I have been thinking recently about getting a 3-Way MTW project going. I want to try an emulate the nice imaging I get from my small desktop speakers, which are 2-ways using 3" drivers in folded TLs, where the 3" midbass is crossed high (about 5000 Hz) to small mylar tweeters. Even though these speakers are built from surplus/bargain bin parts they sound quite good. I know that is not because of the high quality of the drivers used ;-) It must be the point source nature of the small size of the drivers, and probably combined with enclosures that have very effectively stopped internal reflections, and extended bass performance some.
I have been looking for 3" drivers that can produce the sound levels that a midrange in a 3-way needs to have, like 88-92 db sensitivity. There are very few out there. There is a mid unit that is small and plays low that I like the looks of - the Scan Speak 10f/4424g-4, even though it is a 4" unit. They are supposed to be on the small side for a 4-incher. They are about $100 each, which is a high price for a relatively new builder like myself to commit to.
All in all, it looks like I haven't really found a true 3" driver of high quality and high sensitivity. Maybe the Faital 3FE. It is in my price range at least. Then one day I was looking at the PE Factory Buyouts and noticed the little brother of the well reviewed Fountek FR88. The FE83 is on sale for ~$7. It occurred to me that two of those cost less than one 3FE. At 86 db ea., two would also be good for 89 db of sensitivity. And they should have the high resolution accurate sound of good aluminum coned drivers. Another thing about the FE83 is it is tuned more for bass performance relative to high end performance. I like that because I will have a tweeter anyway and I would like to cross low to the woofer, maybe 300 Hz.
Running a pair of FE83s should allow for reduced distortion down in the 150-300 Hz range compared to trying to get that low with a single 3" driver, especially one not of Scan Speak quality. The more I think about it the better the idea of using two of these as a midrange unit looks. The question then becomes just how to do it. How to make this little learning experience turn into a decent listening experience. TMM? Generic MTM? D' Apolitto? I proceeded to read a lot of threads here about these things, thought about it some more, and decided on the MTM idea as the best plan. I even decided to order some FE83s while they were still on sale. I like an extended thought process but I tend to miss the deals when I wander down that road.
I have always shied away from the D'Apolitto array idea because of the problem of comb filtering between the two mids and the (usual) need for a low crossover to the tweeter. The things is, with these little drivers the C - to - C distance can be much smaller than you usually see in MTM and D'Apolitto designs. If I put two FE83s together with a 3/4" flangeless tweeter like a Vifa OX20SC, which can cross low, I might just have the makings of a nearly ideal DA or MTM mid-tweet assembly.
In terms of the bigger picture plan here - what I would like to do is build this MTM unit as well as I can and then use it with the woofers in the old school "monkey coffin" 3-way speakers that currently sit in my living room. I would just disconnect the mid and tweet in those and put a coil on the woofers (paper 10" that run full-range. I think I am hearing too much of the cones flopping around up towards 1000 Hz). Then I would place them upside down on the floor so the woofers were at the top end and set these MTM boxes on top, placed for proper time alignment vs the woofers. I hope that this could serve until I have time to build more ideal woofers one day. The old paper 10" woofers are probably pretty high sensitivity (88-90 dB) but I don't have a proper calibrated mike at this time so I don't know for sure. I can make only relative measurements with my homebrew mike.
OK - I just wrote 6 paragraphs as an introduction! I hope the audience is not winded yet, because _I_ can be kinda long winded...
I tell friends my making motto is "think thrice, measure twice, cut once". Maybe I enjoy the planning stage of projects too much. Time to cut to the chase here and show you my proposal so all you tinkers and thinkers can comment on it. Assuming any of you have any attention span left! I have attached a drawing (.pdf).
(Note
If you are as excited about this experiment as I am and can't wait to see how it all turns out - I am sorry I will have to disappoint you. I will not be starting to build this weekend. Or even this year. Luckliy it is December... Next year is possible.)
Some points that I wonder about that maybe you can help me understand more clearly:
1. Two 8 ohm drivers of 86 db sensitivity in parallel. Do I get +3 db, or do I get more for now having a 4 ohm compound driver instead of an 8 ohm driver?
2. I have read the opinion that a true DA design approximates a point source pretty well. By "true" I mean 3rd order slopes, low XO, and, if memory serves, 90 deg phase at XO. Is that true (about the imaging aspect)?
