I am combining a pro 15 inch driver to a 1.4" compression driver, but I'm crossing as low as possible (400Hz). See the photo.
The drivers are Nippon Gakki (Yamaha) with a very unusual suspension. See the patent:
http://www.google.com/patents?id=DjcsAAAAEBAJ
These are very smooth drivers with no measureable peaking, but they have almost no output above 12kHz. Properly EQ'd (a broad notch filter to boost the less sensitive areas below 600Hz and above 5kHz), they are very flat sounding on the Azurahorns. The directionality may be a little troubling. I'm still not sure about it, but there are other strengths.
The soundfield these things create is astonishing at times. Depth, width, height...you get it all.
Also if something is on the recording above 400Hz, you're hearing it. The LF isn't quite as good, but the chipboard boxes were rescued from the rubbish bin. Properly sized and constructed, plywood cabinets are next. The bass drivers are Altec 515-16G.
The drivers are Nippon Gakki (Yamaha) with a very unusual suspension. See the patent:
http://www.google.com/patents?id=DjcsAAAAEBAJ
These are very smooth drivers with no measureable peaking, but they have almost no output above 12kHz. Properly EQ'd (a broad notch filter to boost the less sensitive areas below 600Hz and above 5kHz), they are very flat sounding on the Azurahorns. The directionality may be a little troubling. I'm still not sure about it, but there are other strengths.
The soundfield these things create is astonishing at times. Depth, width, height...you get it all.
Also if something is on the recording above 400Hz, you're hearing it. The LF isn't quite as good, but the chipboard boxes were rescued from the rubbish bin. Properly sized and constructed, plywood cabinets are next. The bass drivers are Altec 515-16G.
Attachments
CD in rooms
Hi
ScottG, I very much like how you pinpoint the issue of radiating patterns especially in the crossover region I completely agree.
Constant directivity is a must when you want to cover a audience. I once designed a sound system around the electro voice coax MH640 augmented by closed cabinet (!) bass units intended to cover an area of 30 m x 70 m and around 1000 – 1500 persons.
http://www.electrovoice.com/products/167.html
"http://www.electrovoice.com/products/167.html"
after a few month of modding, I was able to get some pretty refined sound from that brute force units. A lot of positive feedback and no complains from around 100 000 customers since make me feel confident that I have reached my goal.
Coax design definitely helped here, though it is not obvious or easy to explain that just very few degrees of deviation from an non-coax setting should make such a big difference. Even if you consider asymmetric directivity with certain crossover types and others.
Above is one extreme, the other end of the spectrum is in room loudspeaker behaviour optimised for a small listening area. This has much more to do with the perception of acoustic events and is nicely outlined on the page of Sigfried Linkwitz and his ORION++
I think SL got the point to a very far extent in this specific enviroment.
Greetings
Michael
Hi
ScottG, I very much like how you pinpoint the issue of radiating patterns especially in the crossover region I completely agree.
Constant directivity is a must when you want to cover a audience. I once designed a sound system around the electro voice coax MH640 augmented by closed cabinet (!) bass units intended to cover an area of 30 m x 70 m and around 1000 – 1500 persons.
http://www.electrovoice.com/products/167.html
"http://www.electrovoice.com/products/167.html"
after a few month of modding, I was able to get some pretty refined sound from that brute force units. A lot of positive feedback and no complains from around 100 000 customers since make me feel confident that I have reached my goal.
Coax design definitely helped here, though it is not obvious or easy to explain that just very few degrees of deviation from an non-coax setting should make such a big difference. Even if you consider asymmetric directivity with certain crossover types and others.
Above is one extreme, the other end of the spectrum is in room loudspeaker behaviour optimised for a small listening area. This has much more to do with the perception of acoustic events and is nicely outlined on the page of Sigfried Linkwitz and his ORION++
I think SL got the point to a very far extent in this specific enviroment.
Greetings
Michael
Hi Michael,
that's great, if Ulrich Mueller implements the Linkwitz style stored energy stimulus. I had thought if this for a while, to just write him and ask for it, because he already had a sine burst implemented, but without options for windowing it with a cosine. Another thing would have been to allow not just saving of stimuli, but using stimuli from external files (so one could create any stimulus and use it in AT).
Yes, Audiotester is great in the sum of its features, I believe the remaining quirks are related to the interfacing to the myriads of soundcards. I especially like the curve operations and saving features. Plus, it's the cheapest shareware of them all.
