Need help on XO frequency for Dayton RS270

[jimbones]

I am building a 3 way system. WMMT. The Mids are 6 1/2 and can go low without a problem so I was thinking of 325-350 first order between the W/M But now that I think of it, the Dayton is an aluminum cone. Does that mean that I should go 2nd order? Advice appreciated

[arlis_1957@yahoo.com]

my bet is 2nd order around 350 should be fine. how wide is your baffle?
check this out.
http://audio.claub.net/software/jbabgy/BDBS.html

[Chris Roemer]

I am building a 3 way system. WMMT. The Mids are 6 1/2 and can go low without a problem so I was thinking of 325-350 first order between the W/M But now that I think of it, the Dayton is an aluminum cone. Does that mean that I should go 2nd order? Advice appreciated

Have you seen the response curve for the RS270?

[mdocod]

Hi Jimbones,

Same offer stands here as on the "other" forum if you want help with the crossover and design. I can more or less tell from your described starting point that you are getting off to a poor initial plan in some regards and would be best served to have a guided build.

Choosing an x-over point can not be based on looking at a single driver in a multi-way system without all of the other factors in play. Where to cross, and HOW to cross, will depend on the other drivers, their box loading, your baffle design, your intended listening position, speaker placement in room. Granted, there are always many ways to skin a cat, but we can't begin skinning until we have a cat, so far, all we have, is a claw. The RS270 could have many potential crossover options below ~800hz that could work well.

The RS270 is a great choice for the bottom of a 3-way, however, depending on your dynamic and bandwidth goals, there is very little reason for it to be mated to any more than a single 2.5-5" "fullrange" as your midrange driver, in fact, there are many 2.5-3" drivers that could hang with the RS270 in a wider bandwidth approach, or doubled up in MTM for a narrower bandwidth, higher dynamic approach. A "pair of 6" " drivers is actually a major setback in such a design.

Regards,
Eric

[jimbones]

my bet is 2nd order around 350 should be fine. how wide is your baffle?
check this out.
http://audio.claub.net/software/jbabgy/BDBS.html

about 12" wide

[jimbones]

Have you seen the response curve for the RS270?

No I wish I could get it. I have seen other DIY'rs use it up to 500hz but I am concerned about any peaking since it is aluminum. I'll have to test it.

[jimbones]

Hi Jimbones,

Same offer stands here as on the "other" forum if you want help with the crossover and design. I can more or less tell from your described starting point that you are getting off to a poor initial plan in some regards and would be best served to have a guided build.

Choosing an x-over point can not be based on looking at a single driver in a multi-way system without all of the other factors in play. Where to cross, and HOW to cross, will depend on the other drivers, their box loading, your baffle design, your intended listening position, speaker placement in room. Granted, there are always many ways to skin a cat, but we can't begin skinning until we have a cat, so far, all we have, is a claw. The RS270 could have many potential crossover options below ~800hz that could work well.

The RS270 is a great choice for the bottom of a 3-way, however, depending on your dynamic and bandwidth goals, there is very little reason for it to be mated to any more than a single 2.5-5" "fullrange" as your midrange driver, in fact, there are many 2.5-3" drivers that could hang with the RS270 in a wider bandwidth approach, or doubled up in MTM for a narrower bandwidth, higher dynamic approach. A "pair of 6" " drivers is actually a major setback in such a design.

Regards,
Eric

OK I was trying to save ink . I plan on reusing 4 Dynaudio 17W75's from a previous project and I know they sounded good then. Also it's "free".

Now as far as the design. I am building something like a Sunflower. I am building the upper baffle so everything is modular. I will be able to "ditch" the Dynaudio's if they don't work and replace with a recommended driver. I am planning on using the SB29 tweeter as it has good reviews and is a good value and meets my technical criteria (XO @2500hz 2nd order). I believe you are correct in that the Dynaudio's is the risk. On the other hand I am not sure they would be all that bad either, there are plenty worse drivers and I could always choose another 4/5 inch mid. I hope that clarifies it a little. My concern for the moment was that I never used aluminum coned drivers and I read that you have to be careful with out of band resonances and peaking. If I use first order the slope may be too shallow.

