Home Forums DIY Speakers and Subwoofers $3500 Speakers for $350? Klipsch KL-650 Inspiration

  • $3500 Speakers for $350? Klipsch KL-650 Inspiration

     123toid updated 1 week, 5 days ago 71 Members · 711 Posts
  • arnout

    January 17, 2022 at 7:58 am

    @123toid Hoi Toid,


    That is what i will do. oh by the way. keep up the good work with the speakers. really love the klipsch thx replica. if there is a design coming with maybe a 12 or 15 inch i will buy the design. and make them.

    So first i will try to make the active croosover work. if not i will make it passive.

    second i will take a good listen and maybe decide to put a 15inch and do a crossover around 200hertz. because the the klipsch thx can sound a little bit tiny (not sure yet). 

  • arnout

    January 18, 2022 at 11:31 am


    so i have been working on the active design yesterday. I first did measurement of the passive design. it is in room so it will not be perfect but gives a rough estimate.

    So what i did then is tried to replicate that with the active crossover. so for the bass-mid i didnot have to change anything. It was set to a 24db linkwitz riley slope at 1450 hertz. it matches perfectly

    But for the highs i ended up with 12db at 2150 hertz. i needed to some boosting in the high frequencies. (i dont like boosting but i wanted to do it quick and dirty). so now frequency response matches perfectly.

    Result: i am very impressed with the result, dialogue in movies is crystal clear. everything in the movies is crystal clear.  now i don’t need my centerchannel any more. (i like it better without centerchannel). there is still room for improvement but my main problem what i always had is gone. that dialogue was always not clear. and that speakers could not create ambiance sound. you know what i mean when a person walks into a church you hear the echos of footsteps bouncing from the walls. with good speakers you get the feeling of the acoustics of the church. that is what my speakers always lacked. Now it is there very cool!!!

    so now what was wrong with my other dsp settings so i maybe learn for the future when i build new speakers 🙂 i know there is more then only frequency response but i need to start somewhere. below is the frequencyplot of the first active version of the klipsch thx.



  • elliottdesigns

    January 18, 2022 at 7:48 pm

    @arnout Measurement looking awesome! Although, it’s never good practice to do a single measurement, since you never sit in exactly the same place with a singular ear on your forehead 😂. There’s two good options for this:

    1. Moving mic method (MMM), a spacial average using periodic pink noise whilst slowly moving your microphone around the listening space. Whilst it creates a great average with many, MANY measurement points, you also lose phase information and it’s typically smoothed out.

    2. (My preference), taking multiple measurement sweeps throughout your listening position, the wider the area you measure, the more measurements you’ll want to take, also the wider the sweet spot becomes (a wider sweet spot also means it’ll be slightly less good overall, as with many things in engineering, it’s give and take). I’d say a minimum of 5 measurements is a reasonable amount.

    The next step (for option 2), is to go into the “all SPL” tab and “time align” the curves, after which you use “average SPL”, to make sure each measurement is as close to one another in terms of SPl. After all of this you then use “Vector Average”, which takes an average of not only SPL, but taking into account phase information as well (this is very important), and finally apply appropriate filtering. When creating filters for room DSP I like to apply different levels of smoothing and check the result with psychoacoustic smoothing (since that’s we can actually hear) to see which gives the best result.

    The psychoacoustic smoothing is probably my favourite filter, it “shows” what a system actually sounds like (in terms of frequency response), and so as long as we aren’t looking for room modes or anything else that might be hidden when we are trying to solve certain problems, I believe it should be used more widely! Anyways, I digress… Maybe take another measurement of the actual speaker so we can see how the drivers sum, otherwise, looks amazing! Maybe try that spacial/phase average whilst your at it too! It should help us see how the frequency response is over a given area.

    P.s. MMM is an average of all the different responses at various positions, Vector averaging is meant to more closely approximate the frequency response at any given point within the space the measurements have been taken

  • 123toid

    February 3, 2022 at 5:57 pm

    Looking good! You have made a ton of progress. I like the idea of recycling the countertops.

    $3500 Speakers for $350? Klipsch KL-650 Inspiration

  • arnout

    February 4, 2022 at 6:31 am

    SO after tons of hours spending on active crossover. i also made a second passive crossover. so now i have 2 channel passive frontstage. the passive is so much better. way more detail in the speakers. (maybe i am doing something wrong with the active or maybe it is the extra digital to analog conversion i dont know) i dont care anymore passive is the way to go for me.

