Need some crossover advice

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • James Dawes
    New Member
    • Sep 2018
    • 6

    Need some crossover advice

    I am rebuilding a pair of Optimus Pro-77s I bought for $15 at a thrift store. I replaced the woofers with Dayton Audio DC-130a and replaced the caps in the crossovers. There was still too much midrange in the sound for me so I thought I might need better crossovers. I was going to keep the original soft dome tweeters because they sounded fine but I *!@#**!! ed em' up removing them, so now I am going to use the Visaton SC5-8 Poly dome tweeter, both woofer and tweeter available at Parts express (of course). And I'm planning on using some Dayton Audio low and high pass filters. My question is what is the best crossover point for these drivers? The woofer has a usable frequency response of 50-5,000 and the tweeter is rated as low as 1,300 but the specs recommend a usable response starting at 4,000. Any help would be appreciated.
  • Chris Roemer
    Obsessed & Proud of It
    • Sep 2005
    • 13476

    #2
    (Woofer) Low Pass: as if drawn from Left (amp) to Right (driver):
    1.0mH iron core series coil, then a 10uF "shunt" cap (to gnd), then a Zobel (across the woofer terms) using an 8ohm resistor and a 15uF cap.
    Also put a tiny ( 0.68uF) cap across the 1.0mH coil. This acts like a notch filter.

    High Pass (on tweeter): 8uF series cap, then a 0.30mH shunt coil (a #20 is fine), L-pad uses 5ohms in series and 3ohms in parallel.

    I'd call it a 6ohm nominal system. Sensitivity is about 81dBv with baffle step.

    Guess you should plug the port w/that woofer (almost reaches 100Hz). Not the greatest woofer to get bass out of a 0.1cf box.

    Comment

    • James Dawes
      New Member
      • Sep 2018
      • 6

      #3
      Thanks, Chris. I'm satisfied with the bass from the little woofer in that little ported enclosure. I plan on using pre-made Dayton Audio high and low-pass filters, I just need to determine a good crossover frequency set-point.

      Comment

      • Kornbread
        Seasoned Veteran
        • Oct 2015
        • 1479

        #4
        The recommended crossover Mr. Roemer provided should get you decent sound from that mix of drivers.
        http://techtalk.parts-express.com/fo...khanspires-but
        http://techtalk.parts-express.com/fo...pico-neo-build
        http://techtalk.parts-express.com/fo...ensation-build

        Comment

        • Geoff Millar
          Seasoned Veteran
          • Oct 2014
          • 1290

          #5
          Originally posted by Kornbread
          The recommended crossover Mr. Roemer provided should get you decent sound from that mix of drivers.
          +1

          To say that Mr Roemer knows his stuff is an understatement: you will most likely be disappointed with the sound of the 'off the shelf' crossover and Chris' will sound far better.

          Chris has given many, many people (myself included) his crossover suggestions.

          Geoff

          Comment

          • djg
            Seasoned Veteran
            • May 2008
            • 8515

            #6
            Ignoring Chris Roemer's advice and using premade crossovers would be a mistake. Not life or death, but you'll never know how good your project could have sounded.

            Comment

            • Chris Roemer
              Obsessed & Proud of It
              • Sep 2005
              • 13476

              #7
              Yeah, "setpoint", interesting concept - if it were only that simple...

              The XO I simmed ends up acoustically crossing the drivers near 3.6kHz.
              If you bought the 3.5kHz 2-way XO (for 8 ohm loads) the tweeter will run AT LEAST +10dB hotter than w/what I simmed.
              The "setpoints" that you see (in pre-made XOs, charts, on-line calculators, and in "textbooks") assume that the driver has a flat response within its passband (which they do not) AND a flat impedance profile (within the passband - and beyond).

              99% of all drivers are typically only at their "nominal" impedance at 3 points on their "Z" (impedance) curve: somewhat below and again above their Fs (resonant freq. - where they'll typically run DOUBLE that for tweeters, and 4x to 8x that for many woofers). They run LOWest (Re values) starting an octave or 2 below their Fs and below that point, and again maybe an octave or 2 above Fs (200-300Hz range for MANY woofers), after which the impedance continually rises (past the nominal value - usually for the 3rd time) up to the top end (20k) where they'll again run 4x-8x the nominal impedance. Tweeters are similar except Fs values are about 10-100 times higher, "ferrofluid" can reduce the Fs peak dramatically, and the T-E-Rise isn't so bad 'cause there just aren't that many octaves left for the value to climb. You CAN (and should) take a look at the published F (freq. resp.) and Z curves for your chosen drivers.

