Greetings,
I was looking for another design project so I decided to augment the bass output of my modified Tang Band 881S three-inch full-range drivers (publication in progress). For this project I wanted to keep the boxes sealed, as small as possible, and use a speaker level passive crossover. I am in my parquet period right now, and since pre-assembled parquet comes in one square foot panels, that meant an enclosure of one by one by one external dimensions, and after lamination and bracing, an internal volume of .6 cubic feet (including driver displacement).
Most of those wonderful full-range three-inch drivers (Tang Band, Fostex, Jordan, Bandor) have sensitivity ratings in the 86 to 87-db range. If you don’t want to run multiple drivers in an array and you don’t want to bi-amp, there are few woofers to match the sensitivities of those three-inch drivers. There are also few drivers with both low Fs and small Vas specifications.
With those design goals I chose to look at the Parts Express #295-335, 7-inch aluminum cone woofer. I don’t even like most metal-coned drivers, but with a free-air Fs of 36 Hz, Vas of .75 cubic foot, and a calculated sensitivity of 85 to 86 db, it seemed a likely candidate. Parts Express, however, publishes a series of BassBox simulations that predicted an F3 of about 63 Hz for my box. That is not as low as I wanted to go.
Fortunately, most box simulation programs are more for entertainment purposes than they are for real designing and I learned long ago not to put too much faith in any of the T/S alignments being better than close. Regardless, despite the simulation prediction, the measured near-field response of the box and driver is 53 Hz Fs (verified by impedance measurement) and 43 Hz F3. In a free space arrangement, peak output is at 100 Hz, down 1.5 db at 60 Hz, and down 3 db at 43 Hz.
In a down firing orientation, with the box elevated 2 inches off the floor, you do get some boundary interface loading of the driver. Near-field peak output is still at 100 Hz, but you get an extra 1.5 db of output. At one meter measuring distance, and allowing that there is room boundary contamination, you get almost 2 db more output centered at 90 Hz, a decrease in output magnitude at 280 Hz and a large peak at 370 Hz. If used in the down firing orientation, this suggests that 300 Hz would be an ideal crossover frequency.
Mark
I was looking for another design project so I decided to augment the bass output of my modified Tang Band 881S three-inch full-range drivers (publication in progress). For this project I wanted to keep the boxes sealed, as small as possible, and use a speaker level passive crossover. I am in my parquet period right now, and since pre-assembled parquet comes in one square foot panels, that meant an enclosure of one by one by one external dimensions, and after lamination and bracing, an internal volume of .6 cubic feet (including driver displacement).
Most of those wonderful full-range three-inch drivers (Tang Band, Fostex, Jordan, Bandor) have sensitivity ratings in the 86 to 87-db range. If you don’t want to run multiple drivers in an array and you don’t want to bi-amp, there are few woofers to match the sensitivities of those three-inch drivers. There are also few drivers with both low Fs and small Vas specifications.
With those design goals I chose to look at the Parts Express #295-335, 7-inch aluminum cone woofer. I don’t even like most metal-coned drivers, but with a free-air Fs of 36 Hz, Vas of .75 cubic foot, and a calculated sensitivity of 85 to 86 db, it seemed a likely candidate. Parts Express, however, publishes a series of BassBox simulations that predicted an F3 of about 63 Hz for my box. That is not as low as I wanted to go.
Fortunately, most box simulation programs are more for entertainment purposes than they are for real designing and I learned long ago not to put too much faith in any of the T/S alignments being better than close. Regardless, despite the simulation prediction, the measured near-field response of the box and driver is 53 Hz Fs (verified by impedance measurement) and 43 Hz F3. In a free space arrangement, peak output is at 100 Hz, down 1.5 db at 60 Hz, and down 3 db at 43 Hz.
An externally hosted image should be here but it was not working when we last tested it.
In a down firing orientation, with the box elevated 2 inches off the floor, you do get some boundary interface loading of the driver. Near-field peak output is still at 100 Hz, but you get an extra 1.5 db of output. At one meter measuring distance, and allowing that there is room boundary contamination, you get almost 2 db more output centered at 90 Hz, a decrease in output magnitude at 280 Hz and a large peak at 370 Hz. If used in the down firing orientation, this suggests that 300 Hz would be an ideal crossover frequency.
An externally hosted image should be here but it was not working when we last tested it.
Mark
Attachments
A crossover to match
Parts Express 295-335 Aluminum cone woofer and 52 mm full range driver frequency response. This graph is the two-meter far field frequency response spectrum showing crossover summing. Crossover frequency would be approximately 350 Hz. There is a slight 1.5 db depression in the low and high pass stop band regions. Frequency resolution and magnitude detail is limited due to limitations of using a quasi-anechoic measurement technique in a practical sized room. The test is useful only for showing level matching and crossover region summing.
