Announcement

Collapse
No announcement yet.

Tweeter padding help please

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

  • Tweeter padding help please

    I just finished a speaker with the DA-175 woofer and D2608/913 (HDS) tweeter..

    It sounds great and the single resistor padding on the tweeter works well as it does not distort the response of the tweeter and phase tracking is under control with a nice reverse null.

    BUT, is the HUGE impedance swing a problem?

    I tested a L-pad, split-pad and various other combinations of resistors, but the nice and flat frequency response goes out the window every time.

    Any suggestion?
    Thanks!
    Attached Files

  • #2
    Since you have the resistor before the high pass filter it doesn't mess with the transfer function of the filter as it would post-filter. so that's OK. The impedance peak looks nasty, but it would only be a problem with a tube amp. An L-Pad post filter shouldn't bother the frequency response, so I suspect something about it wasn't correct. But what matters most is how it sounds. If it sounds good it is good.
    Last edited by billfitzmaurice; 08-04-2020, 09:54 PM.
    www.billfitzmaurice.com
    www.billfitzmaurice.info/forum

    Comment


    • #3
      Try this in your sim...... In front of all the x-over parts, add a cap and resistor in series, and place it across the amplifier connections. Try 10uf and 30 ohms as a starting point, then play with the values a little. Get someone else to tell you if it's safe to do this, or not.

      Comment


      • #4
        rpb's "Zobel" should do the trick. I've seen others add this to an XO design to (supposedly) make it "tube amp friendly".
        If you'd like to try an L-pad; try SR=10n(ohms) / PR=5n (this is at the tweeter end and replaces the 27n in series at the front end).
        I nudged your caps a bit to minimize any changes in FR: 3.3uF to 4uF, and 5.1uF to 6.8. No change to coil.

        Comment


        • #5
          Thanks guys, that's almost black magic!

          The compensation network (Zobel?) seems to work (not tested, only sim'ed).. I used 30uF and 10n across the amp and I can almost make the impedance ruler flat..

          I'm wondering if this will help with the power response.. I can't find any graphs about this in XSim.,

          When I listen to the speaker at normal level, the balance is good, but when I lower the volume, the tweeter seem to take over (louder relative to the woofer).

          More testing soon..
          Attached Files

          Comment


          • #6
            Search 'equal loudness curve'. It's normal for lows to be less audible at lower levels. It's the reason for loudness compensation in receivers. I still think you're too concerned about the impedance curve. SS amps deliver constant voltage into any load.
            www.billfitzmaurice.com
            www.billfitzmaurice.info/forum

            Comment


            • #7
              Do a quick read over at wikipedia on the equal loudness contour. It's all about how the ear and brain perceive equal loudness at different frequencies and absolute levels.

              Comment


              • #8
                That "Zobel" on the amp end doesn't affect your FR. My L-pad/caps suggestion won't affect it MUCH.

                Comment


                • #9
                  Originally posted by Audiotorium View Post
                  Thanks guys, that's almost black magic!

                  The compensation network (Zobel?) seems to work (not tested, only sim'ed).. I used 30uF and 10n across the amp and I can almost make the impedance ruler flat..

                  I'm wondering if this will help with the power response.. I can't find any graphs about this in XSim.,

                  When I listen to the speaker at normal level, the balance is good, but when I lower the volume, the tweeter seem to take over (louder relative to the woofer).

                  More testing soon..
                  My suggested filter should have reduced the impedance up high to about 15 ohms, or so. I chose a larger resistor, and smaller cap to limit the power that would go through, while still improving the impedance. I didn't sim it, I guessed, and did a little math in my head. Was I close?

                  Comment


                  • #10
                    Thanks to all your suggestions and ideas, I started from scratch and came up with a better crossover, no impedance compensation required.

                    Frequency response is ruler flat with the higher mids a bit relaxed.. I will listen to it for the next week before finalizing.

                    One more thing that puzzles me is the power dissipation of the resistors. With the power amp set at 20 watts, XSim predicts
                    that resistor R3 will explode.. see graph below. Not sure how to interpret this.

                    Anyways, its a nice sounding speaker overall, but I am not a huge fan of this Scan Speak D2608/91300 (previously Peerless HDS) tweeter.
                    I have used tweeters that sounded better for less money.

                    Thanks!
                    Attached Files

                    Comment


                    • #11
                      Originally posted by Audiotorium View Post
                      One more thing that puzzles me is the power dissipation of the resistors. With the power amp set at 20 watts, XSim predicts
                      that resistor R3 will explode..
                      With a full range signal maybe, but being high passed filtered it doesn't get a full range signal. Power density drops by 3dB with each octave increase in frequency.

                      www.billfitzmaurice.com
                      www.billfitzmaurice.info/forum

                      Comment

                      Working...
                      X