Hallo Salas and others in this tread. As Salas told a view posts ago, I cloned his design and i like them so much that i make 2 more of these. The tangbands for these are bought already.
I play them with my Hiraga 20W amp and the cabinets are about 80 cm out off the wall.
Amazing sound for such tiny loudspeakers! Nice project and easy to make!
I play them with my Hiraga 20W amp and the cabinets are about 80 cm out off the wall.
Amazing sound for such tiny loudspeakers! Nice project and easy to make!
salas said:
The actual listening result is the heavy blue curve had the least bass (by a long way). All were 3.5 inch drivers tested at around 2.8V using the same test setup. The heavy blue curve driver has an Xmax of 0.4mm, others were over 1mm.
The reason why I'm showing this is that test results might not show what we hear. I did not know this before, and hope that people are aware that this might be the case. Curves do not show up much different if a BSC is used or not, but the sound is different. I also used MLS signals for the tests, and now I'm trying to figure out how to test so that the test results will show more accurately what I hear.
Really want to know if the W3-871S can go to 80 Hz? I think it's really necessary to find a good way to test the driver first.
Dear Bob, your assumption that everything beyond its linear1mm excursion should not be accepted is wrong. THD and compression don’t have a step beyond linear excursion limits from where after they deteriorate in an intolerable manner, but they deteriorate rather progressively, as I will present later on. Your 0,25 Watts are somewhat funny when there are so many people who have constructed such speakers with such full range drivers and work them with 40 times this power enjoying really good music for the given money.
The proposition that a full BSC is needed is questionable. If designed with only one driver (and someone doesn’t want to purchase a second) I would never fully compensate such a tiny driver - enclosure to produce a 3 litter 80dB ‘politically correct’ (dead, for me) speaker, as if I would design a big 3 way system. In my opinion it is perfectly acceptable to design with less (or even minimal, if the speaker is to be placed really near the back wall) correction for such a small speaker, preserve its midband sensitivity, have the capability to give it some 5-10 watts and produce acceptable acoustic volumes, tolerating with the fact that it will sound ‘smaller’ compared to bigger speaker, but still lively and interesting. If such a speaker is to be placed near walls or in a bookshelf, then, as I will present you later, almost no BSC is needed.
Furthermore, if someone really needed a BSC (for placing speakers free-space), your proposed way to compensate for a Baffle step is wrong and leads to bigger problems. In a 3 litter enclosure you can’t easily end up with a bigger than 15 * 25 cm baffle. In such a baffle this driver will have its response declining beneath about 700 Hz.
Producing a peak in the 100-150 Hz region will compensate for nothing, and will only deteriorate its bass response, worsen its step response, and end up with a boomy little speaker, possibly masking the lower mids in favor of 2-3 ‘strong’ upper bass notes.
If someone wants BSC, Salas’s way is the best.
Now, I decided this morning to quickly construct a small speaker with these tangbands in 3,7 Litter and with a Box tuning of 50 (and not even 80) Hz. Also used a cheap Audax tweeter for some specific reasons (I will place these speakers in a high shelf for some time) and worked out a crossover of <12dB slopes in about 3,5 KHz, with almost no Baffle step correction.
I measured this speaker placed in a shelf, in a small 50 cubic meter room (3,7*4,5*3,05m), in a distance of about 1,6 meter from it (in the specific place that I heard it).
I also measured its response in various Wattages to see how it compresses frequencies, and also its distortion vs input voltage behavior
First picture, the speaker itself, constructed with real wood and an aluminum very light and stiff sandwich of 4mm thickness:
Now, its LP crossover section, together with the electrical slopes (minimal BSC, as can be seen)
Here I present you the frequency response in the 40-1000 region for 0,8W and 8W input. If my measurement setup is to be trusted, an overall slight compression can be seen all over the band, mostly because of the V.C. thermal behavior (I had the stimulus running for some time to adjust the voltage), and only slightly more in the 45 to 70Hz region because of the driver’s mechanical restrictions.
Last, here are THD vs. Voltage level measures, for 50, 80, 200, 500 and 1000 Hz, with input voltages from 0,7 to 7V Input (0,061W to 6,1W). In my experience they are perfectly acceptable for such a small driver/enclosure.
Listening impressions: Very enjoyable (especially when looking at it;-)) with a discrete small 15W amp of mine.
Now playing music: Bach’s ‘Die kunst der Fuge’ with Antiqua Koln and R.Goebel
Listening levels obtained with acceptable reproduction quality: 1,5 m apart, 50 c.m. room of about 0,4sec reverberation time, about 95dB max SPL as expected.
Douwe, go ahead and build Salas’s speaker, they are really nice – if you don’t possess 4 drivers and can place the speakers in your bookshelf or really near a wall, you can still proceed with only 2 and be pleased with the outcome.
Regards,
Thalis.
