Be that as it may, I for one am quite exited about 18S products in general. These new PEN/Be 4"ers really seem to aim at the top of what is doable in any 1.4"/1.5" format compression driver. The fierce competition among the six(!) pro speaker companies from Italy for sure isn't a bad thing...
Hi
It would be hard to say what the best driver is, what one is doing with the driver at least partly governs that. For example, if ones target is the highest sensitivity, then one best reads this off a data sheet.
If your designing a loudspeaker you are going to use for music, one can put more qualifications on the target, for example, of one also wants something like flat response, then one finds what matters a lot is what the response / sensitivity between 10 and 20K looks like.
Don’t forget to account for the horn directivity each driver was measured on if possible.
With anything like a CD horn, below that top octave, you will be attenuating the driver to make the response flat anyway but you can’t fake or make it at 10-20k. Start by looking at a drivers response in terms of where the lf response is equal to that at 10-20K, the inverse of the shape above this applied as eq or in the crossover makes the response and phase flat.
All horns and drivers sound very different BUT if you go to the trouble to make them all have the same frequency response, on axis, they ALL sound vastly more similar to each other.
Only at the precise balance of narrowing directivity off setting the drivers falling power response will a horn be anything like flat on axis without compensation.
Since that narrowing directivity is also audible anywhere off axis and undesirable in most A – B listening tests, the constant directivity horn is now more common, especially where more than one person or listening position is desired.
Also, the size of the driver exit, even for a one inch exit is at 20Khz, is acoustically large enough to be defining the radiation pattern, in fact all of the acoustic transformation has taken place WELL before the one inch exit. So if one has a CD horn (where you want the highs to reach the pattern edge} a one inch driver is limited to about 60-80 degrees and a 2 inch about half that. Obviously one can use a much wider combination and sound will still radiate, but this produces an interference pattern / diffraction within the horn and beyond.
Lastly, 18sound makes good stuff but we use hundreds and hundreds of a particular driver at work because it has a high hf sensitively and it’s mechanical design allows it to have much more excursion than dome compression drivers (it can go down unusually low at low powers like hifi) .
Consider the BMS 4550, starting at a small diameter internally, it works very well in conical (CD) horns like we use (which require a partially spherical wave shape at the driver exit entering the horn and doesn't load the driver well at the lower end) .
Best,
Tom Danley
Danley Sound Labs
It would be hard to say what the best driver is, what one is doing with the driver at least partly governs that. For example, if ones target is the highest sensitivity, then one best reads this off a data sheet.
If your designing a loudspeaker you are going to use for music, one can put more qualifications on the target, for example, of one also wants something like flat response, then one finds what matters a lot is what the response / sensitivity between 10 and 20K looks like.
Don’t forget to account for the horn directivity each driver was measured on if possible.
With anything like a CD horn, below that top octave, you will be attenuating the driver to make the response flat anyway but you can’t fake or make it at 10-20k. Start by looking at a drivers response in terms of where the lf response is equal to that at 10-20K, the inverse of the shape above this applied as eq or in the crossover makes the response and phase flat.
All horns and drivers sound very different BUT if you go to the trouble to make them all have the same frequency response, on axis, they ALL sound vastly more similar to each other.
Only at the precise balance of narrowing directivity off setting the drivers falling power response will a horn be anything like flat on axis without compensation.
Since that narrowing directivity is also audible anywhere off axis and undesirable in most A – B listening tests, the constant directivity horn is now more common, especially where more than one person or listening position is desired.
Also, the size of the driver exit, even for a one inch exit is at 20Khz, is acoustically large enough to be defining the radiation pattern, in fact all of the acoustic transformation has taken place WELL before the one inch exit. So if one has a CD horn (where you want the highs to reach the pattern edge} a one inch driver is limited to about 60-80 degrees and a 2 inch about half that. Obviously one can use a much wider combination and sound will still radiate, but this produces an interference pattern / diffraction within the horn and beyond.
Lastly, 18sound makes good stuff but we use hundreds and hundreds of a particular driver at work because it has a high hf sensitively and it’s mechanical design allows it to have much more excursion than dome compression drivers (it can go down unusually low at low powers like hifi) .
Consider the BMS 4550, starting at a small diameter internally, it works very well in conical (CD) horns like we use (which require a partially spherical wave shape at the driver exit entering the horn and doesn't load the driver well at the lower end) .
Best,
Tom Danley
Danley Sound Labs
Hi Tom,
Your expert input is always appreciated!
That one specific driver you mention, being a non-dome, it then is a BMS not an 18S, right? (I'm not aware of annular designs other than BMS, except for some Faitals and the cone section of the B&C DCX50).