[cjd]

IMO the upper limit if you brick-wall filter on the RS270 is about 800Hz. I'm very very picky about breakup, some might be OK pushing it higher. I went around 350Hz and a rather steep filter (mid-tweet is 8th order acoustic C-E, woofer isn't quite so steep.) Two RS270's to two RS180-8's (in series) and a RS28A up top. Breakup around -50dB.

400Hz 4th order is probably reasonable. 2nd order for an octave + a notch or other means of pushing breakup out of band while giving you a range of 2nd order slope is also workable.

The RS270 is a great choice for the bottom of a 3-way, however, depending on your dynamic and bandwidth goals, there is very little reason for it to be mated to any more than a single 2.5-5" "fullrange" as your midrange driver, in fact, there are many 2.5-3" drivers that could hang with the RS270 in a wider bandwidth approach, or doubled up in MTM for a narrower bandwidth, higher dynamic approach. A "pair of 6" " drivers is actually a major setback in such a design.

Well, a pair of 6" with a single RS270 is perhaps questionable depending on topology (it may be ideal - the mentioned sunflower approach suggests it's about right to me) but a pair of 6" with a pair of RS270 is sweet indeed. Or, can be. I would generally not recommend matching the RS270 to a 2.5-3" driver. Heck, any 10" to a 3" is at the outward side of what I would recommend. 5-6" is great. But, that's me. It can work, to be sure.

C

[arlis_1957@yahoo.com]

12 inch baffle 18 inches high, in the middle. bsl starts about 350hz, yes? 2nd order with a 3bw acoustic slope should lower the breakup plenty.
the sb tweeter WILL cross lower if desired.

[jimbones]

OK I was trying to save ink . I plan on reusing 4 Dynaudio 17W75's from a previous project and I know they sounded good then. Also it's "free".

Now as far as the design. I am building something like a Sunflower. I am building the upper baffle so everything is modular. I will be able to "ditch" the Dynaudio's if they don't work and replace with a recommended driver. I am planning on using the SB29 tweeter as it has good reviews and is a good value and meets my technical criteria (XO @2500hz 2nd order). I believe you are correct in that the Dynaudio's is the risk. On the other hand I am not sure they would be all that bad either, there are plenty worse drivers and I could always choose another 4/5 inch mid. I hope that clarifies it a little. My concern for the moment was that I never used aluminum coned drivers and I read that you have to be careful with out of band resonances and peaking. If I use first order the slope may be too shallow.

Specifically why the objection to the (2) 6 1/2 inch mids?

[cjd]

How is 2nd order 3BW? 2!=3 last I checked? Or are you guys just talking electric targets? At any rate, as I said, how far down is "good enough" is different. I usually shoot for -40dB or better, but not always.

The RS270 was one of my least favorite as far as breakup, so I'm perhaps a little more attentive to it. At least, back when I did that design, it was obnoxious. I think in part because it sticks out enough I would note a downward shift in the vertical image if it were excited notably, but even just the 270 by itself is really tough till you knock it down.

[jimbones]

How is 2nd order 3BW? 2!=3 last I checked? Or are you guys just talking electric targets? At any rate, as I said, how far down is "good enough" is different. I usually shoot for -40dB or better, but not always.

The RS270 was one of my least favorite as far as breakup, so I'm perhaps a little more attentive to it. At least, back when I did that design, it was obnoxious. I think in part because it sticks out enough I would note a downward shift in the vertical image if it were excited notably, but even just the 270 by itself is really tough till you knock it down.

Sorry I don't understand the notation. 3BW etc. 2!=3??? Man I have a lot to learn

[Trap]

I'm presently using the RS270S-8 as the bottom of my 3w system crossed at 200hz LR-4; the TM is the Pioneer SP-BS21; I find this to be a balanced sound; voices don't sound thin; I'm crossing these actively and at one time went as high as 350hz LR-4, which was Ok also; the SP-BS21 uses a 4" woofer so I'd thing you'd be fine with a single mid in the 4' to 6" size. If helpful I can measure a couple different crossover freq's and slopes to give you a ballpark of what to expect.