3. Is a true DA really better in the point source sense, and so for imaging, than other ways of rigging up an MTM? I see that Thor, for instance, was not done that way.
In the plan I am proposing I think a true DA alignment could be done with a roughly 2500 Hz XO and still have the mids within one wavelength at XO. Maybe this is a particularly good application for the idealized DA design?
4. In the Statements Holtz put the two mids (also an MTM btw) in tunnels that happened to be of a length corresponding to 1/4 wave at about 300 Hz, which is close to the XO frequency he used. Lots of good reviews of those. I wonder if using a TL on a mid placed at some frequency well above the Fs of the driver results in any roll off below the tuning frequency, or maybe if all it does is to alter the slope of the roll off only below Fs? Which might make it possible to simplify the crossover?
My first experiment might just be to build the baffles, put it the FE83s, and stick pipes on the back to see how it sounds and what the FR looks like.
I have been looking for 3" drivers that can produce the sound levels that a midrange in a 3-way needs to have, like 88-92 db sensitivity. There are very few out there. There is a mid unit that is small and plays low that I like the looks of - the Scan Speak 10f/4424g-4, even though it is a 4" unit. They are supposed to be on the small side for a 4-incher. They are about $100 each, which is a high price for a relatively new builder like myself to commit to.
All in all, it looks like I haven't really found a true 3" driver of high quality and high sensitivity. Maybe the Faital 3FE. It is in my price range at least. Then one day I was looking at the PE Factory Buyouts and noticed the little brother of the well reviewed Fountek FR88. The FE83 is on sale for ~$7. It occurred to me that two of those cost less than one 3FE. At 86 db ea., two would also be good for 89 db of sensitivity. And they should have the high resolution accurate sound of good aluminum coned drivers. Another thing about the FE83 is it is tuned more for bass performance relative to high end performance. I like that because I will have a tweeter anyway and I would like to cross low to the woofer, maybe 300 Hz.
Running a pair of FE83s should allow for reduced distortion down in the 150-300 Hz range compared to trying to get that low with a single 3" driver, especially one not of Scan Speak quality. The more I think about it the better the idea of using two of these as a midrange unit looks. The question then becomes just how to do it. How to make this little learning experience turn into a decent listening experience. TMM? Generic MTM? D' Apolitto? I proceeded to read a lot of threads here about these things, thought about it some more, and decided on the MTM idea as the best plan. I even decided to order some FE83s while they were still on sale. I like an extended thought process but I tend to miss the deals when I wander down that road.
I have always shied away from the D'Apolitto array idea because of the problem of comb filtering between the two mids and the (usual) need for a low crossover to the tweeter. The things is, with these little drivers the C - to - C distance can be much smaller than you usually see in MTM and D'Apolitto designs. If I put two FE83s together with a 3/4" flangeless tweeter like a Vifa OX20SC, which can cross low, I might just have the makings of a nearly ideal DA or MTM mid-tweet assembly.
In terms of the bigger picture plan here - what I would like to do is build this MTM unit as well as I can and then use it with the woofers in the old school "monkey coffin" 3-way speakers that currently sit in my living room. I would just disconnect the mid and tweet in those and put a coil on the woofers (paper 10" that run full-range. I think I am hearing too much of the cones flopping around up towards 1000 Hz). Then I would place them upside down on the floor so the woofers were at the top end and set these MTM boxes on top, placed for proper time alignment vs the woofers. I hope that this could serve until I have time to build more ideal woofers one day. The old paper 10" woofers are probably pretty high sensitivity (88-90 dB) but I don't have a proper calibrated mike at this time so I don't know for sure. I can make only relative measurements with my homebrew mike.
OK - I just wrote 6 paragraphs as an introduction! I hope the audience is not winded yet, because _I_ can be kinda long winded...
I tell friends my making motto is "think thrice, measure twice, cut once". Maybe I enjoy the planning stage of projects too much. Time to cut to the chase here and show you my proposal so all you tinkers and thinkers can comment on it. Assuming any of you have any attention span left! I have attached a drawing (.pdf).
(Note
If you are as excited about this experiment as I am and can't wait to see how it all turns out - I am sorry I will have to disappoint you. I will not be starting to build this weekend. Or even this year. Luckliy it is December... Next year is possible.)
Some points that I wonder about that maybe you can help me understand more clearly:
1. Two 8 ohm drivers of 86 db sensitivity in parallel. Do I get +3 db, or do I get more for now having a 4 ohm compound driver instead of an 8 ohm driver?