My understanding is that a 4" dome within a "classic" CD design like the new 18S will be able to move quite some air down low from sheer diaphragm area, and my hope is PEN will lower distortion around resonance while the Be (or TiN) will provide a clean top end. And 1.5" exit seems a good compromise, too.
Your expert input is always appreciated!
That one specific driver you mention, being a non-dome, it then is a BMS not an 18S, right? (I'm not aware of annular designs other than BMS, except for some Faitals and the cone section of the B&C DCX50).
My understanding is that a 4" dome within a "classic" CD design like the new 18S will be able to move quite some air down low from sheer diaphragm area, and my hope is PEN will lower distortion around resonance while the Be (or TiN) will provide a clean top end. And 1.5" exit seems a good compromise, too.
Hi Tom,
Which driver?
What has been your experience using the Community M4?
Regards,
Bill
Hi KSTR, Bill
The BMS 4550 is a non-dome driver and has a couple nice things going for it.
For one, the impedance peaks at the low end of it’s response are lower than most 1 inch drivers and this makes the high pass crossover easier if passive (like many of ours are).
Second though is something which may or may not be important depending how it’s used. Sound travels at the speed of sound of course and so if one examines the relationship of the radiator, phase plug and exit, one finds that these relationships can vary quite a bit.
That matters because except near breakup, the dome acts like a piston and so when you trace the path lengths from the dome through the phase plug, one finds that some drivers would produce a converging wave front at the exit, while others a plane wave and a few a diverging wave front.
With our horns, the goal is to make them act like they were a single point source driver and so at 20KHz where the exit would be having an effect on the pattern, that is reduced / eliminated by having a source that already has curvature to it.
That is partly why I use conical horns too, the sound does not have to change or refract like can happen if the horn wall angles changes too much / quickly before the pattern is established.
A simple conical horn driven by a point source (a source that is too small to have directivity) radiates a simple portion of a sphere.
That way, the entire angular distribution is more even, has no hf character or an interference pattern when you walk past it playing pink noise. A source that produces an interference pattern will have an audible to very audible combing with pink noise when you walk the pattern.
Hey Bill
The M-4 is an amazing driver, it is REALLY powerful and if I recall, could put out 100 acoustic watts continuously. I attended Bruce’s AES presentation on it’s design and construction in the old days, as well as a really nice guy, he had a real down to earth approach to the development of the radiator.
I remember the first speakers system I heard that used one, was able to play a snare hit louder than I had ever heard, enough to make me wince each time. When I plugged my ears, it sounded absolutely clear and clean, my ears were overloaded.
In the beginning of the Unity /Synergy horn adventure, I had looked at the driver as possibly the mid system but it didn’t lend itself to what I was doing.
It was acoustically easier and cheaper to use multiple small drivers less than ¼ wavelength apart.
At that point, the HF source strength became the limiting factor in system power until last year.
Say, several computers ago and many more years, I think I used to talk to you on an old horn forum. It was something to do with a spiral horn like a snail. When that computer crashed, I lost the bookmarks and the site. Does that ring a bell at all?
Best,
Tom
The BMS 4550 is a non-dome driver and has a couple nice things going for it.
For one, the impedance peaks at the low end of it’s response are lower than most 1 inch drivers and this makes the high pass crossover easier if passive (like many of ours are).
Second though is something which may or may not be important depending how it’s used. Sound travels at the speed of sound of course and so if one examines the relationship of the radiator, phase plug and exit, one finds that these relationships can vary quite a bit.
That matters because except near breakup, the dome acts like a piston and so when you trace the path lengths from the dome through the phase plug, one finds that some drivers would produce a converging wave front at the exit, while others a plane wave and a few a diverging wave front.
With our horns, the goal is to make them act like they were a single point source driver and so at 20KHz where the exit would be having an effect on the pattern, that is reduced / eliminated by having a source that already has curvature to it.
That is partly why I use conical horns too, the sound does not have to change or refract like can happen if the horn wall angles changes too much / quickly before the pattern is established.
A simple conical horn driven by a point source (a source that is too small to have directivity) radiates a simple portion of a sphere.
That way, the entire angular distribution is more even, has no hf character or an interference pattern when you walk past it playing pink noise. A source that produces an interference pattern will have an audible to very audible combing with pink noise when you walk the pattern.
Hey Bill
The M-4 is an amazing driver, it is REALLY powerful and if I recall, could put out 100 acoustic watts continuously. I attended Bruce’s AES presentation on it’s design and construction in the old days, as well as a really nice guy, he had a real down to earth approach to the development of the radiator.
