DiyMobileAudio.com Car Stereo Forum banner

1 - 20 of 30 Posts
W

·
Guest
Joined
·
0 Posts
Discussion Starter #1
Maybe consider this a companion to the A-Pillar line array thread. Seems as though the guys over at carsound already considered pillar arrays ... perhaps not to completion, but since they apparently stole my idea (about a year before I had it ;)), it's time for a new idea. And this one is better ... much better :)

Let's review D'Appolito's invention : the midrange/tweeter/midrange or MTM array. It has three principal merits, as I see it :

1. The lobing error is zero, for ears on-axis with the tweeter, no matter what midrange/tweeter crossover you use. Lobing error is simply a measure of the "tilt" in the radiation pattern. And the physical, vertical symmetry of the array means that the radiation lobe is symmetric around the tweeter. Allows you to design a crossover without a tilted-lobe concern ... in other words, it's easier to maintain a uniform response, for ears on-axis with the tweeter, as frequencies sweep throught the crossover.

2. Vertical radiation is reduced, minimizing floor/ceiling reflections. This is essentially a manifestation of a comb filter. Same reason why ribbons have great horizontal dispersion, but limited vertical dispersion.

3. Two midranges of low-ish efficiency will be easier to passively cross to a higher efficiency tweet, because you get a 6dB improvement in midrange efficiency from using two drivers (twice the cone area gets you 3dB, half the impedance or twice the power gets you another 3dB). There's also a linearity improvement closely tied to this as well ... for any given SPL, each driver will move less air, and therefore generate less distortion, than a single driver.

I have yet to see an implementation in a car that enjoys all of these benefits. For that matter, I've seen plenty of MTM arrays gone wrong in home audio ... floor standing speakers with no hope of placing the ear on-axis with the tweeter, for example.

How can one implement a real MTM array in a car? Let's look at the keys for realizing each benefit :

1. The lobing concern. Forgetting for the moment the possibility of time alignment, which can effectively be employed to "steer" the lobe, what's really required for uniform response (in other words, minimize the comb effects from two midrange drivers) is that the two midrange drivers be placed the same distance from the ear. Each ear, for each passenger in the ideal case. How close is "close enough", we'll develop in a later post.

2. In order to minimize reflections, we want the array line to be perpendicular , rather than parallel, to the most troublesome reflecting surface ... which is undoubtedly the windshield, if we want this array anywhere near ear-level. Which of course we do :) The A-Pillar line array goes along the wrong axis, if minimizing windshield reflections is a concern.

3. Efficiency improvement, and lineaity improvement. Pretty much holds for any two midrange drivers (per channel) inside the car.

After that long-winded intro, here's the idea :

1. Tweeter in the sail panel.
2. One midrange driver high in the door, below the sail panel.
3. Second midrange driver in the A-Pillar, above the sail panel.
4. Alignment : each midrange needs to be the same distance from the ear. A line drawn through all three drivers should be perpendicular to the windshield.

This geometry, and aligment, looks quite possible in the late model Honda Civic. I'll have to take some detailed measurements to verify. Note that, if you can achieve this alignment, the two mids can be on the same amp channel ... no need for individual time alignment on each mid. Hell even the tweet could be passively crossed to the mids ... pretty close to good physical time aligment here as well.

Of course, a midbass low in the door would complete the front stage :)

What drivers to use? Why not the Aura Whisper for the midrange :) Two per channel improves efficiency, fitment in pillars & doors is pretty easy. One may be tempted to ask ... if the midrange drivers are full-range, why even use a tweeter in the sail panel? The reason is this : small "misalignment" of the mids creates a nasty high-freq null (our old friend, the comb filter). We'll show the math, but a small fraction of an inch in midrange alignment can create problems in the treble. So crossing to a small tweet high ... upper octave or two ... is probably a good idea.

What do you guys think? Is it easy to picture what I've tried to describe? A high, MTM array that runs perpendicular to the windshield, as it should :) Pretty easy to implement ... don't need loads of DSP channels. I know there's at least one member here who already has an Aura Whisper high in the door :)
 

·
Registered
Joined
·
578 Posts
If anyone ever pulls it off I wanna see pics :)

Back in 2000 or 01... I remember a white Ford Festiva (or Fiesta), featured in CA&E, owned by a skilled installer, that used Focal drivers to create an MTM setup in the kicks. Used the 5" focal subs up front & 2 15" Focal subs under the back seat (cut into the car floor to fit them).

