Posted by andyr (A) on October 24, 2009 at 02:53:58

OK, thanks to information from several inmates, I can now present a slightly different table of true-ribbon roll-off points - ie. the "@ frequencies" ... as in "12dB ribbon highpass @ XXXHz".

This is much more consistant than my previous table ... so, my apologies for my previous wrong assumptions as to which Maggies had what ribbon resistances. And a small apology to Magnepan for implying that they had no idea what they were doing!! :-)) (Although, they haven't been totally consistent, IMO!)

From modelling the stock 12dB ribbon HP filters (shown under "Tweaks") in lspCAD, I present the following table for the -3dB points (ie. the "@ frequencies"):
* IIIa: 2ohm ribbon - 6,200Hz (NB: this is the later IIIa)
* T-IVa: 2ohm ribbon - 5,600Hz
* 3.3: 3ohm ribbon - 2,600Hz
* 3.5: 3ohm ribbon - 2,300Hz
* 3.6: 3ohm ribbon - 2,300Hz
* 20: 3ohm ribbon - 4,700Hz
* 20.1: 3ohm ribbon - 4,500Hz
* and my own active IIIas: 2ohm ribbon - 3,400Hz.

IMO, the true-ribbon is a magical driver ... the more frequency range that you can cover with this, the better! :-)) I suggest Magnepan believes this also, so they changed back to the thinner, 3ohm ribbon with the 3.3 - and subsequent models - so they could bring it down lower. (I presume the thinner, 3ohm ribbon is more robust?)

Then they experimented a bit more and brought it down to an even lower frequency with the 3.5 & 3.6.

So I suggest to any 3.3/20/20.1 owners who are contemplating rebuilding their passive XOs that they should use the 3.5s'/3.6s' 2,300Hz -3dB frequency point for their ribbon HP filter. Of course, this means also reducing the mid LP roll-off frequency, to compensate.

But why didn't Magnepan use this same frequency for the 20s and 20.1s? Why start rolling the ribbon off at 4,700Hz/4,500Hz for these premium speakers?

WRT the (similar vintage) IIIa/T-IVa, Magnepan brought the wider, 2ohm ribbon down a bit lower with the T-IVa. However, as I didn't know that the 3.3 and the 3.5 had a different ribbon, I chose a -3dB point of 3,400Hz for my current 3-way active IIIa XOs as I wanted a 2,600Hz "crossover frequency" (-6dB point) between mid & ribbon. I've been running this way for at least 8 years so I'd have to conclude that the ribbon can cope with this lower roll-off point! :-)) And I - nor friends who come round to listen - certainly can't hear any negatives.


Regards,

Andy

• RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) - Neuro 20:54:09 10/27/09 (0)
  In Reply to: RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) posted by andyr on October 24, 2009 at 02:53:58

  I'm running a symmetrical 1700hz second-order LR network for my midrange and tweeter on my 3.6Rs. Andy, you can say "I told you so", because I switched back to a near-stock slope setup (250hz 3rd order LP, 280hz 1st order HP on the lower end too) after hearing some live performances and noticing how "thin" the lower midrange felt in comparison. Needless to say, I'm enjoying the music now.
  
  -- Nils
  

  ---
• RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) - Jim_E 11:29:11 10/26/09 (4)
  In Reply to: RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) posted by andyr on October 24, 2009 at 02:53:58

  Andy,
  I finally had a chance to run the full 3.3R high pass filter (Lars values, 3 ohm driver) in Matlab. The component values are not correct for a fourth-order Butterworth filter, and provide some low freq peaking. The actual roll-off point is close to yours.

  Does your analysis tool show the same peaking characteristic as well?

  regards,
  Jim

  
  
  

  
  
  

  ---

  Edits: 10/26/09

  • RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) - andyr 14:16:46 10/27/09 (0)
    In Reply to: RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) posted by Jim_E on October 26, 2009 at 11:29:11

    Hi Jim,

    Further to Davey's comments, if you compare the 3.3 with the 3.5, you can see that the mid LP filters are similar - except the 3.3 has 0.955mH as the first inductor in the 18dB LP filter, whereas the 3.5 has 0.355mH.

    lspCAD produces a very nasty mid-range frequency response for the 3.3... could Magnepan have made a mistake with the 3.3 (and it really should be 0.355mH)? After all, it's the same driver. :-))

    Regards,

    Andy
    
    
    
    

    ---

    Edits: 10/27/09
  • RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) - Davey 13:48:37 10/26/09 (2)
    In Reply to: RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) posted by Jim_E on October 26, 2009 at 11:29:11

    You have to be careful how you analyze that.
    
    The peaking is a result of being connected without the midrange crossover attached. The C1/L1 circuit driving the HP filter without the midrange filter attached will yield a big peak as you've shown.
    
    Try analyzing using just the two high-pass components and nothing else. It'll look good then.
    
