Just taking a lead from Mark H., and starting a new thread on my 2.5-way DIY speaker design project. Yes, please share your HDS measurements!
Specs:
Tweeter - Vifa D25AG05
1.0 Midbass - Peerless HDS 439
0.5 Midbass - Peerless HDS 439
Ported with two 2" wide ports around 6" long for a tune of 50Hz in a 1 ft^3 braced enclosure (10"x21.5x12.5").
The idea behind a 2.5-way is to use the tweeter and one of the midbass driver as the nominal "2-way" speaker, and then you throw in the 0.5 midbass to fill in the bass that you lose from the baffle step loss. My cursory understanding of the baffle step loss: bass radiates in 4pi space all around the baffle, while tweeter radiates in 2pi space in front of the baffle, so there is anywhere from 3dB to 6dB loss in the bass region if you don't account for the baffle step loss. The compensate for it, most people can elect to put a parallel R-L circuit in series with the tweeter and this effectively lowers the tweeter output to match and compensate for the loss in bass from the baffle step. The only drawback is that you lose a little bit of overall speaker sensitivity but you gain a fuller sounding speaker. Here's a link on baffle step loss which better explains it. There are more than one way to skin a cat in this case.
So by using a 2.5-way, I can roll off the 0.5 midbass driver with a 1st order (inductor) around 400-500Hz which compensates for the baffle step loss, without muddying the midrange from 500-2500Hz before the midrange/tweeter crossover kicks in. This also preserves the overall speaker efficiency to match the midrange (around 87.5dB).
Originally I was going for a TMM configuration on my speaker's front baffle. But a kind gentleman, (Paul V) over at the Madisound audio discussion board shared this nifty baffle diffraction frequency simulator (excel spreadsheet) which allows you to simulate various driver placements to get the "best" configuration for the drivers on the front baffle. So after I spec'd out the front baffle dimension (8.5"x20"), he came up with a "optimum" driver layout, and it happens to look like a MTM design, but the drivers are offset from the center vertical line by 0.75" and the drivers are very close to the tweeter (5.4" = center-to-center distance (CtC)) and the tweeter is located around 9.85" from the top of the front baffle.
The other interesting thing is that you get some really nice reduction in baffle diffraction if you also incorporate a roundover on the baffle edges with a radius of 1". I will try to do a roundover with a radius 0.75" if I can find the right roundover bit without costing me an arm and a leg.
Using the Peerless HDS 439 allows me to get the CtC for the tweeter and midbass drivers as close as 5" because the HDS439 has its top and bottom flanges sliced off just for this purpose. Here's a front shot of the driver.
Since I'm using the Vifa D25AG06 which has a Fs of 1500Hz, I needed to crossover near twice the Fs, around 3000Hz. Using the rule of keeping the CtC equal or less than the wavelength of the crossover frequency, I decided on crossing the midrange and tweeter at 2750Hz, and most likely the midwoofer crossover will be a 4th order Linkwitz-Riley filter, while I may play around with the tweeter crossover, using either a 2nd order L-R or 3rd-order L-R and let the tweeter's natural slope help out.
The quick and dirty calculation for the CtC is to use CtC=13585/Fc, where Fc is the desired crossover frequency and the 13585 is the velocity of sound in inches.
I will most likely have to pad the tweeter down around 3dB to match the sensitivity of the midrange (88dB vs. 90/91dB).
Here's a link to the Vifa D25AG05 specs and the site for Vifa main site.
Why this particular model, and not the chambered version D25AG35? Well, I just happen to have one of these tweeters, and I've read mostly favorably reviews of it (though it has the high Fs, resonant frequency), and the D25AG35 isn't available locally for me, while the D25AG05 is.
I have considered the Vifa D27TG45, but it seems to have more critics and unreliable Fs from the current production run (but the variance of the Fs seems to bump it up from the specs of 650Hz to around 1250Hz, so while still a workable tweeter with a ruler flat response - which doesn't seem to be in vogue currently for those who want a super low crossover point near 2000Hz-2200Hz. There are many other choices like the Morel MDT-30, the Seas 27TFFC, the Hiquphon OWI, all of which I considered but I didn't like the price on the Morel ($50 per tweeter), the Seas ($26) would be comparable, maybe a little better than the Vifa D25AG05 ($22) that I'll be using, and the Hiquphon OWI ($75) cost a bit too much but has a remarkably flat response out to 20KHz, and very nice off-axis response. Of course, the highly regard ScanSpeak 9500 ($85), 9700 ($$$) aren't really options for me at this point because of their high costs.
