Line array

[darkkhild]

Does anyone have any suggestions for a cheap but effective Line array? Maybe a 3" or 4" design but on the cheap side, Should have gotten the TB 3" that was a DOTW but I didn't so I am looking for the next best thing.
TIA
-B

[marlboro]

What do you consider cheap? $$$$? $$$? $$?

Have you read the BIBLE of line arrays, Dr. Jim Griffin's white paper on their design?

Marlboro

> Does anyone have any suggestions for a cheap
> but effective Line array? Maybe a 3" or
> 4" design but on the cheap side, Should
> have gotten the TB 3" that was a DOTW
> but I didn't so I am looking for the next
> best thing.
> TIA
> -B

[darkkhild]

Well I really can't say that I have but now I can, I'll read up on it tonight when I get a chance. Thanks for the link!!
As for cheap, per driver.... under $20 or less. Will that be a good starting point?
TIA

> What do you consider cheap? $$$$? $$$? $$?

> Have you read the BIBLE of line arrays, Dr.
> Jim Griffin's white paper on their design?

> Marlboro

[bryanjackson]

The solution to your question is 2 letters and a number.

B3N.

[marlboro]

How were you going to handle the crossovers? If you are going to go cheap on the speakers, then you are looking at flange cut neo domes, instead of 30 bucks each. But if you are using domes then you can't go any bigger than 4 inch mids bcause of comb filtering issues, and you will need to cross them to a big woofer for the bottom end.

You'll need about 34 mids and about 60 neo tweeters.

Biggest problem for you is how to deal with the crossover. Since you have to cross the tweeter at about 2700hz because of horozontal combfiltering, you need at least a LR 4th order 24db slope. This is pretty difficult for any but the best crossover builders to do. Rick Craig will design one for you for about $250, but you'll still have to buy the parts to build it.

The alternative is electronic crossovers and a constant Q eaulizer(about 250 bucks used on ebay), but you be able to make your own adjustments, of course you need an ECM-8000 or so microphone, and phantom power supply and the software to use with your lap top.

Getting complicated? Yeah. But lots of fun and lots of new things to learn. But you can't just slap one together like you can a quick MTM design.

I'd be happy to go over all that I did, and I may eventually describe it in some kind of pamplet.

Thylantyr here knows more than I by a long shot. and Jim Griffin has forgotten more about speaker design than I actually know. Both will help, but neither will suggest that you use dome tweeters. They use ribbons and planars. My next one will probably use B&G ribbons at $55 a pop, for about 600 bucks total. Then I won't have to have such a high crossover. But because of comb filtering issues the only dome I know that you can use is the Dayton 3/4 neo, and even there you have to cut the flanges to make the combing disapear out to about 17K hz.

Marlboro

> Well I really can't say that I have but now
> I can, I'll read up on it tonight when I get
> a chance. Thanks for the link!!
> As for cheap, per driver.... under $20 or
> less. Will that be a good starting point?
> TIA

[marlboro]

I'm apparently too much of a geezer. Can you explain that to me?

> The solution to your question is 2 letters
> and a number.

> B3N.

[tommoore]

"if you are using domes then you can't go any bigger than 4 inch mids bcause of comb filtering issues"

You have stated this several times before, and it makes no sense to me. Please explain. Yes, I have read Jim Griffin's white paper, and I understand what comb filtering is. I just don't see how the choice of a dome tweeter limits the choice of mid to 4 inch.

Thanks, Tom

[tommoore]

[marlboro]

OH....Zaphaudio has a good discussion on them. I think I would prefer to go to a 3.5 inch or 4 incher, but they are surely interesting.

Marlboro

[bpodruzny]

Provided Link: Line Array Crossover Help


The B3N with the Dayton Neo was the project I was going to do. Crossed at around 3000 and….

Well here is the link to the post.

<A HREF="http://www.pesupport.com/cgi-bin/config.pl?read=304428">http://www.pesupport.com/cgi-bin/config.pl?read=304428</A>

However, building a passive crossover for this project is proving to be more than I am capable of. Instead I’m setting aside the money from that project and investing in good testing equipment and active crossover gear.

However if you have testing supplies already and you are willing to spend a bunch of time on the crossover I’m sure you could get it to work.

Cheers
-Brian

[bpodruzny]

I think, and I’m no where near sure on this point, but I think it is because the center to center distance needed to push comb-filtering into the 17,000+ hz range is so small that any tweeter that could do it would have a small diameter.

Going along with that would be the need for a relatively high crossover point (usually in the 3000+ hz range).

Any woofer larger than 3” or 4” would probably suffer a lot of cone break-up if crossed at 3000 hz or higher and generally sound awful.

This is just my best guess; hopefully someone will correct me if I’m mistaken.

Cheers
-Brian

[bobblickle]

Do you really need that many tweeters? There was a guy here showing a design he'd been building for IIRC churches. Again, best recollection, but it seems like his tweeter section was much more minimal. Also, the B3N/S go up reasonably high all by themselves. From a purely subjective viewpoint, my impression is they just need a little more air. Is 3K or so the high limit due to the 3" spacing, is that the point I'm missing?