    Cheers,
    
    Dave.
    

    ---

    • RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) - Jim_E 13:36:40 10/27/09 (1)
      In Reply to: RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) posted by Davey on October 26, 2009 at 13:48:37

      Davey,
      The peak is a classic under-damped response. I agree that adding the midrange network will probably tame that some. Even with reduced magnitude, there will remain a pair of filter poles around 190 Hz introduced by the use of the common mid/ribbon filter. This pair of filters poles introduces an abrupt 180° phase shift which would remain. The roll-off is also limited to 12 dB/octave for most of the ribbon filter's operating range.
      
      The use of the common high-pass LC filter for both the midrange and ribbon reduced the cost, but is not an optimal configuration. Both filters are fourth order(24 dB/octave), but neither even close to classic Butterworth configuration.
      
      Getting rid of the crossover module, and going with the bi-amp configuration seems a far better solution.
      
      regards,
      Jim
      

      ---

      • RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) - Davey 14:58:58 10/27/09 (0)
        In Reply to: RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) posted by Jim_E on October 27, 2009 at 13:36:40

        Yeah, I understand the way the filter works. (I wasn't sure you did in your simulation so was pointing that out.)
        
        There's an alternative tweeter connection in a schematic in the "tweaks" section that moves the tweeter filter connection point upstream of the midrange HP filter. (I think some users have performed this "tweak" previously since the page has been there for a long time.)
        
        http://www.integracoustics.com/MUG/MUG/tweaks/mag3.6xover.html
        
        
        It can be confusing analyzing electrical responses only on the Maggie crossover networks. The acoustic responses have to be factored in to yield approximate final results.
        
        Removing the crossover module and configuring for a bi-amp setup is a better solution, but you still have to design the electrical responses to achieve the proper final acoustic responses. Line-level crossovers require just has much design effort as high-level networks. Maybe a little more. :)
        
        Cheers,
        
        Dave.
        
        
        

        ---
• RE: Have we forgot about the 2.6 - benie 05:35:25 10/26/09 (2)
  In Reply to: RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) posted by andyr on October 24, 2009 at 02:53:58

  being a 2 way, I believe the ribbon was used to cover a much wider band than the III series, much lower than the III's
  

  ---

  • The 2.5 & the 2.6 have a smaller ribbon ... - andyr 12:26:28 10/26/09 (1)
    In Reply to: RE: Have we forgot about the 2.6 posted by benie on October 26, 2009 at 05:35:25

    So I didn't think they would be comparable. But, yes, the ribbons go down lower because of the 2-way configuration.
    
    Regards,
    
    Andy
    
    
    
    

    ---

    • RE: The 2.5 & the 2.6 have a smaller ribbon ... - benie 22:19:11 10/26/09 (0)
      In Reply to: RE: The 2.5 & the 2.6 have a smaller ribbon ... posted by andyr on October 26, 2009 at 12:26:28

      Yes there are the smaller ribbons, but other than the length, they are the same ribbon. When I rebuilt one of the ribbons on my III, I got the Kit off Magnapan with enough ribbon to make 3 ribbons. In the Kit it said it was the same ribbon for the smaller ribbons (2.6-2.5) for me to cut the length I needed. So maybe you can get them down even lower than you think.
      
      Just some food for thought.
      

      ---
• RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) - hemholtz 05:12:20 10/24/09 (2)
  In Reply to: RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) posted by andyr on October 24, 2009 at 02:53:58

  the 20.1r midrange is a little better in the top of it's range. the higher crossover would put the phase shifts out of most of the vocal range? an uncorrected crossover right in the middle of everything is a compromise. with deqx correction, i would lower the 20.1r crossover. the ribbon is still the faster driver. with correction, the top of the 3's and 20's should sound the same.
  

  ---

  • Ah, very good point, Bart. ... - andyr 13:33:23 10/24/09 (1)
    In Reply to: RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) posted by hemholtz on October 24, 2009 at 05:12:20

    Yes, keep the XO point above 3K if possible.
    
    Regards,
    
    Andy
    
    
    
    

    ---

    • RE: Ah, very good point, Bart. ... - hemholtz 16:48:41 10/24/09 (0)
      In Reply to: RE: Ah, very good point, Bart. ... posted by andyr on October 24, 2009 at 13:33:23

      when you can flatten out the midrange and correct the crossover to the ribbon, the crossover frequency can be right where you want it. when you have to do things in the analog world, the 20's specifications make sense and are an improvement over the 3's. when you can fix things up in the digital world, the 3's drivers are more than adequate.
      

