As a starting point for lower cost options I'd recommend a Focusrite Saffire 24 (eBay B stock used to be around USD 225) and Thuneau Frequency Arbitrator + Allocator Lite. The linear phase setup here runs with just under 16% CPU load on a 1.6GHz Core Solo. A 1GHz CPU should be enough for a three way if the PC's dedicated to music serving and running the crossover while you have the stereo going. Decent chance of four way (though you'd probably want the Saffire 40 in that case), but it'll be tight. This intro may also be handy.
Full disclosure: Thuneau's supported a little side project I'm doing with a donation of two of their VSTs.
Full disclosure: Thuneau's supported a little side project I'm doing with a donation of two of their VSTs.
I have built a similar speaker to this
Beryllium ribbon tweeter from pioneer(? series) open baffle
Beryllium dome from yamaha ns 2000 open baffle
Tad TL-1601A U baffle
For the woofer, I bought a class D power amp from Class D Audio CDA-254S Kit
I am also continuing the tuning to find best(as good as it gets) XO and EQ filters
What I want to tell, there is a strange thing that the TAD can go up to 500 Hz without oscillations in FR (U baffle configuration)
After the 500 Hz, I saw a quite deep roll off in my measurements
Maybe, I have a problem with my measurement set up
Beryllium ribbon tweeter from pioneer(? series) open baffle
Beryllium dome from yamaha ns 2000 open baffle
Tad TL-1601A U baffle
For the woofer, I bought a class D power amp from Class D Audio CDA-254S Kit
I am also continuing the tuning to find best(as good as it gets) XO and EQ filters
What I want to tell, there is a strange thing that the TAD can go up to 500 Hz without oscillations in FR (U baffle configuration)
After the 500 Hz, I saw a quite deep roll off in my measurements
Maybe, I have a problem with my measurement set up
Have you seen something like this before?
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.
Oh yes, this is the best version if my dipoles when they still had baffles. My good friend Lars is the owner of these. In fact - he was the one that got me into building dipoles in the first place.
They have integrated Class-D amps mounted inside the woofer U-baffle. All that's needed in addition to these active speakers is a multichannel audio interface (like my RME Fireface, you see it lying on the floor there), and a computer to run the XO and EQ filters. Very neat setup indeed, and whow... it sounds great!
We see the same thing with the 21" Beyma in U-baffle. On-axis frequency response is very flat up to at least 300 Hz. Off-axis its an other story... with peaks around 170 and 340 Hz. Those peaks require some EQ, even if the on-axis response is flat.
Here's a little update! No changes to the speakers, but quite a few changes to the listening room. The front end (which I call "The Dark Side") is unchanged - this end is heavily damped. The news are that the rear end has gotten quite a lot of diffusion, as the pictures below will tell.
The white diffusors are Vicoustic Multifusor DC, made from polystyrene and thus very light. I have a total of 32 of these, where six are stored on top of the cabinets there as you see.... will see where they will be used later on.
The large wood diffusor below the windows are twelwe home-made 2-dimensional diffusors based on this report from BBC:
http://downloads.bbc.co.uk/rd/pubs/reports/1995-01.pdf
They are made from 12 by 12 pieces of 48x48 mm spruce lumber, mounted in a box of MDF to keep it stable. Each diffusor actually has 132 pieces of lumber, as there are 12 pieces that have zero lenght. Total weight is well above 30 kg each. The material cost alone for the home-made diffusors is in the same league as the finished Vicoustic diffusors, and it takes 3-4 hours to manufacture one, including cutting and sanding the lumber, make the MDF box, and glue the lumber pieces into the box. Using MDF boxes was a good idea, as its easy to stack them.
Total depth of the wood diffusors is 300 mm, making them effective over a wide frequency range. Its a lot of work to build them, but their performance is without doubt better than the Vicoustic diffusors.
I also glued four and four Vicoustic diffusors together that I place on top of the wood diffusors to cover the windows when I want to do some serious listening.
There are also two 120x120 cm absorbers in the ceciling, made from two 120x60 cm Rockwool acoutic ceciling tiles each.
I've invested a lot of work and money into this, but feel that its really worth the effort. The listening room is maybe the weakest link in any audio system, and improvements are very easy to hear.
The white diffusors are Vicoustic Multifusor DC, made from polystyrene and thus very light. I have a total of 32 of these, where six are stored on top of the cabinets there as you see.... will see where they will be used later on.