    So after listening i noticed my right speaker was sounding more bright then my left. i did a measurement close to the speaker and it was confirmed by the measurement. so i opened op speaker to see crossover. and i saw i used a different value for the resistor.

    now i am wondering what is the right one. i think the red line is the one to go?? it has a 4 ohm and a 2 ohm resistor.

    i am also wondering about the big bump i have around the 225 hertz(is this normal). i thought the speaker should be kind of flat?? maybe i did something wrong with the crossover??

  • jgs5607

    February 4, 2022 at 7:49 am

    Mentally filling in the summed area, I would think you want to be at the red curve ( or maybe with even slightly more attenuation)

    The 220Hz hump looks odd, maybe a room mode effect? The driver’s published response is relatively flat in that area. Try re- measuring it with the speaker pulled out in the middle of the room as far away from as many boundaries in the room as possible (and on a high stool to avoid being closer to the floor or ceiling) If that seems to improve things, it would point to a room effect. If it looks pretty much the same, it would take adding a notch filter to the crossover to flatten the peak.

  • 123toid

    February 4, 2022 at 11:44 am

    There’s a lot of information missing that really makes it hard to tell you what’s going on. First how far back are you measuring from? Second, are the woofers wired in parallel or in series? Third, it appears that you’re measuring the Tweeter separately from the woofers, what does the combined response look like? Finally, you cut off part of the graph, so we can’t see how far your zoomed out or in. This would help really work towards what the issue is. But definitely something does not appear to be correct.

  • arnout

    February 7, 2022 at 5:49 am

    Oke i will make a measurement outside in the garden when the weathers lets me. so this way i have the cleanest measurement. so we can see where the problems sits

  • jgs5607

    February 8, 2022 at 4:57 pm

    Fired it up for a test run today and I can tell already without any break in, measurement or tweaking that it is going to work well. Since I don’t have an AVR in that system ( yet ), I just let it play with its future playmate subbing in for the left speaker.

    It had no problem keeping up playing Dire Straights Money for Nothing at an average 92 db at the MLP. I have to remember I don’t have a high pass filter protecting it though. Fun stuff.

    I will put REW to work and get some measurements on Thursday to see how it matches up. More to come….

  • 123toid

    February 8, 2022 at 6:52 pm

    That looks nice. I actually like the silver screws in it.

  • jgs5607

    February 9, 2022 at 5:10 pm

    Got some measurements today and decided I need to make some tweaks; the woofers have a broad peak centered around 350 Hz in the 2 positions I tried when checking at 1 meter. At the MLP, it was still there, but with less of a rise; I am pretty sure I know what is going on there, and when things are complete, It should go away. . The JBL was flatter overall and had a dip at the same frequency. Not a big concern for now.

    The horn was ruler flat from the crossover to about 8KHz where it started a gradual roll off until it got to the point where it fell off a cliff like all compression drivers seem to do. I measured with and without the swamping resistor in place and it dropped the tweeter level about .5 db; well worthwhile to damp the impedance peak since this crosses over low. Overall the tweeter level is about 2db lower overall than it should be compared the the woofers; I will attribute that partly to the swamping resistor, partly to the woofers having a bit more output since I lowered the DCR in the crossover with the upgraded inductors and partly due to good old manufacturing variation.

    The SHAPE of the tweeter curve fairly well matched what I had in my XSIM model; I modeled a change to the the value of the parallel leg of the L-Pad to increase the overall level, and a capacitor across the series leg to flatten out the response on the high end to match the JBL which is flat to 15K before it goes off a cliff ( it has 6db more output at 15K).

    I can’t complain about adding 1 EQ element to the crossover (the 6db/octave increase from the capacitor pretty much perfectly compensates for the ~6db/octave roll off from 8 to 15K). The JBL uses 3 notch filters and 1 EQ compensator to tame their horn/driver combo; Ten EQ elements in the high pass vs the One I am adding to match the speakers it will the Center for; I am getting off cheap 🙂.

    I ordered the parts at Parts Express and they will be ready for me to pick up in the morning. More to come!