      ---
• true-ribbon roll-off points for 3-way Maggies - play-mate 04:26:00 10/24/09 (4)
  In Reply to: RE: Musings on true-ribbon roll-off points for 3-way Maggies - corrected!! :-)) posted by andyr on October 24, 2009 at 02:53:58

  hey andyr
  
  I can confirm that the ribbon in my MG3.3R´s can handle a 2kHz crossover (@24dB) without any audible stress.
  
  kind regards
  

  ---

  • Hehe - so it should ... - andyr 04:46:57 10/24/09 (3)
    In Reply to: RE: true-ribbon roll-off points for 3-way Maggies posted by play-mate on October 24, 2009 at 04:26:00

    as:
    a) as I have just found out, the 3.3 has the same 3ohm ribbon as the 3.6, and
    b) the " 2KHz XO with 24dB slopes " (as I'm sure you know, the "XO point" for 24dB L-R XOs is defined as the -6dB point ) of your current 3.3 setup is higher than the -6dB point of the 3.5 & the 3.6 ... which is 1,600Hz.
    
    So you could take it down a tad, if you wanted to. :-))
    
    Regards,
    
    Andy
    
    
    
    

    ---

    • RE: Hehe - so it should ... - play-mate 17:50:05 10/24/09 (2)
      In Reply to: RE: Hehe - so it should ... posted by andyr on October 24, 2009 at 04:46:57

      thanks for the reminder, andy
      
      the beauty of digital XOs are of course one can just set a bewildering number of corrections and try them out.....on the other hand it gets just too much some times.
      
      I have planed to issue a report on my findings with all these variables, because, albeit I´m not a beliver of a "perfect" setting, there are some interesting correlations that would probably be of help for others.
      
      - I´m not quite ready to draw any final conclusions yet, but so far my observations have been that the stock values are not bad, but far from optimal. I believe that magnepan has build-in a (long) range of compromises.
      
      lets state the obvious first : steeper slopes DO allow for less distortion and more transparency !
      downside : they can muck up the phase and thereby imaging.
      doubling the stock values, and keeping the asymmetrical slopes intact, seem to do very well.
      
      the reversal of the polarity on the mid-driver seems to be rectified with "opposite" asymetrical XO slopes....(at least the with and depth returns)
      ....symmetrical slopes are absolutely no good !
      
      the digital XO I use is the computer program "allocator" (running straight from harddisk) which also has a phase correction feature.
      with that enabled, things seem to get quite close to where I hope to get, but only a thorough measurement will confirm if my ears are right.
      
      the dipole nature of maggies does not make it easier to pinpoint what is really "right" !
      
      does anyone by chance have ISPcad measurements of the 3.3R´s ???
      
      
      kind regards
      
      
      

      ---

      • OK, so looking at the enormous variation in the mid/ribbon slopes of 3-way true-ribbon Maggies ... - andyr 20:18:32 10/24/09 (1)
        In Reply to: RE: Hehe - so it should ... posted by play-mate on October 24, 2009 at 17:50:05

        which do you think would be the optimum? The various configurations used are:
        
        * IIIa: 12dB mid LP/12dB ribbon HP (ie. symmetric!!)
        * 20: ditto
        * 20.1: ditto
        * T-IVa: ditto
        
        * 3.3: 18dB mid LP/12dB ribbon HP
        * 3.5: ditto
        * 3.6: ditto.
        
        What did you mean by the " lspCAD measurements of the 3.3Rs "? I have the 3.3 modelled so what would you like me to send you?
        
        Regards,
        
        Andy
        
        
        
        

        ---

        • RE: OK, so looking at the enormous variation in the mid/ribbon slopes of 3-way true-ribbon Maggies ... - play-mate 04:13:41 10/25/09 (0)
          In Reply to: RE: OK, so looking at the enormous variation in the mid/ribbon slopes of 3-way true-ribbon Maggies ... posted by andyr on October 24, 2009 at 20:18:32

          hey andy
          
          sorry,
          -asymmetrical slopes for the bass-/middriver only, of course.
          
          it´s my ambition to see if I can find an alternative slope combination without having the mid-driver in reversed polarity. this is a daunting trial/error, but there seems to be some positive signs when the slope asymmetry is reversed too...: i.e. when the LP is less steep (24dB/oct) than the HP (36dB/oct).
          
          I run the mid-/ribbon at 24dB/oct symetrical (at now 1600Hz)
          
          it´s my feeling that it´s almost impossible to do all this with listening alone, as beside the theoretical reasons there are driver dispersion and room issues too.
          -have not got my meassuring equipment up and working yet, as I have no analog input on my entirely computer driven setup.....
          
          if you have an ispCAD of the MG3.3R´s I would love to study it !
          
          

          ---

Post a Followup:

---

[ Follow Ups ] [ Post Followup ] Thread: [ Display  Email  Next ] [ Planar Speaker Asylum ]
[ Comment ] [ Edit ] [ Delete ] [ Copyright Warning! Click for Details ]

---

[ Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2009, Audio Asylum, all rights reserved ] [ Go Green ] [ Take5 ]

---

[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]

---