2. I have read the opinion that a true DA design approximates a point source pretty well. By "true" I mean 3rd order slopes, low XO, and, if memory serves, 90 deg phase at XO. Is that true (about the imaging aspect)?
3. Is a true DA really better in the point source sense, and so for imaging, than other ways of rigging up an MTM? I see that Thor, for instance, was not done that way.
In the plan I am proposing I think a true DA alignment could be done with a roughly 2500 Hz XO and still have the mids within one wavelength at XO. Maybe this is a particularly good application for the idealized DA design?
4. In the Statements Holtz put the two mids (also an MTM btw) in tunnels that happened to be of a length corresponding to 1/4 wave at about 300 Hz, which is close to the XO frequency he used. Lots of good reviews of those. I wonder if using a TL on a mid placed at some frequency well above the Fs of the driver results in any roll off below the tuning frequency, or maybe if all it does is to alter the slope of the roll off only below Fs? Which might make it possible to simplify the crossover?
My first experiment might just be to build the baffles, put it the FE83s, and stick pipes on the back to see how it sounds and what the FR looks like.
I totally support you in staying with cheap drivers, and using small drivers for midrange and tweeters.
I can see why manufacturers make 3 driver speakers with 2 being identical.
I cant see why DIY people able to source from mulitple manufacturers would, unless fundamentally limited by an insistance on passive cross overs.
We can then get speakers better optimised for frequency ranges, and have to design the cross overs anyway so going for active cross overs is both easier if we go DSP and potentially cheaper if we are carefull and go analogue.
If I was going to go for off the shelf moving coil speakers I would be tempted with active cross overs and chip amps for each driver, even if I was on a low budget. Especially for the Bass. Midrange tweeter could possibly be ok on one amp though, I would go the full way, or play with more expensive full range drivers and with a bass driver. Maybe adding a tweeter later if needed, if I had to avoid the aditional cost of 3 way active.
Parts Express buy out drivers would be what I would get if I lived in the USA. In europe we never see such discounting so for fun I pretended I was in USA and looking to build a 3 way speaker from just parts express buy out drivers:
Tweeter:
https://www.parts-express.com/1-2-mylar-dome-shielded-tweeter-8-ohm--279-002 Just 99 Cents!
Midrange:
https://www.parts-express.com/fountek-fe83-3-full-range-driver-8-ohm--299-020
or
peerless-tc6wd02-04-silver-series-2-midrange--299-008
Bass:
hivi-b4n-4-aluminum-midbass-round-frame--297-429
Or maybe:
wfr0400215-4-cast-aluminum-frame-woofer-4-ohm--299-346
Personally I have been thinking along similar lines. But have yet to come to a conclusion.
I will probaly procrastinate unitll I build some more ESL midrange/tweeters anyway here is my much more expensive european buyers choice for a small desktop 3 way.
Seas 19TFF1 0,75 inch tweeter
Peerless 830970 Full range.
Peerless 830855 Subwoofer
Note us europeans will be paying a lot more for each driver, butr admitedly each of these drivers would proabbly do the job better than the buy out specials I listed above.
In both cases I am not bothering to match sensitivity, as I dont see any point in going for passive crossovers when amplifier modules are so cheap, even if you dont want to salvage amps from second hand gear:
https://www.parts-express.com/t-amp-tripath-ta2024-2x15w-audio-digital-amplifier-board--320-600
I can see why manufacturers make 3 driver speakers with 2 being identical.
I cant see why DIY people able to source from mulitple manufacturers would, unless fundamentally limited by an insistance on passive cross overs.
We can then get speakers better optimised for frequency ranges, and have to design the cross overs anyway so going for active cross overs is both easier if we go DSP and potentially cheaper if we are carefull and go analogue.
If I was going to go for off the shelf moving coil speakers I would be tempted with active cross overs and chip amps for each driver, even if I was on a low budget. Especially for the Bass. Midrange tweeter could possibly be ok on one amp though, I would go the full way, or play with more expensive full range drivers and with a bass driver. Maybe adding a tweeter later if needed, if I had to avoid the aditional cost of 3 way active.
Parts Express buy out drivers would be what I would get if I lived in the USA. In europe we never see such discounting so for fun I pretended I was in USA and looking to build a 3 way speaker from just parts express buy out drivers:
Tweeter:
https://www.parts-express.com/1-2-mylar-dome-shielded-tweeter-8-ohm--279-002 Just 99 Cents!
