Announcement

Collapse
No announcement yet.

Trying to create a crossover filter for the first time.

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

  • Trying to create a crossover filter for the first time.

    Hi Everyone,

    I need another set of speakers and this time I'd like to go one step further in my DIY adventures. My last project were Nano Neo mini towers

    So I got my hands on two noname Chinese 5 inch dual cone speakers. And to be honest only because they look so nice and they were only 50$ total.

    The thing is I still don't like big speakers cabinets so I thought why not use them together with a subwoofer/woofer like the w5-1138smf.

    I created ZMA and FRD files with the SPL trace option and am trying to create a crossover with VituixCAD 1.

    I've been playing the better part of a day with simulating different components but I can't seem to get a good looking result. Either the treble is to high or it shows to many ups and downs.

    I have the attachments ready to go, but for some reason the forum log's me out before I can add them, so to prevent typing this for the third time. I'll try and edit/add them later.

    Ok so I managed to get some screenshots attached. Ill add ZMA and FRD files later. Check...

    My basic idea was to build a 8Liter cabinet and put the full range driver in a closed compartment. That would give the w5-1138smf about 7.5Liters vented. I don't care about sensitivity that much as I will be using these on a desk.

    so to recap and get to a question;

    I can't seem to lower the high frequenties enough and keep them flat.
    If I don't use a filter The graph looks better, but then the two drivers would overlap from 50 to about 1000hz..
    So what should I do? Or what would you do?


    Attached Files
    Last edited by dali; 08-30-2018, 01:45 PM.

  • #2
    For starters, that FR won't be happy (at ALL) in 1/2 L. That pushes its "Q" up well above 2.0 !
    Even in 4L (sealed/stuffed) its Q is still just above 1, but would probably be OK.
    I also wouldn't suggest anything less than 8L for the W5 (vented - 2"id x 16" long - probably need a "Precision Port" to avoid chuffing).
    Cross around 200Hz (MAYbe higher)?

    Your W5- .zma file LOOKs like its for an 8ohm woofer, not a 4ohm?
    Try starting w/this:
    2nd order HP w/L: 80uF series cap, 3.0mH shunt coil (to gnd). L-pad uses 6n series and 4n parallel. (Attenuating about -10dB probably.)
    For woofer, 7.0mH series (iron core) coil, 150uF shunt cap.
    Crosses under 300Hz.
    Your inefficient W5 is pulling down your much more efficient full-range.
    THAT's why you're almost always better off going active on that sub driver. The amp lets you crank its level up to match.

    Comment


    • #3
      Funny thing is that I already HAD that 5" Chinese driver in my database! (You must have mentioned it on TT once before.)
      I'd tune 0.5cf (14L) to the mid 50s using a 2" id by 4" long port.
      It can take 5w RMS before breaching Xmax near 80Hz.
      Last edited by Chris Roemer; 08-09-2018, 08:54 PM.

      Comment


      • #4
        Try inverting the polarity on the full range, it looks like you have a phase cancelation.
        Even though you will probably be making many changes this is something to keep
        in mind.
        Guess xmax's age.

        My guess: 15. His grammar is passable. His trolling is good.

        Comment


        • #5
          This is just a post to let you know how this project is coming along.

          Finally had the time to smack together a prototype cabinet... but I cut the inside diameter... so now its only 11,55L (smiley_facepalm)(smiley_facepalm)

          I continued anyway because I now wanted to hear them, but didn't have time to get some new MDF. I used WinISD to get the new port length -> used 45mm by 90mm.

          I listend to some of my favorite songs and then realized I didn't use any damping material... so I looked in my storage and found a small sheet. It was not enough for the whole cabinets so it only covered the back the top and bottom.

          I then listened some more... These are nice speakers, I listened near field about 1Meter away. It's a bit like listening with headphones without the extended low end.

          After that I used my phone to visualize a sweep played from my computer from 20hz to 20KHz and also some white noise. (Sorry I don't have a calibrated mic). Also this is an in room measurement about 1 meter from the speakers. The speakers are next to my computerscreen. about 15 cm from the wall. The room is 2,3Meters high, 3,5M wide and 5M deep.

          Kind of struggling to find the right words here but here it goes; Looking at the FR I wonder how much I hear the room or how much the room influences the speaker. (I might add that I'm also been looking to buy a set of speakers that play flat or with less coloration, but that's for another thread). Looking at this FR it's not as flat as I would like. As this thread was originally about designing a crossover, would it help to add a filter to smooth the response. *edit* wait. I'll try to simulate some things myself first.

          Then there is XMAX remark about phase cancelation, I need to read about that, but how do you think this would influence the FR?

          This hobby is great by the way! every time you think you're at the end of the rabbit hole you find there is another door...

          Comment


          • #6
            I don't think I see a cancellation in the 200-300Hz range. It might help to also show the response of just the FR, and just the sub.
            Your FR IS in an undersized box, with (probably) an undersized vent (chuffing?). In my WinISD model, it looks like your measured peak (in the 140Hz range) PROBABLY is caused by the smallish box.
            If you feel that the bass is not strong enough, then the FR probably needs even more attenuation. (Lpad = 8n for SR / and 2n for PR ?)

            Comment


            • #7
              Originally posted by Chris Roemer View Post
              I don't think I see a cancellation in the 200-300Hz range. It might help to also show the response of just the FR, and just the sub.
              Your FR IS in an undersized box, with (probably) an undersized vent (chuffing?). In my WinISD model, it looks like your measured peak (in the 140Hz range) PROBABLY is caused by the smallish box.
              If you feel that the bass is not strong enough, then the FR probably needs even more attenuation. (Lpad = 8n for SR / and 2n for PR ?)
              The FR in the shown image is without the sub. I probably frustrate someone with what I'm about to say, but it made me think of the NS10's.