Good to hear that you also didn't get the CSD to work in AT, makes me feel less incompetent
.
that's great, if Ulrich Mueller implements the Linkwitz style stored energy stimulus. I had thought if this for a while, to just write him and ask for it, because he already had a sine burst implemented, but without options for windowing it with a cosine. Another thing would have been to allow not just saving of stimuli, but using stimuli from external files (so one could create any stimulus and use it in AT).
Yes, Audiotester is great in the sum of its features, I believe the remaining quirks are related to the interfacing to the myriads of soundcards. I especially like the curve operations and saving features. Plus, it's the cheapest shareware of them all.
Good to hear that you also didn't get the CSD to work in AT, makes me feel less incompetent
.
MBK and all,
I have an excel worksheet (JoshK helped greatly) that depicts an OS profile combined with a circular transition to the baffle. User input can display any profile of interest. I'd be happy to share if you make an e-mail request.
It would be a good foundation for development to include the LeCleac'h or exponential transition to the baffle.
I have an excel worksheet (JoshK helped greatly) that depicts an OS profile combined with a circular transition to the baffle. User input can display any profile of interest. I'd be happy to share if you make an e-mail request.
It would be a good foundation for development to include the LeCleac'h or exponential transition to the baffle.
Thanks!
My instincts are confirmed, then. One potential awkwardness with the AH-340 Azurahorn is the sheer size - it's twice the depth and radius of the AH-550. By itself, not a huge objection (aside from price and shipping costs) - BUT, and it's a big but, I'd like the system to be reasonably close to time-aligned, without the tweeter being out a wavelength or more at crossover. The crossover design just goes a lot easier when the acoustic centers are close to each other.
It's a safe guess with a 1.2 kHz lowpass, the 12" driver will have an acoustic center about 3" behind the voice coil (based on a net 90-degree phase lag from the lowpass filter and driver acoustic response). This wouid be consistent with my measurements on the vaguely similar Klipsch Chorus. So that tells where the big compression driver is going to sit relative to the 12" woofer.
With an 8-inch-long AH-550, it overhangs the front baffle by a little bit, but not by much. With the AH-340 ... well, that's a different story. Now we have to contend with actual physical shading by the horn. This is the same mess as coaxials, maybe not as bad, but we have fool around with felt damping to get even reasonable performance from the woofer cone. The multiple close-in reflections are not what we want from the woofer, not in the 1 kHz range, no way.
The AH-340 can be moved further away, of course, but with a 1.2 kHz crossover with a 11.5" wavelength there isn't much room to move the horn too far away without getting into lobing trouble at the listening position. So it looks like the AH-550 is the strongest contender, assuming no throat profiles issues arise.
Oh, by the way, I've been thinking of the room energy profile. One of more clever uses of the MLS system is to compare the spectra of the direct sound (floor reflection absorbed by two feet of pillows) with all the room reflections summed over 1 Sec - and with the direct-arrival removed. In practice, I set the dividing line at 6 mSec, when the first set of early reflections started arriving.
This technique compares the first-arrival spectra with the total power into a sphere, since there are thousands of reflections coming from all directions when measuring over an interval of 1 Sec. See the pix below for the measurements of the Ariel. (Yes, I like flat-response loudspeakers.)
Since I've been discussing using one or two 18Sound XD125's pointing towards the rear to create a hypercardioid pattern, the MLS comparison method provides a quick way to get a read on the direct-arrival vs room-energy spectrum. The rear-facing tweeters should ideally be level-set in 1dB steps, with the option of individual adjustment of the Left and Right loudspeakers.
The reason I mentioned two XD125's is the option of pointing the 80-degree dispersion vertically, the 60-degree dispersion horizontally, and placing the XD125's 60 degrees apart. This would illuminate a swath 120 degrees wide by 80 degrees high - with switch-selectable level setting to precisely adjust the shape and size of the hypercardioid rear lobe.
One feature I've wanted for LONG time with MLSSA is the ability to edit out the floor reflection, and replace it with a straight-line (or better yet, splined) interpolation between two sets of visual edit points. Once upon a time I went through an extraordinarly tedious hand-edit of a MLSSA ASCII *.TIM file, and to my amazement, the edit worked quite well without generating obvious spectral artifacts.
