Is it possible to design a different crossover using the measurement found in the audioxpress review audioXpress Thor Review.pdf . I was thinking something along the lines of a cauer-elliptic filter with a lower xover point like the ones for the Dayton kits. has anyone already designed a better xover for this kit? for reference I'm using the stock kit with the upgraded xovers from madisound. thanks
seas thor crossover redesign help
Collapse
X
-
I have the same Seas MTM in a box very similar to the Odin (1.2 cu/ft) reflex tuned to ~35 Hz. I will be doing impedance/FR measurements in box using ARTA/LIMP for the W18s and millenium in the next month or so. I have a fairly recent run of the W18 woofer with the rubber boots.
FWIW, I thought the stock XO sounded terrible. I already modded it significantly to sound much better but I want to do an entirely new design with the XO point at 1500 Hz or so and with the W18 breakup at 4.8 kHz down 50 dB or more. For a quick mod to improve the sound somewhat, remove that 2.2 uF cap parallel with the tweeter. All that does is boost the 2-3 kHz region giving a foward cold sound. The tweeter phase response is not affected by this change so the woofer/tweeter phase integration remains good.- Bottom
-
Hey, not to hijack your thread, but why does the stock XO sound so bad? I know about the inadequate suppression of the metal cone peaks, but didn't this guy "write the book" both literally and figuratively, on testing speakers? He's even written comments about how the guy who designed the "other" Madisound kit for these types of drivers didn't do it right. So, without violating any of the rules here, what gives with all this stuff?
However, reread this:
DIY (Do it yourself): Cabinetry, speakers, subwoofers, crossovers, measurements. Jon and Thomas have probably designed and built as many speakers as any non-professionals. Who are we kidding? They are pros, they just don't do it for a living. This has got to be one of the most advanced places on the net to talk speaker building, period.- Bottom
Comment
-
Half the issue with the Thor (IIRC) is the box... it's quite under-sized.diVine Sound - my DIY speaker designs at diVine Audio- Bottom
Comment
-
Originally posted by speedleHey, not to hijack your thread, but why does the stock XO sound so bad?
Different designers have different priorities. I lean towards maximizing the suppression of the metal cone breakup and lowering the XO frequency at the possible expense of tweeter power handling.- Bottom
Comment
-
I think a better subject here would be simply: crossover design help for these drivers. Forget that it was ever a Thor. It won't be any more.
How many different RS180/RS28A designs are there out there? Yet they're not "variants on XYZ design" - each is unique. It's the crossover that is a design (and of course the baffle layout).
CdiVine Sound - my DIY speaker designs at diVine Audio- Bottom
Comment
-
Originally posted by jimangie1973I have the same Seas MTM in a box very similar to the Odin (1.2 cu/ft) reflex tuned to ~35 Hz. I will be doing impedance/FR measurements in box using ARTA/LIMP for the W18s and millenium in the next month or so. I have a fairly recent run of the W18 woofer with the rubber boots.
FWIW, I thought the stock XO sounded terrible. I already modded it significantly to sound much better but I want to do an entirely new design with the XO point at 1500 Hz or so and with the W18 breakup at 4.8 kHz down 50 dB or more. For a quick mod to improve the sound somewhat, remove that 2.2 uF cap parallel with the tweeter. All that does is boost the 2-3 kHz region giving a foward cold sound. The tweeter phase response is not affected by this change so the woofer/tweeter phase integration remains good.
i will have some measuring equipment available to me in the next few months but this would be my first time using so i'd need some big help, i'll post about that when the time comes though.-Andres- Bottom
Comment
-
Originally posted by cjdI think a better subject here would be simply: crossover design help for these drivers. Forget that it was ever a Thor. It won't be any more.
How many different RS180/RS28A designs are there out there? Yet they're not "variants on XYZ design" - each is unique. It's the crossover that is a design (and of course the baffle layout).