As to why I choose the Peerless HDS439 midbass driver, well, I've used both Vifa and Peerless before, and Peerless spec are really pretty spot on for their production units. The HDS 439 ($47) has a nice flat response from 100Hz to around 3500Hz, and then a bump from 3500-4000Hz (the 4th order L-R lowpass filter at 2750hz is supposed to tame that peak). Here's a link to the Peerless HDS drivers page - click on the "850439" link on that page for the specs.
The contenders for the midbass driver were the Vifa M18WO09 ($50, paper cone, which I like, but had this weird dip around 1400Hz, otherwise not a bad midwoofer), Vifa P17WJ00 ($38, never a big fan of polycone woofers, but still pretty highly recommended), Vifa PL18WO09 (a bit pricey at $63, but didn't do so well in the distortion tests that Linkwitz ran on a bunch of midrange driver, ScanSpeak 18W/8545 (just way too pricey ~$140?), Seas P17REX ($45)(while highly recommended, their frequency response graph might have been too revealing). A quote from the Linkwitz article: "The Peerless HDS 850439 is the definite leader in the performance versus price category for this group. It is well designed and built." So that just about sold me on the drivers enough to order 4 of them today.
Also over at Dillon Acoustics there's a $60 plans for "The Metaphor" speakers that you can buy that specs out the HDS439 for use with the Hiquphon OWI tweeter. You'll note with just the TM configuration, his efficiency is around 86dB, I should gain that 1.5dB with the added 0.5 midbass driver. Now is it worth it? I'll find out! I figure I'm exchanging $92 in 2 extra midbass drivers for that extra bit of efficiency over paying $100 more for the Hiquphon OWI tweeters. Hope I made the right choice.
Just bought the MDF tonight, and hope to start cutting it into panels by the weekend.
For determining the crossover component values, I'll be using the SpeakerMania website, where you can click on their "Filter Design" link for the crossover components, and their "Correction Circuits" link for the L-pad for the tweeter, and maybe a Zobel to normalize the impedance of the HDS midrange if needed.
Well, hope I didn't bore all 3 of you who'll read this. I hope to make some progress with this project, and of course document it for a future webpage. It won't be the cheapest 2.5-way DIY project, far from it. I just don't see the point in using pot-luck parts on sale and have to finesse the crossover enormously to get good sound from the inexpensive parts to tame peaks with notch filters or traps. If I'm going to put them in my house, I might as well make them look and sound good and know that I have quality parts and that others have a shot of duplicating the project without having to worry about getting the same spec'd out drivers.
PatCave; HT Pix;Gear;DIY Projects;DVDs; LDs
Specs:
Tweeter - Vifa D25AG05
1.0 Midbass - Peerless HDS 439
0.5 Midbass - Peerless HDS 439
Ported with two 2" wide ports around 6" long for a tune of 50Hz in a 1 ft^3 braced enclosure (10"x21.5x12.5").
The idea behind a 2.5-way is to use the tweeter and one of the midbass driver as the nominal "2-way" speaker, and then you throw in the 0.5 midbass to fill in the bass that you lose from the baffle step loss. My cursory understanding of the baffle step loss: bass radiates in 4pi space all around the baffle, while tweeter radiates in 2pi space in front of the baffle, so there is anywhere from 3dB to 6dB loss in the bass region if you don't account for the baffle step loss. The compensate for it, most people can elect to put a parallel R-L circuit in series with the tweeter and this effectively lowers the tweeter output to match and compensate for the loss in bass from the baffle step. The only drawback is that you lose a little bit of overall speaker sensitivity but you gain a fuller sounding speaker. Here's a link on baffle step loss which better explains it. There are more than one way to skin a cat in this case.
So by using a 2.5-way, I can roll off the 0.5 midbass driver with a 1st order (inductor) around 400-500Hz which compensates for the baffle step loss, without muddying the midrange from 500-2500Hz before the midrange/tweeter crossover kicks in. This also preserves the overall speaker efficiency to match the midrange (around 87.5dB).
Originally I was going for a TMM configuration on my speaker's front baffle. But a kind gentleman, (Paul V) over at the Madisound audio discussion board shared this nifty baffle diffraction frequency simulator (excel spreadsheet) which allows you to simulate various driver placements to get the "best" configuration for the drivers on the front baffle. So after I spec'd out the front baffle dimension (8.5"x20"), he came up with a "optimum" driver layout, and it happens to look like a MTM design, but the drivers are offset from the center vertical line by 0.75" and the drivers are very close to the tweeter (5.4" = center-to-center distance (CtC)) and the tweeter is located around 9.85" from the top of the front baffle.
The other interesting thing is that you get some really nice reduction in baffle diffraction if you also incorporate a roundover on the baffle edges with a radius of 1". I will try to do a roundover with a radius 0.75" if I can find the right roundover bit without costing me an arm and a leg.
