Announcement

Collapse
No announcement yet.

RS125P-8 + XT25SC40-04 Mini Monitor

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

  • RS125P-8 + XT25SC40-04 Mini Monitor

    The XT25 is loaded into a waveguide from a buyout tweeter. I was hoping it would move the acoustic center of the tweeter far enough back to time align with the woofer, but testing shows there is still a 0.29 inch difference between them. Regardless it makes for a nice visual appeal. The OD of the waveguide is very close to the woofer.

    Directivity curves show that off-axis behavior of the woofer and tweeter match up pretty nicely between 1750 Hz and 3 kHz.

    Click image for larger version

Name:	Directivity Curves.jpg
Views:	2
Size:	325.0 KB
ID:	1381608

    Woofer HD testing seems to show that the breakup at 7.5 kHz might be HD related, but maybe not enough to warrant throwing a bottomless notch at it.

    Click image for larger version

Name:	harmonic distortion 4.0 VDC.jpg
Views:	3
Size:	165.3 KB
ID:	1381609

    Tweeter HD testing lines up pretty well with other ring radiators.

    Click image for larger version

Name:	harmonic distortion 4.0 VDC.jpg
Views:	3
Size:	164.9 KB
ID:	1381610

    All this suggests in my mind that a crossover point between 1750 Hz and 3 kHz would work well.

  • #2
    Back when I spent more time on this forum, the XT25 was known for having good HD performance, but a huge resonance at Fs that needed to be damped with an LCR and still forced a high crossover point. It seems to me that these new versions have addressed that. This tweeter appears to have ferrofluid from the data sheet. My impedance sweeps show a peak at 26 ohms.

    Click image for larger version

Name:	tweeter impedance.jpg
Views:	1
Size:	167.0 KB
ID:	1381614

    Has anybody tracked this issue? Can anyone confirm if these new ring radiators exhibit the same ringing characteristic as the old ones? In my crossover simulations a single cap rolls off the response very nicely with no signs of an impedance hump in the low end roll off. Assuming that is the case, HD performance should be the only thing limiting the bottom crossover point.

    Comment


    • #3
      gregrueff - I'm finishing a build now with an XT25SC90-04 tweeter that is a few years old (2015 maybe?) -
      It shows exactly the same ZMA curve as yours. Planning on HP around 3.1Khz 3rd order, no waveguide per se.

      I think I hear a difference - wow, it's amazing!" Ethan Winer: audio myths
      "As God is my witness I'll never be without a good pair of speakers!" Scarlett O'Hara

      High value, high quality RS150/TB28-537SH bookshelf - TARGAS NLA!
      SB13/Vifa BC25SC06 MTM DCR Galeons-SB13-MTM
      My Voxel min sub Yet-another-Voxel-build

      Tangband W6-sub

      Comment


      • #4
        Cool.
        I took some more distortion sweeps of the tweeter with two different crossovers in place. This first is an approximate crossover point of 1700 Hz LR2:

        Click image for larger version

Name:	harmonic distortion crossed at 1700 Hz.jpg
Views:	1
Size:	195.0 KB
ID:	1381653

        This next one is LR2 around 2100 Hz.

        Click image for larger version

Name:	harmonic distortion crossed at 2100 Hz.jpg
Views:	1
Size:	203.5 KB
ID:	1381654

        Both sounded pretty smooth to me, but it looks like 2100 Hz takes the cake. I couldn't hear any honking, so I'm guessing these new FF ring radiators can go as low as you can bear the HD.

        Comment


        • #5
          I would listen to it, Greg. It will buzz with nasal quality if the Fs is not sufficiently suppressed. When I did the Xenums, and Xenoliths, I just popped a 30 ohm resistor across it to damp the peak. That as all it needed when xoverd at 2.6-2.7kHz. I am in agreement with Don as my impedance info looks similar to yours.