C-Andres- Bottom
Comment
-
The upgraded crossover kit "improves" the component quality, but doesn't change Joseph D'Apollito's design. The big issue for me is that the resonance range for the cone break up is not adequately suppressed just following a classic LR4 roll off at a relatively high crossover frequency.the AudioWorx
Natalie P
M8ta
Modula Neo DCC
Modula MT XE
Modula Xtreme
Isiris
Wavecor Ardent
SMJ
Minerva Monitor
Calliope
Ardent D
In Development...
Isiris Mk II updates- in final test stage!
Obi-Wan
Saint-Saëns Symphonique/AKA SMJ-40
Modula PWB
Calliope CC Supreme
Natalie P Ultra
Natalie P Supreme
Janus BP1 Sub
Resistance is not futile, it is Volts divided by Amperes...
Just ask Mr. Ohm....- Bottom
Comment
-
Well, unless someone has measurements handy, they're the place to start. It's possible to do some subtle tweaking of an existing crossover without measurements, but I'm not sure that'll be what it takes. Particularly if, as mentioned, the drivers are actually a different spec.diVine Sound - my DIY speaker designs at diVine Audio- Bottom
Comment
-
Originally posted by JonMarshThe upgraded crossover kit "improves" the component quality, but doesn't change Joseph D'Apollito's design. The big issue for me is that the resonance range for the cone break up is not adequately suppressed just following a classic LR4 roll off at a relatively high crossover frequency.-Andres- Bottom
Comment
-
Official Seas measurements from Bjorn Idland. There's a readme that tells how they were taken.
- Bottom
Comment
-
Originally posted by cjdI think a better subject here would be simply: crossover design help for these drivers. Forget that it was ever a Thor. It won't be any more.
How many different RS180/RS28A designs are there out there? Yet they're not "variants on XYZ design" - each is unique. It's the crossover that is a design (and of course the baffle layout).
C- Bottom
Comment
-
Originally posted by Dennis HOfficial Seas measurements from Bjorn Idland. There's a readme that tells how they were taken.
http://www.crestviewcable.com/~catapult/Seasdata.zip
I don't have enough fingers to count the people who took that data for the L18 and 27TBFCG, tried to "learn crossovers" by modeling my design, and shortly thereafter complained that there was something wrong with my design because it didn't match their results.
Originally posted by kappa546would anyone like to tackle this? does madisound design CE crossovers, on their site it just talks about 2 and 3way up to 4order slopes.
If I had that driver combination, I'd probably do LR4 at 1500hz, after confirming that the actual drivers had no particular distortion problems. I once came across a Millennium, supposedly never opened up, that had a magnet chip in the gap causing wicked distortion.- Bottom
Comment
-
well like many have accounted, i found the speakers pretty forward sounding with thin bass, which improved some after good break in. i took the fill out of the back part of the line and bottom half of the front line and that really helped fill out the bass. it's hard for me to judge them any further because i admit the temporary room they're in now is far from adequate, it's not very wide and just cramped. the detail and clarity is great but i don't get as enveloped by the music as i have with other systems... i can probably chalk that up to the room. i've heard the crossover is the weak point of the speakers and wanted to get the most out of them.-Andres- Bottom
Comment
-
i've heard the crossover is the weak point of the speakers and wanted to get the most out of them.
- Bottom
Comment
-
yea i read that, hence why i took the polyfill. it seems that for a more pronounced low end improvement i'd have to build new cabinets and thats simply not in the plans right now. i'm ok with the bass as it is, and i'll eventually have a sub filling out the bass better.-Andres- Bottom
Comment
-
So, my guess based on what little I know: they're forward because you're getting a little more grunge from the breakup, AND because that compares so strongly against the weak low end response.
Have you tried "sealing" these? Shore up that bottom end and the forward-ness may no longer be an issue.
CdiVine Sound - my DIY speaker designs at diVine Audio- Bottom
Comment
-
Here's the measured impedance for both the millenium and the W18 pair in my 1.2 cu/ft box tuned to 35 Hz, with polyfill.
If you don't want to measure yourself, I can also get you gated FR measurements after next week. I'll be on vacation starting tomorrow. My baffle width is just shy of 10 inches so I believe it's a tad wider than your's so there will be a small amount of baffle diffraction error relative to your boxes. Again though, it would be ideal for you to get measurements of your particular drivers in your transmission line box.