When the floor reflection, typically at 3.5 mSec, is removed, you can open the window out to 6 to 8 mSec, when all the other reflections start arriving. The longer window size directly corresponds to greater resolution in the frequency domain, something very worthwhile. The waterfall (CSD) displays also get much more usable - with a 3 mSec window, they're almost useless.
As it is now, I have to build a 2-foot pile of pillows in order to get the floor reflection 20 dB down, so it doesn't interfere as much with the direct-arrival measurement. (A floor reflection at 6 to 10 dB generates quite a bit of comb-filtering, filling the measurement with lots of clutter that obscure real frequency data.)
ScottG said:
The short answer then is:
Option 1: If you want to push the crossover from 1.2 down to a limit of 900 Hz,(probably more like 960 Hz), with a 2nd order electrical - then you would spec. the larger horn. (Actually, without regard for the resonance of the driver, the Le Cleac'h roundover exit profile should be able to go quit a bit lower than the mid 900's.)
Option 2: If you are OK with a 2nd order from 1.2 to 1.3, AND want the benefit of a more extended off-axis response at higher freq.s, then go with the smaller horn.
Pair it with either of those 12" drivers and I'd pick option 2 every time.
Of course I'd still opt for a CD design if possible.
EDIT: I suspect the 15" tone tubby would work very well for your open baffle compensation.
My instincts are confirmed, then. One potential awkwardness with the AH-340 Azurahorn is the sheer size - it's twice the depth and radius of the AH-550. By itself, not a huge objection (aside from price and shipping costs) - BUT, and it's a big but, I'd like the system to be reasonably close to time-aligned, without the tweeter being out a wavelength or more at crossover. The crossover design just goes a lot easier when the acoustic centers are close to each other.
It's a safe guess with a 1.2 kHz lowpass, the 12" driver will have an acoustic center about 3" behind the voice coil (based on a net 90-degree phase lag from the lowpass filter and driver acoustic response). This wouid be consistent with my measurements on the vaguely similar Klipsch Chorus. So that tells where the big compression driver is going to sit relative to the 12" woofer.
With an 8-inch-long AH-550, it overhangs the front baffle by a little bit, but not by much. With the AH-340 ... well, that's a different story. Now we have to contend with actual physical shading by the horn. This is the same mess as coaxials, maybe not as bad, but we have fool around with felt damping to get even reasonable performance from the woofer cone. The multiple close-in reflections are not what we want from the woofer, not in the 1 kHz range, no way.
The AH-340 can be moved further away, of course, but with a 1.2 kHz crossover with a 11.5" wavelength there isn't much room to move the horn too far away without getting into lobing trouble at the listening position. So it looks like the AH-550 is the strongest contender, assuming no throat profiles issues arise.
Oh, by the way, I've been thinking of the room energy profile. One of more clever uses of the MLS system is to compare the spectra of the direct sound (floor reflection absorbed by two feet of pillows) with all the room reflections summed over 1 Sec - and with the direct-arrival removed. In practice, I set the dividing line at 6 mSec, when the first set of early reflections started arriving.
This technique compares the first-arrival spectra with the total power into a sphere, since there are thousands of reflections coming from all directions when measuring over an interval of 1 Sec. See the pix below for the measurements of the Ariel. (Yes, I like flat-response loudspeakers.)
Since I've been discussing using one or two 18Sound XD125's pointing towards the rear to create a hypercardioid pattern, the MLS comparison method provides a quick way to get a read on the direct-arrival vs room-energy spectrum. The rear-facing tweeters should ideally be level-set in 1dB steps, with the option of individual adjustment of the Left and Right loudspeakers.
The reason I mentioned two XD125's is the option of pointing the 80-degree dispersion vertically, the 60-degree dispersion horizontally, and placing the XD125's 60 degrees apart. This would illuminate a swath 120 degrees wide by 80 degrees high - with switch-selectable level setting to precisely adjust the shape and size of the hypercardioid rear lobe.
mige0 said:
One feature I've wanted for LONG time with MLSSA is the ability to edit out the floor reflection, and replace it with a straight-line (or better yet, splined) interpolation between two sets of visual edit points. Once upon a time I went through an extraordinarly tedious hand-edit of a MLSSA ASCII *.TIM file, and to my amazement, the edit worked quite well without generating obvious spectral artifacts.