I remember the first speakers system I heard that used one, was able to play a snare hit louder than I had ever heard, enough to make me wince each time. When I plugged my ears, it sounded absolutely clear and clean, my ears were overloaded.
In the beginning of the Unity /Synergy horn adventure, I had looked at the driver as possibly the mid system but it didn’t lend itself to what I was doing.
It was acoustically easier and cheaper to use multiple small drivers less than ¼ wavelength apart.
At that point, the HF source strength became the limiting factor in system power until last year.
Say, several computers ago and many more years, I think I used to talk to you on an old horn forum. It was something to do with a spiral horn like a snail. When that computer crashed, I lost the bookmarks and the site. Does that ring a bell at all?
Best,
Tom
Member
Joined 2003
Interesting tidbit. I'm just curious, how did you find out that the BMS emits a convex wave (vs. planar)? In their patent (#5,878,148) they state that a planar wavefront is one of the design features.
BTW, is there any DSL installation in Germany or nearby one should check out?
Hi Tom,
Thanks for responding. My posts usually are not well received here; so, seldom do I expect a positive reply.
I used to post and contribute software to Daniel Ciesinger's website. He is known for his Cornu Spiral Horn design that back-loads Whizzer-Cone full range drivers such as those made by Lowther, Fostex and others. The spiral is also used by Le Cléac'h in is horn design. The important characteristic here is that curvature is a function of curve path length. So, as acoustic intensity declines boundary curvature is increased and diffraction is minimized as the acoustic wave propagates beyond a curled-back horn boundary.
The M4 is the only reasonably priced large format (diaphragm dia. > 4”) compression driver still around. The others are made by Goto Unit and ALE and are offered at esoteric prices that reflect the use of large Be diaphragms and Permendur in their magnetic circuits. I like this driver, as it permits dividing the audio spectrum into three, one-decade segments for horn loading.
Apparently intense acoustic energy emanating from the surrounding low frequency drivers has only a nominal impact on the gossamer diaphragm of the adjoining high frequency compression driver in your synergy horn. I suspect this problem is mitigated by the stiffness of the small volume of air trapped in the rear chamber of that driver.
Regards,
Bill
Hi Tom
what are the possibly excursion of annular diaphragms?
and annular diaphragm area seems less than dome (for equal coil diameter)
Hi Bill
I don’t know why your posts would be unwelcome, what I remember was you were a horn guy with much experience. That is why I wondered if you were the fellow I remembered.
You are spot on about the website, now that you mention the name, that is the one I remember and had lost in an old computer when the hd crashed.
I can see why you like the M-4, using it so far under it’s maximum output, It would be effortless, no power compression, very very low distortion. Most loudspeaker ickyness gets louder faster than the desired signal so headroom or margin is extremely sensible. For me though, it is still a pretty expensive driver haha.
There are actually a couple things that enter into the pressure distribution /how the drivers feel each other.
If one were producing a flat frequency spectrum of some signal, one finds the pressure within the horn increases proportionally with decreasing area. So, the hf driver radiator is where the pressure is highest.
If that were an area of say 1/10 sq inch, then when the horn area has increased to 10 sq inches, the pressure in psi has fallen by 10 (or -20dB spl).
In the case of the SH-50 (that I can walk over and stick a ruler into), where the mid drivers couple into the horn, the horn is about 3.25 by 3.25 inches and so the operating pressure here is much less than at the hf drivers radiator. For the lower frequency drivers, they couple into the horn where it is about 8.5 by 8.5 inches square and so their operating pressure would be about 14dB less than where the mid drivers are.
As you suggest too, the rear volume comes into play as well, for a horn loaded driver, one wants the suspension / box compliance to push the Fs to a point somewhat above the desired low corner. That allows the throat reactance to push the resistive corner downward (that reactance annulling).
For the mid drivers, they are (4 drivers ) already a sealed back driver and the low drivers are normally a sealed box or Tapped horn.
Electrically, most of our horns start life with a passive crossover so one last step is to use a parallel L across the mid and hf drivers as the dominating high pass . Having a progressive short appears across the VC as the frequency falls, makes the driver well damped and harder to move below it’s operating band.
Anyway, thanks for the “Cornu” hint, I tried a bunch of times to remember what it was but never found it.
Hi Flaesh
I am not sure what the mechanical limits are but I can say that the BMS driver in place of the one I used before that (a very good driver) would tolerate more low frequency content and where it’s phase rotated (important in the crossover) was unusually low for a high power driver with a sensitive top end.