The proposition that a full BSC is needed is questionable. If designed with only one driver (and someone doesn’t want to purchase a second) I would never fully compensate such a tiny driver - enclosure to produce a 3 litter 80dB ‘politically correct’ (dead, for me) speaker, as if I would design a big 3 way system. In my opinion it is perfectly acceptable to design with less (or even minimal, if the speaker is to be placed really near the back wall) correction for such a small speaker, preserve its midband sensitivity, have the capability to give it some 5-10 watts and produce acceptable acoustic volumes, tolerating with the fact that it will sound ‘smaller’ compared to bigger speaker, but still lively and interesting. If such a speaker is to be placed near walls or in a bookshelf, then, as I will present you later, almost no BSC is needed.
Furthermore, if someone really needed a BSC (for placing speakers free-space), your proposed way to compensate for a Baffle step is wrong and leads to bigger problems. In a 3 litter enclosure you can’t easily end up with a bigger than 15 * 25 cm baffle. In such a baffle this driver will have its response declining beneath about 700 Hz.
Producing a peak in the 100-150 Hz region will compensate for nothing, and will only deteriorate its bass response, worsen its step response, and end up with a boomy little speaker, possibly masking the lower mids in favor of 2-3 ‘strong’ upper bass notes.
If someone wants BSC, Salas’s way is the best.
Now, I decided this morning to quickly construct a small speaker with these tangbands in 3,7 Litter and with a Box tuning of 50 (and not even 80) Hz. Also used a cheap Audax tweeter for some specific reasons (I will place these speakers in a high shelf for some time) and worked out a crossover of <12dB slopes in about 3,5 KHz, with almost no Baffle step correction.
I measured this speaker placed in a shelf, in a small 50 cubic meter room (3,7*4,5*3,05m), in a distance of about 1,6 meter from it (in the specific place that I heard it).
I also measured its response in various Wattages to see how it compresses frequencies, and also its distortion vs input voltage behavior
First picture, the speaker itself, constructed with real wood and an aluminum very light and stiff sandwich of 4mm thickness:
An externally hosted image should be here but it was not working when we last tested it.
Now, its LP crossover section, together with the electrical slopes (minimal BSC, as can be seen)
An externally hosted image should be here but it was not working when we last tested it.
Here I present you the frequency response in the 40-1000 region for 0,8W and 8W input. If my measurement setup is to be trusted, an overall slight compression can be seen all over the band, mostly because of the V.C. thermal behavior (I had the stimulus running for some time to adjust the voltage), and only slightly more in the 45 to 70Hz region because of the driver’s mechanical restrictions.
An externally hosted image should be here but it was not working when we last tested it.
Last, here are THD vs. Voltage level measures, for 50, 80, 200, 500 and 1000 Hz, with input voltages from 0,7 to 7V Input (0,061W to 6,1W). In my experience they are perfectly acceptable for such a small driver/enclosure.
An externally hosted image should be here but it was not working when we last tested it.
Listening impressions: Very enjoyable (especially when looking at it;-)) with a discrete small 15W amp of mine.
Now playing music: Bach’s ‘Die kunst der Fuge’ with Antiqua Koln and R.Goebel
Listening levels obtained with acceptable reproduction quality: 1,5 m apart, 50 c.m. room of about 0,4sec reverberation time, about 95dB max SPL as expected.Douwe, go ahead and build Salas’s speaker, they are really nice – if you don’t possess 4 drivers and can place the speakers in your bookshelf or really near a wall, you can still proceed with only 2 and be pleased with the outcome.
Regards,
Thalis.
I am using close to the ELF 1.0 design and I would venture to say they easily hit 80hz.... Now below that there is not levels of volume that are on level with the rest of frequency range. However what I can tell you is with a Neil Diamond LP I have made mine bottom out with some where not even 20wRMS.....
They more more than .3mm FOR SURE, that is not their Xmax. Right now as we speak I can crank the Dave Mathews Band - Everday .flac's to 20wRMS or about and they will move to the point you can clearly see the excursion going on, see them actually go to the point of stressing the surround some. At that level they will distort with certain frequencies that are to low.... However if they where high passed above some thing like 100-150hz they could easily take 20wRMS. Now if this was in a sealed box they might get a little hot though. They can take some abuse.
Personally I would say once you start to exceed 10wRMS the low end tends to fall off a lot. However good desktop speaker listening levels are right in the 5-12wRMS area.
The same thing as my speakers are sold at $100 a piece from some high end audio company for surround sound speakers. They are good enough for the application as long as they are high passed at a decent frequency.
If you have Foobar2000 try downloading the crossover plug in and setting high pass crossovers at different frequencies then try feeding it more power.
They more more than .3mm FOR SURE, that is not their Xmax. Right now as we speak I can crank the Dave Mathews Band - Everday .flac's to 20wRMS or about and they will move to the point you can clearly see the excursion going on, see them actually go to the point of stressing the surround some. At that level they will distort with certain frequencies that are to low.... However if they where high passed above some thing like 100-150hz they could easily take 20wRMS. Now if this was in a sealed box they might get a little hot though. They can take some abuse.
Personally I would say once you start to exceed 10wRMS the low end tends to fall off a lot. However good desktop speaker listening levels are right in the 5-12wRMS area.
The same thing as my speakers are sold at $100 a piece from some high end audio company for surround sound speakers. They are good enough for the application as long as they are high passed at a decent frequency.
If you have Foobar2000 try downloading the crossover plug in and setting high pass crossovers at different frequencies then try feeding it more power.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.