Announcement

Collapse
No announcement yet.

Collaborative Crossover HiVi RT1.3WE and Tang Band W6-1721

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

  • Collaborative Crossover HiVi RT1.3WE and Tang Band W6-1721

    Hi All

    JamesTan started a thread here and wasn't satisfied with his current speaker build and thought he might need to add a better midrange driver. However, the consensus was that he likely needed a more refined crossover, as he designed the crossover with manufactured specs and didn't have measuring equipment. I realized after following the thread for a few days that I actually had both of these drivers in my basement so offered to build a replica of his cabinet and measure the frd and zma. While I have a measurement microphone, I am not experienced with crossover design so we are hoping some of the experienced members can help out.

    jhollander and PWR RYD I'm pretty sure you have both used this tweeter...even if you don't have time to mock-up a xo, any recommendations on how low the tweeter can go and any peaks you needed to tame?

    Some basic inputs:
    • Cabinet: 35 cm (13.75") tall and 22 cm (8.625") wide. Tweeter center 7.5 cm (3") from top an woofer center 22.6 cm (8.875") from top
    • Tweeter and Woofer centered on baffle (0.0 m x-offset)
    • Woofer center-to-center is 6" below tweeter (-0.152 m y-offset)
    • I used Jeff Bagby's method to determine acoustic offset and came up with 0.004 m. Yes, this is positive. However, I built it with a removable baffle and can actually look at the relative distance and I think assuming zero would be good enough.
    • Speakers sit on a TV console about 10" from wall so take that into consideration when implementing BSC.
    • Measurements were with a UMIK-1 USB microphone 25" (0.635 m) from tweeter, in cabinet. FRD files were run through response blender to get minimum phase.
    • However, I did NOT measure nearfield and blend low frequency range to frd curve. If I NEED to do this, please let me know, but I think many people design the xo without this.
    • Images are posted below in addition to the frd and zma files, please let me know if you see anything that looks unusual that suggest errors in measurement.

    HiVi RT1.3 (Min Phase).frd
    HiVi RT1.3WE.zma
    Tang Band W6-1721 (Min Phase).frd
    Tang Band W6-1721 in-box.zma

  • #2
    For a quick view of what those frd files show, see below. Again, if anything looks odd and indicates measurement error please let me know.

    Click image for larger version

Name:	HiVi RT1.3WE.png
Views:	309
Size:	133.4 KB
ID:	1452504
    Click image for larger version

Name:	Tang Band W6-1721.png
Views:	301
Size:	138.2 KB
ID:	1452505

    Comment


    • #3
      Hi a4eaudio, thank you for all the hard work you have done and you have actually gone beyond and above to help out a newbie like me. Using the FRD and ZMA files provided, with my existing XO design the FR curve looks as shown,

      Click image for larger version  Name:	2wayTBHiViFR-True.JPG Views:	0 Size:	113.0 KB ID:	1452508

      Does this explain why the more lay back sounding of the vocal I am hearing from the setup now? I will rework a new XO design based on the new FRD and ZMA files and post the result later. Once again, thank you a4eaudio for all the help you have given.

      Comment


      • #4
        After reworking the XO design, here's what I got

        Click image for larger version

Name:	2wayTBHiViFR-True-1.JPG
Views:	301
Size:	116.6 KB
ID:	1452510

        Do you guys think it will help with the vocal with this new FR? Any suggestion or advice is welcome. Thanks.

        Comment


        • #5
          a4eaudio did you get a RT1.3 plus TB W? I'll derive the total offset for x-sim with that.
          John H

          Synergy Horn, SLS-85, BMR-3L, Mini-TL, BR-2, Titan OB, B452, Udique, Vultus, Latus1, Seriatim, Aperivox,Pencil Tower

          Comment


          • #6
            This probably isn't what's needed. I was just playing around with the files while having my coffee, and waiting on more input. Looks nice though. Probably crossed too low, and uses a few too many parts. It's hard to see much baffle step in the woofer measurement. It may be there, but followed by a dip in the response. I have no experience doing x-overs for speakers near the wall. I presume very little BSC is needed.

            Click image for larger version  Name:	w6 HI VI.jpg Views:	0 Size:	242.4 KB ID:	1452527

            Comment


            • JamesTan
              JamesTan commented
              Editing a comment
              Wow....your simulation looks almost perfectly flat! My simulation below also uses quite a number of components, I believe also 2 notch filter used. But yours looks much flatter than mine.

          • #7
            This is my latest attempt to get a more linear response. Hopefully will get some feedback if the response looks alright, planning to make the purchase of the XO parts.

            Click image for larger version

Name:	2wayTBHiViFR-True-2.JPG
Views:	286
Size:	112.8 KB
ID:	1452529

            Comment


            • JamesTan
              JamesTan commented
              Editing a comment
              Not sure if it is a bad idea to ignore the peck at around the 15 kHz region, I read somewhere saying most likely it will not be audible at that region. How true is that?
              Last edited by JamesTan; 09-26-2020, 11:11 AM.

          • #8
            Click image for larger version

Name:	w6 HI VI revised.jpg
Views:	284
Size:	238.8 KB
ID:	1452540 Here's one with a higher xo point.

            ​​​​​​​

            Comment


            • JamesTan
              JamesTan commented
              Editing a comment
              Both designs look really good to me. The manufacturer recommended XO for the tweeter is at 2 kHz, so I believe both XO designs should work pretty
              good. Care to share your first XO design? Really keen to learn how you managed to create such a flat response curve.

