How do you meassure

[henrikhegh]

Hi,

First i would like to apologies for my bad english. Its not my first language.

I would like to know how you meassure speakers. There are many was do this, and im curious.

I use Clio and LspCAD. Meassure gated Mls at ~1.5m using minimum fase, and mic placed between the drivers. Then i delay the units to be time correct (minimum fase remove time information). The I import it to lspcad and start designing the crossover.

How do you do it?

Hers what im working on now. A 3way Scan-Speak.
http://forum.speakerbuilder.dk/uploads/1/Filter01_HRH_frq.jpg

And with the mid reversed:
http://forum.speakerbuilder.dk/uploads/1/Filter01_HRH_frq_modfase.jpg

[Wolf]

Hi,

First i would like to apologies for my bad english. Its not my first language.

I would like to know how you meassure speakers. There are many was do this, and im curious.

I use Clio and LspCAD. Meassure gated Mls at ~1.5m using minimum fase, and mic placed between the drivers. Then i delay the units to be time correct (minimum fase remove time information). The I import it to lspcad and start designing the crossover.

How do you do it?

Hers what im working on now. A 3way Scan-Speak.
http://forum.speakerbuilder.dk/uploads/1/Filter01_HRH_frq.jpg

And with the mid reversed:
http://forum.speakerbuilder.dk/uploads/1/Filter01_HRH_frq_modfase.jpg

Nice to see you around again, Henrik!
Is it a "Filler-Driver" 3-way? It looks as though you're only running the mid from 1-3kHz, and that is not a very large area. Typical is 3-octave rule, but you can sway from that. But I digress...

I measure the impedance, T/S with a WT2 device, and if I can get a measurement from someone else, or use a reliable chart to trace the anechoic FR plot I will. I'm just not as fluent with the FR rig yet.

So far that has worked getting the data, I typically use Unibox for box alignments, and then I import into the Response Modeler 2.0 for baffle correction, amount of baffle step comp, and nearfield bass-splicing. Then those corrected files are imported in to Jeff's other program, Passive Crossover Designer 6.20 to knock out the xover simulation.

This is not to say it's the method I always use, as I have also voiced by ear with impedance plot as the only aid post box-simulation. It's more rewarding when you get it right this way!

Nice to see you back!
Wolf

[romanbednarek]

Hi,

First i would like to apologies for my bad english. Its not my first language.

I would like to know how you meassure speakers. There are many was do this, and im curious.

I use Clio and LspCAD. Meassure gated Mls at ~1.5m using minimum fase, and mic placed between the drivers. Then i delay the units to be time correct (minimum fase remove time information). The I import it to lspcad and start designing the crossover.

How do you do it?

Hers what im working on now. A 3way Scan-Speak.
http://forum.speakerbuilder.dk/uploads/1/Filter01_HRH_frq.jpg

And with the mid reversed:
http://forum.speakerbuilder.dk/uploads/1/Filter01_HRH_frq_modfase.jpg

I've used Speaker Workshop or SoundEasy to measure drivers with the microphone set at a fixed height so that the phase information is included in the measurement (I will sometimes manually process that data to get a minimum phase response if I had to splice any low frequency near-field driver responses for my crossover simulation and in those cases it was required to determine the amount of offset to use for each driver in the crossover simulation).

When working with three-way speaker designs with measurements that are not based on minimum phase data (as well as those that have had minimum phase data), I usually adjust the driver offsets to match a listening distance that is further away than my microphone measurement distance using a conversion process similar to the one at this link... http://www.rjbaudio.com/Audiofiles/driver%20offset%20calculator.html

http://www.rjbaudio.com/Audiofiles/Driver%20offset%20diagram.JPG



Wolf, I think that the midrange slopes are a bit too steep for a TP filler driver design (I've thought about making one in the past but it is tough to find a good midrange that can cover that wide of a range without too many related or unrelated sacrifices).

[henrikhegh]

Hi,

Im not using a fillerdriver, and not a Duelund. The drivers are not time coherent, so they wont work with my speaker. But i made a calculator for Duelund filters here :
http://pro.speakerbuilder.dk/Synkron/ where you can see the responsecurves and download target curves. Very handy if your drivers are time coherent.

The crossover points are found from what worked best regarding fase and freq. response.

I would like to know, how you "determine the amount of offset to use for each driver in the crossover simulation". Its something ive been working on with lspcad. It can be calculated from groupdelay as some do. I find it by meassuring both drivers individually and summed. Then turn the delay up until the simulated summed response matches the meassured summed response. That has given me the most precise results

Here i have showd how i do, though in Danish.
http://forum.speakerbuilder.dk/forum_posts.asp?TID=102

But would like to know if there are other ways.

Roman > The sound starts where the voicecoil is clued to the cone (approx). So the calculations above is not complete. For a W15CH001 Seas the delay (with a seas Excel tweeter) is 49, the drivers between. Thats the time the basdriver is further delayd then the tweeter, besides mic placement.

[henrikhegh]

Hey,

Wolf > Where do you get frequency meassurements from?
Wolf > Yes the mid is narrow. And to make it worse, its a Scan-Speak 12M

[Wolf]

Hey,

Wolf > Where do you get frequency meassurements from?
Wolf > Yes the mid is narrow. And to make it worse, its a Scan-Speak 12M

I use SPLtrace to trace provided graphs that are reliably taken. Just import and click into oblivion...

My sources have been zaph-audio, augerpro's plots, .pdf sheets, manufacturer's plots, etc. I don't trust all manufacturer's documentation, but the reliable ones have not steered me wrong. As long as they have been taken with the typical 4' baffles, or in an anechoic room or even outside, the results are repeatable and usable.

I use Jeff B's Response Modeler to alter the graphs to the project at hand, and then extract minimum phase for each driver post-adaptation.


Why are you xovering the 12M so high for the HP? I thought it could go lower.
Later,
Wolf