Announcement

Collapse
1 of 2 < >

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
2 of 2 < >

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

Filter help needed for small single driver speaker.

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

  • Filter help needed for small single driver speaker.

    Hi Gentlemen,

    Now that I have the ability to measure frequency response and a little bit more knowledge about building crossovers, I am revisiting a few of my abandoned speaker failures.
    With the help of Chris R's filter, my single driver Mini-Me clones are sounding quite nice. And, as a result of some excellent Tech Talk group input, so are my GRS "el cheapos." Flushed with relative success, I decided to see if I could salvage my attempt to build a small 2 way using the Techtonic Elements 3.5 inch BMR in conjunction with the discontinued Hi Wave BMR12 2 inch extended range driver which has a listed FR of 150 to 20K hz. Well, after many attempts in the past couple of days, I'm thinking that these drivers simply won't play well with each other. However, the 3.5 inch driver does cover a wide frequency range and sounds at least as good as my best effort crossing it over to the little 2 inch BMR.
    I have built a couple of pairs of the single driver BMR speakers which I have given to friends and have a pair of pretty nice walnut veneer cabinets with solid walnut tops from which I swiped the 3.5" BMR drivers to try my two way BMR experiment. So, I think my best bet to salvage some success from all this is to go with a filter to get as much as possible out of the single 3.5 inch driver.
    I am tempted to build the Mariposas, but if I can just use my existing cabinets, my last pair of 3.5 inch BMRs and set the HiWave 2 inch speakers aside for another day I'll consider this project done.
    A long introduction to my request perhaps, but it leads to my hope that, with my measured, in cabinet, FR plot of the TE BMR, some of the crossover gurus can suggest a filter that will smooth out the response.
    The attached photos show my BMR two way failure under test, the in cabinet measured response of the TE 3.5" driver, with no filter, and the actual single driver speaker that was measured.
    As usual, this being an econo-build, I hope to keep the filter as cheap and simple as possible. I feel that I'm asking for a lot, but also have learned that a lot of folks here like to help and like a challenge so hence this post.

    Best,
    Jay

  • #2
    Try a .6mH coil with a 10 ohm resistor bypassing the coil. (Coil, and resistor in parallel, )

    The resistor will tend to get hot, if much power is supplied. I'd keep it where it can be checked to make sure it's not going to get hot enough to catch on fire. A more sophisticated filter might eliminate that problem. You can see how it measures, and sounds anyway. If it gets warm at all, I'd not take a chance with it. At low frequencies, most of the current should flow through the coil, but at higher frequencies, about half will flow through the resistor. (I think.) The speaker is not very efficient, so it may need 30w to get loud. If so, about 15w may be going through the resistor.

    Comment


    • #3
      Here is the comparison of the unfiltered (purple) ,6mH plus 10 ohm R (gold) and .6mH plus 5 ohm (red) frequency responses. To me the best seems to be the 6mH plus 5R which looks better than the unfiltered, but not as good as I would hope for. The BMR has an inherent big dip around 1700 hz which I don't think can be fixed, but the big depression between 1800 and 4000hz looks like a significant problem. Can that be helped?

      Thanks!
      Jay

      Comment


      • #4
        Think of a notch filter as 2, 6dB per octave 1st order slopes and the resistor
        to fill in the gap they create. You need baffle step compensation
        at say 500Hz, add 1st order slope to get it down 6dB at 1K.
        (6dB slope set at 500Hz entering the impedance of the driver).
        This is where it is ok to use text book filters to come up with
        a choke value. Then just a example frequency would be 10K
        for the cap. That will bring everything down 6dB per octave
        from 10k to 5k and 500Hz to 1kHz.
        (picture a Van Halen EQ with the high points at
        500Hz and 10kHz, the lowest slider of the audio control
        eq being around 2.5k around 12dB lower than 500Hz
        and 10kHz.

        The resistor will bring up the 2.5kHz region the most and the freq.
        around it a little less. Remember the
        resistor Wattage is calculated with dissipation, in other words
        a 1 ohm resistor will handle 10x more real world power than
        a 10 ohm.
        Guess xmax's age.

        My guess: 15. His grammar is passable. His trolling is good.

        Comment


        • #5
          You need a filter (or EQ) that acts across 3 different pass-bands. You need BSC below 1000Hz. You'd like to leave the 1.5-3.5kHz band alone. Then you want to shelve the region above 4k.

          Here're 2 LCR filters (ea. w/3 components in parallel) in a string (series) that'll pretty much do what you need, problem being that they co$t WAY more than your, what, $15 driver?
          As has been mentioned, the LC components set the freq. range, and the parallel resistor controls the "depth" of the notch.
          1st) 1.0mH coil, w/16uF cap and 20n(ohm) resistor in parallel. BSC (mostly)
          2nd) 0.40mH coil, w/a 1uF cap and 20n resistor in parallel. (attenuates top end)

          For desktop (computer) use (like iTunes), I'd just try to do this w/the "Equalizer". Basically the xfer fn looks like this:
          Below 100Hz - not much goin' on
          100 to 1-1.2k, gradual drop of up to -10dB or so
          (between 1.5 to 3.5-4k only about -4dB?)
          then a "soft dish" reaching a min of -9dB @ 8k, but only -7dB around 5k and again at 12k

          This plot shows YOUR curve (dim line - adjusted SPL to near "spec"), filtered SPL (lower heavy line), and xfer fn (upper heavy line)
          (I used the "factory" Z-curve to fuel the sim, as you didn't offer one.)
          Click image for larger version

Name:	DrJxferNspl.jpg
Views:	1
Size:	113.1 KB
ID:	1399536

          Comment


          • #6
            Originally posted by Drjay View Post
            Here is the comparison of the unfiltered (purple) ,6mH plus 10 ohm R (gold) and .6mH plus 5 ohm (red) frequency responses. To me the best seems to be the 6mH plus 5R which looks better than the unfiltered, but not as good as I would hope for. The BMR has an inherent big dip around 1700 hz which I don't think can be fixed, but the big depression between 1800 and 4000hz looks like a significant problem. Can that be helped?

