Hello,
I am designing a two-way speaker (7" woofer and the Dayton Audio AMT Mini-8 tweeter). I will likely use a first order x-over (6 DB).
Typically in this type of design, if one considers the woofer as a clock, the tweeter would be above the woofer at the 12:00 position.
Are there fundamental reasons for this?
The reason I ask is that I'd like to position the tweeter at the 4:00 or 8:00 position.
Any feedback would be much appreciated.
Thanks!
Mack
I am designing a two-way speaker (7" woofer and the Dayton Audio AMT Mini-8 tweeter). I will likely use a first order x-over (6 DB).
Typically in this type of design, if one considers the woofer as a clock, the tweeter would be above the woofer at the 12:00 position.
Are there fundamental reasons for this?
The reason I ask is that I'd like to position the tweeter at the 4:00 or 8:00 position.
Any feedback would be much appreciated.
Thanks!
Mack
John DeVore considers a sort of 4 or 8 o'clock position for a tweeter natural for a two way.
http://www.devorefidelity.com/index.html
But I doubt he would waste time on 6dB/octave filters for 7" bass and a tweeter of any sort. That really is next to no filter at all. 🙄
http://www.troelsgravesen.dk/DiscoveryW18.htm
http://www.devorefidelity.com/index.html
But I doubt he would waste time on 6dB/octave filters for 7" bass and a tweeter of any sort. That really is next to no filter at all. 🙄
http://www.troelsgravesen.dk/DiscoveryW18.htm
Attachments
Last edited:
MackK, I don't see any problem, with your 4/8 o'clock positioning. Offset placement on the baffle is good. It just looks assymetrical, so it looks less pleasing to the eye.
It depends on speaker/stand height. People put it above to get tweeter closer to ear level. If you have an unusually high placement of the cabinet it makes sense to keep it lower below woofer. Offset at 4 or 8 o'clock provides the asymmetry to reduce baffle diffraction effects be preventing any two distances from tweeter to baffle edge being identical, thereby preventing deep cancellation interference by baffle edge. The polar response will be asymmetric though.
If you have a well behaved wide band woofer with no nasty breakup peaks, and a wide band tweeter (like a full range driver) a 1st order xo, if done well, with time alignment, it can be used to achieve a phase coherent, transient perfect xo that sounds eerily good for natural soundstage and imaging with realistic percussion like a live performance. In practice, it is much more difficult than a second or fourth order xo to do right. Some very good expensive high end speakers use first order exactly for these reasons. The AMT is not wide band enough.
If you have a well behaved wide band woofer with no nasty breakup peaks, and a wide band tweeter (like a full range driver) a 1st order xo, if done well, with time alignment, it can be used to achieve a phase coherent, transient perfect xo that sounds eerily good for natural soundstage and imaging with realistic percussion like a live performance. In practice, it is much more difficult than a second or fourth order xo to do right. Some very good expensive high end speakers use first order exactly for these reasons. The AMT is not wide band enough.
It's how the sound is originated, then spread and blended.
The first belongs to the speakers, so each has its own kind of emission.
The spreading of the sound 😱 after it leaves the membrane it is confined by the nearby boundaries, in this case the baffle itself in the first instant contributes to radiate sound ( reflection -integration ). For tweeters, the wavelenghts are in the same order to the baffle magnitudes ( from few inches to "some" inches )...
The blending of the soundwaves: Ok, the two emissions at crossover frequency are in phase and before and after they have the same amplitude ( perfect !! ) and the speakers ( tw and wf) are placed at minimum distance according to the crossover point ( and the mini-8 likes to be crossed very high...🙄 ) so the soundwave after being splitted ( electrically, eh !? )
now finds this resistive fluid...
More like 5 or 7, 1 or 11 😉
Keep the tweeter at ear level and at a different distance from each cabinet edge for the smoothest response.
Keep the tweeter at ear level and at a different distance from each cabinet edge for the smoothest response.
Hi,
Using a single cap on the AMT will yield a 3rd order acoustic
high pass around 5.5KHz. Not promising for a 2-way design.
2nd order might yield LR4 acoustic around 4KHz, but
the ripples in response 5KHz to 10KHz are pretty bad.
rgds, sreten.
Using a single cap on the AMT will yield a 3rd order acoustic
high pass around 5.5KHz. Not promising for a 2-way design.
2nd order might yield LR4 acoustic around 4KHz, but
the ripples in response 5KHz to 10KHz are pretty bad.
rgds, sreten.
Last edited:
1200 o'clock will project the +3dB lobe down towards the floor.
0600 o'clock will project the +3dB lobe up towards the ceiling.
This is what you get with a true acoustic single pole crossover.
If you want to avoid pointing the phasing lobe above or below, you need to time align the drivers for the forward beaming. But this still leaves the phasing lobe in the forward direction. Anywhere off true forward will sound different, including reflections.
