Hey!
Do i decided to take a plunge and make my first dipole speaker.
After reading a lot of posts by StigEric, mige0, CharlieLaub, Jihazzi, Rudolf, twest820 and others i may have forgotten and after digging through Siegfried Linkwitz website i decided for myself that it should be a swinging dipole build with an H-Baffle woofer and a closed box subwoofer. Hence, 5-way.
Sketchup mockups:
The insides of the metal profile is filled with sand to prevent 'singing' of the metal.
Here is the look with drivers hung, no wires attached yet, still some tilt of the drivers which i almost fixed later.
I am using:
18 sound 15LW1401 in an H-Baffle 55-160Hz
Beyma 12G40 160-800Hz
B&G Neo8 800-3200Hz
B&G Neo3 3200-...Hz
Kef downfiring subwoofer (PSW2500, IIRC) 20-55Hz.
The whole thing is processed with Reaper application. Audio interface i am using is an old MOTU 828 MKII FireWire. Subwoofer is active, other components use old AV receivers as amplifiers: NAD T752 and Pioneer 819H.
I am using Fabfilter Q2 VS plugin for EQing and x-over. When each passband i EQed i set everything flat and on top of that i am using APL Workshop to correct all other anomalies that might be left. It sets the target curve as well.
The room is only mildly treated - i acquired some DIY 'superchunks' and put them in the corners, but it is not enough - the room has a lot of flutters even after a hand clap and a can see severe combing in the response. Will start making absorbers out of fiberglass. Going to treat the front wall and back wall. Maybe some of the ceiling as well. Will put a rug in front as well.
Here is a link to video:
YouTube
Will provide some measurements a few days later. Not satisfied with the sound yet. The sub and woofer sound a bit boomy, the image is too diffuse and kinda pulls to the left. Need to do more tuning, need to do more room treatment, i guess.
Do i decided to take a plunge and make my first dipole speaker.
After reading a lot of posts by StigEric, mige0, CharlieLaub, Jihazzi, Rudolf, twest820 and others i may have forgotten and after digging through Siegfried Linkwitz website i decided for myself that it should be a swinging dipole build with an H-Baffle woofer and a closed box subwoofer. Hence, 5-way.
Sketchup mockups:
The insides of the metal profile is filled with sand to prevent 'singing' of the metal.
Here is the look with drivers hung, no wires attached yet, still some tilt of the drivers which i almost fixed later.
I am using:
18 sound 15LW1401 in an H-Baffle 55-160Hz
Beyma 12G40 160-800Hz
B&G Neo8 800-3200Hz
B&G Neo3 3200-...Hz
Kef downfiring subwoofer (PSW2500, IIRC) 20-55Hz.
The whole thing is processed with Reaper application. Audio interface i am using is an old MOTU 828 MKII FireWire. Subwoofer is active, other components use old AV receivers as amplifiers: NAD T752 and Pioneer 819H.
I am using Fabfilter Q2 VS plugin for EQing and x-over. When each passband i EQed i set everything flat and on top of that i am using APL Workshop to correct all other anomalies that might be left. It sets the target curve as well.
The room is only mildly treated - i acquired some DIY 'superchunks' and put them in the corners, but it is not enough - the room has a lot of flutters even after a hand clap and a can see severe combing in the response. Will start making absorbers out of fiberglass. Going to treat the front wall and back wall. Maybe some of the ceiling as well. Will put a rug in front as well.
Here is a link to video:
YouTube
Will provide some measurements a few days later. Not satisfied with the sound yet. The sub and woofer sound a bit boomy, the image is too diffuse and kinda pulls to the left. Need to do more tuning, need to do more room treatment, i guess.
Attachments
Left speaker is too close to the corner, bass and low mid get lots of boost from early reflections and stereo turns to the left. This is difficult to eq, any changes to get a more symmetrical positioning? The subwoofer level must be set by listening and most likely L/R must have different eq and gain settings in asymmetrical setup.
Last edited:
I will use tape to measure distances from the front and side walls and will position speakers symmetrically. It may seem that left speaker is too close to the wall, but i was eyeballing it to be the same as the right speaker. Maybe room symmetry plays a role.
I am using 1mm fishing line. Will the elastic band be better in any way?
Woofer is held by the magnet and the screw holes, btw.
