Wife gone last weekend so had a chance to do a “marathon” test of the “slow muffled sound” at 1k cross thing.
System is seal box 8 inch aluminum cone woofer. Cone is flat to about 1k with smooth transition to modal beyond that. Baffle is 10 inch wide. Tweeters are offset and as close to woofer as possible. Tweeters are 2 versions of a ribbon under development. One 3 inches long and the other 4 inches long. The 3 inch version crossed at 1khz 12db shows distortion under 1% to 102 db. The 4 inch gets to about 105 db at 1%. Both are below 0.5% at 90 db to 1k. Woofer show comparable distortions until about 2k where they start to come undone but not really bad. These ribbons are flat down to 200 hz with no peak so crossover is super easy to get right.
Played with crossovers at 3, 2, and 1k and even went to 750 Hz on the 4 inch.
Result… 3 khz was ok but no magic really and overall tonality changes with head moves were apparent. 2k was better. A cleaner sounding midrange (I assume because cone breakup suppressed more) but still not what I would call great and still a rather small listen window. 1 K was noticeably better than the others. Voice was more convincing, imaging was solid, and micro detail was better, noticeably better. AND the listening window was no contest, its huge.
As for the 750 hz cross on the 4 in. Voice seemed a bit more believable than at 1khz. That's about it.
Even went ahead and tried 6 db filters. Response of drivers allows in this case although cannot drive much above 95db. At the low 1 khz cross it was interesting. At 2 or 3 k it was not good at all BUT at 1 k it could pull it off well and I see what the fuss is about with 6 db. Im not sure it was “better” than 12 BUT it was different in a good way.
Conclusion from this super scientific test… I in no way heard a muffled slow sounding bottom end from the lower crossovers as compared to higher. In fact it was better IMO.
BTW just as a note....
there is about 7 inches between driver centers. At 3k cross thats 2 wave lengths, at 2k cross its 1 wave length, at 1 k cross thats 1/2 wave length, at 750hz cross its about 3/8 wave length.
I mention this because once I got down to about 1/2 wave length ( 1k hz cross) is when the image was stabil with larger head moves.
Just curious which type of music you have tried? If played with large scale music like a symphony, I wonder how the tweeter can cope with something like when the mass strings are played in unison?
Last edited:
Good question and that test is up next. The first protos had an issue with long sine sweep power tests in that the polyester film would shrink at about 105 db peak. We have sense moved to kapton. I mention the "long sine sweep" as it hangs in the trouble area that you have mentioned long enough to build some real heat, but I doubt its long enough to simulate massed string energy
This test absolutely shreds traditional ribbon constructions. Mine are a bit different and I cannot go into details BUT they have better motional control and a unique design difference that results in an order of magnitude reduction of 3rd harmonic as we get down to the lower crossover points. As well they have better heat dissipation than typical plastic backed designs. We will see......
Well, I guess it depends a lot on the drivers you're using. I have no scientific theory to add here, sorry, but I can share my last crossover battle, hoping it can be of some use!
My latest project is built upon the SB Acoustics TW29 Beryllium tweeter, the MR16P 6.5" midbass and SEAS L26Roy with the passive radiator, a 3 way active system. I've never had so much difficulties in finding the "right" tweeter Xover frequency, blame it on the TW29 which I now know that it forgives nothing!! I was wondering what I may be doing wrong since crossing the tweeter the lowest seems to be the rule but doesn't work good for me.
I'm a happy owner of the most flexible tool in town for DIYers to design a speaker system and its crossover, the DEQX digital crossover. Sitting at my listening position, I can change the crossing frequency, the attenuation slope and the type of filter just by the hitting the "enter" key and immediately listen to the result ! Moreover, I can "sand" the speaker's frequency response by using the DEQX speaker calibration tool to have a corrected perfectly flat frequency response for each driver.
I did my tests from 1.8 KHz up to 4 KHz, from 12 db/octave to 96 db/octave and a great care was taken to have drivers levels adjusted. While all frequency response measurements at the listening position showed almost the same results, the resulting sound signature was quite different. In the 2 KHz crossing area, which seems to be recognized as THE right spot, in my system, the sound was really harsh, very "metallic" as we use to say, well, not natural at all. At the other end, in the 3 KHz area, while the sound was now natural with no harsh, I was now loosing details resolution. Who said that speaker design is the art of compromises ? Yes, it is !! So, I cut the pie in half, 2.5 KHz and after many hours of listening of various type of music, I can say that this is my "lowest": I lost the harshness without loosing the details resolution, Alleluia !
As a conclusion, I still don't exactly understand why the pro's says that crossing the lowest is THE way to go, but hey, I'm a simple weekend DIY enthusiastic speaker builder who realized that the best rule for me was to trust my ears and my personal tastes !
