Announcement

Collapse
No announcement yet.

Rolls SX21 [24db] crossover question

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

  • #16
    This might not work, but it's what I'd try.

    Measure the woofers from one meter, or a little more, with the mic centered between the woofers. See if there are any response peaks that need attention, and EQ them out. Then add a 1.6k filter of 24dB slope. Now measure the tweeter on its axis, and do the same. What you want is an actual 24dB slope. the DSP may only need 12dB to get there. Now measure about 5" below the tweeter axis, and see how they add. See if adding delay to either helps. Then check for a reverse null. While looking at the reverse null, alter the delay till the null gets deeper. Then swap the polarity back.

    If you don't already have a cap on the tweeter, consider adding one that helps shape the response. It will also help protect the tweeter.

    Comment


    • #17
      Excellent point about a cap on the tweeter. The turn-on thump from amps can damage a directly connected tweeter, sooner if not later.
      Francis

      Comment


      • #18
        Caps on the tweeters, yes was experimenting with that on very similar project with the same wavecor.
        I used sizes 33uf down to 12uf, and could see that they were helping with the cutoff.
        Ended up using a 25uf. I reset this project and try the same thing.

        Thanks!!

        Comment


        • #19
          Originally posted by HIFY GUY View Post
          Caps on the tweeters, yes was experimenting with that on very similar project with the same wavecor.
          I used sizes 33uf down to 12uf, and could see that they were helping with the cutoff.
          Ended up using a 25uf. I reset this project and try the same thing.

          Thanks!!
          I wish you would entertain us a little, by posting more measurements. I'm getting stir crazy, and need some speaker tech to get my mind off all the bad stuff. Looking at HOLM responses of the woofers, tweeter, and sum, etc would allow some more interaction with you on this.

          Comment


          • #20
            Well that would be amazing if the phase could have the classic look.
            This is the TMM is post #14 and now using the standard 2x4 minidsp set to factory settings before starting.

            Loading in the minimum settings:
            1400hz/1600hz with BW 18db.
            Tweeter 4ohm-W11 is only attenuated - 3.6db.
            No helper cap.

            I did look at the phase, still the same.


            Attached Files

            Comment


            • #21
              Secondly with the caps, no other change


              Click image for larger version

Name:	Untitled #2.png
Views:	62
Size:	186.6 KB
ID:	1436357
              Attached Files

              Comment


              • #22
                If you flip the speaker upside down, you can get the woofers further away from the floor, and possibly set the gate lower. I'm interested to see the woofer raw response. What woofer are you using?

                If you measure the woofers raw response centered on a line between them, you should see the baffle step, and any pesky peaks. Those pesky peaks will not show up well if you are not on the center line, due to some cancellation effects caused by the different lengths between the mic, and each woofer.. The first thing I'd do, is add a shelving filter to reduce the baffle step, or eliminate it the full 6dB. In essence, treat the filter like you want to cross higher. (1) Eliminate the baffle step (2) Notch any woofer peak. (3) Using EQ, see if you can shape the woofer response to look like a 2nd, or a 4th order roll-off, without adding a "roll-off" filter. This approach might not work, depending on the raw woofer response.

                Comment


                • #23
                  will do - want to post this before losing my train of thought:
                  24db is tougher, and had to add more driver separation to get this response.
                  Now 1400/1700 but ms is only .32


                  Click image for larger version

Name:	Untitled #3.png
Views:	61
Size:	153.9 KB
ID:	1436362

                  Comment


                  • #24
                    The null looks good. Is that measured on the tweeter axis, or slightly below? You can measure at different heights, and see where it is deepest. The small bump at 700hz could be diffraction, or a measurement issue. Might go away by 30 degrees off axis. If not, maybe EQ it down.

                    Comment


                    • #25
                      Thought it best to listen to a cd before moving on,,,,surprisingly good actually, and I say that because I tried 24db a few day ago.
                      Nice and clear and the bass is robust/nice with no equalization.
                      Sill have the same phase response, but will raise speaker box another 6" which will be 4.5 feet A.F.F. in a 9' ceiling room.

                      Adding woofer manufacturers measurements.

                      Click image for larger version

Name:	Untitled #4.png
Views:	63
Size:	413.6 KB
ID:	1436373








                      Comment


                      • #26
                        Originally posted by rpb View Post
                        The null looks good. Is that measured on the tweeter axis, or slightly below? You can measure at different heights, and see where it is deepest. The small bump at 700hz could be diffraction, or a measurement issue. Might go away by 30 degrees off axis. If not, maybe EQ it down.
                        Tweeter is on axis, mic still at 36" just like the photo in #14

                        Comment


                        • #27
                          Can you save settings, and later compare them? Don't loose the info on something that you like the sound of.

                          I'd try a "wide" notch centered at about 5k, and make it deep enough to give the appearance of a smooth roll-off starting around 2k. Probably a 10dB deep notch.

                          There's a hundred ways to do this. It should be fun. If I make it seem like work, just tell me to back off.

                          Comment


                          • #28
                            I will try that and other ideas about the test set up.
                            Might even go outside.
                            Ill be back soon, huge storm on the way and i need to batten down a few things.

                            Comment


                            • #29
                              While I've been waiting for your return, I looked for some images to show a couple of things. Found this one. The raw response probably was similar to yours. The xo is passive first order, and a notch. Measured in a room with 8' ceiling. This is my measurement of someone else's suggested x-over. The woofer is a Peerless 830874, and has a response much like your woofer. The filter provides full BSC, and notches a 4k peak in the raw response.



                              Comment


                              • #30
                                Sorry for the delay,,,another wind storm here in the East.
                                I think you are suggesting that I I extend the W20RC woofer well above 1,500hz. The W20RC is a budget woofer, although looks impressive with a cast frame, heavy magnet and beautifully made.
                                A great deal for $25 each. Also have read others state it responds well to small changes, and it does. However it may be too complicated to crossover in that range, but i'll try anything....

                                "a wide notch centered at about 5k, and make it deep enough to give the appearance of a smooth roll-off starting around 2k. Probably a 10dB deep notch"

                                Like this?

                                Click image for larger version

Name:	TEMP.png
Views:	41
Size:	52.3 KB
ID:	1436461







                                Comment

                                Working...
                                X