It's not better, it's another world
Though you'd need a taller stand to hold 'em; that's because the speakers ( well, the sub should, either... ) really need to be suspended, i.e. find the right elasticity ( K ) that permits them to have some headroom so that the oscillation is out of working band, so below 20 Hz
Go tell to StigEric, mige0, CharlieLaub, Jihazzi, Rudolf, twest820 and others...
It's not better, it's another world
Though you'd need a taller stand to hold 'em; that's because the speakers ( well, the sub should, either... ) really need to be suspended, i.e. find the right elasticity ( K ) that permits them to have some headroom so that the oscillation is out of working band, so below 20 Hz
Go tell to StigEric, mige0, CharlieLaub, Jihazzi, Rudolf, twest820 and others...
Hm, interesting!
So you would think that fishing line is 'singing' along with the drivers?
Everything vibrates, that's for sure.
You want only the speaker's cone to vibrate. With this OB setup you've already eliminated the classic wooden boxes which tend to add their own resonant modes. I'm not sure what you're supposed to do with the backwave from the midwoofer, there must be something that makes a barrier between the two emissions. Also reflections from the back wall, of course.
You want only the speaker's cone to vibrate. With this OB setup you've already eliminated the classic wooden boxes which tend to add their own resonant modes. I'm not sure what you're supposed to do with the backwave from the midwoofer, there must be something that makes a barrier between the two emissions. Also reflections from the back wall, of course.
One can check unwanted vibrations by taking semi-nearfield (20cm) measurements of a single driver (band-limited signal or with the xo in use) with and without holding a finger touching the frame. When you see a change in response and distortion display - there's your resonance! Check also CSD window in REW.
Then just tweaking, measuring, tweaking etc. - this is a hobby!
My 12" lower midrange is held on place by velcro strip on the back of magnet. The frame is not touching the pole. First it was attached from the outer edge too, but there were more resonances!
Then just tweaking, measuring, tweaking etc. - this is a hobby!
My 12" lower midrange is held on place by velcro strip on the back of magnet. The frame is not touching the pole. First it was attached from the outer edge too, but there were more resonances!
It's not better, it's another world
Though you'd need a taller stand to hold 'em; that's because the speakers ( well, the sub should, either... ) really need to be suspended, i.e. find the right elasticity ( K ) that permits them to have some headroom so that the oscillation is out of working band, so below 20 Hz
Go tell to StigEric, mige0, CharlieLaub, Jihazzi, Rudolf, twest820 and others...
I don't agree that elastic band (whatever that is) would be an improvement. You need the drivers to hang in a certain position and then stay put. There are no up and down vibrations from the driver, only the ones in the front-back direction. Hanging by a stiff wire is fine. You could even use a rod or metal strips as long as the driver can freely move and swing. I use nylon monofilament that is intended for hanging pictures and is rated for up to 25 pounds per line. In my system I install a long threaded bolt thru two of the driver mounting holes near the top. Then I hang the driver from some position back near the top plate of the motor, which is about where the center of mass is found. In my country I used 1/4" bolts that are about 6" long. I used "wing nuts" to tighten the bolts to that they could bear the weight. I was able to suspend a 15" driver with neo magnet in this way very precisely. For heavier drivers this might not work as well due to the lever action on the bolt.
This is a pendulum system and you can easily calculate the natural oscillation frequency. Here is a line to an online calculator:
Simple Pendulum Calculator - Omni
It's really only dependent on the length of the pendulum. For a 30cm length, the natural frequency is 1Hz. I think that is sufficiently below the passband!
Last edited:
When I first saw a system like this a few years back I had the exact same feeling! But it turns out that there are some very good reasons to do this.
When you mount a driver to a baffle, or in a cabinet, it induces vibrations into the cabinet that are re-radiated and cause coloration of the total sound output. When multiple drivers are mounted in this way, the problem is compounded.
By hanging the driver using wires you eliminate all of these resonances and you can decouple the drivers from each other so that the vibration from one does not effect the other(s).
Also, it turns out that the off axis response of the driver has some very attractive characteristics when the driver is not in a baffle. Hanging by wires is one way to achieve a "no baffle" system. For more info, see:
Electro-acoustic models
The "no baffle" response is shown in Figure A3-5, for the case where b=a. Because both on and off axis responses follow each other (there is more or less constant directivity) all you need to do is apply filtering/EQ to flatten the response out and you get a beautiful CD-like response from the loudspeaker. This can be achieved at low frequencies without large horns, etc. In this way you can control the directivity of the loudspeaker.