The second point I wanted to make is that Marlboro Man is an interesting guy and a great asset here but a total empirical evidence junkie. I'm not saying testing is bad, quite the opposite (I am an engineer afterall). I just would hate to see you be dissuaded completely from your project just because you can't do your own measuring.

Third point: This must have already been done! C'mon guys someone knows something about this combination. They might not have actually built it yet, but I guarantee you someone has at least carefully modeled and planned this same basic project already.

bb

> The B3N with the Dayton Neo was the project
> I was going to do. Crossed at around 3000
> and….

> Well here is the link to the post.

>
> <A HREF="http://www.pesupport.com/cgi-bin/config.pl?read=304428">http://www.pesupport.com/cgi-bin/config.pl?read=304428</A>
> However, building a passive crossover for
> this project is proving to be more than I am
> capable of. Instead I’m setting aside the
> money from that project and investing in
> good testing equipment and active crossover
> gear.

> However if you have testing supplies already
> and you are willing to spend a bunch of time
> on the crossover I’m sure you could get it
> to work.

> Cheers
> -Brian

[marlboro]

Tom,

You have at least three problems that have to be addressed, plus the interaction of all three.

1. You have vertical combing. If you take the neos and use them flange to flange you will get vertical combing starting at about 9000hz. Nasty. If you cut the flanges then you can get it to about 17K. But you can only do this for the Dayton Neo domes. There are no other domes at any proce that I can find that allow a .80 center to center distance.

2. So now we know that if we use domes, we have to use the neos. Problem 2 involves the fact that you shouldn’t cross them lower than 3500 to 4000. So now you have to address the vertical combing of the midwoofers. If you cross them at 4000, then to avoid mid combing you have to have the center to center be 3.3 inches or less. At 3500, it has to be 3.8 inches. Below 3000 and to about 2600, its between 4 ½ and 5 ½ inches for the C-to-c for the mid woof. But now you’ve exceeded the comfort level for the tweeter and it will sound awful unless you use a 4th order 24 db slope. Below 2600 to 2000, you can have a 5 ½ to 6.7 center to center for the mid woof’s but unless you have a 8th order 48 db slope, the tweeters will screech, and they may screech anyhow because an octave below 2000 is way below their fs.

So you are looking at with domes a limit of the mid woofer size because of where you have to cross them. Remember that mid ranges have a metal flange of at least 5/8 inches in addition to the cone size that you have to figure in.

3. You have to address horizontal combing between the tweeters and the mid/woofers.
Lets say you have the woofer is a 5: 2.5 c to c for the cone, .5 inch for the woofer flange, .5 inch spacing with the tweeter to look nice, .6 flange on the tweeter, .36 c to c on the tweeter= 4.5 inches or less for the horizontal spacing.

4. Finally Jim Griffin has qualified his white paper to add:

People should cross the tweeters before 2000 because of the on axis amplitude decline equal to one wavelength at crossover in woofers because of crossing too high and not being able to handle the high frequencies adequately. If you keep the woofer to 4 and below you usually solve that problem.

Of course then you have a new problem: you need a real woofer to give you some real bass.

All of this is solved by using ribbons or planars instead of the little neo dome. Or make the mid/woofers smaller, and also have a real woofer.

Marlboro

[bpodruzny]

Do I need that many tweeters? Probably not, one of the points made in the paper was that in ideal situations the woofer line and the tweeter line would extend the length of the array. This had to do with reflections on the floor and ceiling I think, but I would have to refer back to the paper for the specifics.

As for the B3N, if I’ve done the math correctly (13,650 in/s 3.5in) they should be able to be crossed at ~3900 hz before comb-filtering takes place (including the baffle the driver is 3.5 in). So would could give them a bit more head room, 3500 hz for the XO perhaps?

As for someone else who’s modeled these drivers together already...that would be sweet

Cheers
-Brian

[tommoore]

> Tom,

> You have at least three problems that have
> to be addressed, plus the interaction of all
> three.

> 1. You have vertical combing. If you take
> the neos and use them flange to flange you
> will get vertical combing starting at about
> 9000hz. Nasty. If you cut the flanges then
> you can get it to about 17K. But you can
> only do this for the Dayton Neo domes. There
> are no other domes at any proce that I can
> find that allow a .80 center to center
> distance.

That's not quite correct. I agree that a center to center spacing will result in a base frequency of 17000 Hz., but that is not where comb filtering would begin. The first null would occur at 8500 Hz. I don't know why Mr. Griffin fails to mention this in his otherwise excellent paper, but it is the reason some designers recommend a crossover frequency of one half wavelength or less.(based on woofer center to center spacing) If vertical combing that begins at 9000 Hz. is 'nasty', then 8500 would be worse, and render even a hacked dayton neo unusable.

