I asked this aswell,
try :
http://www.diyaudio.com/forums/showthread.php?s=&threadid=6030&highlight=acoustic+centre
The most practicle solution was from Planet 10 :
"" 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. ""
try :
http://www.diyaudio.com/forums/showthread.php?s=&threadid=6030&highlight=acoustic+centre
The most practicle solution was from Planet 10 :
"" 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. ""
This is covered in exquisite detail in Joe d'Appolito's "Measuring Loudspeakers." It requires just a basic FFT or MLS or similar system.
Relative offset (that which you care about) can be estimated by mounting the drivers on a common baffle, putting a test mike at the desired optimal listening spot, and blipping the drivers with an impulse. The times of flight can be measured for each driver and the difference is the relative offset. Getting a precise time of flight by eye is pretty crude and inaccurate, but this will at least get you in the ballpark. But really, if you go to all this trouble, it's probably worth doing correctly using the phase-unwrapping technique d'Appolito recommends.
Relative offset (that which you care about) can be estimated by mounting the drivers on a common baffle, putting a test mike at the desired optimal listening spot, and blipping the drivers with an impulse. The times of flight can be measured for each driver and the difference is the relative offset. Getting a precise time of flight by eye is pretty crude and inaccurate, but this will at least get you in the ballpark. But really, if you go to all this trouble, it's probably worth doing correctly using the phase-unwrapping technique d'Appolito recommends.
Hi Simon, Sy
The seas coax - do they have a built in x-over between the mid/tweet? - I'm pretty sure this isn't mentioned on the seas site..
Sy, I read about the pulse measuring thing somewhere - I'll have a look through my links tomorrow and see if I can dig out something (could have been in speaker workshop, but I'm way too tired now)
Good luck Joz!
Rob
The seas coax - do they have a built in x-over between the mid/tweet? - I'm pretty sure this isn't mentioned on the seas site..
Sy, I read about the pulse measuring thing somewhere - I'll have a look through my links tomorrow and see if I can dig out something (could have been in speaker workshop, but I'm way too tired now)
Good luck Joz!
Rob
Rob,
I don't think the drivers have a built in x-over, although it doesn't specifically say one way or the other.
However, they seem to be recommending a x-over frequency at 3500Hz, 2nd order.
3.5kHz is a bit high for my liking but presumably due to excursion limits / relatively high resonant freq of the tweeter (it's the 25TFFN/G -H615). Maybe could go a bit lower with a 4th order filter ??
Si
I don't think the drivers have a built in x-over, although it doesn't specifically say one way or the other.
However, they seem to be recommending a x-over frequency at 3500Hz, 2nd order.
3.5kHz is a bit high for my liking but presumably due to excursion limits / relatively high resonant freq of the tweeter (it's the 25TFFN/G -H615). Maybe could go a bit lower with a 4th order filter ??
Si
By the way, the nulling technique will not work for a large variety of crossovers. It would be best to do "bare" as Dave implies. The problem with this method is that crossover frequencies are often at points where at least one of the drivers is rolling off. You'll get a result which is off because of the phase lag (or lead) from the driver's frequency response characteristic.
d'Appolito's book will tell all. Basically, you measure the response of each driver at a fixed point, then unwrap the phase by manipulating time delay until each driver shows a minimum phase characteristic. The difference in the two delays times the speed of sound is the difference in acoutical center between the two drivers.
d'Appolito's book will tell all. Basically, you measure the response of each driver at a fixed point, then unwrap the phase by manipulating time delay until each driver shows a minimum phase characteristic. The difference in the two delays times the speed of sound is the difference in acoutical center between the two drivers.
Hi Guys
Well, I had to order the book from audioexpress in the end, as the UK supplier has stopped selling them..
While I wait, what would happen if you used the same sine wave tone through the mid and tweet, and used a mic and pc based 'scope? - would it show the 2 seperate waves, which you could then watch as you moved the tweet back and forwards untill they matched, or would it just show a mess.?
I will be able to try this after I've moved (and had time to format/re-install windows - my pc's not quite right at the mo.
)
But I thought it may be a cheap solution?
Rob
Well, I had to order the book from audioexpress in the end, as the UK supplier has stopped selling them..
While I wait, what would happen if you used the same sine wave tone through the mid and tweet, and used a mic and pc based 'scope? - would it show the 2 seperate waves, which you could then watch as you moved the tweet back and forwards untill they matched, or would it just show a mess.?
I will be able to try this after I've moved (and had time to format/re-install windows - my pc's not quite right at the mo.
)But I thought it may be a cheap solution?
Rob
I think most people here intend to find the acoustic centre of their drivers to get an even response in a multiway system (if they bother at all).
Since the acoustic centre is frequency-dependant: why not determine it for the crossover-range ?
I'd suggest the following:
Take a source of pink noise and pass it through a bandpass filter that covers the crossover range approximately.
Then feed this signal to both channels of an amp. Connect each driver to one channel and adjust balance so that you get approximately the same loudness from both drivers.
Now you invert the phase of one of the drivers and move them relative to each other until you hear an SPL minimum. Now you still don't know the acoustic centre of the drivers but you know how much they have to be electrically delayed or mechanically set off, relative to each other.
This won't work for coaxial drivers of course.
I haven't tried this method personally but logic suggests that it should work.
Regards
Charles
P.S. An alternative would be shifting the phase of the signal by the use of tuneable allpass filters and then measure the delay of the allpass for the desired frequency range with an oscilloscope and a function generator (which then would be suitable for coaxial drivers as well).
Since the acoustic centre is frequency-dependant: why not determine it for the crossover-range ?
I'd suggest the following:
Take a source of pink noise and pass it through a bandpass filter that covers the crossover range approximately.
Then feed this signal to both channels of an amp. Connect each driver to one channel and adjust balance so that you get approximately the same loudness from both drivers.
Now you invert the phase of one of the drivers and move them relative to each other until you hear an SPL minimum. Now you still don't know the acoustic centre of the drivers but you know how much they have to be electrically delayed or mechanically set off, relative to each other.
This won't work for coaxial drivers of course.
I haven't tried this method personally but logic suggests that it should work.
Regards
Charles
P.S. An alternative would be shifting the phase of the signal by the use of tuneable allpass filters and then measure the delay of the allpass for the desired frequency range with an oscilloscope and a function generator (which then would be suitable for coaxial drivers as well).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.