Announcement

Collapse
No announcement yet.

My foolish journey into making a DIY center channel speaker

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

  • My foolish journey into making a DIY center channel speaker

    *Im sorry for the long post, questions at the end*
    Hi all, I recently decided to try building a center channel speaker, after months of trying to score a DIYSG 1099 without any luck. I took a few online courses in speaker building and have been reading a lot. If this turns out to be a total disaster, I figured I can always sell the drivers and at least I kept myself out of trouble. Currently I have an 88-Special as a center and I don't like the form factor. My L and R's are DIYSG Cinema 10's with 2 Eminence Delta 10b's and a SEOS 10 waveguide - IDK what CD is in it. So I set out to build a 2 way MTM and bought 2 Emi Delta 10b's and a BC DE250 compression driver (since it's similar to what's in the 88 Special).
    The first waveguide I tried was the Faital Pro LTH102, because it could be crossed down low and I liked the narrower dispersion pattern (I have a smallish room and thought less room interaction would be better). However I just couldn't get into it. It was just too directional, even though I only have a 3 seat couch where I was listening. So next I tried the Faital Pro STH100 and the B&C ME20. The 100, ME20 and the SEOS10 all measured pretty similarly on the tests I did. I liked the sound of the STH100 over the ME20, but just barely. But, I do like the price of the ME20 and the fact that it comes with a gasket!
    I built a ~2.5ft3 test box with a front baffle that measured 12"x42". I was aiming for a tune of 66hz, but near field port measurements and DATSv3 is saying It's actually around 63hz. I'll have to raise that for the next box I build. I've built a few baffles, and have changed out a lot of speakers and waveguides, this is a lot of work!!! I have tons of appreciation for the people who design the DIY kits for all of us to build.
    I used a miniDSP to find driver time alignment and set the crossovers. I ended up using the new speaker in this config for a week or so as a center channel. It's hard to compare between different horns, or my old 88 special, because everything takes so long to change and I forget how the last set up sounded (except for in cases like the LTH102 which definitely sounded different). Another issue I run into is my old set up has all the Audyssey settings and in room eq set. So that has been a challenge...
    Anyways, my mind kept saying what if... and I decided to try some mid drivers to put under the STH100 to help off axis response in the XO range. I settled on a pair of sealed 6" Eminence drivers (LA6-CBMR 6-1/2") This made things a lot more complicated, and I still don't know if I am going to keep them or go back to the 2 way. Im already concerned with building a 2 way crossover, let alone a 3 way! The 6" drivers are kind of big. So I had to push the 10" drivers further way and I start to get a pretty big null at around 600hz at 25 degrees off axis. On the baffle with just the STH100 and the 2 10's, the null was closer to 1000hz. SO it almost seems like a toss up. On the 2 way I could cross at 1k and keep a decent 50 degree window, OR go with a more complicated 3 way and still have about a 50 degree window since the drivers are now further spaced out. The STH100 says it should be crossed around 1.5k so it should be happier in the 3 way, but I didn't notice anything weird when I had it crossed at 1k and I doubt I will be playing it loud enough for it to matter. But I don't know, this is all new to me!

    So this is where I am at now, and was hoping to get some feedback and/or suggestions.

    I've done measurements in REW to make frd files and I've made zma files with the DATS.
    These are the gated measurements of the 2 10's in parallel, 2 6's in parallel, and the DE250, at 1 meter, at 2.83vdc into an 8ohm load, tested with a 60hz sine wave. I also measured all the drivers in parallel without touching anything to find phase alignment. This speaker is really really heavy so I was only able to get it about 5' off the ground, but it was on end, so the high end was about 9' up and I tested out in my backyard so the impulse response window looked pretty clean.