Midrange:
https://www.parts-express.com/fountek-fe83-3-full-range-driver-8-ohm--299-020
or
peerless-tc6wd02-04-silver-series-2-midrange--299-008
Bass:
hivi-b4n-4-aluminum-midbass-round-frame--297-429
Or maybe:
wfr0400215-4-cast-aluminum-frame-woofer-4-ohm--299-346
Personally I have been thinking along similar lines. But have yet to come to a conclusion.
I will probaly procrastinate unitll I build some more ESL midrange/tweeters anyway here is my much more expensive european buyers choice for a small desktop 3 way.
Seas 19TFF1 0,75 inch tweeter
Peerless 830970 Full range.
Peerless 830855 Subwoofer
Note us europeans will be paying a lot more for each driver, butr admitedly each of these drivers would proabbly do the job better than the buy out specials I listed above.
In both cases I am not bothering to match sensitivity, as I dont see any point in going for passive crossovers when amplifier modules are so cheap, even if you dont want to salvage amps from second hand gear:
https://www.parts-express.com/t-amp-tripath-ta2024-2x15w-audio-digital-amplifier-board--320-600
You will get superior sonic details from a TMW where only one speaker covers each range, than from an MTM with two drivers having slightly different T/S parameters mounted at slightly different listener angles trying to blend.
The best way to save money is to find a great well reviewed design and copy it. No surprises! No wasted time. No wasted passive parts. No wasted wood and finish.
Make a list.... check it twice.... what specs sound nice?
Physical size? Freq range? SPL/watt? Xover freqs?
From your initial goals, maybe look for:
No-frame 0.75" neo silk dome tweeter
+ 3" paper midrange in sealed box for best transients
+ 8" paper woofer
Cool Table of full range speakers for comparison:
http://www.spectrumaudio.de/breit/tabelle.htm
A quick search did not find 3-ways designs for the Faital 3FE25 ($19) or Tang Band W3-1364SA Bamboo Cone ($25). I would look for 3" and 4" cone midrange designs since they can cover most of the vocal range (80-1100Hz).
The best way to save money is to find a great well reviewed design and copy it. No surprises! No wasted time. No wasted passive parts. No wasted wood and finish.
Make a list.... check it twice.... what specs sound nice?
Physical size? Freq range? SPL/watt? Xover freqs?
From your initial goals, maybe look for:
No-frame 0.75" neo silk dome tweeter
+ 3" paper midrange in sealed box for best transients
+ 8" paper woofer
Cool Table of full range speakers for comparison:
http://www.spectrumaudio.de/breit/tabelle.htm
A quick search did not find 3-ways designs for the Faital 3FE25 ($19) or Tang Band W3-1364SA Bamboo Cone ($25). I would look for 3" and 4" cone midrange designs since they can cover most of the vocal range (80-1100Hz).
Hi Datinker,
Two drivers in parallel give +6 dB in SPL. Assuming a (more realistic) sensitivity of 84 dB per driver you effectively get 90 dB at 2.83V/1m. However, baffle step will steal again some dBs at low frequencies.
Whether two 3" drivers are better than a single 4.5" or vice versa depends very much on the actual distortion profile of the drivers chosen. The vertical dispersion pattern of MTM is symmetrical by definition, an MT configuration always is asymmetrical. But that doesn't make MTM superior IMO. The idea of creating a virtual sound source exactly at the location of the tweeter is tempting, no question, but at a listening distance of 3 m it hardly makes a real difference. The error introduced by crossover and mismatch of the acoustic centers probably is far more important.
Allow me a remark to the irregular baffle. Worst case is a round baffle, because all edges are equidistant to the driver. A classic rectangular baffle actually is quite good in smearing edge diffraction. Hence I expect your irregular baffle rather to be worse than better!
Two drivers in parallel give +6 dB in SPL. Assuming a (more realistic) sensitivity of 84 dB per driver you effectively get 90 dB at 2.83V/1m. However, baffle step will steal again some dBs at low frequencies.
Whether two 3" drivers are better than a single 4.5" or vice versa depends very much on the actual distortion profile of the drivers chosen. The vertical dispersion pattern of MTM is symmetrical by definition, an MT configuration always is asymmetrical. But that doesn't make MTM superior IMO. The idea of creating a virtual sound source exactly at the location of the tweeter is tempting, no question, but at a listening distance of 3 m it hardly makes a real difference. The error introduced by crossover and mismatch of the acoustic centers probably is far more important.
