Announcement

Collapse

Midwest Audio Fest

It’s that time audio enthusiasts! Registration for the 2019 Speaker Design Competition is now open! Visit midwestaudiofest.com for details and to list your speaker project. We are excited to see all returning participants, and look forward to meeting some new designers this year, as well! Be sure your plans include a visit to the Parts Express Tent Sale for the lowest prices of the year, and the Audio Swap Meet where you can buy and trade with other audio fans. We hope to see you this summer! Vivian and Jill
See more
See less

KEF C80 crossover redesign

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

  • KEF C80 crossover redesign

    Since I have three pairs of KEF C80s, I thought I might try a little experiment. It seems the premise behind the design of the C80 crossover was to create a uniform 4 ohm resistive load (or as close as they could get) for the amplifier. This produced what some people have described as "the World's most complicated crossover". The amplifiers that I usually use tend to be ones that don't have problems running into low impedance loads (NAD, Proton etc.). So I was wondering, would it be possible to re-design a crossover for the C80 that doesn't concentrate so much on impedance matching, instead just getting the correct range of frequencies to the right driver. Now I realise KEF probably knew best when designing this crossover, but some of us can't help tinkering with this old stuff, right?

    Maybe the new design crossover could mimic something like the KEF Concerto or the Constructor series CS7? The only snag here is, the Concerto and CS7 used the B139/B110/T27 driver compliment, whereas the C80 used the B139/B160/T33 drivers. Not only but also, the drivers in the C80 were all 4 ohms, the B139 itself is the unique SP1212 version.

    Think it's worth trying? I have no idea how to design a suitable crossover though. I'm cross-posting this in a couple of other forums as well.

    Lee.
    KEF C80 crossover KEF Concerto crossover KEF CS7 crossover

  • #2
    Of course you can do it, if you measure carefully, learn to use the SIM tools and are willing to spend a lot of time patiently voicing the result. I will comment that the extra circuitry in that C80 crossover looks to be null networks, probably to reduce annoying peaks.
    Francis

    Comment


    • #3
      Were I you, I'd take all the requisite measurements, get everything into the SIM using the existing crossover, and see if it makes sense, and looks like the actual FR of the existing speaker.
      Francis

      Comment


      • #4
        That orig. XO (most likely) holds impedance (pretty much) between 3-7 ohms (as you thought).
        It's not hard to copy the LP/BP xfer fns w/a LOT fewer components:
        2nd order LP: 6.0mH series coil (low DCR), 150uF shunt cap
        Mixed BP: 4ohm attenuation resistor (depends on coil's DCR), 80uF series cap, 1.5mH series coil, 10uF shunt cap.

        The HP fn is more "complex" - but here's a 4th order that's not too far off (within 1.5dB?):
        4uF series cap, 0.25mH shunt coil (to gnd), 12uF series cap, 2.0mH shunt coil
        (about 10 elements, instead of 25-30 !)

        Comment


        • #5
          Do you mean something like this (see attachment)?

          I could probably recycle some of the original inductors by re-winding them.

          Comment


          • #6
            Yup.
            W/out any F/Z data, I'm making an assumption that by duping transfer functions (at ANY particular load) you should end up at a pretty good point for "voicing" (tweaking).
            Odds are that the mid should end up "reverse" (negative) polarity. In THIS case, both the mid and tweeter show an affinity for being wired reverse.
            I would try to determine proper polarity by ear, 1st between the woof & mid (which MIGHT be really hard to tell - or maybe not, since they cross pretty low - near 200) leaving the tweeter off.
            Then between the mid (or woof & mid) and tweeter. They'll end up crossing near 2-1/2khz (at least electrically).

            Are these 8ohm "nominal" drivers do you know?
            Last edited by Chris Roemer; 05-01-2019, 08:25 AM.

            Comment


            • #7
              They're all 4 ohm nominal drivers, not 8 ohm, I can measure the DC resistance in the next day or so.

              By the way, Chris, thank you for your input on this thread. You helped me out with my aluminum Dayton drivers in an Advent cabinet project a while back. Those speakers still rock when I get them in rotation.

              Lee.

              Comment


              • #8
                Going from 8n to 4n drivers (*indicates a change to previous - which almost all components ARE:

                2nd order LP: *7.0mH series coil (see DCR note), 150uF shunt cap
                note - w/4n woofer, my sim used a DCR of (about) 4ohms on this coil to set the level.
                Not sure if the orig. design had ANY BSC, but I'd use a low DCR (cored) coil and live w/the approx. +3dB MORE BSC (unless these were "in-wall", which could be bad)

                Mixed BP: *2ohm attenuation resistor (depends on coil's DCR), *100uF series cap, *1.2mH series coil, *8uF shunt cap.
                note - need 3ohms between a series cap and the series coil's DCR, I'd go w/a 2n resistor and a #20 1.2mH air core (DCR near 0.8n)

                HP: *5uF series cap, *0.15mH shunt coil (to gnd), *16uF series cap, 2.0mH shunt coil (only 2nd SAME VALUE as init. XO, OTHER part being 150uF cap on woofer)

                Comment


                • #9
                  Here are the DC resistances of the drivers in my "project" C80s :

                  Tweeters T33 : sticker says "4 ohm", DCRs measured as 3.1, 3.2, 3.1 and 3.1. All four are part # SP1210, one pair also say TS1210, other pair says TS1191
                  Midranges B160 : sticker says "6 ohm", DCRs measured as 5.6, 5.4, 5.4 and 5.6. All part # TS1203, one also says SP1203.
                  Woofers B139 : sticker says "4 ohm", DCRs measured as 3.5, 3.4, and 3.5. All part # TS1212, one also says SP1212.

                  Hope this helps.

                  Lee.

                  Comment


                  • #10
                    8 parts:

                    HP (4n): 6.2uF series cap, 0.13mH shunt coil (DA #18 air core OK), 15uF series cap (reverse polarity indicated)
                    BP (6n): 1.2mH series coil (Jantzen #15 air is good), 100uF (npe) series cap, 12uF shunt cap (also reverse indicated)
                    LP (4n): 6mH iron core coil, 150uF (npe) shunt cap

                    (by the xfer fns) the mid looks to cover about 4 octaves, roughly 170-2600Hz. Pretty solid 8ohm nominal (6n, anyway).

                    Comment


                    • #11
                      This is a very interesting thread, I had no idea anyone had gone to such great lengths to keep impedance above 4-ohms.

                      Comment


                      • #12
                        Originally posted by philthien View Post
                        This is a very interesting thread, I had no idea anyone had gone to such great lengths to keep impedance above 4-ohms.
                        Some amps get unhappy with very low impedances, especially if they are reactive. Newer amps tend to be more tolerant.
                        Francis

                        Comment


                        • #13
                          I THINK the idea was to keep from havng the impedance rise too HIGH at any particular point (for tube amps)?

                          Comment


                          • #14
                            Originally posted by Chris Roemer View Post
                            I THINK the idea was to keep from havng the impedance rise too HIGH at any particular point (for tube amps)?
                            Could be. You can get arcing in the output tube if the load is too light.
                            Francis

                            Comment


                            • #15
                              Originally posted by Chris Roemer View Post
                              I THINK the idea was to keep from havng the impedance rise too HIGH at any particular point (for tube amps)?
                              I’ve never seen that don’t before, either.

                              Comment

                              Working...
                              X