Announcement

Collapse
No announcement yet.

First swing @ a big passive 3-way. Input needed.

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

  • #16
    I think your sims may contain 2 key errors.

    First, it doesn't look like your frd files include baffle diffraction. You have said they do but if they did, your final SPL on the sub should be closer to 85dB if you include the full 6dB of baffle step loss and closer to about 88dB if you only include about 3dB of loss. Your sims are still sitting up at about 90/91dB. Baffle diffraction effects should also be included with the tweeter and mid too.

    Second, you need to include the driver delay (found in the 'Tune' menu) in XSim for the phase of the drivers to be correct. In XSim, this is a combination of the difference in acoustic centers between the drivers and the difference in path length to the listener due to the different placement of each driver on the baffle.

    So for baffle step loss with your sub, being so large and therefore needing a rather wide front baffle, I might guess about 3-4dB of baffle step loss might be what you need to shoot for. J Bagby's Frequency Response Modeler - http://audio.claub.net/software/jbabgy/FRM.html - will do the diffraction sims for you and also allow you to alter the amount of baffle step loss. Model your sub in the diffraction section at the bottom, add it to the Register above, find the 'Baffle Step: Simple Step or from Below' up near the top left corner, turn the 'Off' button to 'Inverted', and then change either the '% Step' or the 'Step dB' amount to suit. But it's an Excel spreadsheet, so you'll need MS Office to start off with. Response Modeler will also add the diffraction to your original FR curves, splice in the box curves, modify your impedance curves and extract minimum phase from both your frd and zma files which are all necessary steps in the process.

    The acoustic center (AC) of a driver is more or less the physical point where sound waves originate from on the driver and it is hard to guestimate this right. Generally, it's at about the point where the cone meets the voice coil. For the purpose of sims, you basically have to just physically measure this point based on the manufacturer's drawings. Save the specs to a pdf file and then zoom it out to life size. Then physically measure on your screen from the front edge to the cone/coil intersection, which is often usually where the spider sits too. To actually get the real AC points, you need to measure the drivers on baffle with a mic.

    So with a tweeter, you might use the depth of the face plate as the AC point, but with a waveguide included on your selection, it's going to be further back than that. Looking at the drawing for the Wavecor, I'd probably guess it's somewhere between 25-30mm back. Do the same with the other 2 drivers. Now what you input into XSim is the difference in path length from the listening position to the tweeter and then from the listening position to the mid and to the woofer which are just simple a(squared) = b(squared) + c(squared) equations. Choose a listening point 1m out from the tweeter on axis. This plus the 25-30mm is the distance to the tweeter AC, call it 1028mm. Now draw 2 right angle triangles starting from that same 1m point on the tweeter axis that go to the mid's AC and that go to the woofer's AC and then calculate the respective distances from that same listening point to each driver's AC, which is the length of the triangles' hypotenuse. Now subtract the distance to the tweeter from the distance to each of the other 2 drivers.

    Now in XSim, leave the tweeter delay at 0 but depending on where you place them on the baffle, I expect the extra distance to the mid to be small, perhaps about 1" but the extra distance to the woofer to be fairly large, and I'm just guessing here, maybe between 4-6". Finally, in the FR chart under 'Curves', turn off the System phase and turn on each of the drivers' phase. Now what you are trying to do is get a flat FR while at the same time getting the driver phases to align in the 2 xo regions. Hopefully that should get get you set up more accurately now to start off with.

    Comment


    • #17
      Originally posted by mobius View Post
      I think your sims may contain 2 key errors.

      First, it doesn't look like your frd files include baffle diffraction. You have said they do but if they did, your final SPL on the sub should be closer to 85dB if you include the full 6dB of baffle step loss and closer to about 88dB if you only include about 3dB of loss. Your sims are still sitting up at about 90/91dB. Baffle diffraction effects should also be included with the tweeter and mid too.

      Second, you need to include the driver delay (found in the 'Tune' menu) in XSim for the phase of the drivers to be correct. In XSim, this is a combination of the difference in acoustic centers between the drivers and the difference in path length to the listener due to the different placement of each driver on the baffle.