LR removes the lobing effect. Read Linkwitz.
Either LR2 using cascaded single pole passive filters, or LR4 using at least some active filtering. Passive LR4 requires too much absolute accuracy in component and speaker values.
4/8 o'clock does not remove the lobing, it just rotates (~60degreees) it into a different plane.
The pic in post4 shows ~0515 o'clock. The reduced rotation of the lobing plane may be less noticeable and the gain in different diffraction distances to all the cabinet edges may help in hiding specific colouration peaks in the signal.
Before I knew about diffraction effects, I did a 1245 o'clock to distribute the distances. I did not compare to 1200 o'clock so can never know if it was a success. I copied the idea from the HiFi mag reviewers of the time.
0600 o'clock will project the +3dB lobe up towards the ceiling.
This is what you get with a true acoustic single pole crossover.
If you want to avoid pointing the phasing lobe above or below, you need to time align the drivers for the forward beaming. But this still leaves the phasing lobe in the forward direction. Anywhere off true forward will sound different, including reflections.
LR removes the lobing effect. Read Linkwitz.
Either LR2 using cascaded single pole passive filters, or LR4 using at least some active filtering. Passive LR4 requires too much absolute accuracy in component and speaker values.
4/8 o'clock does not remove the lobing, it just rotates (~60degreees) it into a different plane.
The pic in post4 shows ~0515 o'clock. The reduced rotation of the lobing plane may be less noticeable and the gain in different diffraction distances to all the cabinet edges may help in hiding specific colouration peaks in the signal.
Before I knew about diffraction effects, I did a 1245 o'clock to distribute the distances. I did not compare to 1200 o'clock so can never know if it was a success. I copied the idea from the HiFi mag reviewers of the time.
+1 what AndrewT posted.
Moving slightly away from the 12 or 6 positions will probably improve the smoothness of the tweeter frequency response due to baffle edge diffraction.
This is because in the 6 and 12 position the distance from the tweeter to each side is the same, so the diffraction effects from each side will pile on top of each other. If you move to say 5, 7, 11 or 1 o'clock, now the distance to each side of the baffle is different so edge diffraction effects will partially cancel and the +3dB lobe is still in an almost vertical plane.
"The edge" (Home of the Edge) is a nice simple simulator to play around with baffle diffraction effects.
Moving slightly away from the 12 or 6 positions will probably improve the smoothness of the tweeter frequency response due to baffle edge diffraction.
This is because in the 6 and 12 position the distance from the tweeter to each side is the same, so the diffraction effects from each side will pile on top of each other. If you move to say 5, 7, 11 or 1 o'clock, now the distance to each side of the baffle is different so edge diffraction effects will partially cancel and the +3dB lobe is still in an almost vertical plane.
"The edge" (Home of the Edge) is a nice simple simulator to play around with baffle diffraction effects.
Last edited:
Before you decide where you put your tweeter on the baffle,
decide at what angle off-axis horizontally do you intend to
do the listening. What's not so good on-axis, is much better off-axis.
If you optimize the tweeter position for on-axis then you should listen
to it the same way. I don't bother to offset the tweeter because it
looks simply bad, imo.
decide at what angle off-axis horizontally do you intend to
do the listening. What's not so good on-axis, is much better off-axis.
If you optimize the tweeter position for on-axis then you should listen
to it the same way. I don't bother to offset the tweeter because it
looks simply bad, imo.
Rounding off will attenuate the re-flections. I doubt that rounding removes them entirely.
Offsetting will spread the re-flections over a wider range of frequencies, especially if the three nearest edge distances are all different. eg, 50mm from the top, 70 mm from one side edge and 110mm from the other side edge (=180mm wide).
I see that as meaning that using both techniques add to give better response and are not exclusive to each other.
Offsetting will spread the re-flections over a wider range of frequencies, especially if the three nearest edge distances are all different. eg, 50mm from the top, 70 mm from one side edge and 110mm from the other side edge (=180mm wide).
I see that as meaning that using both techniques add to give better response and are not exclusive to each other.
If you consider the directivity variation with frequency of the combo it seems logical to use either 12 or 6 o'clock positioning.
It is probably less severe to have the beam varying in vertical direction than sweeping horizontally wrt the listening position.
Jan
It is probably less severe to have the beam varying in vertical direction than sweeping horizontally wrt the listening position.
Jan
Offsetting the tweeter mitigates the issue by smoothing the response errors, but only over a small radiation angle (at others it may be worse). Rounding will fix the issue once the round over is sufficient to reduce the diffraction below the threshhold of audibility.
Yes, but separate from the edge diffraction issues there is lobing from the fact that you have the sound moving from one driver to the other with frequency, with different dimensions, so you get changes in the directivity (and sensitivity). You want to keep those changes horizontally aligned.
jan
jan
- Status
- Not open for further replies.