          Later,
          Wolf
          "Wolf, you shall now be known as "King of the Zip ties." -Pete00t
          "Wolf and speakers equivalent to Picasso and 'Blue'" -dantheman
          "He is a true ambassador for this forum and speaker DIY in general." -Ed Froste
          "We're all in this together, so keep your stick on the ice!" - Red Green aka Steve Smith

          *InDIYana event website*

          Photobucket pages:
          http://photobucket.com/Wolf-Speakers_and_more

          My blog/writeups/thoughts here at PE:
          http://techtalk.parts-express.com/blog.php?u=4102

          Comment


          • #6
            Thanks everybody.
            I think I need some help with some fundamentals though.

            What is the impact on my amp if there is a large impedance hump at the tweeter Fs, or say at around 2 kHz? It is the relative same magnitude as the woofer impedance spike, but just higher in frequency. How high a hump is too high? I'm hitting my target slope with a single cap...
            How can damping the impedance in the electrical domain keep from exiting the resonance in the frequency domain when I don't see an immediate impact on the FR curve when I use a LCR circuit targeted on the Fs? I guess this resonance doesn't show up in the FR and HD tests, so maybe that makes sense?

            Let's use the below snapshot as an example. Why is this bad, apart from not damping the tweeter resonance any more than with a simple L-pad?

            Click image for larger version

Name:	question 1.png
Views:	1
Size:	533.5 KB
ID:	1381789

            Comment


            • #7
              I'm not really sure what to tell you except it can have a factor on a lot.
              The amp won't care one iota. The magnitude of the resonance in the tweeter can do a few things:
              1- buzz in the male vocal range.
              2- cause a suckout or cancellation in the woofer response if magnitude is high enough, or even help to create a peak in response.
              3- cause the HD measurement to be higher at said Fs.

              The impedance peak due to xover really isn't so much of an issue unless you're using a tube-amp, or a SS amp that oscillates with a reactive load. Most will be fine.

              I'm really surprised you say your XT has FF, as if it did, the Fs magnitude would not be that high and approach more of a 10 ohm Zmax instead. This means they are similar to the units I used, with a Zmax as you have so did mine, and therefore I think you should listen to it carefully to see if you hear any issues. I recommend using James Taylor for this, as he was easily discernible in this tweeter's outcome. I also know that different filters or topologies are more prone to allowing the Fs to come through at least in modeling. I'm just advising you to be careful, as that buzzing gets annoying.

              Later,
              Wolf
              "Wolf, you shall now be known as "King of the Zip ties." -Pete00t
              "Wolf and speakers equivalent to Picasso and 'Blue'" -dantheman
              "He is a true ambassador for this forum and speaker DIY in general." -Ed Froste
              "We're all in this together, so keep your stick on the ice!" - Red Green aka Steve Smith

              *InDIYana event website*

              Photobucket pages:
              http://photobucket.com/Wolf-Speakers_and_more

              My blog/writeups/thoughts here at PE:
              http://techtalk.parts-express.com/blog.php?u=4102

              Comment


              • #8
                No, you're absolutely right. I'm still learning here. The tweeter clearly does not have FF and has a huge spike . I was confused because the datasheet has a ferrofluid field, so I thought they may have added it.

                I know I'm sounding hard-headed, but I'm just trying to understand the impact of everything. It's clear to me that the impedance spike needs to be damped, so now I'm trying to figure out how much. It sounds like maybe the only way to know is to listen to music since I couldn't see any evidence of it in HD sweeps or FR plots.

                Comment


                • #9
                  I'm continuing my learning process.
                  I put together two different iterations of crossovers, focusing on the tweeter.

                  This one is second order electrical with an L-pad.
                  The roll-off tracks at 1750 Hz with 4th order slopes, but the reverse null shows the crossover point around 2250 Hz. The shape of the knee has been sacrificed for the sake of low component count.
                  The L-pad seems to be taming the impedance spike, but causes a top end rise that isn't offensive, especially with distortion so low up top.
                  I listened carefully for a nasally tone and couldn't detect anything.

                  Click image for larger version

Name:	BW4 Rev00 5 uF.png
Views:	1
Size:	568.3 KB
ID:	1382207

                  This second iteration is also second order electrical with an L-pad.
                  Same characteristics as before, but the slopes track at 1500 Hz 4th order. The knee more closely tracks with an accepted transfer function.
                  The reverse null shows the crossover point around 1750 Hz.

