If the graph is any indicator, then 1500 Hz seems like a fine crossover point. I don't see any point in trying to cross over lower than that. If the woofer sounds good with no filtering, then 4000 Hz also seems like a nice place to cross over. I like simple crossovers, so maybe first order on the woofer and 2nd or 3rd order on the tweeter. If I recall correctly, I think that Jeff Bagby liked 2nd order crossovers.
Last edited:
Hmm, horn VC = ~34400/pi/7.24 = ~1512 Hz, so any lower will be down on the woofer cone behind the horn until out from under the HF horn's shadow and the woofer's is ~34400/pi/10.16 = ~1078 Hz, so while in theory it could be down as far as the HF horn's acoustic terminus size (dia.?) around 600 Hz based on the factory measurement, you're either severely restricting peak SPL and/or going with at least a 4th order XO and maybe even 8th order depending on desired HF horn cutoff, peak power required.
Agreed, me neither and even the pioneer's original huge multicell horn coax used 1200 Hz, then quickly over corrected to 2 kHz, then 1600 Hz and then a smaller horn (WG actually) before locking in at 1500 Hz/2nd, so Beyma apparently decided to 'not mess with success' 😉.
Thanks guys, I appreciate the responses and as usual GM, the lesson in speaker theory and history, it's invaluable!
The outer diameter of the flare is 6", so using your equation I get 1825Hz, so at 1500 I'm actually nice and clear of it. I've been curious about that horn and the diffraction effects, it's a compromise and probably explains some of the anomalies in the original woofer measurements.
1500 get me away from some turbulence in both drivers' responses, and based on what you guys are saying it makes total sense to cross there. I was just curious because the CD has a 2" diaphragm and seemed like it might sound very nice a bit lower, but even if that was true now I understand why the tradeoffs aren't worth it.
Tonight I'll attempt to import the files to WinPCD and start playing around. Thanks again 🙂
This for example, I never thought to take the frequency of the woofer based on where it was out of the shadow of the horn, of course!
The outer diameter of the flare is 6", so using your equation I get 1825Hz, so at 1500 I'm actually nice and clear of it. I've been curious about that horn and the diffraction effects, it's a compromise and probably explains some of the anomalies in the original woofer measurements.
1500 get me away from some turbulence in both drivers' responses, and based on what you guys are saying it makes total sense to cross there. I was just curious because the CD has a 2" diaphragm and seemed like it might sound very nice a bit lower, but even if that was true now I understand why the tradeoffs aren't worth it.
Tonight I'll attempt to import the files to WinPCD and start playing around. Thanks again 🙂
There's also a pipe end correction = 0.613r for free space, so 3" * 0.613 = ~1.84" + 6" = ~7.84" eff. dia.: ~13543/pi/7.84 = ~550 Hz as you see in the measurements. 😉
So I've got all my sweeps in REW, and my impedance data in DATS, no problem. I figured I would just export/import the frd and zma files into WinPCD or Vituix. Three hours later, I threw my hands up. My friend Cam helped me import them using the trace function in Vituix, but wow! Is this really how this has to be done? How do people get their frd and zma files into their crossover software? It seems crazy I had to trace them in! Is there a better way?
Temper your decisions around the actual dynamic range of human hearing above 2khz……….so around 90db. When I see some of these armchair engineers on YouTube touting 2db anomolies at 8khz as problematic I wanna throw my mouse at the monitor to shut these gas bags up………try not to fall into the trap of ‘sighted listening’. While taking measurements and having response graphs available as tools is helpful to the process, it can’t be part of the auditory function. The ear/brain functions works only in the absence of other sensory input.
Your xo targets aren’t going to be about FR primarily,…….you should be targeting directivity matching of the two drivers. It is this interaction where bunching of energy produces confusion to the ear as it tries to listen through phase anomolies…….something that most folks don’t get when voicing a crossover. Look at it this way………your eye/brain when watching film…….the brain doesn’t see individual frames at a certain rate and full motion is accepted……vary that frame rate (phasing) and result isn’t pleasing…..it’s work the brain would rather not do.….these effects are profound. As a mix/recording engineer, phase alignment of stereo signals is part of my workflow…..there’s no way around it and thankfully in this one case study I’m fortunate enough to have sight to back me up……back in the day, the only option was swapping the phase of a microphone or having to measure the distance between mics and locate accordingly.