            Thanks!
            Jay
            You could add a notch just above 1k, but the coil might cost more than the driver. I played with PCD a little, and it looks like 1mH, and 16uf would be close to the values needed to knock down the mid-range spike. This might be what Chris was suggesting as well, but maybe not. I'd try the notch with a 1 or 2 ohms resistor, and see the results. The driver has a spike in its impedance in that area, so I suspect that the notch will need to be fine tuned to get it to work well. A small change in values may make big changes in the response.

            It looks like the 2" driver is decent. Why not cross it to the 3.5" at about 400hz?
            Last edited by rpb; 01-05-2019, 10:23 AM.

            Comment


            • #7
              Thank you very much Gents! I guess the bottom line is that, to my ears, the TE 3.5' BMR driver by itself is never going to provide sound that would meet even "mid fi" standards except perhaps with some fancy DSP equalization. Because I had all the components on hand ( including a bunch of buyout 1 mH inductors at $1.00 each!) I put together the two LCR filter suggested by Chris in post #5, just for grins. The measured FR I got compared to the unfiltered is attached and looks a lot like Chris's sim above. I played a variety of music and it sounded better than the unfiltered version, but still not good enough that I would want to build anything along these lines in the future. Even though I'm sure the Mariposas and John H's 2 ways which use this driver sound good, for me, other driver options look better for future builds.
              Again, thanks for the interest and input. I hope someday I'll get to the point that I can partially repay the favors.

              Jay

              Comment


              • #8
                Hi again,

                RPB's post came in while I was posting #7 above, hence the disconnect. I have tested the 2 inch BMR by itself, in cabinet, and it looked pretty ragged to me, but I'm having fun and would be glad to revisit it if there is some hope for something worthwhile. I was going ti throw the test cabinet for the 2 drivers together in the trash tomorrow, but I'll hang on to it a bit longer.
                We have out of town friends arriving soon, but I'll post my in cabinet FR of the 2" that discouraged me from crossing it low later today or tomorrow.

                Thanks for the continued interest.

                Jay

                Comment


                • #9
                  Originally posted by Drjay View Post
                  Hi again,

                  RPB's post came in while I was posting #7 above, hence the disconnect. I have tested the 2 inch BMR by itself, in cabinet, and it looked pretty ragged to me, but I'm having fun and would be glad to revisit it if there is some hope for something worthwhile. I was going ti throw the test cabinet for the 2 drivers together in the trash tomorrow, but I'll hang on to it a bit longer.
                  We have out of town friends arriving soon, but I'll post my in cabinet FR of the 2" that discouraged me from crossing it low later today or tomorrow.

                  Thanks for the continued interest.

                  Jay
                  Did you try the deeper notch with 1 or 2 ohms? On the notch, try 2mH and 8uf, and compare to the 1mH , 16uf notch that I suggested earlier.

                  You can even take it a step further, and try 4mH, and 4uf. Too costly to actually consider, but as an experiment, you may find it interesting.
                  Last edited by rpb; 01-05-2019, 02:47 PM.

                  Comment


                  • #10
                    Originally posted by Drjay View Post
                    Thank you very much Gents! I guess the bottom line is that, to my ears, the TE 3.5' BMR driver by itself is never going to provide sound that would meet even "mid fi" standards except perhaps with some fancy DSP equalization. Because I had all the components on hand ( including a bunch of buyout 1 mH inductors at $1.00 each!) I put together the two LCR filter suggested by Chris in post #5, just for grins. The measured FR I got compared to the unfiltered is attached and looks a lot like Chris's sim above. I played a variety of music and it sounded better than the unfiltered version, but still not good enough that I would want to build anything along these lines in the future. Even though I'm sure the Mariposas and John H's 2 ways which use this driver sound good, for me, other driver options look better for future builds.
                    Again, thanks for the interest and input. I hope someday I'll get to the point that I can partially repay the favors.

                    Jay
                    i've used the same 3.5's in two builds now, which i have been happy with, Chris provided some good advice for me with the BSC/filter for my boom box and i also was going to try his latest suggestion with my T Line build, although i have a feeling like trying the tweeter from the marapossa's but.....trying to revisit the T Line.

                    Looking forward to seeing where all your pieces fall at the end and what journey you decide to take.

                    Comment


                    • #11
                      Here is the in cabinet FR for the discontinued 2 inch square TE BMR driver. I think you can see why I thought integrating it with the 3.5 inch was pretty hopeless. I didn't measure it below 1.5 K for fear of damaging it, but don't see it as good for anything with even the least pretense of good fidelity. I think the new, round, TE BMR might be a lot better, perhaps that one is what rpb was referencing.
                      I will try the deeper notch filters suggested later this week, but the weather is up to middle 70s, the winds are light and my secret lake, which holds some double digit bass, is calling me for the next few days...

                      Comment


                      • #12
                        Learn what you can experimenting, and move on to better drivers. In looking at the speaker picture, it looks like the tweeter is equal distance from the top, and sides. That can cause some diffraction ripples. If the box is disposable at this point, maybe you could alter the top, or side edges to smooth things out. (Just to see if it does help. Then toss em.)

                        Comment

                        Working...
                        X