What about drivers like the Full Range Dayton Reference 4" driver and the similar Full Range Tang Band? Bearing in mind that in this context, full-range is about 100hz to roughly 20khz.
On the Dayton RS-100 4/8, the off axis response doesn't really start getting flakey until about 4khz.
The Tang Band W4-1052SD specs don't show off axis response, so the best we can do is guess.
Another factor that some might not be aware of, and that is the Power Distribution across the frequency spectrum. According to Badmaieff & Davis - How to Build Speaker Enclosures (now out of print), the peak power band is between 250hz and 500hz, centered on 355hz. There is no reference as to the source of that information.
If the 250hz to 500hz band is running at 10 watts, then between 125hz and 250hz, the power is 4 watts. In the band between 500hz and 1000hz, the power is 2 watts. Farther out from the Peak Power Band, the power drops even more. At 63hz on the low end, the power is 2w. At 4000hz and above on the high end, the power is about 0.5watts. This is for full orchestral music. How that applies to Rock and Roll, I'm not sure.
That makes me wonder if the 250hz to 500hz range is more desirable because the sound is spread between two drivers, or if it is less desirable, in that it would be much wiser to cross above 500hz and leave this power peak to the Low-Bass driver that is more able to handle it. If you cross below 500hz, then that peak powered band, is push off onto the Mid-Bass, Mid-Range driver which is likely a smaller lower-powered driver.
Never ending complications.
Steve/bluewizard
On the Dayton RS-100 4/8, the off axis response doesn't really start getting flakey until about 4khz.
The Tang Band W4-1052SD specs don't show off axis response, so the best we can do is guess.
Another factor that some might not be aware of, and that is the Power Distribution across the frequency spectrum. According to Badmaieff & Davis - How to Build Speaker Enclosures (now out of print), the peak power band is between 250hz and 500hz, centered on 355hz. There is no reference as to the source of that information.
If the 250hz to 500hz band is running at 10 watts, then between 125hz and 250hz, the power is 4 watts. In the band between 500hz and 1000hz, the power is 2 watts. Farther out from the Peak Power Band, the power drops even more. At 63hz on the low end, the power is 2w. At 4000hz and above on the high end, the power is about 0.5watts. This is for full orchestral music. How that applies to Rock and Roll, I'm not sure.
That makes me wonder if the 250hz to 500hz range is more desirable because the sound is spread between two drivers, or if it is less desirable, in that it would be much wiser to cross above 500hz and leave this power peak to the Low-Bass driver that is more able to handle it. If you cross below 500hz, then that peak powered band, is push off onto the Mid-Bass, Mid-Range driver which is likely a smaller lower-powered driver.
Never ending complications.
Steve/bluewizard
Some did it in others threads. The 15" (JBL or not, curviline or not) will have off axis problem at such fhz you talk about !
Very good testimonies about 12" JBL xo at 1000 to 1200 max !
Good 8" to 10" also : B&C, Eminence, PHL, PD, AE, Celestion, Faital pro, Beyma.
One of the problem if my understanding is correct is to find a pro driver withe the precision of the best hifi driver... but with efficienty and not too high or too low mms, and, with a not too high Fs in relation to that : choosing a good trade offs in relation to the design choosed which seems easier here with vented if you don't want to use too large format like the 15" for such high fhz like 1000 hz !
Dr Gedlee testimonied some rare 15" can climb to 700 hz in him design; him second speaker use a 12". In relation to the polar map. They are sealed designs here.
With my por understanding I surmise paper cone fabrication and too hard surround (double fabric surround e.g. to be a problem : breaks up of the material for the cone at some powers, distorsions with the surrounds for micro dynamics...). Anybody tested such drivers with modern laser laboratory Tools for the cone interactions ??? And what sort of motor and BL and coil length for a good trade offs between macro and micro dynamics to allow hifi détails with the dynamics and fast transcient with impacts than the range say below 1000 hz is needing ? Some says : Onken loads are maybe one of the answer or VOT but just 80 hz to 1000 hz for mid bass for this last one !
I saw some good testimonies of Tinitus - (RIP
... this guy was very nice and civilized, re read him posts...) about B&C 8 to 10" drivers when pro was needed !
Very good testimonies about 12" JBL xo at 1000 to 1200 max !
Good 8" to 10" also : B&C, Eminence, PHL, PD, AE, Celestion, Faital pro, Beyma.
One of the problem if my understanding is correct is to find a pro driver withe the precision of the best hifi driver... but with efficienty and not too high or too low mms, and, with a not too high Fs in relation to that : choosing a good trade offs in relation to the design choosed which seems easier here with vented if you don't want to use too large format like the 15" for such high fhz like 1000 hz !