Best,
Tom Danley
Danley Sound Labs
I don’t know why your posts would be unwelcome, what I remember was you were a horn guy with much experience. That is why I wondered if you were the fellow I remembered.
You are spot on about the website, now that you mention the name, that is the one I remember and had lost in an old computer when the hd crashed.
I can see why you like the M-4, using it so far under it’s maximum output, It would be effortless, no power compression, very very low distortion. Most loudspeaker ickyness gets louder faster than the desired signal so headroom or margin is extremely sensible. For me though, it is still a pretty expensive driver haha.
There are actually a couple things that enter into the pressure distribution /how the drivers feel each other.
If one were producing a flat frequency spectrum of some signal, one finds the pressure within the horn increases proportionally with decreasing area. So, the hf driver radiator is where the pressure is highest.
If that were an area of say 1/10 sq inch, then when the horn area has increased to 10 sq inches, the pressure in psi has fallen by 10 (or -20dB spl).
In the case of the SH-50 (that I can walk over and stick a ruler into), where the mid drivers couple into the horn, the horn is about 3.25 by 3.25 inches and so the operating pressure here is much less than at the hf drivers radiator. For the lower frequency drivers, they couple into the horn where it is about 8.5 by 8.5 inches square and so their operating pressure would be about 14dB less than where the mid drivers are.
As you suggest too, the rear volume comes into play as well, for a horn loaded driver, one wants the suspension / box compliance to push the Fs to a point somewhat above the desired low corner. That allows the throat reactance to push the resistive corner downward (that reactance annulling).
For the mid drivers, they are (4 drivers ) already a sealed back driver and the low drivers are normally a sealed box or Tapped horn.
Electrically, most of our horns start life with a passive crossover so one last step is to use a parallel L across the mid and hf drivers as the dominating high pass . Having a progressive short appears across the VC as the frequency falls, makes the driver well damped and harder to move below it’s operating band.
Anyway, thanks for the “Cornu” hint, I tried a bunch of times to remember what it was but never found it.
Hi Flaesh
I am not sure what the mechanical limits are but I can say that the BMS driver in place of the one I used before that (a very good driver) would tolerate more low frequency content and where it’s phase rotated (important in the crossover) was unusually low for a high power driver with a sensitive top end.
Best,
Tom Danley
Danley Sound Labs
K Tom, you are going to have to elaborate a little more about the L across the mid and HF thing...lol. On the couple of the other crossovers that I have seen for some of your older horns, you have a resistor across the first cap in the XO of the HF driver that looks to do something similar for that driver...impedance shoots way up protecting the driver below the XO point.
Ring Radiators
Hi Tom,
Thanks for the detailed response. Comment on ring radiators follows:
BMS must have successfully addressed the problem of getting the inner and outer compliances in elastic balance ([Cms’o] = [Cms’i]). This is a challenge because [Cms] is the most difficult parameter to control in driver manufacture and the amount of suspension material is larger at the outer rim than at the inner rim where the compliances are anchored to the driver housing. The advantage of this arrangement; however, is that low frequency rocking motion exhibited by disk and dome radiators with a single compliance is almost entirely eliminated.
Regards,
Bill
Hi Tom,
Thanks for the detailed response. Comment on ring radiators follows:
BMS must have successfully addressed the problem of getting the inner and outer compliances in elastic balance ([Cms’o] = [Cms’i]). This is a challenge because [Cms] is the most difficult parameter to control in driver manufacture and the amount of suspension material is larger at the outer rim than at the inner rim where the compliances are anchored to the driver housing. The advantage of this arrangement; however, is that low frequency rocking motion exhibited by disk and dome radiators with a single compliance is almost entirely eliminated.
Regards,
Bill
This is a complete stab in the dark - I'm too busy to fire up any software, just doing this off the top of my head:
But here's my take on Tom's statement -
When you high pass a tweeter, you can use a capacitor in series, or an inductor in parallel. Either one works; the inductor in parallel is just something you don't see every day.
So - from a frequency response standpoint, they're basically equivalent.
But what happens in the time domain?
The phase of a speaker (any speaker) begins to increase as we approach the resonant frequency. Since the inductor 'looks' like a short circuit as frequency gets lower and lower, the overall impedance of the system (speaker and inductor in parallel) is lower than just the speaker by itself.
Which begs the question - does this flattening of the impedance peak also flatten out the phase response?
I can't answer that without simulating it, but I'm sure someone here will know off the top of their head.