            • rpb
              rpb commented
              Editing a comment
              If you look at the filter transfer functions, you can see where I added a notch to shape the response. More notches allows for some fine tuning. The downside is a lower impedance, and higher part count. So... try adding a notch on the tweeters top end. Use a cap, a coil, and a resistor in series with each other, and place that across the tweeter terminals. Start with about .3mH for the coil. (That's not what I used.). Just center the notch on the peak, then adjust the resistor to bring down the response. A bigger coil makes the notch sharper. A small coil makes it wider. Like the optometrist. Better, worse, better, worse,... till you smooth it out.

          • #9
            Originally posted by jhollander View Post
            a4eaudio did you get a RT1.3 plus TB W? I'll derive the total offset for x-sim with that.
            Yes, see attachment below. Here is a pic of the drivers in the box. Because of the woofer ring the z-offset is practically zero but not sure how that works in x-sim vs PCD.

            RT1.3 +W6.frd

            Click image for larger version

Name:	box 1.jpg
Views:	281
Size:	59.8 KB
ID:	1452555

            Comment


            • #10
              Originally posted by JamesTan View Post
              This is my latest attempt to get a more linear response. Hopefully will get some feedback if the response looks alright, planning to make the purchase of the XO parts.

              Click image for larger version

Name:	2wayTBHiViFR-True-2.JPG
Views:	286
Size:	112.8 KB
ID:	1452529
              That response looks good, but I'd notch the treble peak.

              Comment


              • #11
                Originally posted by a4eaudio View Post

                Yes, see attachment below. Here is a pic of the drivers in the box. Because of the woofer ring the z-offset is practically zero but not sure how that works in x-sim vs PCD.

                [ATTACH]n1452554[/ATTACH]

                Click image for larger version  Name:	box 1.jpg Views:	0 Size:	59.8 KB ID:	1452555
                Are you going to test a x-over?

                If it's not too much trouble, can you measure from 3' away, instead of 25". Are these gated measurements?

                Comment


                • #12
                  Here's one idea and a few comments. I like crossing the RT1.3 above 2,500 and letting the top end float higher as it falls off pretty fast off axis. This is one you want to listen to for a while and pad / adjust accordingly.

                  Good match on the 3 files when finding the offset, so the measurements are good. Without blending the near field, it's tough to know if this is going to sound thin in the bass.
                  John H

                  Synergy Horn, SLS-85, BMR-3L, Mini-TL, BR-2, Titan OB, B452, Udique, Vultus, Latus1, Seriatim, Aperivox,Pencil Tower

                  Comment


                  • JamesTan
                    JamesTan commented
                    Editing a comment
                    Hi John, thanks for the idea and comments. I tried simulating your XO design and the XO point appears to be around 4,000. If the main purpose is to let the top end float higher as it falls off pretty fast off axis, do you think the effect will be the same if I stick to my XO design in post #7 and keep the peak effect at 15kHz?

                  • jhollander
                    jhollander commented
                    Editing a comment
                    James, this x-o is at 3,500. As I mentioned above I think you should listen to the peak and determine how much you want to pad it down. A small inductor will also roll of the top end but then there are more adjustments and parts needed.

                  • JamesTan
                    JamesTan commented
                    Editing a comment
                    John, I think I will leave the peak alone for the time being. Based on your advice, currently I find not much problem with that peak when I am listening to music. The lay back sounding of the vocal most likely is caused by that dip and some cancelation between 1.2 kHz to 4 kHz region with my current XO design. Hopefully with the revised XO design at post #7, it will correct that vocal problem.

                • #13
                  Originally posted by rpb View Post
                  Are you going to test a x-over?
                  Maybe. I don't have tons of parts on hand but I have one of each of a lot resistor values, capacitors and inductors. The problem comes when I need two of the same value, but I can use the closest value I have. The other problem is I don't have an experienced ear for really listening and voicing, although I'd like to think I'm getting better.

                  Originally posted by rpb View Post
                  If it's not too much trouble, can you measure from 3' away, instead of 25". Are these gated measurements?
                  Yes, I can measure that tonight. They are gated at 3ms.

                  Originally posted by jhollander View Post
                  Good match on the 3 files when finding the offset, so the measurements are good. Without blending the near field, it's tough to know if this is going to sound thin in the bass.
                  Should I measure near field or could we even use box simulations to get a better idea on the bass?

                  Comment


                  • #14
                    If you assemble a xo with jumpers, then we will know it works..

                    Here's my low part count offering.

                    Click image for larger version  Name:	SIMPLE.jpg Views:	0 Size:	245.5 KB ID:	1452596 Click image for larger version  Name:	SIMPLE TWO.jpg Views:	0 Size:	248.9 KB ID:	1452598

                    It would be interesting to see the actual measurements with this one. (Both on, and off-axis. )

                    Comment


                    • #15
                      Should I measure near field or could we even use box simulations to get a better idea on the bass?

                      [/QUOTE]

                      You can measure after the fact for the near field. The box sim doesn't get you the driver frequency response.
                      John H

                      Synergy Horn, SLS-85, BMR-3L, Mini-TL, BR-2, Titan OB, B452, Udique, Vultus, Latus1, Seriatim, Aperivox,Pencil Tower

                      Comment

                      Working...
                      X