Announcement

Collapse
No announcement yet.

DATS V2, Huge Difference between spec sheet and measurement

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

  • DATS V2, Huge Difference between spec sheet and measurement

    I've been eyeing the B&C 18TBW100 for a while, wanting to use them in my home theater. They model very well in 7ft^[email protected] A recent coupon prompted me to purchase a few of the 8 ohm version. I unboxed one to measure, and the results were not what I expected. This is no break in, but I can't feature break-in bringing them anywhere close to specs.

    Parameter: spec, measured
    Qts: .39, .57
    Qes: .41, .6
    Qms: 8, 11.5
    Fs: 35, 46.4
    Re: 5.3, 5.26
    Le: 2.45, 2.73


  • #2
    Break in will help, but not that much. It's not unusual for the engineering department to come up with a prototype that has one set of measurements and then for the actual production model to be different, or for components to change over time due to sourcing issues, without the data sheets being updated. For example, Eminence had a change in all their driver specs when DuPont stopped supplying Kapton for voice coil formers and they had to switch suppliers. They did issue all new data sheets, but some retailers didn't update their information for quite some time. Databases with the old information, WinISD for one, also weren't necessarily updated. I updated my WinISD database myself. These are just a few reasons why discrepancies can occur. Another is the measuring protocol. Did you suspend the driver in free air?
    www.billfitzmaurice.com
    www.billfitzmaurice.info/forum

    Comment


    • #3
      Most of "break in" consists of the spider (adhesives, or coating) "softening" due to a bunch of excursions approaching Xmax limits. Qts (a derivative of Qes) and Fs will drop, and Vas should rise. You often don't NEED to drive the speaker electrically to achieve this, but several (careful) strokes to the Xmax limits by hand will often do the trick.
      You don't show a measured Vas value, but it's one of the "big 3" parms you need for box sizing.
      If used "normally" (speaker somehow held "rigidly" w/out blocking the motor vent or restricting cone movement - even with air pressure), your DATS CAN be trusted to give you accurate Qes (and therefore Qts) and Fs values. Operator "error" (or technique) CAN lead to inaccurate Vas measurements. You'll get better with practice.

      Also, measured values that differ (sometimes even significantly) from "spec" MIGHT still give nearly the same box sim results. You'll never know until you plug them into your modeler (but again, you NEED Vas in order to do that). Have fun!

      Comment


      • #4
        I break in woofers with a sine wave signal lower than Fs, sufficient in voltage to push the cone to just below xmax, applied for at least 12 hours. An Fs reduction of 10% is typical, but in this case even 10% would have the driver still way out of spec.
        www.billfitzmaurice.com
        www.billfitzmaurice.info/forum

        Comment


        • #5
          That's kind of what I expected. Tempted to try and return them.

          Comment


          • #6
            Originally posted by Chris Roemer View Post
            Most of "break in" consists of the spider (adhesives, or coating) "softening" due to a bunch of excursions approaching Xmax limits. Qts (a derivative of Qes) and Fs will drop, and Vas should rise. You often don't NEED to drive the speaker electrically to achieve this, but several (careful) strokes to the Xmax limits by hand will often do the trick.
            You don't show a measured Vas value, but it's one of the "big 3" parms you need for box sizing.
            If used "normally" (speaker somehow held "rigidly" w/out blocking the motor vent or restricting cone movement - even with air pressure), your DATS CAN be trusted to give you accurate Qes (and therefore Qts) and Fs values. Operator "error" (or technique) CAN lead to inaccurate Vas measurements. You'll get better with practice.

            Also, measured values that differ (sometimes even significantly) from "spec" MIGHT still give nearly the same box sim results. You'll never know until you plug them into your modeler (but again, you NEED Vas in order to do that). Have fun!
            Just the tiniest difference in measuring Sd of the driver throws Vas off a small amount I have found. It is tough to know exactly how far into the surround to "count" as driver area for me. I have physically moved the cone upwards to see where the surround flexes, but one can still be off a few millimeters either way.

            Comment


            • #7
              I'd first try breaking them in before jumping to any conclusions. The B&C spec sheet has this: "Thiele-Small parameters are measured after a high level 20 Hz sine wave preconditioning test." They unfortunately don't mention the exact amplitude or length, but I'd imagine it was near Xmax levels. Secondly, I wouldn't get get too concerned about individual absolute T/S values, but look at the Fs/Qts ratio and the Vas X Fs squared value. Paul Kittenger stated this about the Fs/Qts ratio:


              "This ratio also tells you something about a driver's likely performance even if your particular driver has different values for fs and Qts as compared to the manufacturer's published specifications. Over a production of drivers there will be variations in individual values for the T/S parameters, but as long as the ratio is pretty consistent, as well as the value of Vas x fs2, all of those drivers will perform very much the same in the same enclosure.
              Paul"

              PS - Have you measured Vas yet? For Sd I'd use the value from B&C's spec sheet
              Last edited by ernperkins; 10-04-2018, 08:27 AM. Reason: Add PS
              "Everything is nothing without a high sound quality." (Sure Electronics)

              Comment


              • #8
                Originally posted by Mandroid View Post
                I've been eyeing the B&C 18TBW100 for a while, wanting to use them in my home theater. They model very well in 7ft^[email protected] A recent coupon prompted me to purchase a few of the 8 ohm version. I unboxed one to measure, and the results were not what I expected. This is no break in, but I can't feature break-in bringing them anywhere close to specs.

