No announcement yet.

Dayton DS90-8 not measuring up

  • Filter
  • Time
  • Show
Clear All
new posts

  • Dayton DS90-8 not measuring up

    I have a problem with the portable Bluetooth speaker I am building at the moment. I have measured the speakers with ground plane and near field and far field measurements and that endeavour has been documented in another thread. However, I feel the problem I am having now is too far removed from the initial problems I have documented in the other thread and that they warrant a new one.

    To describe my problem. I have build a Bluetooth speaker using the Dayton DS90-8 and the Dayton ND16FA. From my measured responses I have found a quite large discrepancy from my simulated response for the DS90-8 in VituixCAD and my actual measurements.

    I have made a bass reflex enclosure, and to add more confusion Parts Express recommends two different enclosure volumes for the DS90-8. One of the recommended volumes is 3.39 l (0.12 ft3) and a port of diameter 2.54 cm (1 in) and length of 15.2 cm (6 in). This is supposed to give an f3 of 73 hz. On the other hand, a volume of 1.41 l (0.05 ft3) with a f3 of 80 hz is also recommended. Port diameter and length is not specified for this enclosure volume.

    My Bluetooth speakers is made of two separate chambers with a volume of 1.56 l (0.055 ft3) and with a port each with a diameter of 2.4 cm (0.95 in) and length 11 cm (4.3 in). The simulated response of this enclose gives is fb = 82 hz and f3 = 80 hz. However, my measured response of the driver in this enclosure gives fb = 113 hz and f3 = 115 hz.

    This is far from the simulated response. So I though that since the product information page gave two different recommended enclosure volumes, then perhaps the T/S parameters on the product page was wrong and the driver had been updated without updating the information. So I measured the T/S parameters of the DS90-8, but they were quite similar to the product specs. Of note is that I have not broken in the driver.

    Parameter Manufacturer Measured
    Sensitivity 84.1 dB @ 2.83V/1m 82.9 @ dB 1W/1m
    DC Resistance (Re) 6.3 ohms 7.35 ohms
    Voice Coil Inductance (Le) 0.56 mH @ 1000 hz 0.312 mH @ 2643 hz
    Resonant Frequency (Fs) 81.4 Hz 79.2 hz
    Mechanical Q (Qms) 2.19 2.78
    Electromagnetic Q (Qes) 0.53 0.58
    Total Q (Qts) 0.43 0.48
    Diaphragm Mass Inc. Airload (Mms) 3.8 g 3.96 g
    Mechanical Compliance of Suspension (Cms) 1.02 mm/N 1.018 mm/N
    Surface Area Of Cone (Sd) 31.2 cm² 31.2 cm²
    BL Product (BL) 4.76 Tm 5.01 Tm
    Compliance Equivalent Volume (Vas) 1.31 l 1.43 l
    Driver's Mechanical Losses (Rms) -- 0.709 kg/s

    Here are some pictures of the enclosure I have build from the front and the back. Each chamber is 133 mm (5.2 in) W, 95 mm (3,7 in) H, 133 mm (5.2 in) D.

    Click image for larger version

Name:	image_78999.jpg
Views:	2
Size:	80.3 KB
ID:	1409511Click image for larger version

Name:	image_79000.jpg
Views:	2
Size:	85.8 KB
ID:	1409512

    Here is a graph of the simulated and measured impedance of the DS90-8.

    Click image for larger version

Name:	Dayton_Audio_DS90-8_Impedance.png
Views:	2
Size:	25.0 KB
ID:	1409514Click image for larger version

Name:	impedance_ds90-8_ported.jpg
Views:	1
Size:	57.8 KB
ID:	1409513

    The discrepancy between the simulated and the measured responses I find very frustrating. For the past year I have been reading up on the theory behind loudspeaker construction. It is one of the hardest subjects I have tried to teach myself. And it is made even more frustrating if I can't trust my simulations to give me a design which will perform similar to my simulations in real life. Any input or suggestions to solve this is much appreciated. If you need ny more info I will gladly supply it if you ask me in a post.

  • #2
    This IS frustrating !

    Your box/port/tuning is independent of the driver. Is your port "clear" (to breathe) on the inside end (near the back)? Can't really tell from your pics.
    Since your box SEEMs to be tuned 30Hz too high, I'd like you to tune it 30Hz lower.
    While I doubt you could make your port tubes longer, what you CAN do (temporarily) is keep them the same length but make them a smaller I.D. (15mm ought to to it).
    If you can get some kind of (closed cell - like poly foam?) "liner" in them to change the tuning, I'd like to see your measured Fb THEN?

    Something to note is that the rolloff diff. between 53 / 83 / 113 Hz won't be THAT different. The HIgh tuning DOES add +1 to +1-1/2dB "bump" near 200Hz, but . . .


    • #3
      Looks like you have cabinet leakage.
      Don't listen to me - I have not sold any $150,000 speakers.


      • #4
        It looks to me like your baffle is flexing and damping the woofers resonance. Try firmly pressing between the woofers while running an impedance sweep, if that gives you more of a normal double peak response like on your Sim then you just need add some screws to support the center of the baffle.


        • #5
          I would change the port to .75" pvc 2.9" long for about 80hz. I had a pair of these a few years ago and they just dont go very low. Future reference, try building a test box with removable front/rear baffles so you can test different port sizes and tunings. I have three different size enclosures that give me ballpark ideas before committing to builds.


          • #6
            (I wrote this between johnny and jim - but forgot to "Post" it !)

            maybe try shimming and clamping the center of the back down?
            that's a L - o - n - g stretch across there from one end to the other (across the mid divider


            • #7
              If the dip between the two peaks is not clearly defined and almost equal to the driver's Re at the lowest point, then the box is too lossy. Either there is a leak, or it's stuffed too much.
              Brian Steele


              • #8
                Thank you so much for your suggestions. I now have an impedance measurement that looks similar to what I simulated in VituixCAD.

                Click image for larger version

Name:	impedance_ds90-8_ported.jpg
Views:	1
Size:	58.4 KB
ID:	1409873

                I had a leaky cabinet, that's for sure. When I have built my previous enclosure I have glued everything up. This time I wanted to make the baffle and back removable to make working on the design easier. This was also the source of my troubles. I used gasket tape to seal the front and back to the cabinet, but I did not drill the holes large enough in the little wood squares used to secure the panels. The screw did not have the space to turn and pull the panels tight against the gasket tape.

                I also added some more more wood squares to secure the middle of both the front and the back panel.


                • #9
                  Physics wins !
                  (congrats - do they sound any better?)


                  • #10
                    Originally posted by Chris Roemer View Post
                    Physics wins !
                    (congrats - do they sound any better?)
                    Hard to say from just listening to them without a crossover. I will do some frequency responses tomorrow and that will probably tell if they at least have more bass.


                    • #11
                      I measured the frequency response of DS90-8 today. The graph now gives me the correct f3 response of 79 Hz. So everything turned out as it should in the end. The response is flatter without a peak at 200 Hz and I got a bit more bass.

                      Click image for larger version

Name:	bark_tm_FR.jpg
Views:	1
Size:	65.0 KB
ID:	1409919