Announcement

Collapse
No announcement yet.

Help with first-order XO for ND90-8 and AMT Mini-8

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

  • Help with first-order XO for ND90-8 and AMT Mini-8

    I have the ND90-8 running full-range on open baffle with a small sub, and I want to add some air up top. In XSim, the AMT Mini-8 looks like a perfect match...even looks good just using a single cap, but I am absolutely new to XO simulation and design, and I would really like some opinions on this.

    First pic is what my system looks like now.

    Second is the current room response without a tweeter, as crudely measured by a free RTA app called n-Track Tuner on my iPhone, while playing pink noise.

    Last two are screenshots from XSim showing the added AMT tweeter with a series cap. Since my room response at the listening position measures (and sounds) quite a bit darker than what the ND90 factory meausurements suggest, I am wondering if perhaps I should just buy a range of cap values from .47 uF up to around 2 uF, and try them all until I get enough treble to satisfy. Any words of advice?

    Since I am using first order and both drivers are wired in phase, will it be possible to arrive at a time-aligned or "transient perfect" integration with thoughtful tweeter mounting? Judging by the depth of its cone, I estimate the acoustic center of the Nd90 is about 9mm behind the baffle surface. I can't tell from PE's part drawings where the AMT Mini-8 diaphragm is exactly, but it looks to be about halfway back in the housing, so I can guess about 5mm behind the baffle when mounted. In this case, should I offset the tweeters a bit towards the outer edge of the baffles (away from center listening position) to add delay?

    A related XSim question: Under driver options, I see the "mod delay" option. If I want to add time delay to a driver, do I type a positive or negative value (in inches) into this box?

  • #2
    Were it me, I'd try to pad down the AMT a little. A 5dB peak up there is a bit painful to my ears. A series resistor might work. If it then rolls way up high, put a smallish cap across that resistor. You can probably get transient perfect at one spot in space, with the usual caveat that moving the measurement microphone even an inch will result in non-transient perfect.
    Francis

    Comment


    • #3
      Is this better? I tried padding the tweeter with a series resistor, but didn't get much improvement. Finally I realized that the 7K-10K peaks were from woofer breakup. So I added a 1st order low-pass on the woofer and...is that a zobel (I think?), which further sharpened the cutoff. Now the tweeter is really handling much more of the upper range on its own, and although parts cost has risen, I think it'll be worth it for the extra smoothness and dispersion up high. Fpitas, thanks for encouraging me to try again.

      First pic is new XO. Second pic is original for comparison. Still open to suggestions.

      My XSim question still stands for anybody in the know...
      Under driver options, I see the "mod delay" option. If I want to add time delay to a driver, do I type a positive or negative value (in inches) into this box?

      Comment


      • #4
        Your woofer low pass is best described as a damped 2nd order electrical filter. Its not looking bad either. Perhaps try the tweeter resistor in series instead of parallel or maybe an lpad.
        https://www.facebook.com/Mosaic-Audi...7373763888294/

        Comment


        • #5
          Originally posted by stellarelephant View Post

          My XSim question still stands for anybody in the know...
          Under driver options, I see the "mod delay" option. If I want to add time delay to a driver, do I type a positive or negative value (in inches) into this box?
          Mod delay = positive numbers represent moving the mic farther away for that particular driver.

          Comment


          • #6
            Third try on the XO, now available for critique. Second picture is the ND90 full-range for comparison.

            IsaEagle, thank you for setting me straight about the filter now being damped second order. You are probably right that it was still a little hot in the treble. I tried padding the tweeter with a series resistor and with an L-Pad. Both drop the extreme high end too much. BUT...

            The good news is that the Mini AMT is such an inefficient tweeter (is that good news?) that it almost matches the sensitivity of the ND90 perfectly without ANY padding! I found that I can drop the tweeter output very predictably by just lowering the series cap value...this rolls off everything very smoothly above 5K or so. No resistors needed up there now that I have the woofer rolling off smoothly.

            GTPlus, thank you for the mod delay tip!

            Comment


            • #7
              Originally posted by stellarelephant View Post
              Third try on the XO, now available for critique. Second picture is the ND90 full-range for comparison.

              IsaEagle, thank you for setting me straight about the filter now being damped second order. You are probably right that it was still a little hot in the treble. I tried padding the tweeter with a series resistor and with an L-Pad. Both drop the extreme high end too much. BUT...

              The good news is that the Mini AMT is such an inefficient tweeter (is that good news?) that it almost matches the sensitivity of the ND90 perfectly without ANY padding! I found that I can drop the tweeter output very predictably by just lowering the series cap value...this rolls off everything very smoothly above 5K or so. No resistors needed up there now that I have the woofer rolling off smoothly.

              GTPlus, thank you for the mod delay tip!
              I think you may still want your tweeter resistor as a series element.
              Francis

              Comment


              • #8
                Put a combination of a 0.1mh inductor, 3.3uf cap, and 2 ohm resistor ,all connected in series, in parallel with the tweeter. This should tame that nasty 9k peak.