You have a huge hurdle in directivity to overcome because of the driver you chose…..that little waveguide in the middle has absolutely no chance of directivity matching those drive units….it’s simply too small. The driver was designed to beam or cut in a nearfield compact monitoring application. So in lies your major choice on compromise…….either pushing the XO higher to where the CD driver and horn couple within the passband and accept the beaming of the woofer or vice versa, pushing the CD down to 1khz and let the woofer dictate the phase information……you will never get both to play nicely between 1-2khz.……two different frame rates in a matter of speaking.
Your xo targets aren’t going to be about FR primarily,…….you should be targeting directivity matching of the two drivers. It is this interaction where bunching of energy produces confusion to the ear as it tries to listen through phase anomolies…….something that most folks don’t get when voicing a crossover. Look at it this way………your eye/brain when watching film…….the brain doesn’t see individual frames at a certain rate and full motion is accepted……vary that frame rate (phasing) and result isn’t pleasing…..it’s work the brain would rather not do.….these effects are profound. As a mix/recording engineer, phase alignment of stereo signals is part of my workflow…..there’s no way around it and thankfully in this one case study I’m fortunate enough to have sight to back me up……back in the day, the only option was swapping the phase of a microphone or having to measure the distance between mics and locate accordingly.
You have a huge hurdle in directivity to overcome because of the driver you chose…..that little waveguide in the middle has absolutely no chance of directivity matching those drive units….it’s simply too small. The driver was designed to beam or cut in a nearfield compact monitoring application. So in lies your major choice on compromise…….either pushing the XO higher to where the CD driver and horn couple within the passband and accept the beaming of the woofer or vice versa, pushing the CD down to 1khz and let the woofer dictate the phase information……you will never get both to play nicely between 1-2khz.……two different frame rates in a matter of speaking.
Last edited:
Thanks mayhem, that was a great post and very educational, even if some of it is at the edge of my abilities to grok.
I'm starting to add up the obstacles here, and the final one was last night after seeing the models in WinPCD, working from 1500Hz and moving up and down, 2nd order BW, 4th order L/R, and that CD is just not playing nice (My bud Cam was playing with the data while I recovered from tracing it all in). Add to it the impedance spike at 1220Hz, and it's a real fashion disaster.
I guess I can understand how in pro audio you could just squish everything down with DSP and blow a lot of watts through it, but for my goals it's really looking like a bad driver to continue with. I'll post another shot of the CD response below, but besides it being very hard to get the response to comply (turn south) at the crossover point, there's a 6db hump from 3k-4k, followed by a -10db suckout at 5.5k. I'm sure I could start adding correction filters and deal at least with the impedance part, but I really don't want to start piling on notch filters and such just to bandaid this unfortunate driver. I'm pretty positive some of what you're saying is showing in those anomalies, and it's the actual horn that's making things so choppy.
SO, I've spoken with Al at US Speaker, and he'll take them back with an understandable fee. I'm stepping away from this pursuit of a high sensitivity coax for now, and will go a more tried and true route, which means something like the Faital 15PR400, paired with a 1.4" CD and tractrix horn. I have no idea which CD will fit this pairing at the moment, and am open to suggestions. I like the bigger CD because I can cross lower, I'd really like to get below 1k if possible.
I have to say, there's something refreshing about this stuff, as utterly annoying and frustrating as it is sometimes. I'm no hardcore objectivist, but physics never lies, and learning the rules and knowing there's 'no free lunch' is reassuring. Rules are good. I'd rather have some rules I don't like than none at all 🙄
I'm starting to add up the obstacles here, and the final one was last night after seeing the models in WinPCD, working from 1500Hz and moving up and down, 2nd order BW, 4th order L/R, and that CD is just not playing nice (My bud Cam was playing with the data while I recovered from tracing it all in). Add to it the impedance spike at 1220Hz, and it's a real fashion disaster.
I guess I can understand how in pro audio you could just squish everything down with DSP and blow a lot of watts through it, but for my goals it's really looking like a bad driver to continue with. I'll post another shot of the CD response below, but besides it being very hard to get the response to comply (turn south) at the crossover point, there's a 6db hump from 3k-4k, followed by a -10db suckout at 5.5k. I'm sure I could start adding correction filters and deal at least with the impedance part, but I really don't want to start piling on notch filters and such just to bandaid this unfortunate driver. I'm pretty positive some of what you're saying is showing in those anomalies, and it's the actual horn that's making things so choppy.
