I am planning on building a 3-way loudspeaker based on Seas Excel drivers.
Using Seas T25CF-002 tweeter, sealed Seas W18-E001 mid bass and ported Seas W26FX-001 bass.
Madisound has a crossover drawing and measurements for a 3-way using exactly these Seas Excel drivers. The volume is specified, but not the cabinet drawings, baffle wide or driver placement.
Its a 0,75 Ft3 (21,2) Liter sealed mid bass, 3,0 Ft3 (85 Liter) ported for W26FX-001 with 3" diameter and 4" long port.
The measurements look very promising.
How do you think this would work? And any suggestions on the baffle?
I already have all the drivers needed for this project.
I have previously built the Seas Odin mk3 kit, and built two active sealed subs with Seas W26FX-001 in 46 liter enclosure. But now i want to try some other setup, as the one i have is not optimal.
The Odins bass output is almost the same as the sealed subs. And i belive the 3-way speakers would a much better approach.
Using Seas T25CF-002 tweeter, sealed Seas W18-E001 mid bass and ported Seas W26FX-001 bass.
Madisound has a crossover drawing and measurements for a 3-way using exactly these Seas Excel drivers. The volume is specified, but not the cabinet drawings, baffle wide or driver placement.
Its a 0,75 Ft3 (21,2) Liter sealed mid bass, 3,0 Ft3 (85 Liter) ported for W26FX-001 with 3" diameter and 4" long port.
The measurements look very promising.
How do you think this would work? And any suggestions on the baffle?
I already have all the drivers needed for this project.
I have previously built the Seas Odin mk3 kit, and built two active sealed subs with Seas W26FX-001 in 46 liter enclosure. But now i want to try some other setup, as the one i have is not optimal.
The Odins bass output is almost the same as the sealed subs. And i belive the 3-way speakers would a much better approach.
I would try a waveguide on the tweeter and cross to the w18 at 1500hz to avoid the 3rd order distortion peak at 1800hz. Should be a very nice system.
Have you tried contacting Madisound and ask them for the baffle dimensions used in this construction?
Then it's worth waiting for a reply from them. While you do so, have you studied Troels Gravesens designs to see if there is something similar? DIY-Loudspeakers
I have built a Seas 27TDFC/G, L15, L26 (poor mans excel) 3 way. I ended up with the L26 in a sealed 50L enclosure to keep the speaker size smaller with an F3 of ~ 43Hz. I still use my subwoofer. Vented they do model well in an 85 - 95L enclosure as you've already found out and would greatly reduce the need for sub. If I could afford the size, I would probably do vented.
Where you will struggle as Jonasz said is the mid - tweeter XO point. The 27TDFC/G and I also have 27TBFC/G sounded "strained" to me with traditional 4th order acoustic slopes below ~ 2KHz.
I ended up crossing over the L15 and 27TDFC/G at 2,200Hz with ~ 5th order acoustic slopes. This avoided the L15 3rd order breakup at ~ 2,700Hz. The W18 will be lower (take the primary breakup peak and divide by 3 to get the 3rd order peak frequency).
I don't know how the TC25 tweeter will handle < 2,000 Hz XO frequency. I wouldn't cross the W18 over higher than 1,500Hz as suggested by Jonasz. This would require a waveguide on the TC25.
You might be better off choosing another midrange that doesn't have a nasty breakup node. The W18 is a midwoofer anyway.
My L (prestige) system does sound quite nice (subjectively). The L26 sounds very clean in the bass to me, so I can imagine the W26 steps it up a bit. Jed Kunz I think his name did a W26 based design called the Tombstones. You might be able to find it via the wayback machine.
In terms of baffle design, I found this asymetric slope helped with diffraction. I got a very smooth L15 - 27TDFCG integration.
Where you will struggle as Jonasz said is the mid - tweeter XO point. The 27TDFC/G and I also have 27TBFC/G sounded "strained" to me with traditional 4th order acoustic slopes below ~ 2KHz.
I ended up crossing over the L15 and 27TDFC/G at 2,200Hz with ~ 5th order acoustic slopes. This avoided the L15 3rd order breakup at ~ 2,700Hz. The W18 will be lower (take the primary breakup peak and divide by 3 to get the 3rd order peak frequency).