Jim
- Bottom
Comment
-
Originally posted by jkrutkeI'd advise people to be careful with those. Every one of the woofers in that collection has some sort of measurement or enclosure artifact in it causing a lump in the response at 700hz that normally isn't there. The other issue is that the curves, particularly the tweeters, are only good if building the exact same box. They are not good for designing a MTM, or for offset tweeter locations.
I don't have enough fingers to count the people who took that data for the L18 and 27TBFCG, tried to "learn crossovers" by modeling my design, and shortly thereafter complained that there was something wrong with my design because it didn't match their results.
How would you like someone to tackle it? You haven't actually said what you think is wrong with it as is. If you'd like someone to take a guess a new crossover without actual measurements, the results could get worse instead of better. I recommend boxing everything up and sending it to someone who can measure it.
If I had that driver combination, I'd probably do LR4 at 1500hz, after confirming that the actual drivers had no particular distortion problems. I once came across a Millennium, supposedly never opened up, that had a magnet chip in the gap causing wicked distortion.- Bottom
Comment
-
I've finally measured my Odin's and designed a crossover topology. I did all the measurements using ARTA/LIMP software, using Behringer 802 preamp and ECM8000 mic. FR measurements were done in gated impulse response mode. The design was done using Matlab with the imported FR and impedance data.
The attached jpg images show how my design compares with the original Odin/Thor design. In the impedance, electrical response, and acoutic response graphs, the blue/green traces are the original Odin responses and the maroon/red traces are my new design.
I've crossed at about 1500 Hz with high Q fourth order acoustic rolloffs, increasing beyond fourth order as the attenuation increases. The woofer circuit is third order with an additional 4.8 kHz notch circuit. The tweeter circuit is fourth order electrical. The metal cone breakup at 4.8 kHz is now 60 dB down. I've designed for a full 6 dB baffle step.
I'm using the ECM8000 mic uncalibrated. I believe it is up a couple of dB starting at around 5kHz, which is why the tweeter response looks like they do.
The final plot shows how the tweeter level can be modified by changing a cap value in parallel with the tweeter. Values of 3 uF, 4.7uF, and 7 uF are shown.
- Bottom
Comment
-
Looks like you've done your homework, and I'd expect this to "cure" the issues with excessive higher frequency output in the resonance zone of the W18. :T I'd expect you to want the additional tweeter attenuation.
The only other open point you may want to think about is whether you want to engineer in a little bit of a BBC dip, depending on the kind of program material you listen to and your room setup/source gear; the liver your room, and the more close mic material, the more you might want that BBC dip.
If you can upload your woofer and tweeter data files, then others could play with this also in more "conventional" tools like LspCAD or SoundEasy.the AudioWorx
Natalie P
M8ta
Modula Neo DCC
Modula MT XE
Modula Xtreme
Isiris
Wavecor Ardent
SMJ
Minerva Monitor
Calliope
Ardent D
In Development...
Isiris Mk II updates- in final test stage!
Obi-Wan
Saint-Saëns Symphonique/AKA SMJ-40
Modula PWB
Calliope CC Supreme
Natalie P Ultra
Natalie P Supreme
Janus BP1 Sub
Resistance is not futile, it is Volts divided by Amperes...
Just ask Mr. Ohm....- Bottom
Comment
-
;x(
Very nicely done. I've been hanging out for someone to do a optimised design for an boxed W18E/T25CF/W18E MTM.
Can you tell us your cabinet dimensions, driver placement positions, and vent size?
I don't care much for TL of the Thor or the small box of the Odin. IMH a nice big, properly tuned vented cabinet in an appropiate room gives plenty of impact and articulation.- Bottom
Comment
-
Originally posted by JonMarshLooks like you've done your homework, and I'd expect this to "cure" the issues with excessive higher frequency output in the resonance zone of the W18. :T I'd expect you to want the additional tweeter attenuation.