SO, I've spoken with Al at US Speaker, and he'll take them back with an understandable fee. I'm stepping away from this pursuit of a high sensitivity coax for now, and will go a more tried and true route, which means something like the Faital 15PR400, paired with a 1.4" CD and tractrix horn. I have no idea which CD will fit this pairing at the moment, and am open to suggestions. I like the bigger CD because I can cross lower, I'd really like to get below 1k if possible.
I have to say, there's something refreshing about this stuff, as utterly annoying and frustrating as it is sometimes. I'm no hardcore objectivist, but physics never lies, and learning the rules and knowing there's 'no free lunch' is reassuring. Rules are good. I'd rather have some rules I don't like than none at all 🙄
I hope that you are listening to some music through the coaxials before coming to a conclusion. Graphs are not music. In my experience, it is very possible for a speaker to measure bad and sound quite good. And these do not look all that bad to me.
Also, have you tried switching the polarity of the high frequency driver? Based on the graph, it looks like it could be out of phase.
Also, have you tried switching the polarity of the high frequency driver? Based on the graph, it looks like it could be out of phase.
Last edited:
+1 that driver of yours sounded pretty good to me (as far as Youtube videos of speakers playing goes). No idea what xover or anything:
I appreciate the encouragement, and I've seen that video too, no way to say how those really sound of course. But just for an indication, here's an early attempt at the crossed drivers, 4th order L/R:
I'll have another go tomorrow, that one was made by Cam, but he tried for a couple hours and it was really stubborn.
Neon, I'll give that a shot, but my FR looks very close to Beyma's published response, only more exaggerated.
Sorry I didn't respond to your other question - listening to real music is a big step, I'd need to build two finished enclosures, and build two crossovers. Does it make sense to listen to just one? If that could be true, I already have the test enclosure obviously, and a single crossover would take all of 20 minutes to hack together. Does anyone listen to only one?
I usually build just one speaker to start with. If I like it, then I build the other side. I figure that I can cut my losses nearly in half if I end up not liking the basic design.
Lol!….it’s REALLY tough Brian…..everything is an accepted compromise to one degree or another. The ubiquitous bookshelf 2way is STILL the easiest design foundation on the road to success in most rooms. Some thoughts on your project if I might…..
15” for a coax was pushing the envelope. For such a design to work, the CD would need to be very robust……a large diaphragm with a 1” throat. Forget about the horn lens, it would need the full surface of the cone to create any resemblance of a directivity pattern needed. But that would yield a rough response…..1st trade off. 2nd tradeoff would be power handling of course.
12” coax alleviates some of the problems, but the rough response would still exist to a degree. Matching directivity would be less of a challenge. If it were me, I’d have chosen the Eminence coax….but I said that already so no need to revisit.
The path to take here would have been to extend constant directivity as low as possible. Only one real mechanism for that and that’s a Cardioid system. Daniel From Dutch and Dutch is a member here, and the design path was born right here on DIYAudio many years ago…...the thread is still around somewhere and available for review. A 12” coax in a proper cardioid enclosure high passed around 100hz or so……..that would be something. You would absolutely need 3 channels of DSP and amplification per channel though, and quite possibly even a few passive components on the CD for notching. But if you wanna take on a real challenge, that would be a worthy endeavor for sure.………..a two way with separate horn would be easier though.
15” for a coax was pushing the envelope. For such a design to work, the CD would need to be very robust……a large diaphragm with a 1” throat. Forget about the horn lens, it would need the full surface of the cone to create any resemblance of a directivity pattern needed. But that would yield a rough response…..1st trade off. 2nd tradeoff would be power handling of course.
12” coax alleviates some of the problems, but the rough response would still exist to a degree. Matching directivity would be less of a challenge. If it were me, I’d have chosen the Eminence coax….but I said that already so no need to revisit.
The path to take here would have been to extend constant directivity as low as possible. Only one real mechanism for that and that’s a Cardioid system. Daniel From Dutch and Dutch is a member here, and the design path was born right here on DIYAudio many years ago…...the thread is still around somewhere and available for review. A 12” coax in a proper cardioid enclosure high passed around 100hz or so……..that would be something. You would absolutely need 3 channels of DSP and amplification per channel though, and quite possibly even a few passive components on the CD for notching. But if you wanna take on a real challenge, that would be a worthy endeavor for sure.………..a two way with separate horn would be easier though.