    Click image for larger version

Name:	Screen Shot 2021-09-21 at 9.03.55 AM.png
Views:	536
Size:	208.9 KB
ID:	1476381


    I also took near field measurements of one 10" driver and one port. However I haven't merged any of this with the 10" far field measurements because Im not quite sure how to do this with dual woofers and dual ports. I also haven't done any baffle step compensation to the measurements because the speaker will end up sitting flush with two subwoofers and a cabinet below. It will also be up against a wall and the speaker is only 13" deep. So Im not sure how to go about these two calculations, or if it's needed since Im building it for me and a specific location.
    Click image for larger version

Name:	Screen Shot 2021-09-21 at 9.10.22 AM.png
Views:	485
Size:	290.1 KB
ID:	1476382

    These are 25 degree horizontal off axis measurements I took to see what was going on and to help with crossover points.
    Click image for larger version

Name:	Screen Shot 2021-09-21 at 9.14.44 AM.png
Views:	485
Size:	281.5 KB
ID:	1476383

    I also looked a distortion for the crossover points. The 10's basically have no distortion until the go below 100hz. The 6's look like they're good to 500 and start to spike up around 2k.. but still super low.
    Click image for larger version

Name:	Screen Shot 2021-09-21 at 9.16.36 AM.png
Views:	485
Size:	252.5 KB
ID:	1476384

    The DE250 had 2-3% over it's whole range, which I thought seemed weird... Idk if this looks normal to you guys, but its very flat but starts to go up below 1k.
    Click image for larger version

Name:	Screen Shot 2021-09-21 at 9.19.03 AM.png
Views:	489
Size:	282.8 KB
ID:	1476385


    Sorry again for the long post. I have to decide now if I want to keep the 3 way, or go back to 2 way. What to do about baffle step and near field measurements, if needed. And start working on a crossover. Vituix seems amazing with the baffle simulator and auto optimizer, etc, but its really unintuitive and complicated. XSim has been way easier for me to play around in. Right now in my completely unqualified and unexperienced opinion it seems like a good XO points to start would be 500-600 and 1.7-2k.

  • #2
    This is what it looks like as of now with the 6" drivers. Note, I will be able to rotate the 6" drivers 90 degrees to push all drivers closer together but they wouldn't fit in this test box! Right now there's only about 1/8" on the top of the CD and bottom of the 6" drivers. Box is braced and filled with mineral wool. Area looks more cramped than it is.
    Click image for larger version

Name:	65379330431__087C9610-8D79-462E-A9F2-8284267BD8D2.jpg
Views:	558
Size:	580.9 KB
ID:	1476387

    Sitting next to a Cinema 10 with first waveguide I tried.

    Click image for larger version

Name:	65258138332__C636E818-CB9C-4E47-8F86-202A6615C943 2.jpg
Views:	546
Size:	653.4 KB
ID:	1476388

    Comment


    • Steve Lee
      Steve Lee commented
      Editing a comment
      Love your bookshelf wall brackets . . .

  • #3
    Can someone please check my math so far and steps to figure out my woofers response?



    Scale response of port- Area of ports - 1.25” x 10.5” x 2 = 26.25

    Then divided that by area of speakers- Area of speakers = 8.25” measured middle of surround to middle of surround. 53.46 x 2 = 106.92

    Then:

    26.25/ 106.92, √, apply log, then x20 = -6.099db


    Then add 6db because I have 2 ports. = -.1db


    Then I added 6db to my near field woofer measurement.


    Then add near field and port measurements together - generate new curve.


    Then apply BSC to this new measurement.


    Splice new measurement to my existing far field measurement.


    Thank you!

    Comment


    • #4
      I see problems.... The selected mid driver is good only down to 500hz. That's not low enough. You need to be crossing closer to 250hz. Do you have room for a 25" tall center with traditional 3-way driver orientation?

      Comment


      • #5
        Hi rpb, thank you for the reply. If the mid driver is only good to 500, why would I want to cross at 250?

        I do not! Ideally I don't want to go any taller than around 13"

        Thank you!

        Comment


        • #6
          Originally posted by jvphotog View Post
          Hi rpb, thank you for the reply. If the mid driver is only good to 500, why would I want to cross at 250?

          I do not! Ideally I don't want to go any taller than around 13"

          Thank you!
          You wouldn't. You need a different mid.

          Crossing at 250hz would eliminate the off axis nulls in the woofer(s) response.