The only other open point you may want to think about is whether you want to engineer in a little bit of a BBC dip, depending on the kind of program material you listen to and your room setup/source gear; the liver your room, and the more close mic material, the more you might want that BBC dip.
If you can upload your woofer and tweeter data files, then others could play with this also in more "conventional" tools like LspCAD or SoundEasy.
I've ordered all the parts which will arrive on Friday. I will assemble them this weekend and do some listening. I have a 31 band EQ which I use to determine how to improve the responses. I will tune in a BBC dip to see if the overall sound is improved.
I've included the woofer and tweeter frd files. For this driver combination on a flat baffle, the acoustic offset is approximately 2 inches, it is 2.2 inches at 3 feet with mic centered at tweeter level.
I've also include a mic calibration estimation file I made this morning. I generated this based on my measurements done on the millenium as well as the H1212 tweeters. It is flat to 2.5 kHz, then linearly rises to 1.8 dB by 6 kHz, then remains at 1.8 dB. This response is close to what others have found with the ECM8000. It's not perfect, but definately more accurate than uncalibrated.
My final attachment is the response curve with the mic calibration file included, with the 3uF tweeter parallel cap.
- Bottom
Comment
-
Here are the impedance files. I had to add the .txt at the end of the name to make the upload work.- Bottom
Comment
-
Originally posted by tktran;x(
Very nicely done. I've been hanging out for someone to do a optimised design for an boxed W18E/T25CF/W18E MTM.
Can you tell us your cabinet dimensions, driver placement positions, and vent size?
I don't care much for TL of the Thor or the small box of the Odin. IMH a nice big, properly tuned vented cabinet in an appropiate room gives plenty of impact and articulation.
The cabinets are rear ported with a 3" diameter by 10" length port tuned to about 34 Hz. They are 1.2 cubic feet and are filled with acoustistuff to increase effective volume and dampen the BR response.
IMO, the ideal BR box for the W18E pair would be 2+ cubic foot floorstanding.
My cabinets are 1" MDF throughout with dimensions 22h x 10w x15d. The baffles have 0.5" radius roundovers on the sides. The top and bottom of the baffle are not rounded. I believe the Odin/Thor are 9.5" wide with no roundovers (not sure about the roundovers). There will be a slight baffle step difference between mine and the stock Odin/Thor. This should be easy to account for with a small change in value of the 2mH series inductor. I measured on 26" high wooden stands with width/depth dimensions the same as the speaker.
The driver spacing is exactly the same as the Odin/Thor. MT distance of 6", MM distance of 12".- Bottom
Comment
-
Hi, l have been watching this thread with a lot of interest as l have both the thors and odins ( odins for surrounds and centre dutys ) and am keen to get them sounding as good as they can , so hopefully this new crossover design will bring out the best in them as l can tell they need somthing just to "sweeten " them up , bass is'nt a concern as l run a ib sub with them atm ,though l've often wondered how a couple of bass modules with the odins on top would go ........
Lots of good stuff done in these forums , wish l had the knowlege of a lot of people on here ;x(
Lol, cheers for now, ken- Bottom
Comment
-
I too heard the original Thor design about 5 years ago and thought the bad sound was due to them having no break-in time on the drivers. I couldn't figure out how they could have sounded that bad though. So it's nice to know that it wasn't just my hearing.
Jim brought some other speakers over to my house today but I'll have to see if I can get him back over here with these once he has them sounding the way he wants. Of course I'd be more than happy to run on over to his place as well. He lives right by where we work so it's not like it's any real stretch to stop by sometime.- Bottom
Comment
-
Well I've got one crossover completed and have done some simple tests on it. Frequency response, reverse null, and woofer response all look good. It sounds decent as far as I can tell with only one speaker. Hopefully I can get the other build tonight. I've attached a picture of the crossover assembly on the Madisound double matrix board.
- Bottom
Comment
-
I've attached frequency response measurements for the current crossover. I did the measurements with the mic at about 3 feet.
There's a couple dB rise in the 30 deg off axis response at about 2 kHz where the tweeter takes over. They sound very good currently but I need to to some extended listening to determine if I need to pad the tweeter low end slightly to compensate for this rise.