Hi Bill, Patrick, Winslow
Bill I had not given it any thought but I expect your right. I have tried several times to make a simple flat disk passive radiator with only edge suspension. Even though they are uniformly pressure driven, when the level was high enough, they went crazy in a rocking mode. A spider and cardboard tube always fixed it but on a dome compression driver, one has no other point one can connect to. The annular radiator suspension is clamped at the outside and inside leaving a center part to move.
Patrick you have the answer. The choke in parallel or a cap in series do the same thing, one produces a falling impedance with frequency and the other an increasing impedance with increasing frequency.
Since I need (usually) more than a first order slope to eq the high end to be flat, that falling impedance approach does two things, it swamps the impedance curve features which produce peaks and dips normally and makes the driver much harder to move by external air pressure at lower frequencies.
That plus the sealed rear volumes makes the upper drivers pretty immune to the lower pressure the drivers below it produce.
The actual pressure in the horn is proportional to the area so by far the highest air pressure is at the throat inside the compression driver and the lowest, at the woofers.
A crossover hint.
Next time you have your software open, model two band pass filters to make a pretend upper and lower speaker.
Examine the phase response for each. Notice that the phase doesn’t approach zero until quite a ways above the low corners.
Our style xover ideally ends with a flat phase response like there was no crossover at all. So, what one can see is that if one were going to “shove” the two bandwidths closer together to knit them so that the phases were say around 90 degrees apart , that you have to be past the low corner of the upper driver.
Also the less phase shift the corner has (shallower) the less rapidly the phase is changing.
A changing phase is changing time.
This is why I have mentioned the lower response corner and phase of the bms driver, it makes it easier to do.
If you examine the group delay for the sum and displace the upper driver to the rear, one can justify the GD to be equal.
When you have the right electrical part and the right physical offset, it looks acts sounds like one driver covering that band.
I can’t tell you what slopes to use because you end up with something with no name, it has to be adapted to the acoustic magnitude and phase responses of the sources.
I can say that usually I can get away with something reaching a 4th order or higher electrical high pass for the compression driver.
Best,
Tom Danley
Bill I had not given it any thought but I expect your right. I have tried several times to make a simple flat disk passive radiator with only edge suspension. Even though they are uniformly pressure driven, when the level was high enough, they went crazy in a rocking mode. A spider and cardboard tube always fixed it but on a dome compression driver, one has no other point one can connect to. The annular radiator suspension is clamped at the outside and inside leaving a center part to move.
Patrick you have the answer. The choke in parallel or a cap in series do the same thing, one produces a falling impedance with frequency and the other an increasing impedance with increasing frequency.
Since I need (usually) more than a first order slope to eq the high end to be flat, that falling impedance approach does two things, it swamps the impedance curve features which produce peaks and dips normally and makes the driver much harder to move by external air pressure at lower frequencies.
That plus the sealed rear volumes makes the upper drivers pretty immune to the lower pressure the drivers below it produce.
The actual pressure in the horn is proportional to the area so by far the highest air pressure is at the throat inside the compression driver and the lowest, at the woofers.
A crossover hint.
Next time you have your software open, model two band pass filters to make a pretend upper and lower speaker.
Examine the phase response for each. Notice that the phase doesn’t approach zero until quite a ways above the low corners.
Our style xover ideally ends with a flat phase response like there was no crossover at all. So, what one can see is that if one were going to “shove” the two bandwidths closer together to knit them so that the phases were say around 90 degrees apart , that you have to be past the low corner of the upper driver.
Also the less phase shift the corner has (shallower) the less rapidly the phase is changing.
A changing phase is changing time.
This is why I have mentioned the lower response corner and phase of the bms driver, it makes it easier to do.
If you examine the group delay for the sum and displace the upper driver to the rear, one can justify the GD to be equal.
When you have the right electrical part and the right physical offset, it looks acts sounds like one driver covering that band.
I can’t tell you what slopes to use because you end up with something with no name, it has to be adapted to the acoustic magnitude and phase responses of the sources.
I can say that usually I can get away with something reaching a 4th order or higher electrical high pass for the compression driver.
Best,
Tom Danley
See, I did that once in a simulation and thought it was a good idea...but I had never seen anyone else doing it, so I scrapped that crossover (was couple years back and not horns).
Tom, I thank you for the tips...and I wish I could make the drive down from Charlotte and pick your brain once I get everything built.
back to work out v4 of my crossover...darn it. At least I can have something to do while waiting on the fiberglass to dry.
Tom, I thank you for the tips...and I wish I could make the drive down from Charlotte and pick your brain once I get everything built.
back to work out v4 of my crossover...darn it. At least I can have something to do while waiting on the fiberglass to dry.
- Status
- Not open for further replies.