Dr Gedlee testimonied some rare 15" can climb to 700 hz in him design; him second speaker use a 12". In relation to the polar map. They are sealed designs here.
With my por understanding I surmise paper cone fabrication and too hard surround (double fabric surround e.g. to be a problem : breaks up of the material for the cone at some powers, distorsions with the surrounds for micro dynamics...). Anybody tested such drivers with modern laser laboratory Tools for the cone interactions ??? And what sort of motor and BL and coil length for a good trade offs between macro and micro dynamics to allow hifi détails with the dynamics and fast transcient with impacts than the range say below 1000 hz is needing ? Some says : Onken loads are maybe one of the answer or VOT but just 80 hz to 1000 hz for mid bass for this last one !
I saw some good testimonies of Tinitus - (RIP
... this guy was very nice and civilized, re read him posts...) about B&C 8 to 10" drivers when pro was needed !
Last edited:
I don't use or give "testimonials". I use measured data.
All drivers "breakup" some lower than others. Light cones usually breakup sooner than heavier well structured ones.
When you say "off axis problems" do you just mean that the directivity is narrowing? That is not necessarily a problem! I use this fact to significant advantage.
All drivers "breakup" some lower than others. Light cones usually breakup sooner than heavier well structured ones.
When you say "off axis problems" do you just mean that the directivity is narrowing? That is not necessarily a problem! I use this fact to significant advantage.
Sorry, problem of word to word translation from the french to americaglish... In french, the word would mean your "thought" or "own tesmimonials", maybe the nearer reversed English word could be according the sense of "testimonial" translated in french in the dictionnary: "your feedback" (whatever datas or not...). Your "Témoignage" in french could be: "your writted Feedback" (as you talk also of datas...) !
Awsome... I just maid my first reverse enginery here ! (joke)
Yes, I remenmber you or others said that, but what are the datas in relation to the Sd/mms paper material ? It seems paper breaks up very lower in frequencies than all other materials (?) and even if higher fhz are beaming more from trhe center than the middle or surround of the cone, my understandings interactions are bad (?)
What could be a good Newton point here to choose in relation to that problem ? Is a paper 2g mms of a 8" ; 98 DB à la Audax PR17Z0 (discontinued) with foam suround a bad driver for 200 to 700-1000 hz ?
Exactly. What should you advise about constant directivity in relation to degrees when a driver (not the speaker) is measured with IEC plan : e.g. same curve between axial axis to 30° ? (about choosing the driver) then less important if curves are colapsing at 45° or 60° etc...
Or the same at listening distance (not position, not speaking about sweet point) ? Do you like to the result about your research a 0 to 30° constant directivity (3D polar map : horizontal & vertical) e.g. then the lower possible after ? (I would say in french : what patternity about your patern ?
I understood also than you don't worth about surrounding cloth fabric of the cones about pro drivers in a sealed load... it seems to be a minor problem in relation to the other upper ranked factors you choose.
Last edited:
Can no longer say where it's referenced, but Bell Labs already had a large enough anechoic chamber to record a 75 piece orchestra at the time the rush was on to develop cinema sound, so is where the data originated AFAIK and was no doubt instrumental in its quick development that has fundamentally changed hardly at all since then.
AFAIK, once development began on a two way system, a 300 Hz XO was apparently chosen to ameliorate the size Vs performance of lower Vs higher XOs and having had access to some 300 Hz XO multi-cells for awhile Vs my then smaller 500 Hz ones, the choice was clear for me that like with bass-bins, 'BIB' [bigger is better] rules.
I can only guess why these were relatively quickly relegated to pure PA apps and 500 Hz became the norm for cinema, etc., apps regardless of horn size and ultimately disappeared from the catalog.
GM
The crossover must have moved up because of the inability of the horn to control the directivity down to 300 Hz. When this happens the directivity get very narrow and then widens significantly (as F > lower). Crossing higher, where the horn works better alleviates this problem - but not completely unless the horn is still very large. It takes some 24" of width in the horn to control the directivity to 500 Hz. Doable, but not practical.
I do not think that one can say with certainty that there is any fixed relationship between SD, MMS and performance at HFs. It depends on too many other variables that are far more difficult to track. To me if the driver measures smooth and near flat up to the crossover, has a smooth role off off-axis then it is a "good" driver - so long as the more obvious stuff like good VC radius and length, BL, rigid frame, etc. are in order. These later factor are usually obvious from the spec sheet. But that's where spec sheets usually stop. They never show a detailed set of polars so it is impossible to determine a good driver from these sheets. You have to get one and measure it, which is what I do. And more often than not the FR shown on a spec sheet is not what I measure anyways. In many cases they are pure fiction.
I think there is little advantage to having a directivity (to -6 dB) less than 90°, but a significant disadvantage if you go much above 90°. From my work 90° seems to be the sweet spot. I could live with less, but not more.
