into the design from the start.

 

You must more carefully consider the driver (both electrical and acoustic) behaviours before beginning the design, and options to work around them (steep slopes, notch filters) are not as easily realised

[Pulse-R]

Mixture of passive (for the rear/surround speakers) and active for the front...

 

It's a compromise, and as we all know (hahah) you can't get good sound in home theatre.

[Malcolm]

Bodzio Ultimate Equaliser

[john.reekie]

The idea has a lot of merit, I think. The problem with buying/using an off-the-shelf/pre-existing passive crossover is that it does (or may do) things that you won't need if you also have DSP EQ available. Meaning extra components/cost, reduced sensitivity, etc. So that's arguably the worst of both worlds not the best. On the bright side, the passive crossover gets a lot easier to make if you don't need it to do BSC, notch filtering, etc. I agree with Dave that the result is still going to be limited by your passive design skills especially when you look at things like the off-axis response. But hey, if you're talking about a typical tweeter crossover, all it's going to cost you to try is a bag of parts and some time.

Edited July 25, 2014 by john.reekie

[henry218]

If the drivers are easy to work with, its possible to try with bag of parts.

Most of scan speak, sb acoustics and vifa drivers are easy to work with, as the cone breakups usually further up than it's operating frequency. Sometimes for a typical 2ways could be as minimal as 4-5parts.

Typical 8" will have breakups around 1.9 and up, depending on how steep it is, somehow it does need some corrections.

Tweeters like scan air circ 6600, scan 9500 and its family doesnt need any correction. Scan HDS prob need some help around 1khz if it's crossed below 2khz.

I think choosing the right drivers is the key, and i think Peter is pretty flexible for the budget

[Guest Peter the Greek]

Henry, when you're talking 10 speakers, cost is always going to be an issue!....looking at a two way Scan Speak kit like the Nada gets awfully expensive.

 

.....they may be one instance where DIY won't win. Looking at a Tannoy pro model that might work and will dig around some of the other pro brands

 

Long story short, I think I'll just wire the room for fully active and then see what happens (where did I put that powerball ticket....).

[almikel]

Hi Peter,

above you write:

"The main problem I see in "Choice 2" scenario is not being able to add (easily) a delay to align the phase at the crossover frequency (have I said that right?)"

 

Mostly all passive crossovers compromise phase at and near the crossover frequency - theoretically a 4th order Linkwitz Riley acoustic response passive crossover wouldn't, as both the woofer and tweeter are in always phase, but to get the acoustic response to be 4th order Linkwitz Riley requires the drivers to be time aligned, and a very customised/tuned passive crossover (read expensive).

 

If phase and time alignment is a big deal for you, then any passive implementation will be a compromise (commercial, pro or DIY).

 

Sounds like you won't be heading down the DIY path anyway, but Rod Elliot has a great article on designing high quality passives, just to give you some insight into the challenges.

 

http://sound.westhost.com/lr-passive.htm

 

IMO phase and time alignment aren't that critical for surround speakers (compared to your mains), and simpler passives (commercial, pro or DIY) would suffice, unless you're seeking State of the Art

 

 

cheers

Mike

Edited July 25, 2014 by almikel

[henry218]

Henry, when you're talking 10 speakers, cost is always going to be an issue!....looking at a two way Scan Speak kit like the Nada gets awfully expensive.

.....they may be one instance where DIY won't win. Looking at a Tannoy pro model that might work and will dig around some of the other pro brands

Long story short, I think I'll just wire the room for fully active and then see what happens (where did I put that powerball ticket....).

I actually quite agree, if you dont mind to have different speakers for the surrounds, i would suggest to get the new JBL monitors, its max spl about 105-108db, assuming your theatre is not that large, then with the distance of 1-2m, they should be allright.

if you want the absolute SPL, then i would suggest something like 8-10" pro woofers from Eminence like Eminence kappapro or PD woofers from Thomann (50euro flat shipping) with selenium d220ti with smaller waveguide from faitalpro, but again this combo will need some EQ, where if you go active, it wont be a problem.

[A9X]

 

What are people's thoughts on the above?

　

Have you actually heard Atmos in a home? If not, perhaps you should first.

 

Option 3: Buy some JBL LSR305. Can be had for $400/pr and are pretty good.

[Guest Peter the Greek]

Have you actually heard Atmos in a home? If not, perhaps you should first.

 

No one has, but I'm going to do it regardless.

 

The other DIY option I was looking at was this thing (I can get them rather cheap from a place in France):

 

 

 

 

http://www.usspeaker.com/beyma%208cx300ndn-1.htm

[TMM]

I think people make out time alignment of drivers in the same enclosure (e.g. woofer and tweeter) to be a bigger issue than it really is.

You can actually time align with a passive allpass filter or if the misalignment is mild you may be able to correct it with asymmetric crossover points, however neither are easy to design for a beginner.

For a beginner i would recommend buying woofers with a smooth response out to at least 4khz with no serious breakup above that and tweeters with a resonant frequency no higher than 900Hz. You should be able to then get away with a 2nd order Linkwitz Riley crossover at around 2-2.5KHz, a series resistor (before tweeter crossover) to adjust the tweeter volume level to match the woofer and not need anything else.

