Time alignement fundamentals - diyAudio
Go Back   Home > Forums > Loudspeakers > Multi-Way

Multi-Way Conventional loudspeakers with crossovers

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 28th March 2004, 05:40 PM   #1
Kjetil is offline Kjetil  Norway
diyAudio Member
 
Kjetil's Avatar
 
Join Date: Feb 2004
Location: Norway
Question Time alignement fundamentals

As the membrane of a speaker moves back and forth continously, I understand no "exact" reference point can be used to calculate alignement between drivers no matter wether the approach is sloped baffle, stepped baffle, or inverted driver geometry. However, most advices on the topic refers to the voice coil centre, by means the centre of the motor driver frontplate, as a suitable reference point. I would assume the vertical center is at the center of the driver, but the horizontal distance from the frontplate of the driver to the motor driver frontplate is the data I am looking for..

My question is quite simple; Can this point be calculated from any T/S parameters or by any other means, or do I have to contact the driver manufacturer to get the details? I guess measuring it would be an approach, but I do not have access to such equipment.

Any words of wisdom would be greatly appreciated, I'm banging my head against the wall here..
  Reply With Quote
Old 28th March 2004, 05:56 PM   #2
sreten is offline sreten  United Kingdom
diyAudio Member
 
Join Date: Nov 2003
Location: Brighton UK
Simply put you need Bessel or similar crossovers for a start.

Even with these physically aligned voice coils are not
time aligned, this is due to the delay of the crossover,
and phase response of the driver.

(much as some manafactures claim the opposite)

Perfect time alignment is not possible, but can be approximated
by constant delay crossover topologies, and TBH there are
usually far more apposite problems in getting the c/o right.

If you want to go there measurement is the only way AFAIK.

sreten.
  Reply With Quote
Old 28th March 2004, 05:59 PM   #3
diyAudio Member
 
Bill Fitzpatrick's Avatar
 
Join Date: Jun 2001
Location: Eugene, OR
I was just looking at the specs for a Hiquphon tweeter and noticed that thay gave the acoustic center relative to the mounting plate. That may not mean a lot though as the didn't specify frequency.

Fat chance of getting this kind of information from manufacturers.
  Reply With Quote
Old 28th March 2004, 06:14 PM   #4
Kjetil is offline Kjetil  Norway
diyAudio Member
 
Kjetil's Avatar
 
Join Date: Feb 2004
Location: Norway
Actually, I was planning to use the crossover delay of a 1st order series XO (-15) to cancel the delay of the driver geometry (+15), by placing the tweeter below the woofer. To get this right, I obviously need to align the drivers so that tan15=(horizontal offset/vertical offset)...but then ofcourse, to design my box, I need the horizontal offset of the drivers I intend to use, being the Accuton C82-T8 and the C13-6.

Accuton.com did not respond to my email.
  Reply With Quote
Old 28th March 2004, 06:38 PM   #5
Bricolo is offline Bricolo  France
diyAudio Member
 
Bricolo's Avatar
 
Join Date: Nov 2002
Location: Grenoble, FR
I always wonder why it is always told that the AC centre is the centre of the coil

Why isn't it simply the surface of the dome?
  Reply With Quote
Old 28th March 2004, 08:46 PM   #6
Svante is offline Svante  Sweden
diyAudio Member
 
Svante's Avatar
 
Join Date: Feb 2004
Location: Stockholm
Default Re: Time alignement fundamentals

Quote:
Originally posted by Kjetil
My question is quite simple; Can this point be calculated from any T/S parameters or by any other means, or do I have to contact the driver manufacturer to get the details? I guess measuring it would be an approach, but I do not have access to such equipment.
The simple answer is no. I have objections to the importance given to the whole subject of time alignment, see below. The acoustic centre of a loudspeaker would be the position of an equivalent point source, from which the sound radiates. This whole concept is a bit fishy since the sound does *not* radiate from a point, but the entire membrane, so the end result can be nothing but an approximation.

Looking for a special case that the concept *would* work for we can imagine a case where all of the contributions from the cone appear at the listening position at the same time. Let's also study the on-axis response only. In order for this condition to be fulfilled, the outer parts of the cone must be delayed compared to the inner parts. This can be the case, since the cone does not move rigidly, but there is a wave propagation from the centre to the edge. If, and only if, this propagation along the cone has the right velocity with regard to the opening angle of the cone and the velocity of sound in air, the contributions from the different parts of the cone will add up to a coherent image of the original signal. In this case the acoustic centre can be said to be located somewhere at the base of the cone. This is assuming a high wave propagation velocity from the coil to the base of the cone.

The above was a description of the special case when understanding of the acoustic centre is relatively simple. If the wave propagation along the cone does *not* have the velocity as above, an impulse at the voice coil will not yield an acoustic impulse, but one smeared over some time. In this case it is harder to define what the delay is, since the signal at the voice coil and the acoustic signal do not have the same shape. Jumping to the frequency domain can be a solution, and here the phase would tell us what the "delay" is for different frequencies. For sure, this delay will vary with the frequency for a number of reasons, making perfect time alignment of two speakers by positioning impossible.

