Measurement Data for NatalieP??

[gmikol]

Hi--

I don't have access/opportunity to do my own measurements, so I've been investigating doing a simulation-only design. I know it's not the best option, but it's what I've got for now. I'm trying to reproduce the NatP as a test of my methodology. I noticed, though, that the PE measurement data does not seem to match the data in the first post of the NatP thread.

So I'm curious...where does Evil Twin get his measurement data from? Does he get it from someone? Does he do it himself? Infinite Baffle, IEC Baffle, In-Box?

The reason I ask is that I've got a big lump in my FR around 500-2k that I don't see in ET's composite plots, so I'm trying to figure out what's causing it. The transfer functions come out OK, so it's got to be in the measurement data?

Anybody got any insight?

Thanks--

Greg

[Dennis H]

ET rolls his own in the actual box. PE measures in an infinite baffle. Apples-oranges.

[Evil Twin]

As Dennis correctly points out, for designing an actual system, ideally the drivers should be measured in the actual enclosure, especially for an MTM.

Otherwise, one must try to use "infinite baffle data" (which doesn't exist in the real universe, as there are no "infinite" baffles, and do your best to model the enclosure effect with tools such as BDS or EDGE, and then model driver positioning and off axis response combination in LspCAD.

Direct measurements are more reliable, then multiple steps of estimated inferences.

In my case, I use Praxis with a wideband calibrated condensor instrumentation microphone, "borrowed" from the Star Destroyer acoustics lab used to monitor vibration and ultrasonics output from the Turbolasers; they have a specific acoustic signature which develops as they approach the point where a maintenance or total rebuild is required.

[ThomasW]

https://www.htguide.com/forum/showthread.php?t=11700

[Dave Bullet]

gmikol,

You can get close - but you need to do a couple of things.

Each baffle has its own signature. First step. you need to subtract the signature of the infinite baffle (from the FR response plot you download , then add in the baffle ET used in the design. This will give you the approximate in-box response.

2. You need to generate phase data for your "new" FR curves. Since speaker drivers are considered minimum phase, algorithms like hilbert transform can be used to extract the phase from the FR. Without phase - combining frequency response is meaningless in a crossover simulation.

3. Once you have the above - you can import the new FR curves (with phase data) into your xo modeling program. You may also need to enter the acoustic offset of the drivers (usually for the woofer since the acoustic centre is behind the tweeter). For example - this applies if you are using speaker workshop.

If you have time on your hands - you can get a much better tutorial and description from these sites (also links to the FRD consortium's baffle diffraction tool - which does step1. and frequency response combiner tool - which does steps 2. & 3.

http://www.rjbaudio.com/Audiofiles/FRDtools.html

http://www.geocities.com/woove99/Spkrbldg/DesigningXO.htm

There are some cavaets with the above approach. For example - if the woofer was measured on-axis, the upper end rolloff will appear to be shallower than in a design where all measurements are done on / near the tweeter axis (since you are off-axis to the woofer - so rolloff will be greater). what this means in practice is you might see a hump in the woofer to tweeter stop band caused by a more shallow rolloff that doesn't exist at the on-tweeter axis measurement point.

Anyway - the above should be enough to get you started and have some fun.

David.