      So for baffle step loss with your sub, being so large and therefore needing a rather wide front baffle, I might guess about 3-4dB of baffle step loss might be what you need to shoot for. J Bagby's Frequency Response Modeler - http://audio.claub.net/software/jbabgy/FRM.html - will do the diffraction sims for you and also allow you to alter the amount of baffle step loss. Model your sub in the diffraction section at the bottom, add it to the Register above, find the 'Baffle Step: Simple Step or from Below' up near the top left corner, turn the 'Off' button to 'Inverted', and then change either the '% Step' or the 'Step dB' amount to suit. But it's an Excel spreadsheet, so you'll need MS Office to start off with. Response Modeler will also add the diffraction to your original FR curves, splice in the box curves, modify your impedance curves and extract minimum phase from both your frd and zma files which are all necessary steps in the process.

      The acoustic center (AC) of a driver is more or less the physical point where sound waves originate from on the driver and it is hard to guestimate this right. Generally, it's at about the point where the cone meets the voice coil. For the purpose of sims, you basically have to just physically measure this point based on the manufacturer's drawings. Save the specs to a pdf file and then zoom it out to life size. Then physically measure on your screen from the front edge to the cone/coil intersection, which is often usually where the spider sits too. To actually get the real AC points, you need to measure the drivers on baffle with a mic.

      So with a tweeter, you might use the depth of the face plate as the AC point, but with a waveguide included on your selection, it's going to be further back than that. Looking at the drawing for the Wavecor, I'd probably guess it's somewhere between 25-30mm back. Do the same with the other 2 drivers. Now what you input into XSim is the difference in path length from the listening position to the tweeter and then from the listening position to the mid and to the woofer which are just simple a(squared) = b(squared) + c(squared) equations. Choose a listening point 1m out from the tweeter on axis. This plus the 25-30mm is the distance to the tweeter AC, call it 1028mm. Now draw 2 right angle triangles starting from that same 1m point on the tweeter axis that go to the mid's AC and that go to the woofer's AC and then calculate the respective distances from that same listening point to each driver's AC, which is the length of the triangles' hypotenuse. Now subtract the distance to the tweeter from the distance to each of the other 2 drivers.

      Now in XSim, leave the tweeter delay at 0 but depending on where you place them on the baffle, I expect the extra distance to the mid to be small, perhaps about 1" but the extra distance to the woofer to be fairly large, and I'm just guessing here, maybe between 4-6". Finally, in the FR chart under 'Curves', turn off the System phase and turn on each of the drivers' phase. Now what you are trying to do is get a flat FR while at the same time getting the driver phases to align in the 2 xo regions. Hopefully that should get get you set up more accurately now to start off with.
      I think maybe I wasn't clear in my statement about the diffraction being included. What I meant was I tried it but it looked wrong so I took it out. Then I modeled the frd. and .zma files as is to get a flat response. The driver offset (depth estimated) is included in the model I have posted. Now that I have that kind of close I am working on BSC filters and adding them and adjusting. I will post that as soon as I have something I think may be close.
      BSME
      Kannapolis NC

      Comment


      • #18
        Here is the next iteration. I have changed a couple of things.

        1. I changed the midrange from a 150P to a 180P. I did this because there was some notches that I had a tough time working out in the upper crossover region with the 150P. The 180P frd seemed a bit smoother in that region.

        2. I included a 4db baffle step in the model. The baffle is 18 inches wide which correlates to a frequency of about 250hz. I thought with a 15" driver 6 db may be too much. 3db still pushed the tweeter a bit hard @ 200 watts so I settled on 4db.

        This is all trial and error for me anyway so I am sure I will be making adjustments when I build and measure.

        Images below. As always, If errors are noticed or your experience is different than my assumptions don't hesitate to let me know.

        X-Over
        Click image for larger version  Name:	XOVER 12.PNG Views:	1 Size:	444.8 KB ID:	1384626



        Freq/Z
        Click image for larger version  Name:	XOVER 12 2.PNG Views:	1 Size:	385.1 KB ID:	1384627


        Power
        Click image for larger version  Name:	power 12.PNG Views:	1 Size:	457.1 KB ID:	1384628


        Rough- No bracing or mid chamber yet. Just the baffle intent. Vol is about 80l. I need a net between 60-70l for the 390.