The reverse null plot shows the phasing is good.
I will post the crossover schematic either late today or tomorrow. I know it works with my drivers. I have no idea what the driver tolerances are so I can't assume it will be optimal for other runs of the Millenium and W18E, older versions for example.
- Bottom
Comment
-
Attached is the crossover schematic.
The crossover frequency is about 1600 Hz.
The value of C6 can be changed depending on how much top end, or airy, response you want. Raising the value will reduce the top end and vice versa. 4 uF would be good if you want a warmer sound, not my personal preference though.
Thanks TacoD. I'd be interested in seeing the lowpass you designed for the W18EX if you can locate it.
Edit: I'm still messing with the values of the tweeter shunt RC (R6 and C6). The listed values sound good, but I'm playing around with a small BBC dip for a slightly more relaxed sound. I will update these values if necessary.
2nd Edit: I've settled on C6 = 4 uF and R6 = 4 Ohm. This is about a 1 dB BBC dip, which gives really nice detail with no fatigue.
Last Edit: To completely eliminate the off-axis peak just above 2 KHz, I had to add a series resistor to the tweeter circuit to make a voltage divider. The values of the resistors have been changed as a result. The on axis response has a 2 dB or so BBC dip and the power response is now smooth. No cap or inductor has changed. Schematic is updated. Tweaking is complete.
- Bottom
Comment
-
Jim,
That looks pretty good! I can't wait to give them a listen! I'm sure they will sound great.- Bottom
Comment
-
Originally posted by Brian BungeJim,
That looks pretty good! I can't wait to give them a listen! I'm sure they will sound great.- Bottom
Comment
-
Originally posted by kesa32
was the original crossover similar to the thor/odins?
Ken- Bottom
Comment
-
Originally posted by jimangie1973No, the original crossover was much worse. It had a notch, but at the wrong frequency. It was just above 5.5 kHz, rather than 4.8 kHz. The result was the 4.8 kHz peak was only 15 dB down. It also had poor phase integration, a large wide dip centered at 2 kHz, and a wide hump centered at about 600 Hz. It is well received in the audiophile community. :roll:- Bottom
Comment
-
Originally posted by jimangie1973No, the original crossover was much worse. It had a notch, but at the wrong frequency. It was just above 5.5 kHz, rather than 4.8 kHz. The result was the 4.8 kHz peak was only 15 dB down. It also had poor phase integration, a large wide dip centered at 2 kHz, and a wide hump centered at about 600 Hz. It is well received in the audiophile community. :roll:
Yeah thats sounds pretty bad , do you know anyone with oringinal thor/odins you could compare your new crossover to ? it would be good to hear your comments as to the difference (l know everyones tastes differ here.. )
cheers ken- Bottom
Comment
-
Hi,
My 1st post here. I am one of the guys over on diyAudio responsible for the rework of the Thor. I am stoked to see that someone has talked an XO modification.
Joe was certainly not at his best when he designed this speaker (proving everyone is fallible), if you take the time to wade thru the diyAudio thread you can see where he made a couple big mistakes. A number of people have re-built theri cabinets & found that this immensly improves them. The XO is another piece in the puzzle
Jim's XO deserves a better drawing and i have attached one i whipped up.
dave
Edit: 4-sep-07 XO updated to reflect Jim's comments
- Bottom
Comment
-
Thanks Dave for the nice schematic. Did you get my PM? Note that the 10 Ohm in parallel with the tweeter has been changed to 8 Ohm.
Here's simulations and measurements for the final crossover. I've optimized for a smooth power response, using listening tests as a general guide. I'm really happy with the sound. It is very transparent with no listening fatigue at all. Hopefully Brian will be able to make it over soon to listen and review them.
One more thing, I closed up the rear port to make it a sealed response. The impedance plot reflects this. The W18E001s work perfectly in 1.2 cu ft sealed. The bass was a bit too prominent with the port. IMO, a bigger box is needed for ported with the rather high Qts of these drivers.
- Bottom
Comment
Comment