Using the Peerless HDS 439 allows me to get the CtC for the tweeter and midbass drivers as close as 5" because the HDS439 has its top and bottom flanges sliced off just for this purpose. Here's a front shot of the driver.
Since I'm using the Vifa D25AG06 which has a Fs of 1500Hz, I needed to crossover near twice the Fs, around 3000Hz. Using the rule of keeping the CtC equal or less than the wavelength of the crossover frequency, I decided on crossing the midrange and tweeter at 2750Hz, and most likely the midwoofer crossover will be a 4th order Linkwitz-Riley filter, while I may play around with the tweeter crossover, using either a 2nd order L-R or 3rd-order L-R and let the tweeter's natural slope help out.
The quick and dirty calculation for the CtC is to use CtC=13585/Fc, where Fc is the desired crossover frequency and the 13585 is the velocity of sound in inches.
I will most likely have to pad the tweeter down around 3dB to match the sensitivity of the midrange (88dB vs. 90/91dB).
Here's a link to the Vifa D25AG05 specs and the site for Vifa main site.
Why this particular model, and not the chambered version D25AG35? Well, I just happen to have one of these tweeters, and I've read mostly favorably reviews of it (though it has the high Fs, resonant frequency), and the D25AG35 isn't available locally for me, while the D25AG05 is.
I have considered the Vifa D27TG45, but it seems to have more critics and unreliable Fs from the current production run (but the variance of the Fs seems to bump it up from the specs of 650Hz to around 1250Hz, so while still a workable tweeter with a ruler flat response - which doesn't seem to be in vogue currently for those who want a super low crossover point near 2000Hz-2200Hz. There are many other choices like the Morel MDT-30, the Seas 27TFFC, the Hiquphon OWI, all of which I considered but I didn't like the price on the Morel ($50 per tweeter), the Seas ($26) would be comparable, maybe a little better than the Vifa D25AG05 ($22) that I'll be using, and the Hiquphon OWI ($75) cost a bit too much but has a remarkably flat response out to 20KHz, and very nice off-axis response. Of course, the highly regard ScanSpeak 9500 ($85), 9700 ($$$) aren't really options for me at this point because of their high costs.
As to why I choose the Peerless HDS439 midbass driver, well, I've used both Vifa and Peerless before, and Peerless spec are really pretty spot on for their production units. The HDS 439 ($47) has a nice flat response from 100Hz to around 3500Hz, and then a bump from 3500-4000Hz (the 4th order L-R lowpass filter at 2750hz is supposed to tame that peak). Here's a link to the Peerless HDS drivers page - click on the "850439" link on that page for the specs.
The contenders for the midbass driver were the Vifa M18WO09 ($50, paper cone, which I like, but had this weird dip around 1400Hz, otherwise not a bad midwoofer), Vifa P17WJ00 ($38, never a big fan of polycone woofers, but still pretty highly recommended), Vifa PL18WO09 (a bit pricey at $63, but didn't do so well in the distortion tests that Linkwitz ran on a bunch of midrange driver, ScanSpeak 18W/8545 (just way too pricey ~$140?), Seas P17REX ($45)(while highly recommended, their frequency response graph might have been too revealing). A quote from the Linkwitz article: "The Peerless HDS 850439 is the definite leader in the performance versus price category for this group. It is well designed and built." So that just about sold me on the drivers enough to order 4 of them today.
Also over at Dillon Acoustics there's a $60 plans for "The Metaphor" speakers that you can buy that specs out the HDS439 for use with the Hiquphon OWI tweeter. You'll note with just the TM configuration, his efficiency is around 86dB, I should gain that 1.5dB with the added 0.5 midbass driver. Now is it worth it? I'll find out! I figure I'm exchanging $92 in 2 extra midbass drivers for that extra bit of efficiency over paying $100 more for the Hiquphon OWI tweeters. Hope I made the right choice.
Just bought the MDF tonight, and hope to start cutting it into panels by the weekend.
For determining the crossover component values, I'll be using the SpeakerMania website, where you can click on their "Filter Design" link for the crossover components, and their "Correction Circuits" link for the L-pad for the tweeter, and maybe a Zobel to normalize the impedance of the HDS midrange if needed.
Well, hope I didn't bore all 3 of you who'll read this. I hope to make some progress with this project, and of course document it for a future webpage. It won't be the cheapest 2.5-way DIY project, far from it. I just don't see the point in using pot-luck parts on sale and have to finesse the crossover enormously to get good sound from the inexpensive parts to tame peaks with notch filters or traps. If I'm going to put them in my house, I might as well make them look and sound good and know that I have quality parts and that others have a shot of duplicating the project without having to worry about getting the same spec'd out drivers.
PatCave; HT Pix;Gear;DIY Projects;DVDs; LDs
Comment