              I'll construct a new cabinet this weekend to the specs you gave before.

              I'm not sure wat you mean with SR and PR...

              Don't you ever get tired of helping out all the newbies? Anyway thanks so much for helping out Chris!

              Comment


              • #8
                Series resistor and parallel resistor is SR and PR. So 8 ohm resistor in series and 2 ohm in parallel for the Lpad. Good luck!

                Comment


                • #9
                  Don't know what NS10s are, sorry.
                  If your response plots are of your FR only, then I guess there's no polarity to reverse - is there?
                  A LOT of FRs have a rising (or even closer to flat) response that MANY prefer to use a baffle-step filter with.
                  The upper octaves in your plot look to be about +10dB louder than the bottom (before rolloff).
                  A typical baffle-step filter would reduce the top part by -6dB.
                  Also, many like the top end a little "hot" on an FR to sort of compensate (a bit) for the poorer dispersion characteristics of a larger (than a 1" dome tweeter, for example) 3" to 5" FR speaker.

                  Have you built your sub. If so, how does the whole thing sound?
                  If not, then you're basically listening to a 2-way w/out the woofer, right? (Would NOT expect that to sound real good.)
                  Does your FR (in the plots) have its half of the crossover (that I suggested) on it, or are you just showing us its unfiltered output (in the box you ended up with)?

                  Comment


                  • #10
                    Originally posted by Chris Roemer View Post
                    Don't know what NS10s are, sorry.
                    I mean the Yamaha NS10s, The speakers used in most recording studios to listen back the recording while mixing, just to get an idea of how things will sound when average Joe would be listening to it. That is another discusion entirely. I don't want to go there, there is been said more then enough about those, but I just got reminded of them so that's why I mentioned them. I've attached a freq response graph of them just for fun comparison reasons.

                    Originally posted by Chris Roemer View Post
                    Also, many like the top end a little "hot" on an FR to sort of compensate (a bit) for the poorer dispersion characteristics of a larger (than a 1" dome tweeter, for example) 3" to 5" FR speaker.
                    I'm not bothered by to much high, I can't even hear above 14khz, so that's where the analyzer software comes in handy. Still I always try to make things as good as possible, although my enthusiasm takes over and I forget things like the L-pad you mentioned before....

                    Originally posted by Chris Roemer View Post
                    Have you built your sub. If so, how does the whole thing sound?
                    No I did not, I first wanted to finish this set of speakers. In the mean time I see if I can buy a used one.
                    I know this leaves me with a crippled overall frequency response, but as I'm used to switch between speakers like the NS10 and system with more low end, I think I can still judge this speaker when listening to records I know through and through.

                    Originally posted by Chris Roemer View Post
                    Does your FR (in the plots) have its half of the crossover (that I suggested) on it, or are you just showing us its unfiltered output (in the box you ended up with)?
                    No this is just the one full range driver.

                    So I went searching for resistors in my leftover electrical parts box and found some 2W resistors (is that ok or do I need a higher wattage?) and made the L-pad. I ended up with 8.4Ohm and 2.2Ohm. In the attachment you find the new screenshot of the analyzer.

                    That' is indeed a bit of a change, I found some time to get new MDF, I'll be cutting new cabinets tomorrow.

                    Comment


                    • #11
                      Well (w/out the "sub" woofer "in the mix" - meaning taking over the bass duties from your FR via a passive crossover), the "L-pad" won't do ANYthing for you. All it'll do is burn up some of your amp's power by making heat (in the resistors).

                      You MIGHT want a baffle-step filter, to tone down the highs and align them more with the FR's bottom end (although you stated that you don't miss bass that isn't there)?
                      I'll look at my recommendations I made to you (later) to see if they can apply w/out a sub (but I don't think so).

                      continued . . .

                      If you're just running the 5" (w/whizzer) full-range in a box, (based on YOUR measured in-box response) you should try this filter (4 parts).
                      1st you need a Zobel across the FR's terminals (in parallel w/it). This will be a 10ohm (10w) resistor in series w/a 22uF cap.
                      By itself, it does NOthing to the freq. resp., but it DOES alter the impedance curve so the BSC (baffle step compensation) filter can do its job.
                      The BSC filter is a 15ohm resistor in parallel w/a 1.0mH coil (I modeled w/a DCR of 0.8ohms - but it's not really critical), that pair inline (in series) w/the positive term on the 5in FR.

                      Together, the BSC filter and the Zobel will have this effect on the 5"er's freq. response:
                      Below 300Hz, little change.
                      -3dB down near 1kHz.
                      -6dB down near 2kHz.
                      and -9dB down at and above 6kHz.

                      W/out this 4-element filter, the output at 10kHz is about +9dB louder than at 300Hz.
                      WITH the filter(s), the output at 10k is nearly the same as at 300.
                      Last edited by Chris Roemer; 08-31-2018, 11:32 PM.

                      Comment


                      • #12
                        Sorry if I was not clear enough about only using the full range driver, I've ordered the parts you suggested. So I'll have to wait till next week.

                        So I finally got some parts... Enough to build your filter and is exactly as you said! spot on great, but I will not continue this project, After so much waiting I didn't get the (sub) woofers but my money back.... So I started to look for other projects. I think I found one. So I'll have a second try at doing a crossover. and probably lot's more questions.
                        Last edited by dali; 10-11-2018, 01:16 PM. Reason: update

                        Comment

                        Working...
                        X