[jimbones]

I'm presently using the RS270S-8 as the bottom of my 3w system crossed at 200hz LR-4; the TM is the Pioneer SP-BS21; I find this to be a balanced sound; voices don't sound thin; I'm crossing these actively and at one time went as high as 350hz LR-4, which was Ok also; the SP-BS21 uses a 4" woofer so I'd thing you'd be fine with a single mid in the 4' to 6" size. If helpful I can measure a couple different crossover freq's and slopes to give you a ballpark of what to expect.

Basically if I have to avoid the break up I just need the XO freq and slope. Some have recommended a notch filter and lower order so I'd like to go first at 350hz with a notch but thats close so I may need to go 2nd order with notch. Just looking for consensus.

[jimbones]

Sorry I don't understand the notation. 3BW etc. 2!=3??? Man I have a lot to learn

Oh I get it! BW=Butterworth and LR+ Linkwitz Riely?

[Chris Roemer]

No I wish I could get it. I have seen other DIY'rs use it up to 500hz but I am concerned about any peaking since it is aluminum. I'll have to test it.

Sorry. It helps to know the PE# of its predecessor.

The RS270 doesn't really "roll off" on its own, it rolls ON. ;-)

While not exactly the driver you're inquiring about, this is a VERY, VERY close substitute, the shielded version.
I tried attaching the PDF, but I guess it was too big. You should be able to download it off of PE's site, or Dayton's.

Chris

[mdocod]

Basically if I have to avoid the break up I just need the XO freq and slope.

Hello Jimbones,

I'm sorry I must disagree. I used to think the same thing and I'm sure everyone goes through the same phase of having toes dipped into the water just deep enough to start feeling confident, blind to the ocean that awaits .

Desirable results are not likely if you just throw a crossover component calculator at this. This particular crossover point especially challenging for several reasons: being smack in the middle of a baffle step transition that will swallow up a 1st order network completely for the first octave or more above the x-over point, and that's just the tip of the ice burg.

The Dynaudio drivers are fine for use as your mid-range. There are several issues that come to mind against the use of both drivers per side however. With the relatively large mid-range drivers and crossover ~2.5K, a TMM would suffer from unnecessary and undesirable vertical axis lobing issues related to the large distance between the lower midrange and tweeter primarily, and partly between the 2 midbass drivers themselves. In MTM, the relationship of the mid-range drivers to the tweeter is solved, but the distance between the mid-range drivers is then even greater, causing more aggressive vertical axis lobing issues between the midrange drivers. A simpler TMW approach will behave more like the all desirable point source in this particular situation. Modeling suggests a massive suckout in response above 1000hz just by moving the listening height from a traditional sitting to traditional standing position in front of an MTM with this driver size and expected spacing and the open baffle. The single driver has none of that suckout in the intended passband. Having taken a moment to simulate the effects of such a configuration with an open baffle, I can't recommend an MTM OB unless a combination of things were changed to solve the vertical axis issues, (smaller drivers, tighter spacing, lower x-over point, etc etc)

The RS270 is a driver whose characteristics deliver remarkable deep reaching bandwidth, sub-30hz response is no problem. However, the driver is Xmax limited in the sub-50hz range big time. If the system is intended to be functionally flat to ~30hz, then the system will be Xmax limited to ~100dB@1M per side after step losses are factored in. The PAIR of dynaudio drivers used as mid-range drivers would likely out-reach the RS270 by 10dB in their intended passband. A single is the better match dynamically. In otherwords, since the dynamic reach of both drivers is not necessary, the trade-offs (vertical axis lobing problems) that come with implementing both drivers can and should be avoided in this case.

Regarding the speaker design... I can't find response data for the 17W75, only the 17W75 EXT and similar Hi-Vi D6.8 and Morel MW166. It would be best to have the response data for the older drivers you have.

Are you intending to go with the open baffle design up top like the sunflower?

Trust me on this, you WANT a good simulation done for this crossover. A 3 way with an open baffle is going to be a nightmare to try to "voice" by ear and guesswork. I'm happy to do this for you, but I need all of the details about the intended box design, shape, listening position, etc etc.