Allow me a remark to the irregular baffle. Worst case is a round baffle, because all edges are equidistant to the driver. A classic rectangular baffle actually is quite good in smearing edge diffraction. Hence I expect your irregular baffle rather to be worse than better!
Regarding owenhamburg's ideas - I think you are right, it is getting to the point where passive X/Os are not less expensive than electronic bi-amped alternatives. I may well run these proposed MTM boxes and the woofer boxes on separate amps and can then mostly avoid the issue of matching sensitivities.
Still, I think the matter of choosing slopes and X/O points for electronic X/Os will still follow the same rules as for designing passive X/Os so it is worth understanding those filter design parameters either way, so I posed a couple of questions about that.
linesource - that is a nice table of driver specs. You have my initial goals understood correctly. I do in fact ultimately plan to use an 8" woofer. I did start by looking for a 3" wide range driver, as you suggest. Then I became side tracked by the MTM idea. I'm not sure I agree that the two small mids rather close together will necessarily be a noticeable problem when listening from living room distances. But then that is why I posted the idea - to collect and review these opinions.
Dissi - 6 db for two parallel drivers - Roger that. I have read it but I wasn't sure I believed it. That is a point in favor of the idea I proposed. You point out that the errors from crossover and mismatch of acoustic centers will likely make the biggest differences in sound quality. That is my feeling as well. Which is why I proposed to match the tweeter depth to the mids and posed questions about crossover slopes.
The point about the irregular baffle probably not being an improvement over a rectangular one is good. It would certainly be easier to make the simpler shape. What do you think about just using a larger radius on the edges of a rectangle in the area adjacent to the tweeter?
In general, I do not mean to give the impression that I think the MTM(W) design is inherently superior to the MT(W) plan. I am just exploring the idea that it might be almost as good in_the_particular_case where the two Ms are small. Maybe even better when compared to an MT(W) with the _same_ small M, all other things being equal. Because of the greater sensitivity and lower distortion possible when using two and not having to push them as hard as when using just one of the same drivers.
Still, I think the matter of choosing slopes and X/O points for electronic X/Os will still follow the same rules as for designing passive X/Os so it is worth understanding those filter design parameters either way, so I posed a couple of questions about that.
linesource - that is a nice table of driver specs. You have my initial goals understood correctly. I do in fact ultimately plan to use an 8" woofer. I did start by looking for a 3" wide range driver, as you suggest. Then I became side tracked by the MTM idea. I'm not sure I agree that the two small mids rather close together will necessarily be a noticeable problem when listening from living room distances. But then that is why I posted the idea - to collect and review these opinions.
Dissi - 6 db for two parallel drivers - Roger that. I have read it but I wasn't sure I believed it. That is a point in favor of the idea I proposed. You point out that the errors from crossover and mismatch of acoustic centers will likely make the biggest differences in sound quality. That is my feeling as well. Which is why I proposed to match the tweeter depth to the mids and posed questions about crossover slopes.
The point about the irregular baffle probably not being an improvement over a rectangular one is good. It would certainly be easier to make the simpler shape. What do you think about just using a larger radius on the edges of a rectangle in the area adjacent to the tweeter?
In general, I do not mean to give the impression that I think the MTM(W) design is inherently superior to the MT(W) plan. I am just exploring the idea that it might be almost as good in_the_particular_case where the two Ms are small. Maybe even better when compared to an MT(W) with the _same_ small M, all other things being equal. Because of the greater sensitivity and lower distortion possible when using two and not having to push them as hard as when using just one of the same drivers.
YG Audio uses MTM with success in their flagship speaker Sonja. The layout sets some challeges of course
YG Acoustics Sonja 1.3 loudspeaker Measurements | Stereophile.com
YG Acoustics Sonja 1.3 loudspeaker Measurements | Stereophile.com
Maybe just adding to confusion;
But its nice to remind people that you require less power as you go higher in frequancy.
Hence even if you go MTMW it might be worth exploring having the low pass filters for the two M speakers set at different values.