I don't know how the TC25 tweeter will handle < 2,000 Hz XO frequency. I wouldn't cross the W18 over higher than 1,500Hz as suggested by Jonasz. This would require a waveguide on the TC25.
You might be better off choosing another midrange that doesn't have a nasty breakup node. The W18 is a midwoofer anyway.
My L (prestige) system does sound quite nice (subjectively). The L26 sounds very clean in the bass to me, so I can imagine the W26 steps it up a bit. Jed Kunz I think his name did a W26 based design called the Tombstones. You might be able to find it via the wayback machine.
In terms of baffle design, I found this asymetric slope helped with diffraction. I got a very smooth L15 - 27TDFCG integration.
Attachments
Madison sent an answer but not much information from them on this build:
We cannot provide any cabinet drawings, sorry. You just have to make a vented box large enough for the 10" at 3.0 cubic foot with a 3" x 4" -don't make a cube but any other dimensions are fine- and large enough for the other parts and crossover.
Then a 0.75 cu ft sealed chamber for the midrange.
A similar layout as say, the B1371 kit.
B1371 Scan-Speak Kit by Peter Noerbaek - Pair
And happy new year folks. 🙂
We cannot provide any cabinet drawings, sorry. You just have to make a vented box large enough for the 10" at 3.0 cubic foot with a 3" x 4" -don't make a cube but any other dimensions are fine- and large enough for the other parts and crossover.
Then a 0.75 cu ft sealed chamber for the midrange.
A similar layout as say, the B1371 kit.
B1371 Scan-Speak Kit by Peter Noerbaek - Pair
And happy new year folks. 🙂
Once I finish some higher priority work first, I'll try to simulate one reasonably sized enclosure with the specifics from manufacturer datasheet.
Last edited:
I am affraid I won't be able to dive into simulating this one but I can tell you if it ends up not being perfect, there is always hope of modifiying it. If you have any means for measuring speakers, do so, but first decide what kind of a baffle and how large of an enclosure you can tolerate in your room. 3 cuft is a little too small of an enclosure, imo.
Good luck!
Good luck!
What sort of music do you listen to? and what max SPL do you want to achieve?
I found the L15 when crossing over too high cause noticeable distortion on acoustic guitar and other tracks with heavy upper midrange / lower treble (~ 2,700Hz) content. Moving the XO point of my design to 2,300 greatly helped this (although not entirely eliminated).
With the W18 having its primary breakup node much lower (~ 4,700Hz and not ~ 8,100Hz as per the L15), and similar ~ 97dB peak, notching won't be enough and a lower XO point of ~ 1500Hz will probably be needed.
As Lojzek says, a waveguide for the T25 tweeter will almost be essential so it can play without straining (increasing acoustic output ~ 1,500Hz) meaning you can have a higher electrical XO which would be normal for a 1" dome.
I personally would swap the W18 for a midrange driver - such as the Satori MR13P or Scanspeak or Accuton, or be serious about a waveguide... or eliptical slopes (very steep - i.e. 8th order XO) between mid and tweeter.
I found the L15 when crossing over too high cause noticeable distortion on acoustic guitar and other tracks with heavy upper midrange / lower treble (~ 2,700Hz) content. Moving the XO point of my design to 2,300 greatly helped this (although not entirely eliminated).
With the W18 having its primary breakup node much lower (~ 4,700Hz and not ~ 8,100Hz as per the L15), and similar ~ 97dB peak, notching won't be enough and a lower XO point of ~ 1500Hz will probably be needed.
As Lojzek says, a waveguide for the T25 tweeter will almost be essential so it can play without straining (increasing acoustic output ~ 1,500Hz) meaning you can have a higher electrical XO which would be normal for a 1" dome.
I personally would swap the W18 for a midrange driver - such as the Satori MR13P or Scanspeak or Accuton, or be serious about a waveguide... or eliptical slopes (very steep - i.e. 8th order XO) between mid and tweeter.
The cabinet details are very important to the overall design. The cabinet determines the way bafflestep is compensated for and the way diffraction affects the performance of the drivers.
These are not easy drivers to work with and the W18 is one of the worst in the SEAS range for ease of use. It's a 6.5" driver that is obviously intended for two way use, or for a midrange in a three way, but it needs a very low xover point to work properly. As Jonasz says this requires crossing at 1500Hz with a 4th order slope. This isn't just an idea or something that you can be flexible about. 1500Hz or lower.
