How to acquire FR measurement data for 2.5 way vented speaker simulation in VXCAD best?

[Kwesi]

Hi all,

I have a 2.5 way on the bench with the task to renew the drivers and xover, reuse is only the enclosure.
Still a little craftsmanship to do, but I'm thinking how to generate the measurement data for VituixCAD xover simulation best.

I plan to acquire following measurement data:

• 0° and 10° step hor/ver spinorama FR of every driver separately within the enclosure, at least half space +/-90°. Ideally, I can measure outside with >2m mic height in 1,5-2m distance on a little turntable.
• Near-field data of both woofers and the port to merge below 200-300 Hz after gating the spin data.

Challenge is how to handle the two midwoofers acting to one enclosure and resonator, but driven via two seperately filtered electrical pathes.

I think the nearfield data and merge is easiest:

1) Connect both woofers in parallel to the measurement amp. Run the sweeps.
2) Measure port nearfield response, and nearfield response randomly* of one of both woofers.
3) Prepare diffraction response of both woofer positions in Vituixcad baffle simulator.
4) In Vituixcad merge tool:
- load random woofer nearfield response and scale to 1x woofer cone area
- load port nearfield response and scale to 1x port area, and then substract -6dB
- load specific woofer diffraction response
- load specific woofer gated anechoic response --> merge and export
- repeat with second woofer

*) Rationale: Both woofers shall act same in the nearfield at these low frequencies and driven in parallel. I would nevertheless measure both, but more for integration check that the cones are healthy and no standing wave resonance within the enclosure introduces disturbance - so to verify that they act same.


Spin data gets more interesting.

a) Option "lazy":

• Connect both woofers in parallel, go to somehow ~2m measurement distance, and do only 1 spin for the woofer path
• Substract -6dB from the responses, and just use the same for both channel xover path simulations.

For the horizontal data it might get not that wrong, but how will that affect overall vertical system response simulation, especially in the range they are filtered differently it would be messed or? Or I am even "super-lazy" and only take horizontals for the filter tuneset in simulation...


b) Option "Cover":

• Build a tight and stiff cover, that fits exactly one rebate hole for the woofer
• Close/plug the port.
• Mount one woofer, and close the rebate for the other with the cover. Do the spin.
• Exchange woofer and cover. Do another spin for the second woofer data.

One woofer will not interact with the port as two will do, so better close it as the measurements were anyway gated below 200-300Hz.

Sounds like a solid solution?
The cover may have a bit other shape than a woofer cone, so there might be some difference in baffle reflections between the measured and later used system.
And I'm too lazy to build that cover ;-)


c) Option "electrical damping":

• Close/plug the port.
• Mount both woofers to the enclosure. Connect one to the measurement amp, and shotcut the clamps of the other. Do the spin.
• Exchange amp connection and shortcut. Do another spin with the second woofer.

Could work, or?
Ideally, the short-circuited driver will see somehow similar conditions like in full speaker operation (there with zero ohm drive impedance from the voltage amplifier at the clamps)?
--> Have to do some math and thinking, anyone concepts/equations/models for input?

Anyhow, could be a solution for the gated measurements above 200-300Hz, with the resonance frequency behavior changes between the different "pneumatic" operating conditions well damped and below that range.
Will try that out first in comparison to both drivers connected parallel (and with one open clamp driver also for interest) before building a cover....


Do you have experience, hints or thoughts about measurement data acquisition from some 2.5 way?

Thanks a lot and best regards
Peter

 

[AllenB]

Connect both woofers in parallel to the measurement amp. Run the sweeps.

Normally this is a good way to obtain directivity data for the pair, however it will also create restrictions. Since this is a 2.5 way, they won't play together at all frequencies so you want separate data for each. Vituixcad can simulate the pair's interaction according to their positions, but only if you have separate data.

Prepare diffraction response of both woofer positions in Vituixcad baffle simulator.

Don't double the diffraction response. If you measure it on the baffle then you don't need to simulate it.

 

[Kwesi]

Vituixcad can simulate the pair's interaction according to their positions, but only if you have separate data.

Year this is option b) and c) then i mentioned, but do you have experience if better measuring with b) just one woofer and the other woofer's rebate plugged/covered, or if it works to c) just shortcut the driver channel you don't want to measure?

Don't double the diffraction response. If you measure it on the baffle then you don't need to simulate it.

For correction of the nearfied responses before merging with gated far-field data, because nearfield data does not contain baffle information.

 

[AllenB]

For both of these cases, they will only be a problem at bass frequencies. This region is sometimes away from the crossover and it's easier to not chase exact measurements, so you have to decide how much leeway you give yourself.