Announcement

Collapse
No announcement yet.

Z-offset importance in a 3-way

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • fpitas
    commented on 's reply
    Yes, the off-axis is surprisingly good with my setup.

  • dlr
    replied
    Originally posted by fpitas View Post
    Given DSP, digital delay allows you to time align easily. That was important to me, since the horn voice coil is 17 inches behind the woofers.
    That's probably one of the best uses, since the on-axis delay is rather large and the off-axis from the horn output will probably still be close the same delay as the on-axis relative to the woofers. With dynamic drivers, however, significant excess-phase correction on-axis will introduce more change to the off-axis that will have more impact to the power response since the delays with axis will vary much more than for the horn. How detrimental that may be I don't know. I use DSP in my 3-way, but the corrections on-axis are relatively small as my drivers are close to vertically aligned.

    dlr

    Leave a comment:


  • fpitas
    replied
    Given DSP, digital delay allows you to time align easily. That was important to me, since the horn voice coil is 17 inches behind the woofers.

    Leave a comment:


  • dlr
    replied
    Originally posted by jhollander View Post
    I'd not considered phase alignment in a 3-way. Would the M-W network be simplified? I typically avoid complex high pass on the mid, which can be difficult to accomplish
    Unlikely that you'll see much change in the M/W crossover complexity. I've never offset a woofer because in my cases the crossover Fc is low enough that the phase rotation from the woofer excess delay was insignificant. I'm not trying to time align. I only consider the midrange excess delay to be an issue.

    dlr

    Leave a comment:


  • jhollander
    replied
    I'd not considered phase alignment in a 3-way. Would the M-W network be simplified? I typically avoid complex high pass on the mid, which can be difficult to accomplish

    Leave a comment:


  • dlr
    replied
    Originally posted by jhollander View Post
    If I lump your questions into time alignment of the drivers, then driver spacing, stepped and slant baffles are ways to time align the drivers. I'm not sure there's any advantage to time align the drivers. I suppose that time alignment is meant to allow the designer to create a simple crossover with minimal delay. In practice the simple crossover has created (at least for me) vertical off axis nodes. I can rationalize this easily with slanted baffles due to the baffle response, not sure about the stepped baffles as the edge diffraction ripples makes the response more difficult to understand.
    Aligning drivers can be for time alignment such as for first order crossovers if that's the goal, but it's effective for good phase alignment in higher orders and even second as well, more so in my opinion. I offset the midrange in a 3-way LR2 for that reason. It's more likely that the primary lobe will be on-axis as well as possibly reducing the crossover network required for flat on-axis response, less shaping required.

    dlr

    Leave a comment:


  • jhollander
    replied
    Originally posted by a4eaudio View Post
    So theoretical question.
    I know a lot is made (at least by some) regarding z-offset. Tweeters in waveguides, slanted baffles, maybe even a spacer on the woofer to reduce the z-offset. In a TMW 3-way, assuming you cannot do anything to force the acoustic centers to be equal for all three drivers, what would be your priority? e.g., get the tweeter and midrange as close as possible and ignore the woofer, minimized the average distance between TM and WF, etc.?
    If I lump your questions into time alignment of the drivers, then driver spacing, stepped and slant baffles are ways to time align the drivers. I'm not sure there's any advantage to time align the drivers. I suppose that time alignment is meant to allow the designer to create a simple crossover with minimal delay. In practice the simple crossover has created (at least for me) vertical off axis nodes. I can rationalize this easily with slanted baffles due to the baffle response, not sure about the stepped baffles as the edge diffraction ripples makes the response more difficult to understand.

    Leave a comment:


  • dlr
    replied
    I was intrigued by the Bessel Filter paper by Rane. I've toyed with the idea of a page at my web site to cover filters in general in more than one way, however. The Bessel graph on my WinFilters page and Bessel target descriptions were inspired by that Rane paper. My takeaway from that was that what is often described as Linkwitz-Riley filters in practice devolves into a more general filter set on the design axis. Prime example is a 2-way with a "relaxed" woofer to try to ameliorate the problem of differing driver delays to the design point. Experimenting with that in WinPCD one day I noticed that in "relaxing" the woofer, then changing the target to find the closest fit to the final woofer lowpass response, the actual response was very close to Butterworth. Target began as LR4/LR4, but the result was much closer to BW3/LR4 for closest to flat response on-axis.

    Also, as the Rane paper showed, offset Bessel filters can sum nearly flat on-axis. The issue with both of these situations is the off-axis, primarily the vertical. Any on-axis that isn't pure LR4/LR4 will have nodes in the vertical off-axis. There will also be an impact on the power response. But on the design axis the closest to flat response really isn't LR4 if there's much driver offset to try to correct. It is often close, but in extreme cases of a large midwoofer/tweeter combo, the individual driver slopes may be much closer to some other filter type.

    dlr

    Leave a comment:


  • fpitas
    commented on 's reply
    Nice to see someone else fooling around with Bessel filters.

  • dlr
    replied
    Originally posted by a4eaudio View Post
    So theoretical question.
    I know a lot is made (at least by some) regarding z-offset. Tweeters in waveguides, slanted baffles, maybe even a spacer on the woofer to reduce the z-offset. In a TMW 3-way, assuming you cannot do anything to force the acoustic centers to be equal for all three drivers, what would be your priority? e.g., get the tweeter and midrange as close as possible and ignore the woofer, minimized the average distance between TM and WF, etc.?
    T/M offset is primary, due to wavelength. M/W crossovers are usually at frequencies with wavelengths where the time delta results in a small rotation of phase between the two drivers.

    You can use my program WinFilters to experiment with offset between two drivers using combinations of filters on each at user selected frequencies and see the resulting summed response. It's a simple, fairly intuitive small program.

    WinFilters

    dlr
    Last edited by dlr; 07-28-2021, 12:44 PM. Reason: Moved las sentence to where I intended to add it.

    Leave a comment:


  • rpb
    replied
    It doesn't matter a whole lot, but I prefer the sloped baffle anyway. The x-over can make everything line up on a standard baffle.

    Leave a comment:


  • a4eaudio
    started a topic Z-offset importance in a 3-way

    Z-offset importance in a 3-way

    So theoretical question.
    I know a lot is made (at least by some) regarding z-offset. Tweeters in waveguides, slanted baffles, maybe even a spacer on the woofer to reduce the z-offset. In a TMW 3-way, assuming you cannot do anything to force the acoustic centers to be equal for all three drivers, what would be your priority? e.g., get the tweeter and midrange as close as possible and ignore the woofer, minimized the average distance between TM and WF, etc.?
Working...
X