What advantage from short arrays?

[Dumbass]

Been perusing Jim Griffin's line array whitepaper, and thinking about some of the "short" arrays you see, for example the Jordan and Bandor arrays (4 drivers per, about 8" long), also the MMTMM arrangement, etc.

AFAIK these setups don't put you in near-field distance as defined by Griffin, so what exactly is happening?

What advantages are gained by shorter arrays?

Sorry if these questions seem a bit broad, I'm just mulling some things over.

[Pbassred]

I wish I understood the mathematics ..... or the phisics. ...... and is the effect frequency dependent?

[rcavictim]

For one thing a short array still gives the benefit of requiring each driver to develop less cone excursion and subsequently less distortion! Lowering the excursion requirement allows one to tap into less expensive drivers, some which are absolutely fabulous in line arrays but would not be considered serious hi-fi candidates used individually. Also, efficiency and sensitivity are improved, expanding the choice of suitable drivers to the DIYer.

[Pbassred]

yeah but ....... how?

[Dumbass]

Originally posted by Pbassred
I wish I understood the mathematics ..... or the phisics. ...... and is the effect frequency dependent?

Yes. Figure 5 from the whitepaper gives the frequency-dependent "transition distance" from near-field (cylindrical wavefront) to far-field (spherical wavefront):
http://www.audiodiycentral.com/resource/pdf/nflawp.pdf

[johninCR]

You should note that the nearfield to farfield transition is a gradual thing not sharp and definite line.

Short arrays work great for desk top speakers, so you are in the nearfield for array purposes. Other than that you are going to measure and tune after they are built if you want a flat FR, because you aren't going to be able to reliably predict what will happen. Why? If you have enough drivers in a line to exhibit array behavior at some distance (which is frequency and line length dependent), then you will have a speaker that changes in tonality depending your listening distance.

IMHO, the easiest way to predict your response with an array is to build a tall array of sufficient length that your listening position is in the nearfield. There's no reason not to do so, since it only takes up unused vertical space. If you don't want very tall speakers, then don't build an array, build something else. I think only going half way is a big mistake.

[Dumbass]

Originally posted by johninCR
IMHO, the easiest way to predict your response with an array is to build a tall array of sufficient length that your listening position is in the nearfield. There's no reason not to do so, since it only takes up unused vertical space. If you don't want very tall speakers, then don't build an array, build something else. I think only going half way is a big mistake.

To me it is a compromise between array size and quality of drivers. If you are using high-quality drivers, there is a big reason not to go floor to ceiling, namely money.

The Jordan array, which uses 4 50mm drivers in vertical array, has an excellent reputation. It obviously shares something in common with the big arrays, and single driver setups. Both the Jordan and Bandor drivers are rather pricey.

An alternative might be to use 3" Tangband drivers, say 6 to 8 per side.

But to get back to your post, you say, "I think only going half way is a big mistake," but why? (As several highly regarded systems do indeed "go halfway".)

[DSP_Geek]

You can combine short line arrays with baffle step compensation if you low pass the top and bottom drivers at 0.707 of the baffle step frequency. I have the derivation at:

http://www.diyaudio.com/forums/showt...752#post650752

With this technique you get

LM
LM
FM
FM
FM
FM
LM
LM

where LM = low mid and FM = full mid.

Since BSC is a first-order lowpass, it turns out that for realistic baffle widths the transition distance (TD) will stay roughly the same over an octave or two if you don't cross over to a tweeter, and about two or three octaves if you do. Room effects from the LM drivers coupling with the floor and ceiling will extend that a bit more.

I have another trick to smooth TD, but it's as yet untested so I won't inflict it on you yet.


Cheers,
Francois.

[DSP_Geek]

Originally posted by johninCR
Other than that you are going to measure and tune after they are built if you want a flat FR, because you aren't going to be able to reliably predict what will happen. Why? If you have enough drivers in a line to exhibit array behavior at some distance (which is frequency and line length dependent), then you will have a speaker that changes in tonality depending your listening distance.

John speaks TRVTH.

On the other hand, if one takes care to ensure the transition distance is well controlled over the range of the speaker, then the tonality also will stay reasonably even with distance. Changing line length with frequency is a possible way of doing that, but it's a non-trivial problem, particularly in the upper mids. I'm still messing with frequency contouring networks - as a matter of fact I'll be running another test tonight.


Francois.

[nerd of nerds]

somebody mentioined using 4 bandor or jordan 2" drivers in a mini array...

Why not try this with (still excellent but) cheaper drivers such as the 2" hivis at PE or the 2"auras at madisound?

PE even has some pretty spiffy (visually) new drivers in...

I have long contemplated the idea of a mini (4 drivers per side) array using the 3" hi-vi full range...

One could make some VERY WAF friendly speakers using an array of mini 2" or 3" drivers...

I was thinking something like the seven veil loudspeakers only w/ hivi drivers