Acoustic centres of drivers/measuring

SpeakerBob

Member
2002-08-15 6:21 pm
UK
Hello all,

I've got to the point in my speaker project where I've got all materials to begin construction, but still need to design the mid/tweeter enclosures. (BAF in the post mon/tues btw,Dave )

I know what they're gonna look like, but need to know how to find the acoustic centres of the drivers, so they can be physically time aligned accurately....(s-speak revelator tweeter/ focal audiom 6wm mid/ sspeak 6585-01 bass)

Does anyone know for sure how to measure this.?

I searched and got this thread:

http://www.diyaudio.com/forums/showthread.php?s=&threadid=742&highlight=time+alignment

But they ended up deciding it was easier to use a co-ax than measure the centres....:bigeyes:

I'm about to e-mail the manufacturers too, to see if they can give me a figure..

Thanx in advance,

Rob
 

roddyama

diyAudio Moderator Emeritus
2002-01-19 9:25 am
Michigan
Hi Rob,

The "acoustic center" of a driver is not one of those hard numbers that you can measure and have it work for all situations and speaker configurations. The “acoustic center” of the driver depends on the shape of the polar dispersion pattern of the driver in the finished box. Actually, the polar dispersion pattern, and therefore the “acoustic center”, will almost certainly change with frequency, size and shape of the box, baffle arrangement, and driver type, along with a list of other factors. The polar dispersion pattern has to be measured in an anechoic chamber with the speaker in its final configuration. Even with this information, the “acoustic centers” will be an approximation that will change off axis because the shape isn’t a perfect circular arc.

What can you do?

First, don’t get fixated on the “acoustic centers”. This is just an approximate starting point. What you really need to do is to make the sound emanating from each driver reach the listener at the same time. This takes us to the following thread.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=5699

The easiest place to start is where a few post in your link said to start. Align the voice coils first. You can experiment by mounting the driver(s) for each range in separate boxes. It will make the task of making changes much easier. As you add x-overs or any other components that may have an effect on the alignment, you have the flexibility to make the necessary changes to give you the best sound. Once you got everything where you want it, you can build the final boxes incorporating what you’ve learned.

Hope this helps,
Rodd Yamas***a
 

SpeakerBob

Member
2002-08-15 6:21 pm
UK
Hello Rodd ,and thanks for your input....

The speaker will be : 2 x scan speak 8565-01 in their own enclosure, with an audiom 6wm in its own enclosure above, and a scanspeak 9500-9900 revelator in its own enclosure above.

I have followed the thread you posted , - btw I believe that a conventional driver would produce a square wave if you super cooled the voice coil - it would also be more linear...... does'nt really mean much in the real world though.....

btw I will be using a 3 way active x- over (esp boards linkwitz-rieley)

The x-over will be phase aligned, which I am assuming means I only have to work out the offset to each driver from where I'm sitting, hence my question...)

you said "What you really need to do is to make the sound emanating from each driver reach the listener at the same time. "

This is what I'm trying to acheive:confused:

I thought if I got the mid point (zero bit, when no signal is present) of each driver the same distance from me, then the drivers would be phase aligned.....

Just trying to find out where on each driver the centre is?

Cheers

Rob
 
RobWells said:
the enclosures will be cylindrical wilth the driver at the end, and no baffle at all....

How big is the baffle, err, the faceplate on the revalator.

Acoustic centres are a moving target that as well as being somewhat nebulous move around as a function of frequency. Somepart of the voice coil is a very rough approximation. You really need at least an SPL meter and a sine wave generator ('puter into amp works fine) to get it right easily. Run a sin wave at exactly the XO frequency. Flip one of the drivers out of phase. Move a driver (top one cause it is easier?) until you get minimum output. The less reflections and/or deader the room the easier it is to do. I guess you could get closer than the voice coil guess with a sentitive ear.

dave
 

SpeakerBob

Member
2002-08-15 6:21 pm
UK
Hello Dave,

I've borrowed the drawings for each driver from their web sites..

The 2x scanspeak 8565-01 per side are in approx 750 high x 280 wide baffle (thats the straight front before it goes into 55mm radius corners..)