--------------------



To elaborate on my prior suggestion that even a 2.5" midrange could "keep up" with the RS270...If I were implementing a 2.5" midrange driver, the crossover point would be set higher than the design being discussed here. In fact, it would be centred somewhere around the peak of the baffle step transition. That "route" wouldn't really work in this build anyways since the layout here demands a larger spacing from the woofer to midrange driver.

Regards,
Eric

[jimbones]

Hello Jimbones,

I'm sorry I must disagree. I used to think the same thing and I'm sure everyone goes through the same phase of having toes dipped into the water just deep enough to start feeling confident, blind to the ocean that awaits .

Desirable results are not likely if you just throw a crossover component calculator at this. This particular crossover point especially challenging for several reasons: being smack in the middle of a baffle step transition that will swallow up a 1st order network completely for the first octave or more above the x-over point, and that's just the tip of the ice burg.

The Dynaudio drivers are fine for use as your mid-range. There are several issues that come to mind against the use of both drivers per side however. With the relatively large mid-range drivers and crossover ~2.5K, a TMM would suffer from unnecessary and undesirable vertical axis lobing issues related to the large distance between the lower midrange and tweeter primarily, and partly between the 2 midbass drivers themselves. In MTM, the relationship of the mid-range drivers to the tweeter is solved, but the distance between the mid-range drivers is then even greater, causing more aggressive vertical axis lobing issues between the midrange drivers. A simpler TMW approach will behave more like the all desirable point source in this particular situation. Modeling suggests a massive suckout in response above 1000hz just by moving the listening height from a traditional sitting to traditional standing position in front of an MTM with this driver size and expected spacing and the open baffle. The single driver has none of that suckout in the intended passband. Having taken a moment to simulate the effects of such a configuration with an open baffle, I can't recommend an MTM OB unless a combination of things were changed to solve the vertical axis issues, (smaller drivers, tighter spacing, lower x-over point, etc etc)

The RS270 is a driver whose characteristics deliver remarkable deep reaching bandwidth, sub-30hz response is no problem. However, the driver is Xmax limited in the sub-50hz range big time. If the system is intended to be functionally flat to ~30hz, then the system will be Xmax limited to ~100dB@1M per side after step losses are factored in. The PAIR of dynaudio drivers used as mid-range drivers would likely out-reach the RS270 by 10dB in their intended passband. A single is the better match dynamically. In otherwords, since the dynamic reach of both drivers is not necessary, the trade-offs (vertical axis lobing problems) that come with implementing both drivers can and should be avoided in this case.

Regarding the speaker design... I can't find response data for the 17W75, only the 17W75 EXT and similar Hi-Vi D6.8 and Morel MW166. It would be best to have the response data for the older drivers you have.

Are you intending to go with the open baffle design up top like the sunflower?

Trust me on this, you WANT a good simulation done for this crossover. A 3 way with an open baffle is going to be a nightmare to try to "voice" by ear and guesswork. I'm happy to do this for you, but I need all of the details about the intended box design, shape, listening position, etc etc.

--------------------



To elaborate on my prior suggestion that even a 2.5" midrange could "keep up" with the RS270...If I were implementing a 2.5" midrange driver, the crossover point would be set higher than the design being discussed here. In fact, it would be centred somewhere around the peak of the baffle step transition. That "route" wouldn't really work in this build anyways since the layout here demands a larger spacing from the woofer to midrange driver.

Regards,
Eric

Agree, these are shark infested waters and I was warned it won't be easy, haha.

OK,only one mid, yes sunflower up top 1 mid 1 tweeter. here is the dynaudio link to specs
http://www.gattiweb.com/images/dynaudio/17w75_data.pdf

Please please model it. I'd love to know what reasonably priced program i could buy so i could learn over time.

BTW, the tweeter
http://www.madisoundspeakerstore.com...-dome-tweeter/

I notice that Nola uses a ribbon dipole tweeter. The dynaudio is ok up to 2500hz I don't know if I could get that type of ribbon to go that low (without blowing up) I know steep slopes!


BTW note that i was only going to use 1 tweeter so it is only dipole for mid. Maybe I should change the tweeter. This might be easier to discuss on the phone/skype.

Jim

[mdocod]

Hi Jim,

I use all free modeling software. As do most people here...