When the floor reflection, typically at 3.5 mSec, is removed, you can open the window out to 6 to 8 mSec, when all the other reflections start arriving. The longer window size directly corresponds to greater resolution in the frequency domain, something very worthwhile. The waterfall (CSD) displays also get much more usable - with a 3 mSec window, they're almost useless.
As it is now, I have to build a 2-foot pile of pillows in order to get the floor reflection 20 dB down, so it doesn't interfere as much with the direct-arrival measurement. (A floor reflection at 6 to 10 dB generates quite a bit of comb-filtering, filling the measurement with lots of clutter that obscure real frequency data.)
Attachments
MBK said:
MBK - do you have a horizontal polar plot (synthetic or real) of the 500 Hz horn? That might provide some insight.
Now a 2" exit 500 Hz 120 degree constant horizontal coverage (constant from 7-800 Hz to 10-12 kHz) with a Le Cleac'h roundover would be awesome! (vertically restrictive to about 40 degrees resulting in a moderately low vertical profile to keep the CTC spacing reasonable, and reducing troublesome vertical room reflections.)
But we are all a bunch of really cheap SOB's at heart, and unless you have access to really cheap labor and materials, the fiberglass horns would be expensive (..and even then you have to consider shipping which would be costly).
Consider instead a light weight form that can be pressed to shape, uses cheap material, and is lightweight for shipping.
Perhaps a pressed thin-wall cellulose with a light coating of silicone (for an easy "release"). Form wall support can be a cheap high expansion foam (either "filled" by you or the customer). The purchaser could then go to any "home improvement" store and cheaply purchase a bucket, mixing drill attachment (assuming they have a power drill), and ready mixed cement. A little bit of water and mixing, some pouring, and a week + of curing and the customer would have one AWESOME non-resonant horn for a pretty low cost.
Manufacturing & Shipping costs could be VERY low and allow for quite a bit of profit. (..well at least when considering manufacturing expenses.)
Just a thought!
(..though in this instance it IS a self-serving thought.. )Re: Thanks!
Just thinking "out loud"..
I've often wondered what that would be like for the "main" horn itself.. i.e. an improved version of a multi-cell design with multiple compression drivers and horns. Of course matching dispersion would be critical to avoid combing. Still, I forsee a bit of a problem with such a design relating to acoustic center - i.e. where imaging starts to become "bloated" from a large apparent acoustic center (..not unlike the larger ribbon midbass pod for the MBL's).
Lynn Olson said:
Just thinking "out loud"..
I've often wondered what that would be like for the "main" horn itself.. i.e. an improved version of a multi-cell design with multiple compression drivers and horns. Of course matching dispersion would be critical to avoid combing. Still, I forsee a bit of a problem with such a design relating to acoustic center - i.e. where imaging starts to become "bloated" from a large apparent acoustic center (..not unlike the larger ribbon midbass pod for the MBL's).
Hi Ed,
I used JohnV's OS worksheet from this post (Patrick Bateman) for the OS part, and J-M LeCleac'h's worksheet from this page . I combined both profiels manually. Don't quite know how to integrate both, the LeCleac'h calculation is quite complex. Curious to see your worksheet, I'll PM you.
ScottG:
to clarify, the 500 Hz horn which I drew and which has been sent to production by my frieds, is a classic hypex LeCleac'h, not a CD / OS. No status on completion yet, so no measurements.
My own project a few posts above, the 120 degree coverage OS / LeCleac'h hybrid, is not in the physical works yet. I plan to ask a friend with a CNC machine for that, but he's been moving shop the past few weeks and so I felt shy to ask.
Making axially symmetric horns the easy way would be great, but how do you get the cellulose/foam etc, into a proper symmetrical 3D shape, with little deviation and roughness? If there was a way to make a mould in an easy way it would be nice to pour the horn in gypsum / alabaster (actually one of the LeCleac'h enthusiasts uses synthetic plaster, apparently the best he found during extensive testing). The thing one would have to get right is the passage from a profile cutout to a 3D form, say, rotating the profile in clay or such.
I used JohnV's OS worksheet from this post (Patrick Bateman) for the OS part, and J-M LeCleac'h's worksheet from this page . I combined both profiels manually. Don't quite know how to integrate both, the LeCleac'h calculation is quite complex. Curious to see your worksheet, I'll PM you.