                  Click image for larger version

Name:	BW4 Rev01 10 uF.png
Views:	1
Size:	585.8 KB
ID:	1382208

                  Conventional wisdom would say this is too low a crossover point for the XT25, so I took a combined system sweep of this second iteration:

                  Click image for larger version

Name:	HD sweep.jpg
Views:	1
Size:	216.0 KB
ID:	1382209

                  This looks really good to me?
                  Again, with music listening I could detect no harshness that would be expected from crossing low or insufficiently damping the tweeter's impedance peak.

                  Finally, I put together a crossover the way I thought I probably should. I used a 12.5 ohm resister shunted across the tweeter to damp the impedance peak (it seemed to accomplish the same thing as a full blown LCR) and used a zobel to hit the target slope at 2500 Hz. But it's third order electrical and ups the component count on the tweeter from 4 to 7.

                  Click image for larger version

Name:	BW4 Rev02.png
Views:	1
Size:	582.9 KB
ID:	1382210

                  Feedback and comments welcome, but I'm struggling to see why I shouldn't stick with the simpler crossovers.

                  Comment


                  • #10
                    Your first one there looks like an Fc closer to 3.2k, because if you are using an in-phase LR method, it should be -6dB at the summation/xover point. I agree that an L-pad is sufficient to damp the Fs, as that is what I did, and yours does seem to be doing the job. You are also correct that I would think the second is too low of a xover point. I have heard the SC90 used to about 2k vs my 2.6K, and you can just hear that the tweeter is a hair more uncomfortable. Some people do not find this offensive, and it suits them fine. I still feel that 12dB electrical slopes yielding 4th order acoustic rolloff of 2.5k+ is really the sweet spot as long as you damp the Fs to least 40% of it's initial magnitude. It doesn't have to be nulled out completely to reduce the nasality.

                    Later,
                    Wolf
                    "Wolf, you shall now be known as "King of the Zip ties." -Pete00t
                    "Wolf and speakers equivalent to Picasso and 'Blue'" -dantheman
                    "He is a true ambassador for this forum and speaker DIY in general." -Ed Froste
                    "We're all in this together, so keep your stick on the ice!" - Red Green aka Steve Smith

                    *InDIYana event website*

                    Photobucket pages:
                    http://photobucket.com/Wolf-Speakers_and_more

                    My blog/writeups/thoughts here at PE:
                    http://techtalk.parts-express.com/blog.php?u=4102

                    Comment


                    • #11
                      So I ran into something interesting while testing my woofer's new x-over; excess capacitance in the LCR circuit resulting in an off-center notch.
                      Zaph talks about this in his blog, but I've now verified his results.

                      The woofer filter in question below is first order electrical with a notch on the breakup to provide a nice roll-off. The values in the sim below are measured values per DATS. Damping the breakup results in a super smooth power response and allowed me to hit my 2nd order target slopes. You can see that the notch completely flattens the breakup, but does so through the use of a diminutive 1.3 uF cap.

                      Click image for larger version

Name:	predicted model.png
Views:	1
Size:	88.0 KB
ID:	1385465

                      However, when I take a sweep of the notch filter in DATS, look where the null appears:

                      Click image for larger version

Name:	sweep per model.png
Views:	1
Size:	35.6 KB
ID:	1385466

                      Right smack at 6.5 kHz which is not where WinPCD was predicting. It should be centered at 7.5 kHz.
                      I also took a frequency sweep (didn't save a screen shot) with this filter in place, and sure enough the woofer breakup is not suppressed as the notch was not centered on it.

                      I played around a little and found that if I increased the capacitance in WinPCD by around 0.5 uF it closely matched what I was seeing in my FR sweep.
                      But this is not a equipment tolerance issue because the capacitor was measured at 1.3 uF. It was actually supposed to be 1.5 uF +/- 1% and just measured low. So I have excess capacitance in the circuit from an unknown source.

                      So then anticipating the excess capacitance, I adjusted the inductor value to try and shift the center of the notch.