Frequency responses of 52 mm full range driver in box, in box with first order filter, and in box with two components in the high pass filter. Driver free air resonance is 109 Hz and Qts is .63. Mounted in a .02 cubic foot enclosure, Fs is 190 Hz with output rising toward resonance. Add a series capacitor and while the response flattens between 260 and 400 Hz, the output at resonance is significantly up. Add a parallel inductor for a “second” order crossover and this resonance bump is eliminated.
While the small box is used as part of the high pass filter, there are consequences to the technique. Midrange ripple is increased and an equalizer network is required to maintain plus or minus three-db response in the far field pass-band. Even with the ripple, response is still plus or minus two db through the important 500 Hz to 2 kHz midrange region.
Ten-component crossover network.
As stated in my first post, I wanted to design a system with a larger diameter and higher power low frequency unit and a passive speaker level crossover for the three-inch full range drivers with sensitivities ranging from 86 to 88-db. Here is my design. While the system shown uses one of my prototype drivers, you can substitute a Tang Band or Fostex driver of equal diameter, impedance, and efficiency with good results. All the limitations of the other drivers will be preserved, except at the lower frequencies now handled by the seven-inch woofer.
The loudspeaker system is a modular design, with separate enclosures for the full range driver and woofer. The full range driver enclosure is intended to be part of the high pass filter and that fact determines the size. The examples shown have outside measurements of 130 mm by 130 mm by 87 mm deep. The sides and back are made from ¾ inch mdf and the front baffle of ¼ inch veneer plywood. The enclosure is liberally stuffed with poly fill material.
When I first listen to a new design, I am trying to hear what is wrong. This means that I am not deciding if I like a system right away. The liking part comes only after an initial evaluation and critiquing period. If I come to like a system, then I will like if for a long, long time. Indeed, I have some of my older designs that are more than 25 years old and I still enjoy listening to them. I have spent a number of hours listening to this system and I am beginning to like the sound.
I see no reason to be falsely enthusiastic about the sound, but it is quite good. I can also still critique it even though I built it with my own hands. For example, the low bass is a bit lush to my taste and the lower pitched male vocalists sound a little dry. Otherwise, the loudspeakers are clear and detailed, with excellent soundstage, extended range. French horns are reproduced with good accuracy. Piano is also excellent with preservation of tonal accuracy across the entire range of the instrument. Except for the already mentioned dryness in the lower vocal range, vocal reproduction is excellent, with great clarity.
There are many ways to achieve accuracy of reproduction and sound you can listen to for years. This design seems to be one of them.
Although I will put in more time listening before I make any changes, I am likely going to add more poly fill to the woofer cabinet. I will probably increase the poly fill from the current 8 ounces to 12 to 14 ounces.
Mark
Parts Express 295-335 Aluminum cone woofer and 52 mm full range driver frequency response. This graph is the two-meter far field frequency response spectrum showing crossover summing. Crossover frequency would be approximately 350 Hz. There is a slight 1.5 db depression in the low and high pass stop band regions. Frequency resolution and magnitude detail is limited due to limitations of using a quasi-anechoic measurement technique in a practical sized room. The test is useful only for showing level matching and crossover region summing.
An externally hosted image should be here but it was not working when we last tested it.
Frequency responses of 52 mm full range driver in box, in box with first order filter, and in box with two components in the high pass filter. Driver free air resonance is 109 Hz and Qts is .63. Mounted in a .02 cubic foot enclosure, Fs is 190 Hz with output rising toward resonance. Add a series capacitor and while the response flattens between 260 and 400 Hz, the output at resonance is significantly up. Add a parallel inductor for a “second” order crossover and this resonance bump is eliminated.
An externally hosted image should be here but it was not working when we last tested it.
While the small box is used as part of the high pass filter, there are consequences to the technique. Midrange ripple is increased and an equalizer network is required to maintain plus or minus three-db response in the far field pass-band. Even with the ripple, response is still plus or minus two db through the important 500 Hz to 2 kHz midrange region.
An externally hosted image should be here but it was not working when we last tested it.
Ten-component crossover network.
An externally hosted image should be here but it was not working when we last tested it.
As stated in my first post, I wanted to design a system with a larger diameter and higher power low frequency unit and a passive speaker level crossover for the three-inch full range drivers with sensitivities ranging from 86 to 88-db. Here is my design. While the system shown uses one of my prototype drivers, you can substitute a Tang Band or Fostex driver of equal diameter, impedance, and efficiency with good results. All the limitations of the other drivers will be preserved, except at the lower frequencies now handled by the seven-inch woofer.