Last edited:
Interestingly JBL deliberately chose a 120° horizontal beamwidth for its M2, maintained from about 700Hz up to 8kHz.
Olive hasnothing but praise for them, measurements look impressive (no polar though), and for what it is worth I have yet to read a negative review on them ("best speakers I have ever heard" being the most frequent citation).
Attachments
For what it is worth, I can provide the Power Distribution chart. This seems to have been confirmed by a couple other people from other source, one was Bell Labs, and the other was Altec Lansing.
I'm still torn on the Low/Mid crossover. Would it be an advantage to cross at 355hz because that peak load would be shared by two drivers, or would it be better to cross above that peak at 500hz or higher thereby pushing the peak load onto the driver that can most handle it.
Also, isn't a bit of stating the obvious to say you crossover in the 1khz to 4khz range. Isn't that the range where virtually every speaker out there crosses over?
Also, the thread is 85 pages long, so I haven't read everything, but if you are not crossing over in the 1khz to 4khz range, then where are you crossing over?
In a 2-way system, in the 2khz to 3khz range is about the only place you can get Mid-Bass and Tweeters to meet. In a Bass/Mid/High 3-way system, the Low/Mid tends to be between 350hz and 1khz, and that puts the Mid/High in roughly the 3khz to 6khz range.
So to the basic question - Why crossover in the 1-4khz range? I think the only answer can be - where else? Circumstance, mainly the nature and limitation of drivers gives you no other choice.
It seems to me the real question, is - Where is the best/ideal place to crossover? I speak of the theoretical best place to crossover in both 2-way and 3-way designs. If we are not constrained by the limitations of drivers, and we further assume we must still use a 2-way or 3-way systems, where are the hypothetically ideal locations to cross over?
If there are idea locations to cross, it would seem wise for someone to make drivers that conformed with that ideal.
Just a few thoughts.
Steve/bluewizard
I'm still torn on the Low/Mid crossover. Would it be an advantage to cross at 355hz because that peak load would be shared by two drivers, or would it be better to cross above that peak at 500hz or higher thereby pushing the peak load onto the driver that can most handle it.
Also, isn't a bit of stating the obvious to say you crossover in the 1khz to 4khz range. Isn't that the range where virtually every speaker out there crosses over?
Also, the thread is 85 pages long, so I haven't read everything, but if you are not crossing over in the 1khz to 4khz range, then where are you crossing over?
In a 2-way system, in the 2khz to 3khz range is about the only place you can get Mid-Bass and Tweeters to meet. In a Bass/Mid/High 3-way system, the Low/Mid tends to be between 350hz and 1khz, and that puts the Mid/High in roughly the 3khz to 6khz range.
So to the basic question - Why crossover in the 1-4khz range? I think the only answer can be - where else? Circumstance, mainly the nature and limitation of drivers gives you no other choice.
It seems to me the real question, is - Where is the best/ideal place to crossover? I speak of the theoretical best place to crossover in both 2-way and 3-way designs. If we are not constrained by the limitations of drivers, and we further assume we must still use a 2-way or 3-way systems, where are the hypothetically ideal locations to cross over?
If there are idea locations to cross, it would seem wise for someone to make drivers that conformed with that ideal.
Just a few thoughts.
Steve/bluewizard
Last edited:
Steve,
I'll repeat what I recently wrote in post #19 here:
http://www.diyaudio.com/forums/multi-way/261440-comment-favor-when-bass-not-bass-2.html
If crossovers are properly designed for the transducers employed (assuming adequate bandwidth of the transducers), the result is a flat frequency and phase response through the crossover region, the crossover points undetectable either by measurement or hearing (as long as polar response is also consistent), regardless of the crossover frequencies chosen.
The "ideal crossover location" is transducer dependent, and also level dependent, what may be a completely acceptable transition at low SPL can fall apart at higher levels if the transducers large signal response is not adequate for the SPL desired.
Answers to your specific questions:
1) The average power distribution chart for orchestral music is not indicative of many forms of pop music, where peak SPL demands often are similar to the spectra of pink noise, equal peaks from 20 to 20 kHz. Average spectra of a particular musical genre will determine the average power needs for it's reproduction, but other than from a thermal standpoint (avoiding excessive voice coil heating, which changes multiple driver parameters), the spectra of the music is of little use in determining crossover frequencies.
2)No.
3)Anywhere from as low as 20 Hz to 1000 Hz, over a five octave range.
4) If we were not constrained by the limitations of transducers, crossovers would not be required. As previously stated, the ideal crossover frequencies are dictated by transducer (and horn/waveguide) limitations.