        Click image for larger version  Name:	rough dim.PNG Views:	1 Size:	374.2 KB ID:	1384629
        BSME
        Kannapolis NC

        Comment


        • #19
          You should not expect your measured response to be this smooth. Part of the baffle response is going to include edge diffraction. IMO it's best to add the baffle and edge diffraction to each FRD for this type of sim, but move on to measuring as that's another learning curve.
          John H

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

          Comment


          • #20
            Originally posted by jhollander View Post
            You should not expect your measured response to be this smooth. Part of the baffle response is going to include edge diffraction. IMO it's best to add the baffle and edge diffraction to each FRD for this type of sim, but move on to measuring as that's another learning curve.
            Agreed. My next move is to build the test box and measure the drivers. I have no idea what I will see. By the way, really interesting work on the horn.

            Matthew
            BSME
            Kannapolis NC

            Comment


            • #21
              Have you established what the maximum SPL at a certain distance you want from these speakers because you are worrying me when you reference 400W for your mid and 200W for your tweeter which are well well beyond either of your chosen drivers' Power ratings?

              Also, besides adding in baffle diffraction to each driver as mentioned, you need to account for phase alignment between drivers in the xo regions. Again, turn off the system phase and turn on the phase for each driver. This is also something you need to do with the measured FR's so you might want to give it a shot now in your sims.

              Comment


              • #22
                Can someone please take a look @ this phase plot and help me understand it? I think I know what I am looking at but I am not sure how to read it. My understanding is that the goal with phase is to get the driver phase plots to cross at or near the crossover points. I have circled what I think are the phase cross points. I seem to have pretty good agreement with the lower cross point but the upper looks off but to be clear I am not sure what I am looking at. Is my interpretation correct?

                Click image for larger version

Name:	phase 12.JPG
Views:	1
Size:	522.4 KB
ID:	1384868
                BSME
                Kannapolis NC

                Comment


                • #23
                  I turn off the summed phase and the other driver's phase to see the alignment of the two drivers. Imo there's not much to be gained in messing with the phase at this point. Some look for crossing points others like aligning slopes and reverse nulls for even ordered x-o slopes.
                  John H

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

                  Comment


                  • #24
                    I was all set to chime in that I've used the RS150P for something very similar to what you're doing (RS225 + RS150P + RS28) and I love it. In fact, in a WMTMW center channel, it's a ferociously good mid driver. Easy to work with and sounds terrific.

                    But, then I noticed you switched to the RS180

                    Comment


                    • #25
                      Originally posted by 50 watt head View Post
                      I was all set to chime in that I've used the RS150P for something very similar to what you're doing (RS225 + RS150P + RS28) and I love it. In fact, in a WMTMW center channel, it's a ferociously good mid driver. Easy to work with and sounds terrific.

                      But, then I noticed you switched to the RS180
                      The 150 may be a better solution. I just changed to the 180 because the published frd file was a bit smoother up top. This doesn't tell me much about how the driver sounds. Once I get the baffle cut and do my own measurements with a test box the study will likely change. I have not counted the 150 out.
                      BSME
                      Kannapolis NC

                      Comment


                      • #26
                        Here's a filter sim that might give you an idea of how to deal with the RS150 peak. It is basically a 4th order lowpass with a cap in parallel with the first coil. I would probably add a 3 or 4 ohm resistor in series with the .68uf cap to lessen the effect of the notch just a little. PCD does not have a way of showing this in the sim. With a litle fine tuning, you could probably match the target curve at 6k if you wanted to.

                        Comment


                        • #27
                          From your last sim, in no particular order:

                          - the woofer phase looks wonky, something may not be right there
                          - you want to align phase in the xo zone, so you are correct for the woofer/mid but for the mid/tweeter, you should be looking at the phase between about 1400-2000Hz, not up around 5kHz. You could probably get them a little closer together but for the purposes of simulation, that's probably close enough.
                          - adjusting the mid and tweeter levels 4dB down is a step in the right direction but it isn't quite the right way to sim baffle diffraction effects. They should be added in to each of the driver's original frd's and then you extract minimum phase from those. Phase is a function of the FR, so when you get the FR wrong you are also getting the phase wrong.
                          - if you go with a larger mid, you may want to bring the woofer/mid xo point down lower, maybe closer to 300Hz or perhaps even a little lower than that when you consider the large moving mass that that sub is pushing around compared to your mid.

                          Comment

                          Working...
                          X