Announcement

Collapse
No announcement yet.

BMR Box Tuning not matching Simulation

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

  • BMR Box Tuning not matching Simulation

    I must be missing something or making a mistake - I have a box that's about 4L net with the TEBM65C20F-8 in it, and the box tuning isn't really close to the WinISD sim. I'm using 1.25" PVC (ID = 1.38") with an elbow, and I thought I'd need a port of +-6.125" to tune the box to about 63Hz. I tried with simply the PVC connected to the baffle and elbow ( 5" or so plus a 1/2" roundover) and box tuning was ~80Hz, so I went to longer ports than planned.
    a port that's 7.25" tuned the thing to about 67Hz (green line)
    8.25" was about 65Hz (purple line)
    9.25" got me to 63Hz (blue line)

    9.25" tunes the box to what I want, but it doesn't make sense to me, and the low first impedance peak just seems odd - I've seen mismatched peaks, but never quite this different - although I've never tuned a box this far below Fs before.

    Why is there such a mismatch between the sim and reality? According to WinISD, that box should be tuned at about 52-53 Hz. I am running the wires through the port, but I can't imagine that's significant given the small wire cross section.

    I measured the driver (minus vas- lazy) but the other parameters were remarkably close to the manufacturer and the original sim, so I don't think it's driver variation.

    Does it make sense to just move forward with the empirically determined tuning?





  • #2
    The impedance is what you're going to get tuning well below Fs. Change the tuning to 104Hz and you'll see the peaks are the same. Using an elbow probably explains why the port needs to be longer.
    www.billfitzmaurice.com
    www.billfitzmaurice.info/forum

    Comment


    • #3
      I get 7.22" for 63hz. tuning. Figuring that driver is so small the rebate on the rear of the baffle probably makes up the difference in physical driver displacement.

      Click image for larger version

Name:	boxtning63hz.JPG
Views:	135
Size:	90.1 KB
ID:	1439231 Click image for larger version

Name:	portvel1.38.JPG
Views:	125
Size:	59.6 KB
ID:	1439232
      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


      • #4
        You shouldn't count port-roundover/flair sections as port lengths, so that might account for a 0.5"-1.0" discrepancy if you've been measuring the port end-to-end.
        My first 2way build

        Comment


        • #5
          Oddly enough, the actual driver has nothing to do w/the box tuning. No matter what driver you use (given identical Vbs and vents), the Fb will be the same.
          OTOH, your DATS IS showing you your actual Fb (at the dip).
          Any discrepancies could be due to the roundover (which should not be included in the port lenth - for ISD, as lout mentioned), or the elbow (your port length should be what you'd get if you could suspend a paper ruler (flexible) right down the middle of the port (through the elbow) w/out counting the additional roundover), or the proximity of the vent's interior opening to other internal driver/box parts (like the walls).

          Does your DATS curve correlate w/ISD's predicted Z-plot?
          For you Win model, are you plugging in your DATS parms?
          Lastly, how does it sound (to you)? That's really what you should go with.

          Comment


          • #6
            Might want to check for leaks.
            John H

            Synergy Horn, SLS-85, BMR-3L, Mini-TL, BR-2, Titan OB, B452, Udique, Vultus, Latus1, Seriatim, Aperivox,Pencil Tower

            Comment


            • #7
              Thanks for all of the responses!

              I'm not including the roundover in the length of the port - so the straight section of the port is the lengths I listed. The length of the elbow is measured as if a paper tape were placed through the center of the elbow.

              There's certainly leaks. Since the box is so small, I don't have the the second section of the port glued in - just press fit (I measured the length past the end of the elbow) . Similarly, I don't have all of the screws on the driver in. That could account for some of the discrepancies. The port is about a port diameter from the back wall, which I thought was the rule of thumb from obstructions, but who knows, that could also come into play.

              There's a difference between WinISD and Unibox in one of the default leakage parameters if I remember, so that probably accounts for the difference between WinISD and Unibox.

              The impedance doesn't quite match WinISD's z plot - the lower peak was higher in winISD.

              And I was thinking about the elbow. Since the PVC doesn't meet in the middle, there's a bit of a lip in the elbow, meaning there's a bit more air in the port air slug, which probably adds a little more to the port length again.

              Sounds like the moral of this story is trust DATS.

              Chris - I didn't listen to any music last night. My intent was to put the port tube in, confirm the tuning, run a nearfield measurement of the driver and port, splice them together to confirm sim matches reality again, and listen to the tuning but the tuning discrepancy led to me messing with different lengths of PVC for hours. There should be music tonight. Followed by ...messing with different lengths of PVC for hours to see what sounds best.



              Comment


              • #8
                Click image for larger version

Name:	NS3leakyVb.jpg
Views:	100
Size:	179.3 KB
ID:	1439281
                Here's an AURA NS3 (like Dayton's ND90 series) in a 0.15cf ported box. The green curve is what I got after I did a better job at sealing the 3" driver's (flimsy) frame.
                The blue curve was my 1st mounting attempt - no special care taken. NO changes to the box or the port (tuned to low 50s - per WinISD) between the 2 z-runs, just the seal on the driver/box.

                The 1st (leaky) attempt LOOKs more like a leaky (closed) box - you don't really see the "signature" valley that a vented box shows - maybe just a "hint" of it starting to form. I suppose there's a whole spectrum of leakyness plots you could end up with.

                Comment


                • #9
                  It was the box leaks.

                  I fully tightened the terminal cup, tweeter, and woofer screws (and used all of them) and it made a drastic difference. I've got a +-7.625" port in it now and it's tuned at 57Hz, which I"m going to bring up a little at some point to see how it sounds. I actually got caught up listening to just the BMR with no tweeter - that thing is incredible for the cost. I ended up pushing it a bit and it didn't complain as long as I kept the voltage under about 8-9 volts, which was plenty loud. The slightly undersized front port wasn't objectionable either - I guess the roundover helped. I got every bit as much joy out of listening to this thing for a while as I have any of the more expensive drivers I've had. I'm really excited for this project - it's been a while since I've done a low-cost build and the BMR seems to have a ton of potential. That off axis too...

                  Comment


                  • #10
                    If you do a Google search for: Filter help forTechtonic Elements 3.5" in single driver speaker, please. it will take you to a post where in response #6 John Hollander suggests a filter to greatly flatten the frequency response of the BMR when used as a single driver. As I posted in a follow up response, it works extremely well juat as he posted it. Might be worth a try.

                    Comment


                    • #11
                      I was surprised how good it sounded full range, but I'm going to cross it to a 3/4" tweeter. Part of the reason I'm doing this is experimentation on directionality. I had plans on crossing it just under 4K, CTC is ~3.5", but the sim's I'm running just keep gravitating to about 3.3k. The tweeter, SB19, should be able to handle it.

                      With that said, I tried to a ton of variants on the notch filter to tame the peak just under the suckout, but ended up back at John's.

                      Comment

                      Working...
                      X