A 3.5 way speaker so to say
For my self:
My ideas are more based on size than yours, More a book shelf speaker for my work desk. So for me a 8" speaker is too big for my plans. I am probably going to go for a 3"-5" bass driver crossing over around 300Hz to 600Hz to a segmented home made electostatic driver, built with ideas taken from the Quad ESL63.
My main system:
For my main system, I have Quad ESL63's, and a pair of nice sounding 10" subwoofers using a sealed box. In my opinion thsi system has two floors, size, and complexity. The subwoofers are Peerless SLS 830668 based that sound very good, with an 2nd order active crossover at 120Hz as my main system for maximum sound quality. This has shown me sub 100Hz sound is not so important for most of my music, but nice to have when I want to hear some Reggae / Rap / Funk and shake the flat with the bass turned up.
My impresion is the Peerless SLS 830668 subwoofers are very good to about 400Hz, but I prefure the Electorstatic sound in the 120Hz - 400 Hz regon, and for quiet accoustic music listening prefure the Quads full range. So the next plan for the main system is to have more than one cross over point, for different music moods, and times of day.
My rational:
Quad ESL63's being dipole are far better late at night listening than the subwoofer combo as the bass does not carry out of my room so badly and at low volumes thiers plenty. So next is to put a switch in the cross over.
While other times I woudl like to give the bass drivers everything below 400Hz for the moments when you want to crank up the volume with an electric bass or feel the a blast of an explosion in a movie hit you in the chest. (hopefull not upsetting the neighbours)
But its nice to remind people that you require less power as you go higher in frequancy.
Hence even if you go MTMW it might be worth exploring having the low pass filters for the two M speakers set at different values.
A 3.5 way speaker so to say
For my self:
My ideas are more based on size than yours, More a book shelf speaker for my work desk. So for me a 8" speaker is too big for my plans. I am probably going to go for a 3"-5" bass driver crossing over around 300Hz to 600Hz to a segmented home made electostatic driver, built with ideas taken from the Quad ESL63.
My main system:
For my main system, I have Quad ESL63's, and a pair of nice sounding 10" subwoofers using a sealed box. In my opinion thsi system has two floors, size, and complexity. The subwoofers are Peerless SLS 830668 based that sound very good, with an 2nd order active crossover at 120Hz as my main system for maximum sound quality. This has shown me sub 100Hz sound is not so important for most of my music, but nice to have when I want to hear some Reggae / Rap / Funk and shake the flat with the bass turned up.
My impresion is the Peerless SLS 830668 subwoofers are very good to about 400Hz, but I prefure the Electorstatic sound in the 120Hz - 400 Hz regon, and for quiet accoustic music listening prefure the Quads full range. So the next plan for the main system is to have more than one cross over point, for different music moods, and times of day.
My rational:
Quad ESL63's being dipole are far better late at night listening than the subwoofer combo as the bass does not carry out of my room so badly and at low volumes thiers plenty. So next is to put a switch in the cross over.
While other times I woudl like to give the bass drivers everything below 400Hz for the moments when you want to crank up the volume with an electric bass or feel the a blast of an explosion in a movie hit you in the chest. (hopefull not upsetting the neighbours)
Last edited:
I have thought about the idea of arranging the MTM unit as a 2.5-way. It seems like a difficult thing to work out in the sense that I would be getting the higher sensitivity (+6 db) of the parallel drivers below the 0.5 way X/O frequency but not above it. It is easy to imagine how that works for baffle step compensation if the MTM unit was going to be the final product, a full range speaker. It seems trickier when it is going to be part of a 3-way system. I suppose I could get comfortable with a good modeling program and try to work that out.
Which brings me to the next comment from jReaves about diffraction simulation programs - no, I am not familiar with those. Is there one you particularly like?
Which brings me to the next comment from jReaves about diffraction simulation programs - no, I am not familiar with those. Is there one you particularly like?
Yes, yes, yes - get comfortable with a good modeling program, like PCD or WinPCD.
For baffle diffraction, these 3 programs will handle slightly different scenarios:
Baffle Diffraction and Boundary Simulator
Baffle Diffraction Simulator
The Edge
To design your box and driver placement, these 2 tools are indispensable:
SPLTrace
Frequency Response Modeler
What you do is convert a driver's FR into data points (frd files) with SPLTrace. Then enter that into the Frequency Response Modeler. Down at the bottom, this program includes the 1st baffle diffraction simulator I listed above. Now you can start playing around with cabinet dimensions, driver positioning and edge radius and then see immediately how that will sum together with the FR of your imported driver up at the top of the program. Your goal is to create the flattest summed response as you can.