                      Click image for larger version

Name:	correct notch.png
Views:	1
Size:	197.8 KB
ID:	1385467

                      Sure enough, by swapping in a different inductor measured as 0.27 mH, and experimentally adjusting the resistor as well, it relocates my notch to 6.5 kHz and suppresses the woofer breakup. FR measurements (again, were not saved) show a fairly nice roll-off that matches my target slope pretty well.
                      But loading these values back into WinPCD?

                      Click image for larger version

Name:	last one.png
Views:	1
Size:	87.9 KB
ID:	1385468

                      Looks terrible, but measures beautifully.
                      Anyone have a root cause explanation by chance?


                      Comment


                      • #12
                        I think I'm happy enough with the crossover now to solder the components on the board.
                        LR2 at around 2 kHz so the tweeter is hooked up in reverse polarity:

                        Click image for larger version

Name:	Design B directivity curves.png
Views:	1
Size:	257.1 KB
ID:	1385616
                        The directivity curves are on axis, 30 deg., 60 deg., and 90 deg horizontal. The green trace is the reverse null. I tried and tried to get better phase alignment for a deeper and broader null but this is the best I could accomplish while still keeping the on-axis curve respectable.

                        Full system HD sweep taken on the tweeter axis:
                        Click image for larger version

Name:	HD with Design B filter.jpg
Views:	1
Size:	264.0 KB
ID:	1385617

                        So even with only a cap on the tweeter, I've somehow managed to keep the tweeter low end HD in line with the system.

                        Component count is four (4) on the woofer and six (6) on the tweeter. Both the woofer and tweeter are first order electrical with some other stuff thrown in. Somehow the tweeter doesn't need a padding resistor with the impedance notch and zobel in there.
                        I'm using a lot of baffle step compensation with the woofer's main inductor, but it sounds right.

                        Click image for larger version

Name:	Woofer Circuit.png
Views:	1
Size:	244.3 KB
ID:	1385618
                        Click image for larger version

Name:	Tweeter Circuit.png
Views:	1
Size:	206.5 KB
ID:	1385619
                        Sounds nice and clean!

                        Comment


                        • #13
                          Model this (w/your files) if you'd like - tell me what you think.
                          Woofer LP: 2.0mH low DCR series coil (w/a 4ohm resistor + tiny 0.10uF cap in series, w/BOTH across that coil - for a notch), and a 5uF shunt cap (to gnd).
                          HP: 12uF series cap and 3ohm series resistor, followed by 2 shunts (to gnd): a 0.10mH coil (#20 fine) and a 2 ohm (parallel) resistor.

                          I see similar xfer fns using 7 parts total.

                          Comment


                          • #14
                            You mean 8 parts per side?
                            Craig

                            Comment


                            • #15
                              Fascinating woofer topology Chris:

                              Click image for larger version

Name:	woofer and FR.png
Views:	1
Size:	574.8 KB
ID:	1385701

                              Seeing how I had excess capacitance issues earlier with my notch, I'd have to physically build this to see where that filter null ends up.
                              Would we consider that a bottomless notch? I had ruled one out since the woofer break-up looked too broad. This doesn't completely flatten it, but it'd be interesting to see if it was audible.

                              Tweeter filter:

                              Click image for larger version

Name:	tweeter.png
Views:	1
Size:	420.4 KB
ID:	1385702

                              If I understood you right, basically it's just a second order electrical filter with a pseudo L-pad?

                              I appreciate your thoughts!

                              At first I thought I was going to go for the lowest component count possible, but then I got hung up on actually being able to make second order acoustic slopes work.
                              The filter I settled on and posted has a simulated better power response and phase tracking through the crossover region, admittedly at the expense of a greater component count and a big freakin inductor on the tweeter. Below:

                              Click image for larger version

Name:	gregs.png
Views:	1
Size:	252.8 KB
ID:	1385703

                              I was really focusing on the power response and trying to keep it as smooth as possible, since it looked possible.
                              I don't have a whole lot of experience hearing the difference in power response, but it seemed like a worthy goal to pursue.

                              Comment

                              Working...
                              X