Another thing that complicates the matter is that towards higher frequencies, when the cone no more acts rigid, the radiation will be dominated by the inner parts of the cone.

From the above it appears clear that the T/S parameters has very little to do with the position of the acoustic centre. Key factors would be the wave propagation speed along the cone, and other parameters like that. It is unlikely, but not impossible that manufacturers have good data on this. It is more likely that you will get a nonsense response from the sales department.

I stated initially that I think that this whole subject has been given too much importance. The reason for this is that we have a time resolution that is in the order of 1 millisecond at the most. Between the ears the sensitivity is better, but here we are talking about delays between different frequencies, not between the ears. The wave propagates 0.345 m in 1 ms, so that should be about the distance difference that we could detect based on time delay only.

However, I do not claim that moving the tweeter backwards a few cm is not audible. It will affect the phase relation at the crossover frequency and the signals from the two drivers will interfere differently with each other at these two positions, and this will give different frequency responses. This can be very audible. It is, however not given that the "correct" time alignment (typically with the tweeter shifted backwards) is the configuration that yields the flattest response. The phase responses of the drivers and the crossover filters will add to the delays and the net effect is not only from the shifts of acoustic centres.

This is why I think that the whole concept of time alignment is a bit inappropriate. It is seemingly easy to understand, but when the other properties of the drivers and filters are taken into consideration, it all comes down to the resulting frequency response of the system. Time alignment is a side track that does little good, IMO.

Now I think I'd better stop here. There are other factors as well (like off-axis response) but I think this post is long enough already. If you managed to get through this post, I hope it helped at least a little.

/S
__________________
Simulate loudspeakers: Basta!
Simulate the baffle step: The Edge
  Reply With Quote
Old 28th March 2004, 11:38 PM   #7
angel is offline angel  Norway
diyAudio Member
 
Join Date: Oct 2003
Location: Norway
When trying to time align, you must also bear in mind that what you are trying to achieve, is not expressed in degrees, but in terms of delay vs frequency. You want a constant delay vs frequency, which means a phase that varies linearly vs frequency. An acoustic first-order filter gives you this. The L/R filters give a decent approximation.

As pointed out, it is a rather vague concept all in all, since the radiating surface produces a wave front where parts of the wave arrive at different times than other parts of the wave.

If you want to experiment, construct a 1st order allpass filter, as per the active L/R crossovers from Linkwitz Labs, and implement it with a variable delay, so that you can find the exact point of best sonic alignment by using your ears for measurement in listening position.
  Reply With Quote
Old 29th March 2004, 12:08 AM   #8
diyAudio Member
 
Bill Fitzpatrick's Avatar
 
Join Date: Jun 2001
Location: Eugene, OR
Richard Heyser did an JAES paper where he went into the problem of acoustic centers. It is worth a read for the mathmatically inclined.
  Reply With Quote
Old 29th March 2004, 12:20 AM   #9
Kjetil is offline Kjetil  Norway
diyAudio Member
 
Kjetil's Avatar
 
Join Date: Feb 2004
Location: Norway
Quote:
Originally posted by angel
If you want to experiment, construct a 1st order allpass filter, as per the active L/R crossovers from Linkwitz Labs, and implement it with a variable delay, so that you can find the exact point of best sonic alignment by using your ears for measurement in listening position.

Right, but now we're talking bi-amping. Intentions are, to run these babies on tubes, wich would make bi-amping outrageously expensive as I would be looking for a separate power stage for the tweeters, although I can clearly see the advantages.

My approach to time-aligning would not be to obtain "true" time alignment, but rather as a "detail" in my project, to minimize the delay caused by difference in the horizontal offset of the drivers. In other words not perfectionism but improvement.

Just a thought. It will probably create more problems than it solves, but do you think it is possible to get a good sonic result by actively splitting the signal into high pass and low pass using a heck of a steep filter (6th order+), then implement the variable time alignment, throwing both signals (high and low pass) together again and into the amp, and then filter passively in the speakers at *exactly* the same frequency, again with a verry steep filter...? or am I talking nonsense here...?!?
  Reply With Quote
Old 29th March 2004, 12:21 AM   #10
Ap is offline Ap
diyAudio Member
 
Ap's Avatar
 
Join Date: May 2002
Location: Sydney, Australia
Offsetting drivers of course affects freq. & phase repsonse - if you are intent on 'time-aligning" (I hate that term - consider you hear sounds from the drivers at the same time, so therfore all systems are time-aligned)
- what you want to measure is step response.

In terms of imprtance design in this order:
frequency/spl response
phase response
step response
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Fundamentals of Linear Regulators jan.didden Parts 2 5th June 2009 10:30 AM
bipolar transistor fundamentals... help help! oldeurope Solid State 5 7th June 2008 07:52 AM
Fundamentals of tube amp design rokali Tubes / Valves 6 17th December 2007 05:55 AM
Please help me understand speaker fundamentals jnb Multi-Way 5 20th July 2007 03:26 AM
Of Harmonics, Fundamentals, Tubes, Etc. kelticwizard Everything Else 3 12th November 2003 11:30 PM


New To Site? Need Help?

All times are GMT. The time now is 11:53 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2