ScottG:
to clarify, the 500 Hz horn which I drew and which has been sent to production by my frieds, is a classic hypex LeCleac'h, not a CD / OS. No status on completion yet, so no measurements.
My own project a few posts above, the 120 degree coverage OS / LeCleac'h hybrid, is not in the physical works yet. I plan to ask a friend with a CNC machine for that, but he's been moving shop the past few weeks and so I felt shy to ask.
Making axially symmetric horns the easy way would be great, but how do you get the cellulose/foam etc, into a proper symmetrical 3D shape, with little deviation and roughness? If there was a way to make a mould in an easy way it would be nice to pour the horn in gypsum / alabaster (actually one of the LeCleac'h enthusiasts uses synthetic plaster, apparently the best he found during extensive testing). The thing one would have to get right is the passage from a profile cutout to a 3D form, say, rotating the profile in clay or such.
MBK said:
Well basically you are making a "user" negative.
For a horn it would require 2 parts for the negative: 1. the exterior of the horn; 2. the interior of the horn.
Functionally the user pours the material, in an appropriate volume amount, into the exterior form. The interior form then gets placed "into" the exterior horn form cavity where the poured material rests. It compresses the poured material to fill in the surrounding air space and the additional amount purges from a "weep hole" from one of the negative portions. (Note: I've done this before.. not for a horn, but for a baffle. With cement the form requires "tapping" the sides of the form to evacuate any trapped air.)
That process isn't likely to be a lot different than for your "master" forms, but inverted.
In effect you need to create a positive (your horn..not really though, see below) from a very rigid/stable material that also will release from the material you are using for the user's form (..perhaps a cellulose mixture as I mentioned before). Lets say a polished and anodized aluminum. You will also need to create the negative surround to the positive that will encapsulate cellulose mixture.
Of course you are actually making 2 separate forms here one for each of the user's forms. Moreover considering support extensions and weep holes and factoring in uniform compression, you are likely looking at 3 form pieces for each of the 2 separate "user" forms. I.E 6 master form pieces of polished and anodized aluminum:
User form "exterior" composed of 3 master form pieces:
1. Positive exterior of horn with support
2. 1/2 of Negative surround of #1 with screw/vice compression clamping system
3. 1/2 of Negative surround of #1 with screw/vice compression clamping system
User form "interior" composed of 3 master form pieces:
(basically as above).
An alternative to aluminum for the master forms could be with UHMWPE, which is also easy to machine and doesn't need polishing or anodizing.
Cellulose in this instance is nothing more than paper and starch, which after compression and dry time will be extremely stable (..as long as its not around moisture).
You will of course also add in things like some support bracing allowing for "in-fill" of high expansion foam,(and its weep holes), for each section of the user form.
Finally, you will also need to factor in the small amount of silicone needed to coat the interior of both user forms.
And yes, it will require a LOT of thought to get it right. But probably not as much effort as it will to get the right horn profile (..unless you have a software package that will do it.. IF you know how to use Akabak, or will learn, (now free for download) - you could almost certainly do it with that.)
The beauty of it is though that once the master forms are made, (and you only need one set), labor *could* be cheap (..in that it requires no special knowledge to make the user forms), and the master forms would last as long as you cared to continue making the product.
The product would basically look like a box (from its exterior when the two user forms are together), and weigh maybe 5? times the weight of the cardboard box its shipped in.
It would be serious effort though, so don't entertain it unless you have a viable business model.
Oh.. forgot to mention - you have 2 potential markets here if you want it.. 1. for DIY'ers like us for the user forms and a restrictive agreement not to use commercially or available for resale/trade, 2. for making your own forms to use to make a horn for the commercial "pro sound" market (..likely from a pour-in polymer though).
Well, umm, I sort of imagined it would be complicated... this only makes sense for series production. Anything else it's probably best to go to a CNC shop and have the horn simply carved from wood.
Regarding mouldings - my question was actually simpler: how to produce a horn positive (the actual horn used to make the form) *without* access to a lathe or a CNC machine. Because if you do have access to one of those, it seems mightily simpler to just have a custom job done on a pair of "one-of-a-kinds". If there was a simple way to make the template, the outside box for the mould could then be poured from this original positive. I imagined one could laminate it from stacked plywood disks, and then filling in the "steps" with concrete or such, using a precise horn profile cut from sheet aluminium or wood, turned around the horn axis. But all of this still looks like a lot of trial and error...