          Comment


          • #7
            Originally posted by rpb View Post

            You wouldn't. You need a different mid.

            Crossing at 250hz would eliminate the off axis nulls in the woofer(s) response.
            How come? I measured off axis at 25 degrees and it seemed alright. What am I missing? Thank you

            Comment


            • #8
              Originally posted by jvphotog View Post

              How come? I measured off axis at 25 degrees and it seemed alright. What am I missing? Thank you
              You said... So I had to push the 10" drivers further way and I start to get a pretty big null at around 600hz at 25 degrees off axis.

              Your posted 25 degree measurement doesn't look that bad.

              My preference would be to cross lower, but it's your design, and compromises are part of the design decisions.

              Comment


              • #9
                Originally posted by rpb View Post

                You said... So I had to push the 10" drivers further way and I start to get a pretty big null at around 600hz at 25 degrees off axis.

                Maybe crossing at 500hz, the mids would cover the dip. My preference would be to cross lower.
                Got it. I probably didn't explain that part fully since I was already writing so much. My mlp is about 15' back and its a 3 seater couch, and about 95% of the time it's just me. My window is about +/- 10 degrees. So I made those 25 degree measurements thinking that was wide enough. I also figured that less energy bouncing around in my smallish room would be a good idea.

                -

                I've been playing around with XSim all evening. I've gotten a flat ish response, with 15-20db dips when i invert drivers. I got there by lowering the mids by -6db. Im going to try running them in series and starting over. :( System impedance is not good right now, im hoping this will help.


                Question about z offset and phase. I derived minimum phase by applying the Hilbert transform to the driver frd files. In XSim I was able to match up the driver frd’s to the overlay frd of all the drivers in parallel at -3.5in for the 6’s and -2.9 for the 10’s. Does that seem too far??

                Last week when I was using it as a center with the miniDSP as the crossover I got around .2ms delay for the 10’s and about .3ms for the 6’s when testing in REW and the impulse response.

                .2ms is ~2.7 inches and .3ms is ~4.1 inches. Does it seem like I screwed something up, or is there a tolerance for error here? The more I read about offset and how its relative the more I get confused...

                Comment


                • #10
                  I like the looks of the 6" speakers flat like the first image, but if I rotate them I can get the big drivers to sit an inch closer. Do you think it's worth it? Thanks!

                  Click image for larger version

Name:	Screen Shot 2021-09-16 at 11.17.00 AM.png
Views:	392
Size:	310.3 KB
ID:	1476542 Click image for larger version

Name:	Screen Shot 2021-09-16 at 11.15.31 AM.png
Views:	384
Size:	332.4 KB
ID:	1476543

                  Comment


                  • #11
                    One inch will not make any difference.

                    Comment


                    • #12
                      Is there music that uses a center channel? I thought it was bandwidth limited and intended for dialog.
                      Francis

                      Comment


                      • billfitzmaurice
                        billfitzmaurice commented
                        Editing a comment
                        Some formats allow for music having center channel content. I don't use them myself, they don't sound right to me compared to stereo. In HT, where center channels do get used, they handle most of the content above where the subs operate, which is mainly dialog or anything else that should seem to be coming from the center of the screen. But when the L/R kick in they're usually louder than the center, so the center doesn't need the output capability of the L/R speakers.

                      • fpitas
                        fpitas commented
                        Editing a comment
                        Thanks Bill

                    • #13
                      Originally posted by fpitas View Post
                      Is there music that uses a center channel? I thought it was bandwidth limited and intended for dialog.
                      Concerts on dvd are often 5.1.

                      Comment


                      • fpitas
                        fpitas commented
                        Editing a comment
                        OK. Didn't know that.

                    • #14
                      I'm about 95%+ movies and video game. The center channel speaker is by far what I hear the most. Followed by my surrounds.

                      Comment


                      • #15
                        Measurements will get you in the ballpark, but then ideally, you make changes to the response, and find a response that sounds clear, and natural. I like "the Bucket List" for voicing.

                        Comment

                        Working...
                        X