That is the only time I seen anyone attempt an MTM setup in a car. Not much of a car to brag about, but the install was top notch & with the price of gas now days... the ultimate commuter ;) Wish I still had the mag...
 

·
Registered
Joined
·
434 Posts
your ideas are always very intriguing mr werewolf
 
W

·
Guest
Joined
·
0 Posts
Discussion Starter #4
Thanks guys :) In order of importance :

1. Whenever you have multiple mids per channel, they need to be the same distance to your ear. Or, use time alignment ... but that requires a separate dsp/amp channel for each mid. I think keeping the same distance to the ear is preferrable, if at all possible. As we all know, the car presents many challenges ... on the list of positives, listeners don't wander around the listening space much :)

2. A straight line drawn through both mids (two points do define a straight line, right? ;)) should be perpendicular to the windshield.

3. The tweet should be between the two mids, also equidistant to the ear. This point is least critical, I think, because it's (at least somewhat) common to devote a separate dsp/delay/amp channel to the tweet.
 

·
Registered
Joined
·
164 Posts
WOW talk about crazy coincidence, i was thinking of this idea just a couple of days ago. Here is my proposal on my cardomain website. I wrote TMM there, but was actually wanting to do an MTM instead, mounted in the corners of the dash (perpendicular to the dash), where the windscreen meets the dashboard itself. What are your thoughts on this form of mounting wolf?

Here is the link to the idea i posted. http://www.sounddomain.com/ride/751397/2

Note also that included there is a schematic of a passive crossover designed for me using LEAP by madisound to link the auras to an LPG 26NA. Crossover point is approximately 4.5KHz. I tweaked the original circuit a little to add bypass switches an a variable L-Pad so i can at least tune the system a little better. I'm stuck going with passive because i only have a 3 way crossover on my head unit and i dont want to waste money upgrading it

I don't know enough about MTMs though. Does anyone know of any white papers (sort of like jim grifffin's line array whitepaper) which describe how an MTM works?

I would probably be your first guinea pig too, but i wont have a chance to get the system installed till december
 

·
Registered
Joined
·
12 Posts
What a great idea. :D

One question though. How could the midranges ever be the same distance from you ear? It seems that one would be at least 4-6 inches further forward, unless you drastically lowered the one in the door?
 

·
Registered
Joined
·
127 Posts
Time alignment. You delay the midranges so that the sound from each hits your ear at the same time, making your brain believe they are all equadistant.

(threadjack, where are you in MN?)
 
W

·
Guest
Joined
·
0 Posts
Discussion Starter #8
blackreplica .... are the mids in your plan the same distance to your ear? If so, i'd say run widdit.

drumah .... take a measuring string out to your car, you may be surprised. If I run a string from my left ear to the upper/forward part of the door (just below the sail panel), then move that same length of string up to the A-Pillar, I locate a spot just above the sail panel. The two points so located form a line almost perpendicular to the windshield. A sail panel tweeter will not be exactly on that same line ... but's it's close.

Or, you can move each mid a bit as you please, and use time alignment as suggested.
 

·
Registered
Joined
·
1,335 Posts
Now what about having all three drivers on the same plane?

I was thinking about building a small to med sized baffle/pod to sit up on my dash. One on each side of the dash. Maybe triangular in shape. While keeping the perpendicular alignment to the windshield for the mids. The tweet wouldn't be in-between the two mids. It would be off to the side.

Any thoughs?
 
W

·
Guest
Joined
·
0 Posts
Discussion Starter #11
The ideal case :

- each mid on the same channel should be same distance to ear.

- line drawn through the two mids should be perpendicular to the windshield ... measured from any angle.

The second point would be hard to achieve for mids on the dash it seems? I wouldn't worry too much about the tweet ... get it close, time align the rest :)
 

·
Registered
Joined
·
164 Posts
werewolf said:
blackreplica .... are the mids in your plan the same distance to your ear? If so, i'd say run widdit.
Yo wolf, yeah they will ideally be vertical off the dash so the distance to the ear will be approx equal for both drivers. I guess its worth a shot eh? Only trouble is that mounting it this way will mean it wont be pefectly perpendicular to the windscreen. Oooh can't wait to try it out
 
W

·
Guest
Joined
·
0 Posts
Discussion Starter #13
Perhaps the greatest virtue of the Aura Whispers (again, due credit to npdang & DS-21 for car use :) ) is that they are small enough, and cheap enough, to experiment!
 