Art
Last edited:
Art,
1.) equal peaks from 20hz to 20khz - sorry but that seem unlikely, when I watch the bar graph display of music (FooBar), there is not that much content above about 12khz, and it seems to start tapering off about 4khz. Which makes sense as 4khz is about the highest fundamental. Most of the most intense content, though not exclusively, using Stevie Ray Vaughan as an example, seems to be below about 300hz or 400hz.
The question isn't that the most power demand in the low end, but rather WHERE in the low end? Knowing that it peaks substantially at 355hz, should be of value to speaker designers.
2.) Crossover in 1k to 4k range - Where else? - NO - thanks for the detailed explanation. All have to do is look at the spec on anyone one of a vast majority of commercial speakers, and I can see the answer is YES, a vast majority of them cross in the 1k to 4k range. Not 100% but a very very substantial majority.
3.) It depends on what you are crossing over. You certainly do not cross tweeters between 20hz and 1khz. The range itself (1k to 4K) implies crossing something to a tweeter. It could be a Mid-Bass or it could be a Midrange, but it is definitely to a tweeter.
4.) I qualified by statement by saying that the result would be a 2-way or 3-way design. Perfection would be a single driver that is clean and smooth from 20hz to 20khz, but we all know such perfection does not exits.
But we have two choice, we can find drives and work out what crossover will work for them, or we can consider what the best ideal crossover would be and try to find driver to match that ideal. My question was framed in the second option.
On the assumption that we can find appropriate drivers, what are the ideal crossover points for a 3-way system? That is the whole point of this discussion, to see if the ear naturally hears crossover points in the spectrum.
I've not done this test in a long time, and I generally find my hearing varies from day to day and from mood to mood. I suspect even the weather effects my hearing. But when I ran the test on several occasions before, roughly 320hz and 3200hz were where I heard the transitions.
I'm not saying you are wrong in general, but you seem to be ignoring the context of the discussion.
But ... thanks for your input, everything expands the discussion.
Steve/bluewizard
1.) equal peaks from 20hz to 20khz - sorry but that seem unlikely, when I watch the bar graph display of music (FooBar), there is not that much content above about 12khz, and it seems to start tapering off about 4khz. Which makes sense as 4khz is about the highest fundamental. Most of the most intense content, though not exclusively, using Stevie Ray Vaughan as an example, seems to be below about 300hz or 400hz.
The question isn't that the most power demand in the low end, but rather WHERE in the low end? Knowing that it peaks substantially at 355hz, should be of value to speaker designers.
2.) Crossover in 1k to 4k range - Where else? - NO - thanks for the detailed explanation. All have to do is look at the spec on anyone one of a vast majority of commercial speakers, and I can see the answer is YES, a vast majority of them cross in the 1k to 4k range. Not 100% but a very very substantial majority.
3.) It depends on what you are crossing over. You certainly do not cross tweeters between 20hz and 1khz. The range itself (1k to 4K) implies crossing something to a tweeter. It could be a Mid-Bass or it could be a Midrange, but it is definitely to a tweeter.
4.) I qualified by statement by saying that the result would be a 2-way or 3-way design. Perfection would be a single driver that is clean and smooth from 20hz to 20khz, but we all know such perfection does not exits.
But we have two choice, we can find drives and work out what crossover will work for them, or we can consider what the best ideal crossover would be and try to find driver to match that ideal. My question was framed in the second option.
On the assumption that we can find appropriate drivers, what are the ideal crossover points for a 3-way system? That is the whole point of this discussion, to see if the ear naturally hears crossover points in the spectrum.
I've not done this test in a long time, and I generally find my hearing varies from day to day and from mood to mood. I suspect even the weather effects my hearing. But when I ran the test on several occasions before, roughly 320hz and 3200hz were where I heard the transitions.
I'm not saying you are wrong in general, but you seem to be ignoring the context of the discussion.
But ... thanks for your input, everything expands the discussion.
Steve/bluewizard
Glad you mentioned that little detail of the polar response
Electrical filters are not my expertise, but I was under the assumption all causal filters cause a group delay. So you can get flat response but the impulse response will be smeared. Do you understand differently?
When I did thermal test on two-ways - FR at various levels - it was the crossover region that changed the most, followed by the tweeters overall response. But the crossover area went all-to-hell.
Except with a 1st order filter, which has zero phase shift and GD. Right ?
Isn't it plausible that the crossover region changed the most because that's where the tweeter is operating in its most "stressful" area in terms of voice coil thermal effects ?
It's entirely possible that I don't understand some deeper problem, but if I charged myself with designing a speaker that didn't cross over in this range, the first step would be to select a midrange driver that had good behavior in this range (1 - 4 khz) and then build the rest of the speaker around it, crossing outside of that range.
Maybe a 3" full range would make a good mid driver that dodges having to be crossed in that range, like a Fostex FF85WK? 500hz to 4500hz?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.