Heres the tweeter one :

Rob
 

Attachments

  • draw_d2905_990000.gif
    draw_d2905_990000.gif
    5 KB · Views: 380

SpeakerBob

Member
2002-08-15 6:21 pm
UK
I've just got what you're on about with the centres..... Its the signal addition thing isnt it? ie, as the cone is moving out as part of a 400Hz signal, it's already going in and out from the higher frequencies......

Does this mean there is only phase coherance at the x-over frequencies?

Its looking like I've got to build my active x-over and buy the 3 amps before I finish constructing the cabs....

I'll do the top of the bass unit as a seperate piece, which will let me get started anyway..

Rob
 
Originally posted by RobWells
I've just got what you're on about with the centres..... Its the signal addition thing isnt it? ie, as the cone is moving out as part of a 400Hz signal, it's already going in and out from the higher frequencies......

That too. Even without signal addition the acoustic centre moves as the radiation frequency changes.

Does this mean there is only phase coherance at the x-over frequencies?

You want to get them coherent at the "join",

Its looking like I've got to build my active x-over and buy the 3 amps before I finish constructing the cabs....

Yes.

dave
 

SpeakerBob

Member
2002-08-15 6:21 pm
UK
Can you give me a pointer to these issues?Do the diffraction issues mean I would be better going for a triangular mid/tweeter enclosure a la watt/puppy?

I chose the no baffle route as I thought it was meant to be the best type.

These speakers were originally conceived as a watt/puppy rip off.

I then thought that the b+w 's use the teardrop type shape with no baffle as such, and that this would improve things a bit.

As I'm all set on the bass enclosure, I only need to get this finalised and I can begin construction. - I'll start the bass bins tomorrow anyway.

I'd better hassle the electronics section on components to buy for my esp boards.

Cheers

Rob
 
RobWells said:
Can you give me a pointer to these issues?Do the diffraction issues mean I would be better going for a triangular mid/tweeter enclosure a la watt/puppy?

I chose the no baffle route as I thought it was meant to be the best type.

These speakers were originally conceived as a watt/puppy rip off.

I then thought that the b+w 's use the teardrop type shape with no baffle as such, and that this would improve things a bit.

The problem comes in only with the tweeter and only because it has such a big baffle (for a tweeter) built-in. No matter what you do with this baby it is not going to be baffle-less. The B&W tweeter has no faceplate to speake of and that is why it works. Your tweeter with the 130 mm faceplate and the dome sitting right in the middle will have a very strong diffraction component caused by the equidistant 57mm travel distance from the dome to the face-plate edge.

Ways to minimize this include using felt-damping on the face-plate, extending the existing "baffle" into a more appropriate shape, or hacking/replacing the face-plate to integrate it better with a small pod.

dave
 

SpeakerBob

Member
2002-08-15 6:21 pm
UK
""Ways to minimize this include using felt-damping on the face-plate, extending the existing "baffle" into a more appropriate shape, or hacking/replacing the face-plate to integrate it better with a small pod.""

I'm with you now, Dave..I've just spent a couple of hours browsing through baffle effects and compensation stuff on the web, which has got me a bit nervous about the x-over:bigeyes:

I'm poss gonna have to adapt the esp boards to include for baffle step compensation..I kind of like the idea of running the pre through an eq curve before it hits the linkwitz-reilley x-over. Think thats going to be another thread though!


I'm going to use the "layers" thing to do the tweeter enclosure, so I can flush mount it into any shape I want, and radius the edges with a router. How about an egg/oval type shape for a start - this will give different distances from the driver to the edges.

The mid will have radiused edges and be flush fitted also. I can change the shape of that a bit to a more oval shape if it helps things.

Once more thanks for your input..

Rob
 
RobWells said:
I'm poss gonna have to adapt the esp boards to include for baffle step compensation ..I kind of like the idea of running the pre through an eq curve before it hits the linkwitz-reilley x-over

You could use either an active circuit as in Alex Megann's article at the above link, or adapt the passive RL circuit.

I'm going to use the "layers" thing to do the tweeter enclosure, so I can flush mount it into any shape I want, and radius the edges with a router. How about an egg/oval type shape for a start - this will give different distances from the driver to the edges.

The right kind of direction. I would consider hacking the face plate too...

Download the BDS and see what it predicts for whatever shape you choose.

The mid will have radiused edges and be flush fitted also. I can change the shape of that a bit to a more oval shape if it helps things.

The mid should be fine with the smallest shape you can put behind it that is consistent with a smooth shape.

dave