  • 123toid

    February 9, 2022 at 10:54 pm

    That is great! I am glad to hear the H290C is working out for you. I’m excited to see the final result. I’m a little jealous that you can just stroll over to Parts Express. I wish I had that luxury. Actually, it is probably good that I don’t 🤣

  • jgs5607

    February 10, 2022 at 9:41 pm

    Today was mod and test day as well as finalizing the build. First pics of My C6

    This was a “Green” project; The cabinet was made from our laminate countertop that was 2 years old when we replaced it with granite 10 years ago; it was aging gracefully in the basement. First project with it was a router table and it turned out so well, it gave me the idea of doing the Cinema 6 build with it. The Baffle was made from what was a cantilevered section that was 2 layers of plywood; plenty strong. I left the edges unfinished for now and will likely use the iron on veneer edging when this is is out with my JBL Studio 590’s in a HT setup; I specifically built these for when I expand this currently music only system to a 2nd HT setup. I wanted a center that can keep up and complement the towers.

    In the mean while, it will be moved to the HT system replacing the JBL Studio S Center that is built in to our entertainment center until I get another AVR for the S590 system.

    I did some deviations from the as designed Cinema 6:

    1) Cabinet dimensions were modified to fit the entertainment center; so it is slightly shorter and wider

    2) Upgraded the LP inductors; air core for 1 and an ERSE Super Q for the other to keep the cost and the total DCR low; it ended up .8 ohms less

    3) Used Wayne Parham’s (Pi Speakers) H290C waveguide instead of the Eminence H290B. Wayne developed this years back when Eminence stopped producing The H290 which is what he used in his speakers; It is an Oblate Spheroid to Elliptical Conical waveguide. The flange matches up so that guys that had the original could upgrade to the new version. He offers the DE 250 as an upgrade to the base Eminence CD and his site is loaded with measurements of this combo; I had no worries about it working well.

    4) I added a swamping resistor on the high pass to knock down the impedance peak at resonance; Wayne uses one in his crossovers when they are crossed over low.

    That was all the intentional changes made up front

    Once I got measurements; I decided to make some changes to match up with the JBL’s flatter response on the high end; changing one of the resistors in the L Pad to raise the output of the High pass a few db; and adding a Capacitor to flatten out the curve; I modeled it yesterday in XSIM and ran up to Parts Express today to pick them up.

    The first plot from REW shows the progression from Baseline to adding the Capacitor and then modifying the LPAD; results matched up pretty well with the simulation.

    The second plot is a comparison of the Cinema 6 to the Studio 590 in the same relative location to account for room effects

    The third plot is the average of the Cinema 6 in the Left Center and Right position. All plots have the Microphone at 1 meter and the REW pink noise reference level at 75db. Height wise the mic was right at the bottom of the waveguide, which I figured would be close to the acoustic center ( I tried 1 scan with it on the tweeter’s axis; there was not a whole lot of difference)

    The only weird thing in the response is the peak in the 200-300Hz range and the broader rise in that general range. I initially though it might be a resonance from having the rear panel only attached at 4 places. I ruled out a room location affect as it looks pretty much identical no matter where the speaker is measured. I added the vertical braces based on “knock testing” and when I finalized attaching the back to the cabinet and brace, that panel was no longer lively. The other thing I thought of was an internal resonance leaking through the ports, so I plugged them but no impact at all in that area. So whatever is causing it is real, but it really does not have a detrimental impact listening to it. This speaker is very clear, and at higher volumes it seems cleaner than the towers It should, The DE250 has a 13db headroom advantage over the 2414H driver in the tower; 10 db higher sensitivity and 3db in power handling, so it is just cruising.

    In reality, when I get it in a HT setup, a little Parametric EQ on the center will fix that if it bothers me; even my older AVR has that built in and it is ancient by today’s standards; so it is a non issue.

    Bottom line is that I am real happy with the result; this thing does exactly what I wanted in a Center and it will work well with either of my systems. Thanks for this project Nick

  • mupp3t

    February 20, 2022 at 2:04 pm

    So I got a associate that can CNC the cinema 6 for me. But on the design there is nothing mentioned about the bit that needs to be machined out for the drivers & horn to “lay in”. Since I still have to wait quite a bit for the parts to arrive to measure; any of you can give me the measurements for that?

    • This reply was modified 3 months ago by  mupp3t.
  • 123toid

    February 20, 2022 at 2:19 pm

    That is up to the machinist on what bit they want to use when they throw it in the cam software. Assuming the do the CAD drawing, they will then decide what bit they want to use based off what they have.

Viewing 676 - 690 of 700 replies