              For your 5-1/4" woofer (in "free air") - 8ohms "nominal".
              at 20Hz - about 9n(ohms)
              at Fs (near 50Hz) - around 27n
              around 140Hz - 8n
              near 350Hz - 6n (the "Re" value in the broad valley between "nom. above Fs" and the "top end rise")
              around 1.2khz - 8n, for the last time
              20kHz - about 29 ohms

              So, impedance is NOT "flat". A big woofer problem is that the "top end rise" usually means that the "textbook" coil (to roll off the top end) needs to be considerable larger - since the impedance is a MULTIPLE of the nominal value. A simple "Zobel" circuit (usual co$t - about $1) can be put in parallel (across) the woofer's terminals to effectively FLATTEN out the top end rise.

              Biggest problem w/tweeters is that they're more efficient (louder) than the woofers they run with. Your "off-the-shelf" XO does not include any "padding" (attenuation) to correct the issue. If you'd like "baffle-step compensation" (BSC) in your speaker (and most often it IS needed), then the tweeter typically will need an ADDitional -3 to -6dB of attenuation.

              So . . . if you NEED to buy premade XOs:
              #260-1872 (2.5k 8ohm LP - ASSuming there's a 1.0mH coil on it) would be OK on the woofer - provided you ADD +6uF of capacitance (we need 10uF total) AND run the Zobel (I recommended) across the woofer terms.
              #260-1898 (4.0k 4ohm HP - assuming the coil is around 0.32mH) will work on the tweeter IF you add +3uF more capacitance (figure it comes w/5uF, and we want 8uF) and you add the L-pad I recommended (for attenuation).

              These 8n drivers are rated at about 88dB (woofer) and 90dB (tweeter), - but the tweeter's "rising top end for "SPARKLE" " is the killer here. Also, it would seem that the tweeter needs about -2dB of padding (-4 to -5 to counteract some of the sparkle), but when you add in maybe -5dB more for BSC, it ends up needing a total of more like -10dB.

              Comment

              • Chris Roemer
                Obsessed & Proud of It
                • Sep 2005
                • 13476

                #8
                Imagine if you'd paid $80 for a pair of "Pro" pre-made Eminence XOs (#290-636). Spec'd for 8n loads, crossing at 3500 Hz.
                The grey curve is the "stock" DC130A-8 driver. The BLUE curve shows the effect of the -636 LowPass section on that woofer.
                Not only isn't the woofer rolled off (typically down -3dB to -6dB @ 3.5k) at ALL, the LP filter actually worsens the cone breakup distortion by about 50% ! !
                (It's making the sharpness of the resonance peak even greater.) In general, Dayton's "Classic" woofers don't work well w/ 2nd order BW filter alignments (not in 2-ways, anyhow). Click image for larger version  Name:	Classic5breakup.jpg Views:	1 Size:	74.7 KB ID:	1387994 The white curve (bouncing around the 81dB line) shows your system FR w/my simmed XO. The green-ish curve is the woofer w/my 2nd order filter and Zobel. (Notice how the cone breakup is down by around -20dB using MY filter, while the Pro filter is actually +1 !) The red curve shows the filtered tweeter plot. Baffle-step will gradually suck about -6dB off the curve (roughly in the 1000 to 200Hz range) with it ending up relatively "flat".

                BTW - your "too much midrange" is most likely the cone breakup on your new woofers (it's almost +10dB higher than "nominal" approaching 4khz). NOthing wrong w/that little guy, as long as you put a custom XO on him!

