The ways to determine the offset of drivers in a two way project???

[ergo]

Hey.

I have Vifa XT25 ring radiator tweeter and Peerless HDS134 woofer ready for my next project.

I would like to try this time to construct the box so that the front panel has a step and tweeter's acoustic centre is alined with woofers acoustic centre.

Now I have a test box for woofer and I would make up a rig that would hold the tweeter on top of this box. This way I would be able to move tweeter on "Z" axis.

How would you guys go with this on measurement side? How to find out the point where the acoustic centres are on one plane?

Preferably all this with either JustMLS, Speaker Workshop or LspLab.

Regards,
Ergo

[SY]

Follow the procedure outlined in d'Appolito's book on testing loudspeakers. Basically, you'll put the mike in a fixed position, put the two divers on a common baffle (or fix their distance from the mike), take two separate pulse response measurements, then unwrap the phase of each one (putting in a delay in most software) until you have a minimum phase response. The difference in time between one unwrap and the other can be converted to relative zero delay planes.

This procedure avoids conflating the zero-delay plane with the phase shift caused by driver rolloffs.

[catapult]

Follow the procedure outlined in d'Appolito's book on testing loudspeakers.

That works better in theory than in practice. The problem is determining when "you have a minimum phase response." The easiest way, especially if you have a DSP crossover box handy (no swapping caps and coils), is to build a standard LR4 acoustic crossover. Then reverse the lead to the tweeter and tilt the baffle until you get the deepest null at the XO frequency.

[ergo]

OK, the following is a way I finally solved the problem. Or at least I think I did

http://f5.infonet.ee/ergo/offset/offset.html

Ergo

[5th element]

There is another way I can think of doing this but my explanation might not be very clear.

When setting up drivers in LSPcad for accurate simulation, you take a combined response of both the tweeter and woofer playing without a crossover, like you would measure the tweeter normally but with the woof in parallel. You also take the individual responses of the woof and tweet. Then you import both the woof and tweet responses into LSPcad & you use the combined response of the two you made before as a target response.

Then you alter the Z axis until the simulated combination and the measured combination are the same.

Now if the woof and tweet were "time aligned" I would imagine zero modification of the Z axis would be rquired.

This would take quite a bit of time I think but should get the job done.

Also bear in mind that when you actually put crossovers onto the drivers you do add phase shifts which will actually alter the Z axis so to speak, see linkwitz work on the orion for this. But if you already know what xover freq and slope you are going to use for this then you could work out the additional delay to add or take off in your design for the cabinet contruction.

Matt