Commercial production, I wouldn't dream of that. People will expect to buy this at say $50-100 (max), 80% of which will go to dealers and marketing. That is, *if* wou manage to find someone to sell that thing, the harder to find the more idiosyncratic it is. So anything less than a 1000 pcs. run wouldn't make sense.
Regarding mouldings - my question was actually simpler: how to produce a horn positive (the actual horn used to make the form) *without* access to a lathe or a CNC machine. Because if you do have access to one of those, it seems mightily simpler to just have a custom job done on a pair of "one-of-a-kinds". If there was a simple way to make the template, the outside box for the mould could then be poured from this original positive. I imagined one could laminate it from stacked plywood disks, and then filling in the "steps" with concrete or such, using a precise horn profile cut from sheet aluminium or wood, turned around the horn axis. But all of this still looks like a lot of trial and error...
Commercial production, I wouldn't dream of that. People will expect to buy this at say $50-100 (max), 80% of which will go to dealers and marketing. That is, *if* wou manage to find someone to sell that thing, the harder to find the more idiosyncratic it is. So anything less than a 1000 pcs. run wouldn't make sense.
I don't think you would try this without a CNC machine. You could CAD the thing with very thin (almost veneer like) sheets of plywood and spend an absurd amount of time tracing and cutting each sheet. 
Then, after gluing it together, spending even more time sanding down the edges and praying that you don't screw up the profile. The "laminated" approach isn't one I'd recommend though!
For a conical design its a LOT simpler, a lathe based system.. and it doesn't even have to be a real lathe. I've seen horns formed with plaster, a bucket, and the horn profile cut from the edges of a single board of plywood (that the user turned in the bucket to "route out" the horn.. true, but ).
Then, after gluing it together, spending even more time sanding down the edges and praying that you don't screw up the profile. The "laminated" approach isn't one I'd recommend though!For a conical design its a LOT simpler, a lathe based system.. and it doesn't even have to be a real lathe. I've seen horns formed with plaster, a bucket, and the horn profile cut from the edges of a single board of plywood (that the user turned in the bucket to "route out" the horn.. true, but
).
The worksheet has been forwarded to 3 more (happy, I hope) customers! It was developed in the interest of getting people on the same page, speaking of the same parameters, leading to a "finished" profile.
I recognize the difficulties involved in making one of these things...just consider the trouble getting to a consensus for what it looks like!
That being said, my toolbox is open. I've been a woodturner for years and intend to continue until I can't support the tools any longer. How big is this thing going to be?
It was developed in the interest of getting people on the same page, speaking of the same parameters, leading to a "finished" profile.I recognize the difficulties involved in making one of these things...just consider the trouble getting to a consensus for what it looks like!
That being said, my toolbox is open.
I've been a woodturner for years and intend to continue until I can't support the tools any longer. How big is this thing going to be?
I look at it this way:
For my own tastes I would build it from hardwood...I'm just nutty about natural materials....ok?
For multiples I see a 2 part negative form.
1) The horn that includes a stub tenon @ the throat and
2) A flange that recieves the stub tenon from the throat and
includes the mounting appointments for the driver.
The user/builder would work with their fiber/binder composite of choice. After lay-up the flange would be removed, allowing the form to be removed from inside the horn.
For my own tastes I would build it from hardwood...I'm just nutty about natural materials....ok?
For multiples I see a 2 part negative form.
1) The horn that includes a stub tenon @ the throat and
2) A flange that recieves the stub tenon from the throat and
includes the mounting appointments for the driver.
The user/builder would work with their fiber/binder composite of choice. After lay-up the flange would be removed, allowing the form to be removed from inside the horn.
Member
Joined 2003
Several people have written requesting test results for the 10NDA520, so...
For a fairly well known "reference" I included the Seas W26...besides I have another use for it coming up. Both drivers are centered on very similar 19" wide baffles, closed across the top, but open to the sides and back...so the baffles influence the low end and likely introduce a little diffraction on the top end. No smoothing, no baffle extensions, basically "un-fooled-around-with" driver measurements at 500mm.
First, the W26:
Violent breakup at 4k which is why I like to stay below 800Hz with these guys. However, good correlation with the Seas factory curves which have a reputation for brutal honesty.