For box simulation, I like winISD pro and hornresp. I like to use winISD for quick and dirty, hornresp for more control over the little details. Hornresp can simulate basically any box design you can dream up, from sealed, vented, up to very elaborate horn and transmission line arrangements. It can also export response, impedance, and phase data that can be simply converted to .frd and .zma files (change filename to .frd and presto).

For baffle response modelling, I like "Edge." This program can model the effects of a traditional front baffle or open baffle designs, taking into account driver diameter, position on baffle, and can even model the effects of multiple drivers playing the same frequency from various listening positions. This program can export the uncorrected response and phase data which can later be blended with driver response for crossover simulation. Just change the extension on the file output to .frd and it's ready to go. (export the uncorrected response data for this).

I use SPLTrace to create the .frd(response) and .zma(impedance) files from driver data published online and to trace from. Screen capture the response, import, trace, and export.

For crossover simulation, I use Speaker Workshop. This program allows you to blend the response data that you will gather and create with the programs above, then create drivers and build circuits to see their effect. It's a older program that can be glitchy.

Most people on the forums here use the many free excell programs to do the baffle and crossover simulation. This is fine if you have excell, the programs above are independent and don't require a working copy of excell (I have no interest in owning it, I am a linux user, so I just run the windows programs above under wine in linux and that works well for me).

------------------

How tall did you have in mind for these speakers? Are you planning on a U-shape frame? That could be tricky to simulate (I have some ideas though).... The U-shape frame I fear would behave both like a larger baffle, especially around the dipole peak, and like the normal baffle width above the peak to some extent, which complicates simulations. At that point I won't lie, I'd be throwing in some guesswork to splice and combine a few different simulations to form what I believe would best represent the real world response.

Do you have the 4 or 8 ohm version of the midbass drivers?



Regards,
Eric

[jimbones]

Hi Jim,

I use all free modeling software. As do most people here...

For box simulation, I like winISD pro and hornresp. I like to use winISD for quick and dirty, hornresp for more control over the little details. Hornresp can simulate basically any box design you can dream up, from sealed, vented, up to very elaborate horn and transmission line arrangements. It can also export response, impedance, and phase data that can be simply converted to .frd and .zma files (change filename to .frd and presto).

For baffle response modelling, I like "Edge." This program can model the effects of a traditional front baffle or open baffle designs, taking into account driver diameter, position on baffle, and can even model the effects of multiple drivers playing the same frequency from various listening positions. This program can export the uncorrected response and phase data which can later be blended with driver response for crossover simulation. Just change the extension on the file output to .frd and it's ready to go. (export the uncorrected response data for this).

I use SPLTrace to create the .frd(response) and .zma(impedance) files from driver data published online and to trace from. Screen capture the response, import, trace, and export.

For crossover simulation, I use Speaker Workshop. This program allows you to blend the response data that you will gather and create with the programs above, then create drivers and build circuits to see their effect. It's a older program that can be glitchy.

Most people on the forums here use the many free excell programs to do the baffle and crossover simulation. This is fine if you have excell, the programs above are independent and don't require a working copy of excell (I have no interest in owning it, I am a linux user, so I just run the windows programs above under wine in linux and that works well for me).

------------------

How tall did you have in mind for these speakers? Are you planning on a U-shape frame? That could be tricky to simulate (I have some ideas though).... The U-shape frame I fear would behave both like a larger baffle, especially around the dipole peak, and like the normal baffle width above the peak to some extent, which complicates simulations. At that point I won't lie, I'd be throwing in some guesswork to splice and combine a few different simulations to form what I believe would best represent the real world response.

Do you have the 4 or 8 ohm version of the midbass drivers?



Regards,
Eric

I have the 8 ohm version. The Bass section is a sealed cabinet with volume about 1.6 cu ft and the top is a baffle with wings (the sides do not extend past the front surface and extend back starting at 4 inches up top to about 6 inches at the bottom (tapered sides). If you are wondering why I chose a sealed cabinet? I like the sound of sealed over vented. i have a sub to dig really low if I need it (Velodyne HGS 12).

This is like I'm going to school. It's great to learn this, part science, part art.

Thanks Much, Jim