                Comment


                • #9
                  Originally posted by fpitas View Post

                  I think you may still want your tweeter resistor as a series element.
                  I don't have any resistor in there at all at the moment. Can you explain why? Still looks too hot?

                  Originally posted by dynamo View Post
                  Put a combination of a 0.1mh inductor, 3.3uf cap, and 2 ohm resistor ,all connected in series, in parallel with the tweeter. This should tame that nasty 9k peak.
                  Thanks for that. I'll try it tonight in XSim when I get home.

                  Comment


                  • #10
                    Might need to tweak your values a touch based on your measurements but those were the value for mine. Adjust the resistor to adjust level, just watch your net impedance..

                    Comment


                    • #11
                      Finally got around to finalizing my XO design, so I ordered the Mini AMTs and XO parts this week. This is what I arrived at. Very excited to put this together soon. I did buy some extra resistor values to allow some flexiblity with voicing.

                      A question for anybody who knows about time aligning drivers: Why does the phase shift get steeper when I simulate pushing the AMT backwards, to roughly align acoustic centers of the two drivers? I set the AMT's mod delay value to 0.25 and was surprised that the phase response was worse. I had planned on routing shallow waveguides into my baffles for the AMTs, but maybe not!

                      Click image for larger version  Name:	Final XO.png Views:	0 Size:	136.0 KB ID:	1442121

                      Comment


                      • #12
                        I can't recommend a shunt resistor on the woofer (33ohm to ground).
                        I end up w/: 0.60mH series coil, 2uF shunt cap, Zobel w/ 10ohms+4.7uF
                        I can't get a good rolloff on the AMT going 1st order.
                        I get: 4ohm series resistor and 1.5uF series cap, then a 0.15mH shunt coil (to gnd).

                        Don't know where you're at on phase. I'd try normal/reverse on AMT and use what sounds best to you.

                        ?
                        How come your latest plot of the ND90 doesn't look (anything) like your initial "system" response, which was just the ND90 - FR?

                        Comment


                        • #13
                          Originally posted by Chris Roemer View Post
                          I can't recommend a shunt resistor on the woofer (33ohm to ground).
                          I end up w/: 0.60mH series coil, 2uF shunt cap, Zobel w/ 10ohms+4.7uF
                          I can't get a good rolloff on the AMT going 1st order.
                          I get: 4ohm series resistor and 1.5uF series cap, then a 0.15mH shunt coil (to gnd).

                          Don't know where you're at on phase. I'd try normal/reverse on AMT and use what sounds best to you.

                          ?
                          How come your latest plot of the ND90 doesn't look (anything) like your initial "system" response, which was just the ND90 - FR?
                          Hi Chris. Thanks for your detailed response.

                          Perhaps I made a silly mistake... Is a shunt resistor on a woofer a big no-no? Enough that I should put on the brakes with my build and exchange my parts?

                          Your recommended low-pass gives a nearly identical slope to mine, the biggest difference being the impedance curve. And one less part.

                          As for why my system measurements don't look similar to the published ND90 specs...maybe its because I used a freeware app on my iPhone. The sub was on, too.

                          Comment


                          • #14
                            It's not done because:
                            Shunt resistors are used to attenuate louder drivers - a woofer's usually the quietest one. If you could UN-attenuate a woofer (say, 2dB), then you could also unattenuate your tweeter - gaining +2dB system sensitivity.

                            A woofer shunt resistor (OR series resistor, for that matter - which isn't done EITHER) has to handle W-A-A-Y more power than ones in a tweeter filter. They're probably not (thermally) up to it unless used in multiples.

                            Comment


                            • #15
                              Originally posted by Chris Roemer View Post
                              It's not done because:
                              Shunt resistors are used to attenuate louder drivers - a woofer's usually the quietest one. If you could UN-attenuate a woofer (say, 2dB), then you could also unattenuate your tweeter - gaining +2dB system sensitivity.

                              A woofer shunt resistor (OR series resistor, for that matter - which isn't done EITHER) has to handle W-A-A-Y more power than ones in a tweeter filter. They're probably not (thermally) up to it unless used in multiples.
                              Yikes. Well I just got off the phone with PE and my order already shipped out this morning

                              Chris, here is the response with your recommended low-pass (Red) versus mine (Blue) for comparison. I didn't change the tweeter side. Only a hair of difference in sensitivity that I can see. It looks like the 33R (10W rated) resistor will hardly be dissipating any power. Am I wrong? Also the impedance...is there any benefit to taming that peak? Also my eyes still like the way the blue FR curve seems every so slightly less peaky in the midrange (more damped, I guess?)

                              I'm not trying to get you to tell me that my newbie mistake was a good one... I'm just wondering if there could be any benefits, or if it is just benign. OR if it is going to burn up a the resistor OR sound worse...then of course I am open to exchanging some parts. Seeking advice from you (and others) here! What would you do?

                              Attached Files

                              Comment

                              Working...
                              X