Very, very useful.
I would skip the .5-way idea for the mids in a 3-way. Just go with the MTM idea. With small mids like that, you want both drivers to be handling the mid's whole frequency range to keep distortion down and SPL's up. There are plenty of excellent sounding MTM's out there; it's not an arrangement that you need to avoid. If you try to keep the cabinet quite narrow though, you may have to cross to the woofer just a little bit higher than usual (maybe 500Hz-700Hz), but that may be a good thing with those drivers as the harmonic distortion will start rising as you go lower in frequency.
For baffle diffraction, these 3 programs will handle slightly different scenarios:
Baffle Diffraction and Boundary Simulator
Baffle Diffraction Simulator
The Edge
To design your box and driver placement, these 2 tools are indispensable:
SPLTrace
Frequency Response Modeler
What you do is convert a driver's FR into data points (frd files) with SPLTrace. Then enter that into the Frequency Response Modeler. Down at the bottom, this program includes the 1st baffle diffraction simulator I listed above. Now you can start playing around with cabinet dimensions, driver positioning and edge radius and then see immediately how that will sum together with the FR of your imported driver up at the top of the program. Your goal is to create the flattest summed response as you can.
Very, very useful.
I would skip the .5-way idea for the mids in a 3-way. Just go with the MTM idea. With small mids like that, you want both drivers to be handling the mid's whole frequency range to keep distortion down and SPL's up. There are plenty of excellent sounding MTM's out there; it's not an arrangement that you need to avoid. If you try to keep the cabinet quite narrow though, you may have to cross to the woofer just a little bit higher than usual (maybe 500Hz-700Hz), but that may be a good thing with those drivers as the harmonic distortion will start rising as you go lower in frequency.
Generally it is always better to go TMW rather than MTMW if you can help it. I am a large proponent of an even off axis response and constant directivity designs. YG Acoustics use machined wave guides along with their MTMs , which is always a good idea. If you can cross over low enough and keep the MM's close enough then these problems are not anywhere as severe.
You need less average power as you go up higher in frequency. It is still possible to find recordings that contain a lot of high peak amplitude transients at high frequencies, where high average power isn't required, yet an amplifier with a high voltage swing capability is.
The SS 10F really isn't a 4" driver and is more like a 3" in regards to it's cone surface area.
The trouble you are going to run into is finding small 3" drivers that have high sensitivity. This is something you've already found out and the idea of using an MTM of 3" drivers, with a crossover of around 2kHz with that dinky Vifa is a very good idea if you go this route.
Now if you want to create a little MTM speaker that can stand on its own two feet, rather than needing the bass module, for decent listening, then the 10F is not a good idea.
I am trying to think off the top of my head of some 3" 8 ohm driver that would be ideal for your situation and I am struggling to. The only ones that jump out at me from memory are the RS75 or RS100 drivers from Dayton Audio, both the paper and the aluminium coned versions. These are low distortion and high value drivers, they look awesome and some also come with truncated frames, specifically designed for tight C2C situations. I've used both the RS125 and RS225 in past designs and they are brilliant drivers, especially for what you pay.
You need less average power as you go up higher in frequency. It is still possible to find recordings that contain a lot of high peak amplitude transients at high frequencies, where high average power isn't required, yet an amplifier with a high voltage swing capability is.
The SS 10F really isn't a 4" driver and is more like a 3" in regards to it's cone surface area.
The trouble you are going to run into is finding small 3" drivers that have high sensitivity. This is something you've already found out and the idea of using an MTM of 3" drivers, with a crossover of around 2kHz with that dinky Vifa is a very good idea if you go this route.
Now if you want to create a little MTM speaker that can stand on its own two feet, rather than needing the bass module, for decent listening, then the 10F is not a good idea.
I am trying to think off the top of my head of some 3" 8 ohm driver that would be ideal for your situation and I am struggling to. The only ones that jump out at me from memory are the RS75 or RS100 drivers from Dayton Audio, both the paper and the aluminium coned versions. These are low distortion and high value drivers, they look awesome and some also come with truncated frames, specifically designed for tight C2C situations. I've used both the RS125 and RS225 in past designs and they are brilliant drivers, especially for what you pay.
Just angle two small(er) fullrangers horizontally so the primary lopes for the highest frequencies doesn't intersect (where compfiltering is most objectionable).