The large wood diffusor below the windows are twelwe home-made 2-dimensional diffusors based on this report from BBC:
http://downloads.bbc.co.uk/rd/pubs/reports/1995-01.pdf
They are made from 12 by 12 pieces of 48x48 mm spruce lumber, mounted in a box of MDF to keep it stable. Each diffusor actually has 132 pieces of lumber, as there are 12 pieces that have zero lenght. Total weight is well above 30 kg each. The material cost alone for the home-made diffusors is in the same league as the finished Vicoustic diffusors, and it takes 3-4 hours to manufacture one, including cutting and sanding the lumber, make the MDF box, and glue the lumber pieces into the box. Using MDF boxes was a good idea, as its easy to stack them.
Total depth of the wood diffusors is 300 mm, making them effective over a wide frequency range. Its a lot of work to build them, but their performance is without doubt better than the Vicoustic diffusors.I also glued four and four Vicoustic diffusors together that I place on top of the wood diffusors to cover the windows when I want to do some serious listening.
There are also two 120x120 cm absorbers in the ceciling, made from two 120x60 cm Rockwool acoutic ceciling tiles each.
I've invested a lot of work and money into this, but feel that its really worth the effort. The listening room is maybe the weakest link in any audio system, and improvements are very easy to hear.
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.
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.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
The pattern looks like this, where the numbers in the squares is the relative lenght of each element. I used 7,5 cm = 1, equal to a total depth of 30 cm.
Its much simpler than RPG's Skyline, but its effective enough, and farily simple to make as you have to make only four different lengths of elements.
Its much simpler than RPG's Skyline, but its effective enough, and farily simple to make as you have to make only four different lengths of elements.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
I see now that I linked to the wrong BBC report... the diffusor shown there is obviously not using the same sequence as I do. I will try to find the correct one....
EDIT: Found it:
http://downloads.bbc.co.uk/rd/pubs/reports/1990-15.pdf
EDIT: Found it:
http://downloads.bbc.co.uk/rd/pubs/reports/1990-15.pdf
Last edited:
Nice !
So we need a total length of 84*max-length - in your case 25.2m
Those 25.2m to cut down into 84 pieces of 30cm
15 pieces we put apart and form the rest we cut 19 pieces into half and the rest (40 pieces) into 1/4 3/4
Leaves us a rest of two 15cm pieces we give the children to play with ..
Needing 4 of those for one m2 - ui, ui, ui - pretty much work and some cost too.
This is about 100m of material and roughly 500 cuts for a 1m2 diffuser
Stig Erik you are a really dedicated man !!!
Michael
The net amount of lumber for one 30 cm deep diffusor is actually 22.05 meter. I cut all pieces to their right lenght without first cutting into 30 cm lenght and so on.
You'll need (for one diffusor):
38 pcs 7,5 cm length
38 pcs 15 cm length
40 pcs 22.5 cm length
15 pcs 30 cm length
That's 132 pieces. All pieces should have the edgdes of the cutted ends sanded, thats 132 x 8 = 1056 edges... for 12 complete diffusors we get 1584 pieces of wood to cut and 12672 edges to sand... dont even consider anything like this without a table saw and a belt grinder!
You'll need (for one diffusor):
38 pcs 7,5 cm length
38 pcs 15 cm length
40 pcs 22.5 cm length
15 pcs 30 cm length
That's 132 pieces. All pieces should have the edgdes of the cutted ends sanded, thats 132 x 8 = 1056 edges... for 12 complete diffusors we get 1584 pieces of wood to cut and 12672 edges to sand... dont even consider anything like this without a table saw and a belt grinder!
And now some words about why I did all this. The most important is of course to show pictures and brag about it here.....
Seriously - the dipoles became so good that my room was getting more and more the weakest link in the chain. The front end of the room is damped very hard, and the rear end was reflective (no acoustic treatment). A damped front end provides a very clearly defined ITD (Initial Time Delay) before reflections start to come back from the room. The problem here was these reflections. Because of the heavy front end damping, the later room reflections became easier to hear, and they were also not as dense or diffused as we could want.
Installing diffusors in the rear part of the room will not reduce the late arriving energy in the room, but rather spread them out in time and direction. This will make the "ambience" in the room more pleasing and less disturbing.