Now the 10NDA520:
Again, good correlation with the 18Sound factory curves. Bottom line is that the 18Sound factory curves seem trustworthy.
For a fairly well known "reference" I included the Seas W26...besides I have another use for it coming up
. Both drivers are centered on very similar 19" wide baffles, closed across the top, but open to the sides and back...so the baffles influence the low end and likely introduce a little diffraction on the top end. No smoothing, no baffle extensions, basically "un-fooled-around-with" driver measurements at 500mm. An externally hosted image should be here but it was not working when we last tested it.
First, the W26:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Violent breakup at 4k which is why I like to stay below 800Hz with these guys. However, good correlation with the Seas factory curves which have a reputation for brutal honesty.
Now the 10NDA520:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Again, good correlation with the 18Sound factory curves. Bottom line is that the 18Sound factory curves seem trustworthy.
Nice data! Thanks for rolling out the cart and getting the measurements!
Usually I'd grumble about using a 85 by 115 cm IEC baffle with the driver off-center, but there appear to be almost no points of correlation between the two drivers, despite similar baffles. The Seas driver is certainly dominated by high-Q resonances - man, those won't be easy to filter out, a 24 dB/octave filter will just be getting started when it hits the big one at 4 kHz. (Back when I was wrestling with the KEF B139, I found it took 40 ~ 50 dB of filtering to audibly get rid of the severe resonance at 1.5 kHz.)
Usually I'd grumble about using a 85 by 115 cm IEC baffle with the driver off-center, but there appear to be almost no points of correlation between the two drivers, despite similar baffles. The Seas driver is certainly dominated by high-Q resonances - man, those won't be easy to filter out, a 24 dB/octave filter will just be getting started when it hits the big one at 4 kHz. (Back when I was wrestling with the KEF B139, I found it took 40 ~ 50 dB of filtering to audibly get rid of the severe resonance at 1.5 kHz.)
Jeff,
interesting design, surprising that a 1.4" would go that low. No excursion problems / THD / IMD? I sort of assumed around 800 Hz as the lower practical limit for a 1.4".
Ed,
thanks for the worksheet, will compare the data for similar design specs to see if there are any significant differences with my composite attempt. Not sure if I can patch the LeCleac'h formulas into it, they way he build his spreadsheet is quite complex (plenty of correction factors) and besides expansion ratio and wavefront shape you only really can choose the cutoff. So I had to iterate before I found values where diameter, angle and depth would line up nicely with the OS WG throat.
Paul,
impressive data, again! The 10NDA520 seems to set a very tight standard. BTW the Seas W26, is that the metal dome? The breakup certainly looks like it. If yes the result is even more impressive for the 18Sound, much less peaking *and* better CSD, and that from paper vs metal.
interesting design, surprising that a 1.4" would go that low. No excursion problems / THD / IMD? I sort of assumed around 800 Hz as the lower practical limit for a 1.4".
Ed,
thanks for the worksheet, will compare the data for similar design specs to see if there are any significant differences with my composite attempt. Not sure if I can patch the LeCleac'h formulas into it, they way he build his spreadsheet is quite complex (plenty of correction factors) and besides expansion ratio and wavefront shape you only really can choose the cutoff. So I had to iterate before I found values where diameter, angle and depth would line up nicely with the OS WG throat.
Paul,
impressive data, again! The 10NDA520 seems to set a very tight standard. BTW the Seas W26, is that the metal dome? The breakup certainly looks like it. If yes the result is even more impressive for the 18Sound, much less peaking *and* better CSD, and that from paper vs metal.
MBK said:
Altec used this size driver at 500Hz in it's A5 systems using 2nd order filters. They also made 300Hz horns for some of their 1.4" drivers. I haven't taken measurements of these things, but I'll eventually get around to it.
The Yamaha drivers I use have a very unusual suspension - see the patent I linked above - resulting in fantastic compliance and a reduction in side to side motion that allows a narrower gap and a surprisingly small magnet. Another feature of the diaphragm is that the suspension is formed by a photo-resistive etching process out of berylium copper and doesn't suffer from work hardening problems as the pressed diaphragms do.
Meyersound used these drivers in a couple of their loudspeakers. When Meyersound started making their own compression drivers, Yamaha stopped making these. I'm not sure Yamaha ever used them in their own loudspeaker products.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.