Announcement

Collapse
No announcement yet.

Cheap 8-inch two-way to try

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

  • Cheap 8-inch two-way to try

    I recently started building a subwoofer for my son’s college house. He was sent home from college due to the COVID-19 outbreak, and we started talking about what to do for his main speakers. Too much has been invested in the subwoofer, and he needs to go with something inexpensive. I started doing some design work using the Dayton DC200-8 and DC28F-8. I have used this tweeter before, and it is quite nice for the low price. I have not used the woofer. While it is not ideal above about 1kHz, it can be made to work for an inexpensive system.

    Normally, I would set up measurement gear and perform the design based on actual measurements of the purchased drivers. I have not purchased the drivers yet, and I am not likely to put that much work into this inexpensive system. In this case, I downloaded the stock files from Parts Express and converted them to minimum phase. I used the resulting frequency response and impedance files to model the system in PCD-8. I assumed the acoustic center of the woofer is 32mm behind that of the tweeter based on previous work I did using the reference series drivers. Again, this is not very precise, but this is not s critical listening application. Also, +/- 5mm or even 10mm does not make a huge difference in this design. I selected a crossover with approximately 4th order LR acoustic response at approximately 1.8kHz. The filters are asymmetric to help align the phase in the crossover region. The crossovers were designed to provide slightly decreasing amplitude with increasing frequency. These will be installed in a sealed enclosure with Q of approximately 0.8 to 0.9. This will give it another slight rise in amplitude below 300Hz.

    I have attached images of the WinISD frequency response simulation, crossover circuit, and system frequency response. It is worth noting that the system frequency response is generated from the stock files and does not include the effects of the enclosure. The actual system response will increase slightly below 300Hz. These will always be positioned near boundaries and will not need a great deal of baffle step compensation.
    At first glance, I would normally think the crossover is a little expensive for an overall inexpensive system. In this case, the extra expense helps to compensate for the woofer’s weaknesses. These speakers will have a difficult life, and some of the drivers will likely need to be replaced before their time. It’s worth a few extra bucks to get palatable performance and still have cheap driver replacements.

    Since most of the world has been locked up in their houses trying not to catch the Coronavirus, I thought I would put this information out there. Maybe somebody out there is itching to build a cheap two way to pass the time.
    Attached Files

  • #2
    What are the target frequencies for the two notch filters in the woofer crossover?

    Comment


    • #3
      Originally posted by Brij View Post
      What are the target frequencies for the two notch filters in the woofer crossover?
      The woofer has a really nasty peak at 2kHz and an unpleasant dip at 1kHz. The filter reduces the peak, but you kind of have to live with the dip. I have attached an image that shows the raw woofer response, the total filter response, the target response, and the actual response of the woofer and filter combo.
      Attached Files

      Comment


      • #4
        Originally posted by Brij View Post
        What are the target frequencies for the two notch filters in the woofer crossover?
        Sorry, I missed part of your question. There is only one notch fiter to get the peak at 2kHz. The other is actually an impedance equalizer to improve crossover filter performance by flattening the impedance rize above the crossover frequency.

        Comment


        • #5

          Here's an old PETT thread which you might find useful:

          http://techtalk.parts-express.com/fo...28f-8?t=209087

          HTH

          Geoff

          Comment


          • #6
            Originally posted by Geoff Millar View Post
            Here's an old PETT thread which you might find useful:

            http://techtalk.parts-express.com/fo...28f-8?t=209087

            HTH

            Geoff
            Thanks for sharing this thread. I didn't really look before I made the original post, but I'm sure there are many good designes here that use the same drivers. I had originally considered using the Noah 8 crossover. When I modeled it, I ended up making some changes to lower the peak at 2kHz and center the main lobe by adjusting the phase at the crossover frequency. Mostly, I was entertaining myself during the lockdown.

            Comment


            • #7
              I only see 1 notch, around the 2kHz peak.

              pick, try this (saves 3 elements) ...
              LP: RLC notch = 40n + 0.60mH + 12uF, and a 10n+8uF Zobel
              HP: 4n series resistor, 5uF series cap, 0.30mH shunt coil

              Comment


              • #8
                Originally posted by Chris Roemer View Post
                I only see 1 notch, around the 2kHz peak.

                pick, try this (saves 3 elements) ...
                LP: RLC notch = 40n + 0.60mH + 12uF, and a 10n+8uF Zobel
                HP: 4n series resistor, 5uF series cap, 0.30mH shunt coil
                Chris, thanks for the input. I have been away for a while but just modeled this last night. The natural roll-off of the woofer is pretty close to 4th order LR except for the anomaly at 2kHz. Using a simple notch in combination with the Zobel is a nice, simple solution. I'm actually a little upset I didn't think to do that. This approach also reduces the cost significantly. I plan to use this LP filter as is. I have attached the frequency response plot. I made some tweaks to the HP filter to flatten the tweeter response. Phase matching is about the same. I will post this version separately.
                Attached Files

                Comment


                • #9
                  Here is the version I plan to run with.
                  Attached Files

                  Comment


                  • #10
                    If you're wanting to be really cheap and have the time, unwind an electric motor and roll your own coils. Some washing machine motors have ~14 gauge windings.
                    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


                    • #11
                      Your HP filter (post #9) looks accurate w/the "DC28F-8" F/Z files, but your FR (from post #8) using MY rec. filter does NOT.

                      And now I think I see WHY! My 4n(ohm) series resistor is on the amp side of the shunt coil - but YOU simmed it on the back end (same place you put your 8n resistor).
                      TOTally diff. FR curve. (So, you really can't do that.). Putting it on the wrong end (in back) adds about a +4dB hump @ 5kHz or so.

                      Comment

                      Working...
                      X