·
Registered
Joined
·
151 Posts
Quick noob question. If the lobing error is zero in an MTM, does that imply that the center to center spacing at the x/o frequency is no longer a concern?
 

·
Registered
Joined
·
164 Posts
PlanetGranite said:
Quick noob question. If the lobing error is zero in an MTM, does that imply that the center to center spacing at the x/o frequency is no longer a concern?
Hardly a noob question i think! That was the main question in my mind. I'm reasonably familar with line array theory, but it kinda goes against MTM theory apparently since it does not advocate a minimised center to center spacing for the mid, otherwise it would be TMM
 
W

·
Guest
Joined
·
0 Posts
Discussion Starter #16
Lobing error is very carefully defined, something like the vertical slope of the dispersion pattern on-axis with the tweeter. Doesn't mean there's no lobe, just that the lobe is symmetric around the tweeter ... no matter what electrical or acoustical phase relationship exists in the xover between mid & tweet.

If the ear is exactly centered bewteen two mids, you won't have to worry about center-to-center spacing. As soon as you have three or more mids in a linear array, your ear can no longer be exactly centered between any two mids in the array ... your ear cannot be centered between mid number 2 from the top and mid number 5 from the bottom, for example. Know what I mean? So center-to-center spacing matters as soon as you have three or more mids (in a linear array), if my brain is working well tonite. If the mid array is circular, or some other non-linear geometry, you can achieve equal spacing to the ear with more than two mids.

CTC spacing may also be a consideration if you want a large sweet spot for even a simple MTM array. But fortunately, cars don't need a very large sweet spot for listening :)

Make sense?
 

·
Registered
Joined
·
164 Posts
Or here's another idea. Lets assume that center to center distance does matter. What if the tweeter was offset from the midrange, allowing the midrange to be positioned right next to each other? I got the idea from here :



where the tweeter, as you can see, is slightly offset from the mid.

I wonder if mounting it this way would minimise problems associated with the c-to-c distance without removing the benefits of using an MTM.

This basically explains what i mean:

This:



Instead of this (my original idea):

 
W

·
Guest
Joined
·
0 Posts
Discussion Starter #18
Remember that whenever you have two drivers playing the same freq, you get a dispersion pattern that falls "off axis". This pertains to two mids on the same channel, or a tweet & mid on the same channel at crossover. Actually, it also pertains to all the tiny elements that make up a cone of a single driver. Only a true point source radiates unifromly in all directions for all frequencies.

So let's think about the MTM with the misaligned tweeter. Below crossover, the tweet isn't even playing. So we just have two mids, vertically aligned. And they are closer together ... which will widen the dispersion pattern vertically, or make the sweet spot "taller" through the midrange. At higher freqs, only the tweet is playing ... so the high freq dispersion is the same in both cases ... only offset, with the tweet misaligned. So at crossover, we'll have a horizontal lobing error that depends on the xover phase characteristics.

How audible? Who knows ... it's late. :(
 

·
Registered
Joined
·
164 Posts
werewolf said:
Remember that whenever you have two drivers playing the same freq, you get a dispersion pattern that falls "off axis". This pertains to two mids on the same channel, or a tweet & mid on the same channel at crossover. Actually, it also pertains to all the tiny elements that make up a cone of a single driver. Only a true point source radiates unifromly in all directions for all frequencies.

So let's think about the MTM with the misaligned tweeter. Below crossover, the tweet isn't even playing. So we just have two mids, vertically aligned. And they are closer together ... which will widen the dispersion pattern vertically, or make the sweet spot "taller" through the midrange. At higher freqs, only the tweet is playing ... so the high freq dispersion is the same in both cases ... only offset, with the tweet misaligned. So at crossover, we'll have a horizontal lobing error that depends on the xover phase characteristics.

How audible? Who knows ... it's late. :(
Thank you sir! given the fact that we are trying to reduce vertical dispersion it might be a better idea to stick to the traditional MTM layout perhaps. it might be a bit tricky trying to make sure the listening position is between the 2 mids though.
 

·
Registered
Joined
·
415 Posts
i dont know whats really going on because i just skimmed through the thread, but this was what i was trying to explain to someone over on ECA.



Blue circle = mids locations [set at equal distances from the ear] ( Seas Lotus Reference's 8) )
Red = possible tweeter locations ( LCY 110's :p ) [tweeters can be angle towards the listener for more on-axis, and the tweeters in the doors will be at a greater angle to the listener (then kickpanels), for a possible wider soundstage]
 
1 - 20 of 30 Posts
Top