Announcement

Collapse
No announcement yet.

Crossover design methods

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

  • Crossover design methods

    I:'ve been spending a lot of time the past month or two doing a lot of in depth reading about crossover design.* It's definitely a slow learning curve for a beginner like me.* From what I can tell, there seems to be roughly four different ways to design crossovers:

    1) Buy a pre made crossover and hope for the best.

    2)* "Look" at a drivers manfacture data sheet.* Make an educated "guess" for a crossover point. Using a generic textbook calculator, enter the crossover frequency you picked.* Build the crossover and again, hope for the best.

    3)* Input a drivers manufacturer data sheet info (T/S parameters, frd's, zma's, etc) into different free/paid simulation software ( baffle diffraction simulators, box modeling simulators, crossover simulators, etc.)* Combine everything and create the crossover.**

    4) Discard the driver's manufacturer data sheet. Instead, take your own measurements. Using your data, do everything in method 3. Build the crossover AND build the enclosure.* Use a mic to measure the completed project. Make any changes/tweaks to the crossover.

    *Even as a beginner, methods 1 and 2 seem like a complete waste of time.* Method 4 seems to be the best way to go about things (but the most expensive with measuring equipment thats needed)* Method 3 seems like its a good choice since it can all be done free, without having the physical drivers or enclosures.* Is method 3 gnod enough to make a quality design?* Or is method 4 really the only way to achieve a really good speaker?**

    *

  • #2
    number 3 can be darn good!

    https://sites.google.com/site/undefi...d-measurements

    *

    Comment


    • #3
      I also vote for number 3 - 1 and 2 are not worth the trouble and expense (i.e. your expensive drivers will sound awful). I've used Xsim free software and the manufacturer supplied frd and zma files to work out a crossover design for a two way Peerless/VIFA speaker.

      I used the PETT forum to get valuable input from other members, this is a great way to learn and reduce the intimidation factor in trying to learn about what is really quite a complex subject. But, and this is a big but, I haven't built the speakers yet! When I do, I'm sure I'll need to do a bit of fine tuning to come up with a reasonable speaker, but it's been a great learning process.

      With number 4, unless you intend to design lots of speakers I really think you would be better off to save the $$ and go for an established speaker whose designer uses those techniques and equipment.

      Geoff

      Comment


      • #4
        Duplicate post, sorry

        Comment


        • #5
          Option 3 can get you very close, but not all supplied curves/data are created equal. You're also stuck making an educated guess at the driver offset when modeling. Option 4 isn't as expensive as it used to be. Less than $100 buys the UMM-6 (when they're back in stock) and a shorty boom mic stand. T/S parameters from the manufacturer can also be suspect. But there's probably someone out there who used the woofer you have chosen, saving you from buying a DATS.*
          Co-conspirator in the development of the "CR Gnarly Fidelity Reduction Unit" - Registered Trademark, Patent Pending.

          Comment


          • #6
            #4
            Get a cheap measurement mic and phantom power. There's free software and ARTA which can be used without save option. You don't need DATS. Build a $5 jig for ARTA.
            http://www.diy-ny.com/

            Comment


            • #7
              I use 4. 3 can work, but the biggest problem is that factory spec are done on a large baffle. There is no baffle step shown in their FR plots. I think Jeff has a program to add this to factory large baffle plots. This becomes a requirement if you plan to use 3.

              -Bob
              -Bob

              The PEDS 2.1 mini system
              My A7 Project - another small desktop speaker
              The B3 Hybrid Dipole - thread incomplete and outdated

              Comment


              • #8
                Originally posted by ceiol View Post
                The link you provided was very convincing, I enjoyed reading it.* *I try to do as much research as I can but sometimes what may seem like asking simple questions is a big help for me.

                Comment


                • #9
                  Originally posted by tom_s View Post
                  Option 3 can get you very close, but not all supplied curves/data are created equal. You're also stuck making an educated guess at the driver offset when modeling. Option 4 isn't as expensive as it used to be. Less than $100 buys the UMM-6 (when they're back in stock) and a shorty boom mic stand. T/S parameters from the manufacturer can also be suspect. But there's probably someone out there who used the woofer you have chosen, saving you from buying a DATS.*
                  Correct, of course, thank you. I 'cheated' by working out a new crossover for an existing commercial kit design which had a less than optimal response. So, assuming the commercial product had taken the driver offset etc into account, (hopefully) these things were worked out in the original design. If not, we'll see when I eventually build it!