The subjective impression is that the room sounds much smoother with less tonal coloring. Recordings that used to sound harsh suddenly became much more acceptable. At the same time the resolution, dynamics and stereo image presicion and scale improved. I started to hear tiny details on records I've never really heard before.
Seriously - the dipoles became so good that my room was getting more and more the weakest link in the chain. The front end of the room is damped very hard, and the rear end was reflective (no acoustic treatment). A damped front end provides a very clearly defined ITD (Initial Time Delay) before reflections start to come back from the room. The problem here was these reflections. Because of the heavy front end damping, the later room reflections became easier to hear, and they were also not as dense or diffused as we could want.
Installing diffusors in the rear part of the room will not reduce the late arriving energy in the room, but rather spread them out in time and direction. This will make the "ambience" in the room more pleasing and less disturbing.
The subjective impression is that the room sounds much smoother with less tonal coloring. Recordings that used to sound harsh suddenly became much more acceptable. At the same time the resolution, dynamics and stereo image presicion and scale improved. I started to hear tiny details on records I've never really heard before.
Last edited:
Kudos for a no compromise-project Stig!
A bit off topic Stig: with your experience with the Beyma TPL150 and TPL150H: do you believe a TPL150H with horns on both sides could interesting in a dipole-speaker? Maybe down to 1000 Hz and crossed over to a baffel-less/hanging 6" (or even maybe 8")? Uniform directivity in the crossover is a must
Best regards
Gisle
A bit off topic Stig: with your experience with the Beyma TPL150 and TPL150H: do you believe a TPL150H with horns on both sides could interesting in a dipole-speaker? Maybe down to 1000 Hz and crossed over to a baffel-less/hanging 6" (or even maybe 8")? Uniform directivity in the crossover is a must
Best regards
Gisle
I think that would be very interesting!
Michael (mige0) made a dipole horn for the B&G Neo3 which worked very well:
http://www.diyaudio.com/forums/multi-way/159731-beautiful-swingin-speaker.html
If you manage to run the TPL150H as low as 1000 Hz, it should be possible to mate it with even a 10", and still maintain Constant Directivity.
Michael (mige0) made a dipole horn for the B&G Neo3 which worked very well:
http://www.diyaudio.com/forums/multi-way/159731-beautiful-swingin-speaker.html
If you manage to run the TPL150H as low as 1000 Hz, it should be possible to mate it with even a 10", and still maintain Constant Directivity.
Michael has a disclaimer on the beautiful swinger that he doesn't recommend the horn profile. Which begs the question of what he does recommend; I'm sure we'll find out in not too long. See also the TPL150 and AMT1 horn threads. Something I'd note is all three of these horns operate horizontally but make little or no effort vertically. Personally I'm unconvinced controlling horizontal directivity while neglecting vertical is wise---in my listening room the ceiling is the reflection I notice the most, since I'm renting and hence lack a diffuser---so you might consider more complex horn shapes.
With nude drivers and flat baffles directivity typically narrows only gradually above the dipole peak and horns generally produce some directivity narrowing at the lower end of their working range. So with a 1kHz cross to a nude driver I'd suggest looking at 12s if your equalization capacity is up for it and you can cross in the LR6 to LR8 range to avoid cone breakups. Check the off axis curves of the driver against the crossover slope. 15s might also be an option depending on horn and driver selection; I don't have a good feel for how much the rear SPL reduces above the dipole peak but with 12s and larger the cone's a good bit larger than the magnet so the rear polars shouldn't be too bad.
See also the TPL150 and AMT1 horn threads. Something I'd note is all three of these horns operate horizontally but make little or no effort vertically. Personally I'm unconvinced controlling horizontal directivity while neglecting vertical is wise---in my listening room the ceiling is the reflection I notice the most, since I'm renting and hence lack a diffuser---so you might consider more complex horn shapes.With nude drivers and flat baffles directivity typically narrows only gradually above the dipole peak and horns generally produce some directivity narrowing at the lower end of their working range. So with a 1kHz cross to a nude driver I'd suggest looking at 12s if your equalization capacity is up for it and you can cross in the LR6 to LR8 range to avoid cone breakups. Check the off axis curves of the driver against the crossover slope. 15s might also be an option depending on horn and driver selection; I don't have a good feel for how much the rear SPL reduces above the dipole peak but with 12s and larger the cone's a good bit larger than the magnet so the rear polars shouldn't be too bad.