The waveguide + tweeter idea is perfect for this driver. You can do far better than the T25CF-002 too by the way. This was never a great tweeter to begin with, the only thing it has going for it is its healthy amount of xmax, something you can get from others out there. For an off the shelf tweeter my preference would either be for the DXT tweeter from SEAS, the SB26ADC from SB acoustics, or the D2905/99 from Scanspeak if you want to spend large, all three will work well with a 1.5kHz crossover.
This isn't the kind of design, with these drivers, that you jump into, without an excellent design to start with, or without the knowledge to design it yourself.
These are not easy drivers to work with and the W18 is one of the worst in the SEAS range for ease of use. It's a 6.5" driver that is obviously intended for two way use, or for a midrange in a three way, but it needs a very low xover point to work properly. As Jonasz says this requires crossing at 1500Hz with a 4th order slope. This isn't just an idea or something that you can be flexible about. 1500Hz or lower.
The waveguide + tweeter idea is perfect for this driver. You can do far better than the T25CF-002 too by the way. This was never a great tweeter to begin with, the only thing it has going for it is its healthy amount of xmax, something you can get from others out there. For an off the shelf tweeter my preference would either be for the DXT tweeter from SEAS, the SB26ADC from SB acoustics, or the D2905/99 from Scanspeak if you want to spend large, all three will work well with a 1.5kHz crossover.
This isn't the kind of design, with these drivers, that you jump into, without an excellent design to start with, or without the knowledge to design it yourself.
i listen to most kind of music except classical. I have also thought about making the Odins sealed, and ported cabinets for the W26. Would be the easiest and cheapest solution if i am to "recycle" these drivers. But even with half the the volume of the Odin this means the speakers would be large. Considering approx. 100 liter for sub and 35 liter for sealed Odins.
The crossover i have used for the Odins are the latest Jimangies crossover, crossover point of about 1500 Hz.
Each sub would then be powered by Hypex DS 4.0.
This should sound better than the ported Odins and closed Subs i have now.
The crossover i have used for the Odins are the latest Jimangies crossover, crossover point of about 1500 Hz.
Each sub would then be powered by Hypex DS 4.0.
This should sound better than the ported Odins and closed Subs i have now.
Interesting discussion. The mid-to-tweeter crossover here was one of the key challenges I had to tackle with the Darbari: The Darbari | DHANDANOUGHT.ORG | TCPIP which I built in 2014. Active xo with a MiniDSP 4x10HD. It's still one of the best speakers I've heard, as per my personal tastes.
I didn't have money for Seas drivers, but I used Dayton RS metal cone drivers whose breakup is as nasty. I used the RS150, and did a 48dB/oct crossover at about 1500Hz to the Peerless HDS soft-dome tweeter. It's worked out beautifully -- no distortion, harshness, glare. A lot of warmth, including when playing slightly hard pieces like strings, trumpets, violin, etc. Just a fourth order isn't good enough for this kind of cone breakup, IMHO -- you need to get the peaks down 70dB below the base line. I could do that for the RS270 woofer with a 4th order because I had many more octaves to play with before hitting the cone breakup. Not for the mid. I had read the 70dB remark in Jon Marsh' descriptions of his Dayton RS Modula MTM, some 8-9 years ago, and I've tried to be conservative about that cone breakup ever since.
I have a basic question: what's the difference, in audio terms, between the Excel W26FFX-001 and the Prestige L26RFX/P? Both are metal cone. Remarkably similar SPL curves: the Excel graph:
and the Prestige graph:
Parameters like Mms etc are broadly similar. They model well into boxes, one 50% larger than the other, but that's ok. Both have phase plugs, not dust caps. The Excel has what seems like a smart rubber boot over its magnet, the Prestige is naked.
What exactly does one get for the 3x price of the Excel? How does magnesium justify such a premium over aluminium? Are the distortion profiles very different? The Excel description mentions "heavy copper rings mounted above and below the pole piece."
After looking at the data and reading all the descriptive text, I'm asking: how much difference does it make, audibly, if both are put into 3-ways crossed over at, say, 300Hz 4th order? I'm not asking either driver to reproduce delicate mid-frequencies where, I believe, our ears are very sensitive to distortion...