Then you will have made the tweeter unnecessary (the main reason for the use of a tweeter is to reduce beaming) and have eliminated the crossover.
Then you will have made the tweeter unnecessary (the main reason for the use of a tweeter is to reduce beaming) and have eliminated the crossover.
Some of the programs you listed have versions for different versions of Excel, 2003 and 2007 I believe. I have not tried those yet because of the problem described below. Should I expect problems if I have a more recent version of Office installed? I may be able to find a PC that is not fully up to date if I need to.
I downloaded and installed SplTrace yesterday but I could not get it working. Is it possible that there is a compatibility issue with Windows 7 (64)? It opens and runs, it allows me to set high and low limits on a graph imported from the clipboard from an Acrobat Reader image, but when I set it to tracing the graph it just runs, or should I say "runs", forever. If I stop trace and/or save file I get a 0kb file.
I am looking forward to trying this out. Thanks for the tips!
I had wondered about that as well. It is one of the things I hope to address in the modelling programs.
I downloaded and installed SplTrace yesterday but I could not get it working. Is it possible that there is a compatibility issue with Windows 7 (64)? It opens and runs, it allows me to set high and low limits on a graph imported from the clipboard from an Acrobat Reader image, but when I set it to tracing the graph it just runs, or should I say "runs", forever. If I stop trace and/or save file I get a 0kb file.
I am looking forward to trying this out. Thanks for the tips!
I had wondered about that as well. It is one of the things I hope to address in the modelling programs.
I haven't heard of compatibility problems with SPLTrace. But it did stump me the 1st time I tried to give it a go as well. See if one of these threads help:
spl trace question
FRD/ZMA files
I'm not really much of a computer guru so I'm not sure about the other compatibility issues. I think I've heard that Excel no longer works, or to be more precise, Excel VBA code no longer works in the 64 (bit?) systems. I thought Windows 7 was ok. I could be wrong. I'm still on XP and Office 2007 and that works for me. You need to have the right Add-ins enabled though. That should be in the relevant instructions.
A few more helpful links:
FRD Consortium
Unibox
WinISD
The Speaker Building Bible
spl trace question
FRD/ZMA files
I'm not really much of a computer guru so I'm not sure about the other compatibility issues. I think I've heard that Excel no longer works, or to be more precise, Excel VBA code no longer works in the 64 (bit?) systems. I thought Windows 7 was ok. I could be wrong. I'm still on XP and Office 2007 and that works for me. You need to have the right Add-ins enabled though. That should be in the relevant instructions.
A few more helpful links:
FRD Consortium
Unibox
WinISD
The Speaker Building Bible
What about multicell horns and perhaps more apt linearrays that employ only FR drivers? No problems there. Nothing wrong with cone resonances or breakup modes if they are controlled. Some of the best drivers and speakers ever has used them to their advantage. And it's not as if they are absent from commonly used drivers either.
Have you ever really tried it? You will be surprised.
Last edited:
I wonder why some are so taken by coax/concentrics like KEF UniQ speakers? Perhaps it is good imaging, clear transients, brilliant highs?
-
Double mids, double tweeters, BMR radiators, large fullranges or line arrays have some advantages but they try to make a "wall of sound" of "cylindrical wavefront" which is mostly just a network of criss-crossing soundwaves at high frequencies.
Comb Filtering, Acoustical Interference, & Power Response in Loudspeakers | Audioholics
-
Double mids, double tweeters, BMR radiators, large fullranges or line arrays have some advantages but they try to make a "wall of sound" of "cylindrical wavefront" which is mostly just a network of criss-crossing soundwaves at high frequencies.
Comb Filtering, Acoustical Interference, & Power Response in Loudspeakers | Audioholics
Last edited:
^^ This is your proposed design, using dual 3" woofers flanking a tweeter
Here's a speaker project I published a couple months ago, using the same configuration
Here's the project that I'm listening to at the moment.
Personally, I prefer the latter over the former. Here's why:
1) It doesn't take much of a waveguide to horn load a midrange. A waveguide just 8.25cm deep will load a midrange down to about 1khz
2) Even with dual 3" woofers, I found that the former project could be pushed to it's limits. Basically it was possible to 'bottom out' the woofers
3) The use of unity summation aperture on the new project improves the intelligibility, the imaging, the polar response, etc
Either project is good, but the second one is better IMHO
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.