> 2. So now we know that if we use domes, we
> have to use the neos. Problem 2 involves the
> fact that you shouldn’t cross them lower
> than 3500 to 4000. So now you have to
> address the vertical combing of the
> midwoofers. If you cross them at 4000, then
> to avoid mid combing you have to have the
> center to center be 3.3 inches or less. At
> 3500, it has to be 3.8 inches. Below 3000
> and to about 2600, its between 4 ½ and 5 ½
> inches for the C-to-c for the mid woof. But
> now you’ve exceeded the comfort level for
> the tweeter and it will sound awful unless
> you use a 4th order 24 db slope. Below 2600
> to 2000, you can have a 5 ½ to 6.7 center to
> center for the mid woof’s but unless you
> have a 8th order 48 db slope, the tweeters
> will screech, and they may screech anyhow
> because an octave below 2000 is way below
> their fs.

A fourth order acoustical rolloff isn't that diffulcult. Sometimes it can be achieved with a second order circuit. I used the neo in my boombox project and had a 18db/octave acoustical slope with a 2nd order filter. Furthermore, the reduced distortion provided by so many drivers helps alot. I think a crossover point of 2500 Hz. is not unreasonable.

> So you are looking at with domes a limit of
> the mid woofer size because of where you
> have to cross them. Remember that mid ranges
> have a metal flange of at least 5/8 inches
> in addition to the cone size that you have
> to figure in.

It's okay to cut the flange on a tweeter but not a woofer?

> 3. You have to address horizontal combing
> between the tweeters and the mid/woofers.
> Lets say you have the woofer is a 5: 2.5 c
> to c for the cone, .5 inch for the woofer
> flange, .5 inch spacing with the tweeter to
> look nice, .6 flange on the tweeter, .36 c
> to c on the tweeter= 4.5 inches or less for
> the horizontal spacing.

The woofer center to center spacing will almost always be greater than the woofer to tweeter spacing, and should be used to determine crossover frequency maximum. Horizontal combing becomes a non-issue.

> 4. Finally Jim Griffin has qualified his
> white paper to add:

> People should cross the tweeters before 2000
> because of the on axis amplitude decline
> equal to one wavelength at crossover in
> woofers because of crossing too high and not
> being able to handle the high frequencies
> adequately. If you keep the woofer to 4 and
> below you usually solve that problem.

I'm not quite sure what that means. Is that an opinion regarding the high frequency output of all woofers everywhere?

I think too much importance is placed on the effects of vertical comb filtering of tweeter lines when using dome tweeters. A nearfield array's output is overwhemingly on axis, and any filtering effects would be heard as reflections only. An equalizer could easily deal with that. Comb filtering is real and should be minimized, but I see no reason why tweeters with slightly larger diameters than the dayton neo should be dismissed outright.

Tom

[hulan]

The B3N - Dayton Neo line array is one I would love to try if I wasn't on such a budget. It will probably be a while before I'm able to take on a project of that cost. Even though both drivers are exceptionally priced, buying that many of them, then the crossover components, etc. Too much for me right now. Would still like to sink my teeth into a challenge like that, though! I bet it would sound and look sick!

[jerrald]

> The B3N - Dayton Neo line array is one I
> would love to try if I wasn't on such a
> budget. It will probably be a while before
> I'm able to take on a project of that cost.
> Even though both drivers are exceptionally
> priced, buying that many of them, then the
> crossover components, etc. Too much for me
> right now. Would still like to sink my teeth
> into a challenge like that, though! I bet it
> would sound and look sick!
If you want something really small, try the Pyle PLST6 1/2" tweeter. This should really cut down on the amount of audible comb filtering.

[marlboro]

Not necessarily. Just because its 1/2 inch doesn't guarantee that you can raise the start of combing higher. It still depends on the center-to center distance, and that depends on the housing and the width of the magnet structure.

I can't find a photo of one or a description of its size inserted into the baffle.

Marlboro

> If you want something really small, try the
> Pyle PLST6 1/2" tweeter. This should
> really cut down on the amount of audible
> comb filtering.

[jerrald]

> Not necessarily. Just because its 1/2 inch
> doesn't guarantee that you can raise the
> start of combing higher. It still depends on
> the center-to center distance, and that
> depends on the housing and the width of the
> magnet structure.

> I can't find a photo of one or a description
> of its size inserted into the baffle.

> Marlboro
I responded to Malboro's question through email but I wanted others to know the response as well. I don't know if I'll rember everything I said but here it goes:
Go to Parts Express web site and type 267-748 in the search box and it will take you to the Pyle page. The overall diameter of the driver is 1 5/8" so: 13560/1.625"=8344.62 Hz. If the housing is reduced to say 1" then: 13560/1"=13560 Hz. This eliminates any audible comb filtering, increases sensitivity, and according to some provides better off axis response than ribbons. (I think my math and conclusion(?) is correct. If not, I'm sure someone will correct me.) I look forward to any responses this generates.
By the way, if anyone decides to use these tweeters in a line array just remember they swivel so fill the bottom of the housing with something like putty to prevent this from happeneing.