I didn't have money for Seas drivers, but I used Dayton RS metal cone drivers whose breakup is as nasty. I used the RS150, and did a 48dB/oct crossover at about 1500Hz to the Peerless HDS soft-dome tweeter. It's worked out beautifully -- no distortion, harshness, glare. A lot of warmth, including when playing slightly hard pieces like strings, trumpets, violin, etc. Just a fourth order isn't good enough for this kind of cone breakup, IMHO -- you need to get the peaks down 70dB below the base line. I could do that for the RS270 woofer with a 4th order because I had many more octaves to play with before hitting the cone breakup. Not for the mid. I had read the 70dB remark in Jon Marsh' descriptions of his Dayton RS Modula MTM, some 8-9 years ago, and I've tried to be conservative about that cone breakup ever since.
I have a basic question: what's the difference, in audio terms, between the Excel W26FFX-001 and the Prestige L26RFX/P? Both are metal cone. Remarkably similar SPL curves: the Excel graph:
and the Prestige graph:
Parameters like Mms etc are broadly similar. They model well into boxes, one 50% larger than the other, but that's ok. Both have phase plugs, not dust caps. The Excel has what seems like a smart rubber boot over its magnet, the Prestige is naked.
What exactly does one get for the 3x price of the Excel? How does magnesium justify such a premium over aluminium? Are the distortion profiles very different? The Excel description mentions "heavy copper rings mounted above and below the pole piece."
After looking at the data and reading all the descriptive text, I'm asking: how much difference does it make, audibly, if both are put into 3-ways crossed over at, say, 300Hz 4th order? I'm not asking either driver to reproduce delicate mid-frequencies where, I believe, our ears are very sensitive to distortion...
Last edited:
Don't kill me for suggesting this, but if your heart is set on using a Seas Excel magnesium cone midrange for this 3-way, I'd suggest the 4.5" W12CY instead of the 6.5". The cone breakup immediately gets pushed out to much higher frequencies, and you can actually hope to use your midrange like a midrange.
And if you're open to the idea of not using a magnesium cone, may I suggest one of the Excel Nextel cone drivers, smaller than 6.5"? They actually have a purpose-built midrange, which is not a midbass driver. I feel the whole point of trying to extract that extra performance from a 3-way comes when you use a midrange which has been designed for midrange duties, which means it's been optimised to not be so good for bass duties. The 5" Nextel cone midrange seems superb on paper -- others here may be able to talk about their experiences, since I have none.
Keep the 6.5" midbass drivers for another project. Make a lovely TM stand-mount, or buy two more 6.5" drivers and make a Thor killer. (The work done on these forums in re-designing the Thor crossover is superb -- I have no doubt that it's improved the sound quality hugely. The original Thor crossover was too primitive in taming the midbass cone breakup.)
Strange there are so few who have made a TM of the Excel W18E001 and the Millenium tweeter. On my current Odins i have the upgraded xover by Jimangie, that made the speakers much better overall. I am now leaning towards remaking the Odins cabinets according to Seas specs with the upgraded xover i already have, and adding a 100 liter ported W26 sub for each channels also working as stands...
Long time ago SEAS published the THD figures on it's datasheets. I own several W12CY's and would not use them substantially higher than 2kHz. My own measurements are more or less the same like the ones formerly published by SEAS. Beware of THD's of metal domes at frequencies > f_conebreakup/4 ... this is not a select problem of this driver.
On who's paper? Have a look at
SEAS M15CH002 (E0043) | HiFiCompass
The H2 peaking @ 1.1 ... 1.2kHz does not really seem appealing to me
Yes, I learned a new term from Jon Marsh and gang over at HTGuide: resonance distortion amplification. Yes, you're right, a lot of the metal cone drivers show this problem.
So, this means that even if I use the W12, I guess I'll have to cross over fairly low.
I hadn't seen this data. Yes, it seems strange and I'd hesitate to use it as a midrange.
Not strange. Statistically, there are fewer and fewer takers for anything as prices go to the top of the pyramid of drivers. Secondly, there may not be that much of an audible difference between a Millennium tweeter and something half its price.
It's interesting that Linkwitz did not see any reason to continue to use the Millennium tweeter for his later designs -- his LX521 uses a far less expensive tweeter. If I had to build a TM with the Excel W18 midbass, I'd probably go to the Peerless HDS (called Scan-Speak HDS these days I believe), but may not see any reason to go beyond that.
- Status
- Not open for further replies.