                  Geoff

                  Comment


                  • #10
                    Originally posted by Geoff Millar View Post

                    Correct, of course, thank you. I 'cheated' by working out a new crossover for an existing commercial kit design which had a less than optimal response. So, assuming the commercial product had taken the driver offset etc into account, (hopefully) these things were worked out in the original design. If not, we'll see when I eventually build it!

                    Geoff
                    Option 3, of course, allows you evaluate different driver combos before purchase.

                    Comment


                    • #11
                      You can get excellent results with #3.* Honestly, it's probably worth doing a design or two like this before investing in measurement equipment.* If you get bit by the bug like the rest of the poor souls here, then plunk down the cost of a nice set of drivers for an OmniMic and a DATS and move on to #4 to reach your full potential.* It's true you can get into measurements for cheaper, but the convenience of the aforementioned devices are worth the cost if you will continue to design.

                      Dan
                      _____________________________
                      Tall Boys
                      NRNP Computer Sub
                      The Boxers
                      The Hurricanes
                      The Baronettes
                      Conneccentric
                      UX3

                      Comment


                      • #12
                        Originally posted by Millstonemike View Post

                        Option 3, of course, allows you evaluate different driver combos before purchase.
                        Just beginning to roll my own, but this is what I like about option #3; I'll pick some drivers that 'might' play well with each other.* Then use their supplied .frd and .zma files to build a theoretical crossover.* If that went well, then save those parts/design in a file for later consideration.* This way one gets a good idea if the drivers are going to be a good match before laying out the cash.***

                        ... then

                        ....

                        Sorry, OSI just came up on rotation.* If you haven't heard the group, give them a listen, they kick buttt.****

                        Back on track.*

                        If, or when, the drivers get ordered I'll build the enclosure, mount all drivers, and use REW (<free) and a usb mic (<$60) to take freq. measurements.* Then DATS for the .zma files.*

                        REW works with the ARTA box to make .zma files, so why DATS, for $100,* instead of using the ARTA box?*

                        Because you have to buy a decent soundcard (or whatever it's called) for the ARTA box to work.* In my case I wasted ~$40 on a Behringer UCA222 that was too noisy to take measurements, plus the ~$20 on the ARTA box, and a few extra rca's.* With time, cost, and headaches considered, it was cheaper to just buy DATS, and DATS is super easy to use.

                        After getting all those 'real' .zma and .frd files gathered, I'll play with Xsim and WPCD for the final time.* Crossover parts get ordered and while waiting the speaker exterior can be finished.

                        Just my $.02, YMMV. * ** ** * * * ** * ** * *

                        * ** **
                        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


                        • #13
                          Drjay* and I have been doing a bunch of simulations lately (#3). What we learned is the spl match is always off unless you measure. Then the baffle response is highly influenced by how you mount the drivers. I think the simulations are great fun and good experience but you should measure to determine what you are hearing.
                          Last edited by jhollander; 04-30-2018, 10:54 PM.
                          John H

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

                          Comment


                          • #14
                            What works best for me is to use manufacturer sheets to find what might work (usable frequency ranges, impedance, sensitivities, directivity match).* Might do a quick quasi-design using baffle or driver placement software to see how plausible it all seems.* Then get the drivers, build the box (or at least the baffle), mount them together and measure actual drivers on actual baffle - then do actual design with a crossover simulator and buy parts to build.
                            Free & Free-form simulator/designer for Passive Crossovers
                            SynergyCalc 5: design spreadsheet for Wooden horns and DIY Synergy Waveguides
                            Super easy and cheap to make high performance sound diffusers

                            Comment


                            • #15
                              So in layman's terms, #3 would be like taking specs out of a hotrod catalog fnr various engine components (crank, cam, rockers, rods, etc). Then inputing those specs into an engine simulator.* You will prob end up with a decent engine. However, you will never know for sure if it's running 100%.

                              Doing #4 would be like testing that custom engine on a dyno.* From there you can see what tweaks need to be made (adjust timing, adjust fuel mixtures, swap cams, etc). The end result will turn the decent running engine into a mean tire smoking beast.*

                              having measurement equipment is tempting.

                              *

                              Comment

                              Working...
                              X