                Parameter: spec, measured
                Qts: .39, .57
                Qes: .41, .6
                Qms: 8, 11.5
                Fs: 35, 46.4
                Re: 5.3, 5.26
                Le: 2.45, 2.73
                The low-level signal from the DATS v2 would be barely enough to tickle that 18" driver, and then there's the questions of break-in (mentioned previously) and the issue semi-inductance to consider (DATS does not currently support measurement of the semi-inductance parameters).

                1. Use a high-level signal at or below Fs to break in the driver before measurement.

                2. Export the impedance curve and use the latest version of the semi-inductance spreadsheet (downloadable from the "Hornresp" thread on the DIYAudio Subwoofers forum) to determine the t/s parameters and the semi-inductance parameters for that driver.

                More about "semi-inductance" here - https://www.diysubwoofers.org/misc/semi-inductance.html

                Brian Steele
                www.diysubwoofers.org

                Comment


                • #9
                  Originally posted by Mandroid View Post
                  I've been eyeing the B&C 18TBW100 for a while, wanting to use them in my home theater. They model very well in 7ft^[email protected] A recent coupon prompted me to purchase a few of the 8 ohm version. I unboxed one to measure, and the results were not what I expected. This is no break in, but I can't feature break-in bringing them anywhere close to specs.
                  Hello everyone! This thread was brought to my attention, let's see if I can make myself useful to you.

                  As others have mentioned, I would absolutely start with break-in. A factory fresh woofer will always measure quite differently from our spec sheet, because the glues and treatments on especially the spider and surround have not been worked. When we put numbers down on our spec sheet they are taken after a high level Klippel LSI break in process, which exercises the suspension quite enthusiastically. The suspension parameters will quickly approach their long term stable figures, and shouldn't change much absent abuse once broken in. Klippel has a few papers on this topic, here is a relevant one: https://www.klippel.de/fileadmin/kli...on_Klippel.pdf

                  The second culprit is probably differences between either how DATS determines TS parameters or how you're taking those parameters and how we do it in our R&D facilities. This is some of the fun behind the scenes action in loudspeaker development, no two programs which curve fit to impedance data and predict TS parameters agree. Small measurement differences, rounding in our spec sheet figures, and math differences all add up. Not to mention you're measuring a physical woofer sample subject to production tolerances! As you can imagine, we have chosen the method which we feel best represents the real world performance of our products, and while that should be consistent across our line you'll need to apply a different "windage" to correct your DATS measurement to our spec sheet numbers. Just like measuring the same voltage source with two multimeters, which one do you trust? If you're a subscriber to Voice Coil magazine (and you should be if you like transducers!) you'll see Vance Dickason provides parameters from two different T-S models for two samples of the same woofer from us... and none of them match our spec sheet. Example image attached from the Test Bench review of our 8MBX51 woofer.

                  Click image for larger version  Name:	20170720131139_Table1-BCSpeaker8MBX51MidbassWoofer.jpg Views:	1 Size:	90.0 KB ID:	1389644

                  Long story short, it will be quite challenging to measure a woofer in the real world using a convenient portable system like DATS (which I use all the time to take impedance measurements of whole cabinets to develop processing settings, thanks Parts Express!) and reliably match our spec sheet numbers. The more important thing is whether our spec sheet numbers are believably representative of the long term performance of the woofer. Of course to verify that you'd need to use some prediction software to design a box, and there's a whole new set of assumptions in play there. This is a very deep rabbit hole, and a big part of why there are very few well known and successful cabinet designers - and why they keep their cards pretty close to their chest.
                  Test Bench of the 8MBX51 from B&C Speakers' MBX series of mid-bass woofers.
                  Last edited by Bennett Prescott; 10-04-2018, 06:45 PM.
                  -- Bennett Prescott
                  Sales & Operations Director
                  B&C Speakers North America
                  [email protected]
                  +1 (973) 248-0951

                  Comment


                  • #10
                    Had a chance to perform a bit of break-in yesterday evening. Applied a 20Hz sine wave for about an hour. Cone excursion was a bit below the point where distortion noticeably increased. Remeasured while motor was warm, then again after it cooled down. I make no claims to the accuracy of the Vas Measurement.

                    Click image for larger version

Name:	18TBW100-8.jpg
Views:	1
Size:	210.0 KB
ID:	1389747

                    New numbers appear to model very close to specs. Also trying the semi-inductance model in hornresp. I'm not very familiar with hornresp, but here is the alignment I was originally planning to use. Black line is with semi-inductance turned on.

                    Click image for larger version

Name:	18TBW100 semi inductance.jpg
Views:	1
Size:	57.8 KB
ID:	1389748

                    Comment


                    • #11
                      Those look much better.
                      Personally, I'd probably tune higher. Mid 30s maybe.

                      Comment


                      • #12
                        Just be careful when doing break-in tests to check on the cone excursion regularly. A safe amount of X when the suspension is cold can become unsafe once it warms up.
                        -- Bennett Prescott
                        Sales & Operations Director
                        B&C Speakers North America
                        [email protected]
                        +1 (973) 248-0951

                        Comment

                        Working...
                        X