My latest project is built upon the SB Acoustics TW29 Beryllium tweeter, the MR16P 6.5" midbass and SEAS L26Roy with the passive radiator, a 3 way active system. I've never had so much difficulties in finding the "right" tweeter Xover frequency, blame it on the TW29 which I now know that it forgives nothing!! I was wondering what I may be doing wrong since crossing the tweeter the lowest seems to be the rule but doesn't work good for me.
I'm a happy owner of the most flexible tool in town for DIYers to design a speaker system and its crossover, the DEQX digital crossover. Sitting at my listening position, I can change the crossing frequency, the attenuation slope and the type of filter just by the hitting the "enter" key and immediately listen to the result ! Moreover, I can "sand" the speaker's frequency response by using the DEQX speaker calibration tool to have a corrected perfectly flat frequency response for each driver.
I did my tests from 1.8 KHz up to 4 KHz, from 12 db/octave to 96 db/octave and a great care was taken to have drivers levels adjusted. While all frequency response measurements at the listening position showed almost the same results, the resulting sound signature was quite different. In the 2 KHz crossing area, which seems to be recognized as THE right spot, in my system, the sound was really harsh, very "metallic" as we use to say, well, not natural at all. At the other end, in the 3 KHz area, while the sound was now natural with no harsh, I was now loosing details resolution. Who said that speaker design is the art of compromises ? Yes, it is !! So, I cut the pie in half, 2.5 KHz and after many hours of listening of various type of music, I can say that this is my "lowest": I lost the harshness without loosing the details resolution, Alleluia !
As a conclusion, I still don't exactly understand why the pro's says that crossing the lowest is THE way to go, but hey, I'm a simple weekend DIY enthusiastic speaker builder who realized that the best rule for me was to trust my ears and my personal tastes !
Im not sure the pros always say crossing low is the way to go. They would know that crossing low is a can of worms for the vast majority of domes. The pk pk travel necessary to go even a little lower can easily upset the tiny motor and or suspension and as u experienced the harsh will rule.
If a tweeter can do it then there are certainly rewards , but tiny motors, stiff suspensions, and limited surface area are BIG issues and small domes have to wrestle with these
If a tweeter can do it then there are certainly rewards , but tiny motors, stiff suspensions, and limited surface area are BIG issues and small domes have to wrestle with these
That's why waveguides exist. They enable dome tweeters to cross over lower than usual with lower distortion, they move acoustic center closer to the midwoofer, control off axis response, reduce diffraction and enables us to use woofers in their pistonic region. Not to mention improvement in imaging.
Only if done poorly.
Only if done poorly.
Last edited:
I achieved what are, in my opinion, most satisfactory results with Visaton WG148R. It is easily one of the best sounding loudspeakers i've made.
Seas L15RLY/P with Morel DMS30s in Visaton WG148R
Cabinet is just a tiny bit wider than waveguide so there is no trace of tweeter diffraction, crossover point is low enough to avoid as much as possible diffraction products of mid-woofer and low enough to avoid distortion peak in 5th harmonic.
Seas L15RLY/P with Morel DMS30s in Visaton WG148R
Cabinet is just a tiny bit wider than waveguide so there is no trace of tweeter diffraction, crossover point is low enough to avoid as much as possible diffraction products of mid-woofer and low enough to avoid distortion peak in 5th harmonic.
Last edited:
I find 1" dome tweeters work best between 2.2kHz ~2.7kHz. Anything below 2.2kHz, I switch to compression drivers and horns.
I think there's a lot of misunderstanding in this. If it's hifi dome tweeters, there's a limit as to how low it can go. Crossing a dome tweeter at 1,600Hz is not advisable.
Like your listening tests have shown, the dome tweeter becomes increasingly harsh below 2kHz.
I've made couple of speakers with compression drivers (B&C DE250-8, B&C DE500-8 and Selenium D220Ti) and a number of loudspeakers with dome tweeters with crossover points well bellow this arbitrary 2KHz. I never had problem with tweeter sound. I've used Morels, Vifa XT, Dynaudio and SB tweeters. Not a single problem.
Maybe it is implementation that is the real issue ?
Yep. And even when I used a 96 dB/octave filter that has all required to limit the excursion of the tweeter below the crossover point, the sound was still harsh. I saw that the crossover in the kit offered by SB acoustics using this tweeter is set to 3 KHz so, maybe I'm not that offroad
They don't beat CD's. Nothing, that i listened, does in my opinion. SB29 is just better sounding direct radiator tweeter from other tweeters i used. B&C DE500 sounded the best in 18" waveguide with lipover to me - crossed to 15" midwoofer at 950Hz. Sound that came out of that is something unique.
Last edited:
A well designed speaker should not suffer from that.
LOL ... XO is horrible no matter how good it is.
the ideal Xo should be a 7th order lol
You can hear 10hz difference in the XO point at 2Khz, 0.25R , I cannot even imagine Xo's at 1Khz... a real **** sound. Perfect though for loud rock concerts, led zeppelin, aero smith, deaf lepard, metallica, you are right.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.