Announcement

Collapse
No announcement yet.

WinISD sim issue (Microbes by Roman)

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

  • WinISD sim issue (Microbes by Roman)

    Hi Guys,
    I was playing around trying to get some experience with the SW and ended up simulating a few speakers.
    Results are quite poor as I cannot get even close to the simulations done by the people who built them!

    Let's look at the Microbes by Roman for example:

    Here the official sim comparing RS125S-8 and RS125-8:

    Click image for larger version  Name:	Microbes.gif Views:	0 Size:	31.0 KB ID:	1463240

    Here my sim:

    Click image for larger version  Name:	Microbes SIM.JPG Views:	0 Size:	91.4 KB ID:	1463242

    Vol: 5.5L
    Tuning: 57Hz
    Vent Dia.: 3.49cm (1.375 in)
    Length Calculated: 13.49cm? (Should be 4.5in 11.43cm)

    I understand some misalignment in the measures, but this looks like a different driver completely!
    When I sim the "Stance" by Wolf il looks even worse and very different from what Wolf posted in the build tread, but in this case also the T/S are different (RS125-4 in this case)
    I got those results using PE CLIO data and various downloaded WinISD files of this driver and the result is always the same, also tried unibox same result again.
    Changing the Ql etc modifies the curve, but doesn't bring it any closer.

    Am I doing something wrong?

    Here the T/S I'm using (Dayton RS125-8)
    :
    Click image for larger version  Name:	Driver 1.JPG Views:	0 Size:	61.6 KB ID:	1463243

  • #2
    Sometimes the same driver will have it's T/S parameters shift a little over a production run, or Roman or Wolf may have used their own measured parameters, or included the effects of the crossover in the box sim. Also predicted vent length and the actual length required to achieve that tuning frequency vary. It looks like all 3 have an f3 in the mid 50's.

    Comment


    • #3
      Here's the "deal" - you're running WinISD well. Roman called his box "5L", but it WAS a little larger.
      Your "sag" (around 100Hz) indicates that a slightly smaller box might be in order (for YOUR data).
      What you call his "official" driver comparison was the (new - at the time - circa 10 yrs ago?) unshielded RS125 vs. the (orig.) RS125S.
      That driver has gone through at lest FOUR iterations - that I know of.

      These came from MY ISD DB:

      Qts 0.44, Fs 72, Vas 0.12cf <-- this "orig." RS125S sort of had "middle of the road" T/S values
      Qts 0.49, Fs 70, Vas 0.15 < - - "new" (unshielded) RS125-8, around 2010? (highest "Q" suggested a larger box)
      Qts 0.43, Fs 74, Vas 0.15 < - - 2013 "revision" same Vas, but a lower Q brought the box vol. back down (near Roman's)
      Qts 0.34, Fs 59, Vas 0.17 < - - 2019 (my newest data) the smallest Q, but also the lowest Fs (these "compliment" ea. other - AFA F3 goes,
      meaning that although their Fs values are 15Hz apart, their F3 values (in-box) are only more like 6Hz apart).
      It ALSO has the largest Vas - but that is offset by the lower Q.

      YOUR drivers might match any of these data sets, or NONE of them.

      Comment


      • #4
        But if that's the dispersion and the drivers could be so different what's the point of running sims?

        I understand that if you have chosen a driver, bought it and measured it properly you are good to go and hopefully you are simulating what you are going to get, but if you are choosing a driver, sims look pretty useless to me...unless I can use a measurement from a reliable source...for example PE CLIO. Are those up to date and reliable?

        Also what's the point of building a proven design if it's not going to use a driver of the same batch? You have to pick a design from last year at most and hope...

        Comment


        • #5
          Well, if you plug my 4 datasets into WinISD using Roman's box and tuning, the modeled F3s all fall between 57 and 63 Hz. That's just not that much difference. If you had my 4 sets of data (and used Roman's box and tuning) you could say (reasonably) that the Microbes have an F3 "near" 60Hz, and odds are your "error" would be +/- 3dB. That's 5%. That's pretty good. So running a box sim gives you a good "estimate" of the required box volume, tuning, and F3.

          In general, PE's specs (on Dayton drivers) are pretty accurate - AND they typically "update" them every few years or so. Building most designs using Dayton drivers, you can end up pretty sure of the outcome. If the mfr. specs for (Dayton as well as other brands of) drivers were always exactly perfect, there'd be no market for something like DATS/WT.

          Comment


          • #6
            Thanks for the answers...and sorry for the questions that may sound stupid, but I'm still learning, just trying to understand.
            For the tuning frequency it works and the error is reasonable.
            But for box volume and output at a certain freq. an error of 3db is 100% error on pressure level, so it doesn't look that trascurable. +/- 3DB should not be acceptable...by far...but I cannot tell in the actual effect for a listener, not enough experience there.

            Looking at the graphs I'd say Roman design should sound full-bodied with a nice rolloff, while my sim should sound empty from 200hz down. We have a difference of 2DB between the 2 simulations at about 100hz.

            Just to be clear, If I wanted to build The Microbes, my best bet would be to:
            1. get the drivers,
            2. measure them (or take PE measures)
            3. simulate the response,
            4. define a new volume that gives me the most similar response to Roman design
            5. adapt the depth of the cabinet design in order to get to the new volume.
            6. Build the speaker.
            Right?

            Comment


            • #7
              I would sim using PE's latest measurements (the ones you've already modeled), and reduce the volume by about -10% (shooting for around 0.17cf). So, you could reduce the depth by maybe 1/2" or 3/4".
              I'd also use the same tuning (which means a 5" long port - as a box gets smaller the port lengthens to maintain tuning), and I think you'll find the dBs will vary by <1dB across the board - from Roman's orig. design).

              The very minor change in your Microbes' bottom end can/will most likely be changed W-A-A-A-Y more by your room size, wall/floor/ceiling coverings (or "treatments"), furniture, and (ultimately) by your positioning of the speakers in your room, as well as your listening position.

              Most say that it's pretty hard to detect changes of <3dB by ear, anyway.

              There ARE other Dayton drivers that can give fuller (deeper) bass in boxes that are actually smaller than the Microbes. The TCP115s come to mind.

              A great book on speaker design is "SpeakerBuilding 201".

              Comment


              • #8
                Thanks!
                Everything is clearer now...Got a mic now: imm-6 starting small...I'm going to take some measurements...so I can understand how my room performs! If I'm lucky I'll have a perfectly flat boost from 30 to 80HZ...

                It is quite difficult to imagine how a line on a graph will sound...I suppose experience would/will fill the gap.

                "SpeakerBuilding 201" is in the pipeline...have to finish "Loudspeaker Design Cookbook first!
                Learn to mesure my room and I'll be back on the speaker building topic...In the meanwhile I'll waste some time on sims just to understand the differences between the various makers and productlines...

                Any other driver you would suggest to simulate? (Noted the TCP115s) More as training than building ATM. A good EU alternative eventually? Dayton get quite expensive here in the UK...

                Cheers!

                Comment

                Working...
                X