The loudspeaker system is a modular design, with separate enclosures for the full range driver and woofer. The full range driver enclosure is intended to be part of the high pass filter and that fact determines the size. The examples shown have outside measurements of 130 mm by 130 mm by 87 mm deep. The sides and back are made from ¾ inch mdf and the front baffle of ¼ inch veneer plywood. The enclosure is liberally stuffed with poly fill material.
When I first listen to a new design, I am trying to hear what is wrong. This means that I am not deciding if I like a system right away. The liking part comes only after an initial evaluation and critiquing period. If I come to like a system, then I will like if for a long, long time. Indeed, I have some of my older designs that are more than 25 years old and I still enjoy listening to them. I have spent a number of hours listening to this system and I am beginning to like the sound.
I see no reason to be falsely enthusiastic about the sound, but it is quite good. I can also still critique it even though I built it with my own hands. For example, the low bass is a bit lush to my taste and the lower pitched male vocalists sound a little dry. Otherwise, the loudspeakers are clear and detailed, with excellent soundstage, extended range. French horns are reproduced with good accuracy. Piano is also excellent with preservation of tonal accuracy across the entire range of the instrument. Except for the already mentioned dryness in the lower vocal range, vocal reproduction is excellent, with great clarity.
There are many ways to achieve accuracy of reproduction and sound you can listen to for years. This design seems to be one of them.
Although I will put in more time listening before I make any changes, I am likely going to add more poly fill to the woofer cabinet. I will probably increase the poly fill from the current 8 ounces to 12 to 14 ounces.
Mark
Attachments
Thanks for the comment. I am sorry to say that I did not pay much attention to Kloss' last loudspeaker company and I am unfamiliar with his "ensemble" loudspeakers. There is, however, little that has not already been done in the way of two-way configurations. Also, implementation counts for something.
Just knowing a little of Kloss' past design portfolio I would believe these sound a bit different. Still, Kloss was a solid designer and I am sorry there will be no new designs from him.
Mark
Just knowing a little of Kloss' past design portfolio I would believe these sound a bit different. Still, Kloss was a solid designer and I am sorry there will be no new designs from him.
Mark
Very nice work there Mark, how did u get the edges so smooth?
I am also looking into a 2 way setup using a 3" TB full range and a woofer. w3-871s + w5-610s in seperate enclosures in a ordinary 2 way could sound nice, problem is getting my hand on the w5-610s, I havent found any resellers of this driver.
I am also looking into a 2 way setup using a 3" TB full range and a woofer. w3-871s + w5-610s in seperate enclosures in a ordinary 2 way could sound nice, problem is getting my hand on the w5-610s, I havent found any resellers of this driver.
Edgar,
The panels are composites. The parquet is bonded to a composite panel and then cut. I have stated before that I am a follower of building layered composites (materials of differing mechanical impedance) for resonance control.
I also use a very sharp 60+-tooth 10-inch carbide tipped saw blade. If the wood still chips out, covering the edge with tape can help. Also, be patient and cut very slowly.
I am also building several cabinets with the composite as the outside laminate. Gives an interesting all black look. Very striking with the metallic look of the cones.
Back to an earlier comment...My inspiration was the Bandor Trident loudspeakers. But I wanted to have better controlled bass and better seaming of the two drivers than they achieve with a band pass enclosure and "natural" roll-off. Of course, the Bandor woofer is much better controlled through breakup than the parts express AL cone and so doesn't impact the midrange as much.
Most humbly, I believe the finished loudspeaker does sound better and does not require sitting precisely on axis to hear the highs.
Lastly, if you want to have the woofer front firing and time-align the drivers, then with a center to center separation of 30 cm, set back the full range driver by 10 cm.
Regardless of where you find your inspiration, be brave and don't put to much trust or faith in any of the simulator/modeling programs.
Mark
The panels are composites. The parquet is bonded to a composite panel and then cut. I have stated before that I am a follower of building layered composites (materials of differing mechanical impedance) for resonance control.
I also use a very sharp 60+-tooth 10-inch carbide tipped saw blade. If the wood still chips out, covering the edge with tape can help. Also, be patient and cut very slowly.
I am also building several cabinets with the composite as the outside laminate. Gives an interesting all black look. Very striking with the metallic look of the cones.
Back to an earlier comment...My inspiration was the Bandor Trident loudspeakers. But I wanted to have better controlled bass and better seaming of the two drivers than they achieve with a band pass enclosure and "natural" roll-off. Of course, the Bandor woofer is much better controlled through breakup than the parts express AL cone and so doesn't impact the midrange as much.
Most humbly, I believe the finished loudspeaker does sound better and does not require sitting precisely on axis to hear the highs.
Lastly, if you want to have the woofer front firing and time-align the drivers, then with a center to center separation of 30 cm, set back the full range driver by 10 cm.
Regardless of where you find your inspiration, be brave and don't put to much trust or faith in any of the simulator/modeling programs.
Mark
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.