Last edited:
At least 1m away from the back wall is good practice for dipoles. The same distance, or more, should be observed to the side walls. You also need to sit at least 1.5-2m away from each speaker. The designer should consider whether they have sufficient space to do that.
This, and other factors, make dipoles a bad choice for small listening spaces. The bass response of the dipole will actually be WORSE in a small room than in a larger one. A dipole cannot pressurize the room, so the room must be able to support the lowest frequency "wave", that is be at least 1 wavelength long at the lowest bass frequency you wish to use. For instance: 35Hz is about 10m = 33 feet! As a result, in rooms less than 20 feet long I would use a closed box subwoofer below 80-100Hz and dipole above. There will be a cardioid formed in the crossover region between sub and woofer, but that is not a problem.
Last edited:
I have a open baffle DIY speaker friend in Florida who suspends his drivers via wires in a open frame structure. He has woofers, midrange in a waveguide, and tweeters all hanging in space. Think of a large upright locker as the frame and add speakers to suit. He has been using this technique for a few years with changes and updates as he switches drivers to improve his sound. Very good results from his efforts.
Hi Stab,
I also don´t agree with the elastic band. I used 3 mm Dyneema (due to heavy drivers) for my swinging OB. Elastic bands would have made the whole thing even more complicated, so I have to admit I never tried it.
Regarding free-swinging bass driver, I found (of course subjectively - found measuring bass OB in room quite tricky) that bass went clearer as I attached counterweights to the motor (non-magnetic stainless steel of 22 kg), maybe that applies to upper bass as well?
As you have a subwoofer, I would as well try the 18sound free swinging without H-baffle.
Room symmetry and symmetrical placement of OBs is critical. Hard to judge from photos, but if possible, you should try bigger distances from all reflecting surfaces. A carpet between speakers and sofa should help as well.
All the best
Mattes
I also don´t agree with the elastic band. I used 3 mm Dyneema (due to heavy drivers) for my swinging OB. Elastic bands would have made the whole thing even more complicated, so I have to admit I never tried it.
Regarding free-swinging bass driver, I found (of course subjectively - found measuring bass OB in room quite tricky) that bass went clearer as I attached counterweights to the motor (non-magnetic stainless steel of 22 kg), maybe that applies to upper bass as well?
As you have a subwoofer, I would as well try the 18sound free swinging without H-baffle.
Room symmetry and symmetrical placement of OBs is critical. Hard to judge from photos, but if possible, you should try bigger distances from all reflecting surfaces. A carpet between speakers and sofa should help as well.
All the best
Mattes
Will keep the fishing line for a while then, since elastic bands kinda lose the battle
Can't blend the sub with mains fr some reason. Have a huge peak at about 45Hz, REW suggests a 17Db cut REW Room Sim tool kinda agrees:
Adding more subs in this Room Sim tool doesn;t "correct" the problem.
The one i am using is an old sealed KEF PSW2500. I have an old Yamaha ported sub and a car audio sub with 23Hz port tuning. If i add those 2 subs across the room, is it more likely to screw up things even more or there are chances of evening out the sound?
Ordered some fiberglass panels, need to buy some fabric to put them into and will be putting them on the front and rear wall. Still have a lot of flutter in the room.
Can't blend the sub with mains fr some reason. Have a huge peak at about 45Hz, REW suggests a 17Db cut
REW Room Sim tool kinda agrees:Adding more subs in this Room Sim tool doesn;t "correct" the problem.
The one i am using is an old sealed KEF PSW2500. I have an old Yamaha ported sub and a car audio sub with 23Hz port tuning. If i add those 2 subs across the room, is it more likely to screw up things even more or there are chances of evening out the sound?
Ordered some fiberglass panels, need to buy some fabric to put them into and will be putting them on the front and rear wall. Still have a lot of flutter in the room.
Attachments
Haven't noticed any noise as of yet.
BTW, my initial attempt was to use 12G40 from 160 up to 800 Hz, but measurements show that it is useless until about 200Hz - so this is the new crossover point.
Yes, a nude 12" dipole is a midrange, no point trying to eq it flat below 100Hz. even a shallow LS2 xo slope is a challenge!
Here is Kreskowsky's graph
http://musicanddesign.speakerdesign.net/NaO_Note_II_RS_Details.html
Here is Kreskowsky's graph
http://musicanddesign.speakerdesign.net/NaO_Note_II_RS_Details.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.