                Comment

                • James Dawes
                  New Member
                  • Sep 2018
                  • 6

                  #9
                  Mr. Roemer,

                  I appreciate the amount of time and effort you have put into answering my question, it's obvious you know your stuff. I believe completely that your custom crossover network will do the trick and I have no intention of ignoring such sage advice and I'll even plug that port if you say that's whats needed. I am deep into this project now and don't want to quit ......... BUT.......... here's my dilemma, I am a novice enthusiast at best and while I am trying to learn electronics I don't quite yet posses the technical knowledge to tackle building your XO from just your description alone, most of the people on this forum could probably do it blindfolded but not this little guy, not yet. So if you will permit me a few hundred more follow up questions I'd like to proceed. I've been trying to do this project "on the cheap" but I am willing to drop a few more $$ to obtain the desired outcome which is good sound quality (for thrift store speakers), so I am thinking of using the Morel MDT12 tweeter ($35ea. vs. $10 ea.) in which case the values of your XO components would change. I would need to know what type of caps (electrolitic, ceramic, etc.), resistors, I think I can figure out the coils. Thank you again for your input.

                  Comment

                  • Chris Roemer
                    Obsessed & Proud of It
                    • Sep 2005
                    • 13476

                    #10
                    Tell me about your (little) port. What's its I.D. and length?
                    Is there any way you'd consider something like the 0.23 cf Denovo cabinet?
                    I'd rather recommend a Peerless tweeter to you.

                    Comment

                    • James Dawes
                      New Member
                      • Sep 2018
                      • 6

                      #11
                      The port is 5/8" in Dia. and 2 5/8" long. I'll check out the Denovo cabinet but I'd really like to make this die cast box work if possible. I know Peerless tweeters are good but it's been a challenge finding any tweeter to fit this box but I will look into this as well. Also I have been watching a lot of videos on crossover design and construction and feel I could make your crossover if I had a schematic.

                      Comment

                      • Chris Roemer
                        Obsessed & Proud of It
                        • Sep 2005
                        • 13476

                        #12
                        *note that nearly all values have been changed a bit (to better match your cabinet and the little Peerless tweeter)

                        (Woofer) Low Pass: as if drawn from Left (amp) to Right (driver):
                        *1.5mH iron core series coil, then a 6.8uF "shunt" cap (to gnd) - I "cancelled" the Zobel - but went to a bigger coil (nicer curve shape).
                        Also put a tiny ( 0.68uF) cap across the 1.5mH coil. This acts like a notch filter.

                        I'd like you to use the Peerless DX20BF00-04 (#264-1472). It's only $11, but is light yrs. ahead of the 1/2" poly.
                        Can you make that fit in your cabs?
                        *High Pass (on tweeter): 5.6uF series cap, then a 0.30mH shunt coil (a #20 is fine), L-pad uses 3ohms in series and 6ohms in parallel.

                        (Part nos. : #257-552 027-336(027-424 optional) 027-408 027-425(027-334 optional) 257-028 016-3.3 016-6.2)

                        I'd call it an 8ohm nominal system. Sensitivity is about 83dBv with baffle step.

                        I figured you'd use the woofer and (vented, undersized) box "as is". That port tunes the box to around 50Hz, which kinda means nothing (in THIS case), but I think the response is better WITH it, than w/out it.
                        Doubling the box volume (and porting) would only gain about 1/2 octave on bottom end (from 105Hz to 75Hz).
                        These should sound pretty good crossed to a small sub (like TangBand's W5-F or W6-F) at 100Hz.

                        Comment

                        • James Dawes
                          New Member
                          • Sep 2018
                          • 6

                          #13
                          Parts have been ordered. Yes, the Peerless tweeter you chose will fit just fine. Glad we're keeping the little port, it did puff air like a fat chick on a stair-master. You've been a ton of help so far on this project Chris, thanks again. Any chance I could get a simple schematic? I know you gave what would be for most guys here an adequate explanation but may need a little more guidance.

                          Comment

                          • Kornbread
                            Seasoned Veteran
                            • Oct 2015
                            • 1479

                            #14
                            Is this correct?

                            Click image for larger version

Name:	xsim p.JPG
Views:	262
Size:	60.3 KB
ID:	1388196
                            http://techtalk.parts-express.com/fo...khanspires-but
                            http://techtalk.parts-express.com/fo...pico-neo-build
                            http://techtalk.parts-express.com/fo...ensation-build

                            Comment

                            • Chris Roemer
                              Obsessed & Proud of It
                              • Sep 2005
                              • 13476

                              #15
                              Exactly !

                              Thanks Kornbread

                              Comment

                              Working...