Edited July 27, 2014 by TMM

[davewantsmoore]

I think people make out time alignment of drivers in the same enclosure (e.g. woofer and tweeter) to be a bigger issue than it really is.

You can actually time align with a passive allpass filter or if the misalignment is mild you may be able to correct it with asymmetric crossover points, however neither are easy to design for a beginner.

 

The 'big deal' (for a passive crossover) is that you need to design the crossover and the cabinet (driver locations) together....

[henry218]

The 'big deal' (for a passive crossover) is that you need to design the crossover and the cabinet (driver locations) together....

Isnt that what every speaker designer will do to design speakers???

[davewantsmoore]

With more sophisticated (eg. electronic) crossover it is possible to have delay independent of response shaping....  If you are designing a passive crossover, it is not as easy to do this.

 

 

 

You encountered exactly this issue when designing the 3 way I gave you some advice on.....  there was a compromise between filter shape, and achieving phase alignment.

 

Designing the driver location in conjunction with the filter(s), would mean that the drivers could be located where they produced phase alignment (when using the most desirable filter shapes).

[henry218]

With more sophisticated (eg. electronic) crossover it is possible to have delay independent of response shaping.... If you are designing a passive crossover, it is not as easy to do this.

You encountered exactly this issue when designing the 3 way I gave you some advice on..... there was a compromise between filter shape, and achieving phase alignment.

Designing the driver location in conjunction with the filter(s), would mean that the drivers could be located where they produced phase alignment (when using the most desirable filter shapes).

Unfortunately none of your advice work (in my speakers)....

I think we know your preference for active speakers, and we can just leave it at that. Its not necessary to put down passive xover on every occasion.

Even the most sophisticated OB active by John K or Sigfried Linkwitz and Geddes speakers still operated under the same rule and limitation

[davewantsmoore]

Unfortunately none of your advice work (in my speakers)

 

Exactly.... of course it didn't

 

....     this is because the cabinet and the crossover must be created together.    Because you had already built the cabinet, then the only crossovers which would cause time alignment, did not produce an ideal frequency response.

 

 

I think we know your preference for active speakers, and we can just leave it at that

 

I'm not trying to spruik active speakers....  I'm trying to discuss the way to design a passive speaker.   Trying to help anyone who wants to understand how to design a (passive) speaker.

 

 

Its not necessary to put down passive xover on every occasion

 

I'm not.   I'm discussing how to design them.   The design process is more complex as you have less freedom.

 

 

 

Even the most sophisticated OB active by John K or Sigfried Linkwitz and Geddes speakers still operated under the same rule and limitation

 

I don't know which 'rule and limitation' you mean sorry.   If you mean that the drivers must be located in space where they achieve time alignment .....   then the point is simply that electronic delay gives you more freedom.   That's all.

 

In a passive speaker you must choose a combination of driver location and filter shape which combine to achieve time alignment.    That simply means a passive speaker can be more complex to design.

[henry218]

Exactly.... of course it didn't

.... this is because the cabinet and the crossover must be created together. Because you had already built the cabinet, then the only crossovers which would cause time alignment, did not produce an ideal frequency response.

.

As saying goes... Many ways to Rome

I honestly dont give a damn if the crossover is time aligned or not, as long as the timing between drivers is correct

Im more interested on real result, listening experience coupled with enough data to make sure that the speakers performs as i wanted them to.

[Guest Peter the Greek]

FWIW Dave, you convinced me in your first post it was a bad idea

 

I'll do one of several things now:

- buy a passive speaker

- pay someone to design me a DIY passive

- buy a passive kit

 

I am leaning towards the first option. JBL have a real nice coax as do Tannoy. I am also waiting for someone on AVS to build a passive for that Beyma, so it'll be interesting to see the outcome (its the cheapest option). 

 

P.S. I am not fixed on Coaxes, but there are several good ones out there that suit......there is also a lot of ****...

[TMM]

With more sophisticated (eg. electronic) crossover it is possible to have delay independent of response shaping....  If you are designing a passive crossover, it is not as easy to do this.

 

You encountered exactly this issue when designing the 3 way I gave you some advice on.....  there was a compromise between filter shape, and achieving phase alignment.

 

Designing the driver location in conjunction with the filter(s), would mean that the drivers could be located where they produced phase alignment (when using the most desirable filter shapes).

DSP is great (i have a fully active/processed system in my car...) but in a lot of cases it is hardly worth the cost of having to buy two or three times as many amplifiers to fix what is a relatively minor issue, especially in a theater system with many speakers and many listening positions.

An allpass filter is far cheaper and will achieve an equal result if done properly.

[davewantsmoore]

As saying goes... Many ways to Rome

 

Not really   ;-)

 

 

I honestly dont give a damn if the crossover is time aligned or not, as long as the timing between drivers is correct

 

Isn't that the same thing? .....   My point (just so it's clear) is that if you fix the driver is space by designing the cabinet on it's own ..... then you have very little choice over the crossover slopes and frequencies if you want a filter which gets the timing in between drivers correct.

 

 

Im more interested on real result

 

I am talking about the "real result".

 

If you think I'm talking about "theoretical" only, or something which doesn't apply directly to how well the speaker sounds .... then that's your loss     ;-)

[henry218]

Ok... I cant be sure how to put your response, but yeah you can have it.

I know that i dont know everything