It can be clear cut, as they can be viewed exclusively.
When each driver has a flat response (and you have to think of it in these terms in order to analyse the situation correctly), then physical alignment and response phase alignment are exactly the same. Ie: when you achieve one, you get the other.
Looking at the next step, when you get the responses correct for the crossover (and whatever the natural response of the driver was is irrelevant, as you've now transformed it completely with a crossover), the result is in accordance with the theory for that kind of crossover. That includes response, phase and lobing concerns provided physical alignment was done first and you stick with the chosen measurement position.
When each driver has a flat response (and you have to think of it in these terms in order to analyse the situation correctly), then physical alignment and response phase alignment are exactly the same. Ie: when you achieve one, you get the other.
Looking at the next step, when you get the responses correct for the crossover (and whatever the natural response of the driver was is irrelevant, as you've now transformed it completely with a crossover), the result is in accordance with the theory for that kind of crossover. That includes response, phase and lobing concerns provided physical alignment was done first and you stick with the chosen measurement position.
Only when no crossover is fitted and both drivers are fed the same signal. The purpose of baffle step/tilting is to align the EFFECTIVE acoustical centres at the crossover frequency at a listening position/distance. This will include the phase contribution of the crossover.
I can do the same with dsp!
Some basics here
Phase, Time and Distortion in Loudspeakers
http://linea-research.co.uk/wp-cont...Tech Docs/CrossoverFilters White Paper -C.pdf
Last edited:
Measure with the crossover installed,
the type of crossover also has an effect on the phase,
for my speakers I have an electrical 2nd order on the mid and a 3rd order on the tweeter, both resulting in a 4th order acoustical.
Measure with one driver's wires in reverse connected, so you get deepest null measurement at xo point.
Also measure at different heights to get an idea of where the max null is,
1cm can make a lot of difference.
You should find something like this.
I've optimized the drivers/xo for best phase alignment between mid and tweeter at 110cm height, that's about listening height.
Also measure at different distances to see that the phase alignment between the drivers stays at the same height (if desired).
Danny
the type of crossover also has an effect on the phase,
for my speakers I have an electrical 2nd order on the mid and a 3rd order on the tweeter, both resulting in a 4th order acoustical.
Measure with one driver's wires in reverse connected, so you get deepest null measurement at xo point.
Also measure at different heights to get an idea of where the max null is,
1cm can make a lot of difference.
You should find something like this.
I've optimized the drivers/xo for best phase alignment between mid and tweeter at 110cm height, that's about listening height.
Also measure at different distances to see that the phase alignment between the drivers stays at the same height (if desired).
Danny
In order to isolate the time difference you eliminate the phase differences and to isolate the phase differences you eliminate the time difference.
I don't think we disagree, however if you consider the drivers without phase differences then only the time difference will show and they could be aligned in a physical sense. In theory then any textbook symmetrical crossover could be implemented from this starting point, and retain the expected properties.
Although I could see aligning the acoustic center and aligning the phase may be related, we shouldn't be conflating the two together.
For example, let's say you use a step baffle to align the tweeter and the mid so that the acoustic center is at the same plane on the baffle. After that, however you design your xover, the acoustic center shouldn't change, that is regardless of the xover topology, the acoustic centers of tweeter/mid are still on the same plane and nothing has changed.
On the other hand, when you tilt the baffle angle, not only you change the acoustic center, you also change how the phase of the tweeter and the mid are combined. And in this case, the acoustic center and the phase are somewhat mixed together and difficult to separate them.
For example, let's say you use a step baffle to align the tweeter and the mid so that the acoustic center is at the same plane on the baffle. After that, however you design your xover, the acoustic center shouldn't change, that is regardless of the xover topology, the acoustic centers of tweeter/mid are still on the same plane and nothing has changed.
On the other hand, when you tilt the baffle angle, not only you change the acoustic center, you also change how the phase of the tweeter and the mid are combined. And in this case, the acoustic center and the phase are somewhat mixed together and difficult to separate them.
This is something I have difficulty with. Can someone explain to me why it is not the position of the cone but the position of the voice coil.
"by voice coil" comes from PA world where compression driver+horn is used in 2-ways.
Acoustic centre is what really matters, but it can be determined only by acoustic measurements at design axle and specified distance.
Effective acoustic center redefined
andy2, when you tilt the baffle all you do is move the wavefront, up, down, left, right just like my flashlight example. You're looking to position the null lobes off axis. You are not (can not) change the acoustic center position (speaker is built, drivers mounted be they right or wrong as far as AC line-up is concerned).
You are not (can not) change the phase relationship either. If the xo is not correct with regard to phase alignment between multiple, mounted drivers, moving that "group" around will not change how they function as a group.
Any response irregularities will still exist.
AllenB,
"In order to isolate the time difference you eliminate the phase differences and to isolate the phase differences you eliminate the time difference."
These are not "exclusively" separate issues was/is my point.
^ I'm not sure whether we agree or disagree. My point is that they can be viewed or measured separately, analysed separately, and manipulated independently.
And yes, while listening we hear both at the same time and there's no getting around that, the point was about design time, and there are properties that can be controlled.
And yes, while listening we hear both at the same time and there's no getting around that, the point was about design time, and there are properties that can be controlled.
How are these two cases different, (apart from the big bump in the middle of the baffle and the very small vertical separation reduction) unless you are changing the theoretical vertical axis of rotation?
Tilting the baffle changes "arrival time", to your ear, of each driver. It does not change timing/phase between the drivers.
The first case has you physically/electrically aligning AC/phase on the baffle assembly through placement and/or XO.
The second case has you adjusting wavefront arrival time (to our ears), between the drivers assembled, works for one axis, as you know.
If case one is done incorrectly, case two will become an exercise in compromise.
What am I missing here?
The first case has you physically/electrically aligning AC/phase on the baffle assembly through placement and/or XO.
The second case has you adjusting wavefront arrival time (to our ears), between the drivers assembled, works for one axis, as you know.
If case one is done incorrectly, case two will become an exercise in compromise.
What am I missing here?
Last edited:
I think the confusion begins with those graphics of a vertical line drawn through drivers AC's where upon tilting the baffle brings a tweeters AC in vertical line-up with a woofer. Problem is the vertical line drawn should be considered always parallel to the baffle. Not plumb with the earth.
eg. difference between slanted baffle and tilted speaker with regard to driver(s) AC.
eg. difference between slanted baffle and tilted speaker with regard to driver(s) AC.
Last edited:
This is I believe, what I meant by the axis of rotation. So think about the AC's as two points in space and delete the boxes. The two points are there whether the baffle is tilted or stepped.
What is the real definition of "correct" for that axis of rotation? ... The AC's (in either case) line up vertically, the listening axis is horizontal, I know what I'd choose.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.