Both driver series evidently share the same vented baskets. Difference seems to be use of magnesium on the Excel line (except W26 which is an aluminum-magnesium alloy – probably the same as all L drivers), distortion killer faraday rings and copper vs. Plastic phase plug. Really?
W26FX001 vs. L26RFX/P:
Magnet and coil data are identical. Flux density in gap is given as 1.23T in both cases. Conventional wisdom has it that this is impossible with a ferrite magnet and regular iron because the iron saturates below. In fact, it might be possible with very clever iron return system design. This is good news, because the pole piece will be satured, so inductance, inductance variation with coil travel and flux modulation will be low. Given even the identical Bxl values, and the very low inductance of 1.43 vs. 1.48 mH(this is within measurement error), either both systems have distortion killer rings, or more likely, they both don’t because these are not needed with proper pole piece design.
Impedance curves look identical, even to the small resonance at 4 kHz. Moving mass is 56.3 vs 58 g, probably also within production tolerance. It is surprsing that the Excel cone is nominally ligher, knowing that is is painted with the stuff that gets put onto the smaller Excel drivers. Response curves look almost identical. The W26 might have a tad less droop between 1 and 2 kHz. Position of the peaks at 4 kHz and beyond is identical, actually the L26 peaks look a little less pronounced, but this might be smoothing.
=> Differences seem to be the paint on the cone and the phase plug material, and their effect does not seem to be big. The motors and suspensions seem to be identical, so there is hope that the L26 will have the same good distorion performance as the W26. On a sinde note: the W22 has better 2nd harmonic behavior between 100 and 1000 Hz than ist larger cousin! On the other hand, distortion measurements on the W22 were probably made in a sealed box which lowers excursion and hence distortion.
W22EX001 vs. L22RN4X/P:
Magnet weight is 0.64 kg which is not much compared to the Scan 21W8555. Nominal gap geometry is identical, but the L22 has the longer coil (more travel but less efficiency). Flux in gap is 1.0 vs. 0.9T. If this is not a rounding issue, this is surprising, because if the W22 has any copper inside the gap, its flux should be lower. Inductance is 0.5 vs. 3.5 mH, clearly a sign that the L22 does not have the copper rings. Cone masse is 29 vs. 43 g, clearly a plus for the magesium cone.
Response data are a little hard to compare because the W22 was measured in a test box and the L22 on an infinite baffle. The first peak is a 5 kHz for the W22 vs. 4 kHz for the L22. The response droop begins at 2 kHz for the W vs. 1 kHz for the L, probably a result of the inductance (the only way to cure this in the L22 is to use a current drive amplifier).
=> Advantages of the W22 are clearly there, but still I would not want to use it out to 2 kHz.
Big 18: W18EX001 vs. L18RNX/P:
Both share the big magnet used also in the W22/L22. Nominal gap geometry is identical, so is the flux of 1.0T. The W18 has the shorter coil and hence higher efficiency and Bxl. Inductance is 0.4 bs. 1.0 T, not big, but clearly due to the copper rings.
Cone mass is 15.5 vs. 14 g, i.e. in this case the magesium cone + coil is heavier. Response is again hard to compare because of box vs. Baffle. Both can equally well be used out to 2 or 2.5 kHz. The W18 sustains response in the 3-4 kHz region a little better, a result of the lower inductance. The first peak occurs at 4.5 kHz ans is very strong, subsequent peaks are very weak, but this probably does not matter. The L18 has a micro-peak at 4.5 kHz, the first major peak occuring at 7 kHz and being 5 dB lower than the 4.5 kHz peak of the W18. There are several almost equally high peaks beyond 7 kHz, but again, this is probably not as important.
=> The W18 might well be the lower distortion driver in terms of motor nonlinearity, but the L18 might be easier to use.
Small 18: W18E001 vs. L18RCY/P:
Both have the smaller 0.42 kg magnet. Due to its small VC diameter, the L18 has a high flux density of 1.25 T. The inductance of 0.9T is surprisingly high considering that the pole piece should be saturated. The W18E seems to have the larger VC diameter of 39 mm used by the W18EX and the W22. This results in a lower flux of 0.88 T. It is not clear whether these figures are correct, because the data sheet contains other inconsistencies. The distortion plot is identical (down to the measurement date and time) to the plot found in the EX data sheet.
Response results are similar to the big 18 findings.
=> Seas should have a look at the W18E data sheet.
W26FX001 vs. L26RFX/P:
Magnet and coil data are identical. Flux density in gap is given as 1.23T in both cases. Conventional wisdom has it that this is impossible with a ferrite magnet and regular iron because the iron saturates below. In fact, it might be possible with very clever iron return system design. This is good news, because the pole piece will be satured, so inductance, inductance variation with coil travel and flux modulation will be low. Given even the identical Bxl values, and the very low inductance of 1.43 vs. 1.48 mH(this is within measurement error), either both systems have distortion killer rings, or more likely, they both don’t because these are not needed with proper pole piece design.
Impedance curves look identical, even to the small resonance at 4 kHz. Moving mass is 56.3 vs 58 g, probably also within production tolerance. It is surprsing that the Excel cone is nominally ligher, knowing that is is painted with the stuff that gets put onto the smaller Excel drivers. Response curves look almost identical. The W26 might have a tad less droop between 1 and 2 kHz. Position of the peaks at 4 kHz and beyond is identical, actually the L26 peaks look a little less pronounced, but this might be smoothing.
=> Differences seem to be the paint on the cone and the phase plug material, and their effect does not seem to be big. The motors and suspensions seem to be identical, so there is hope that the L26 will have the same good distorion performance as the W26. On a sinde note: the W22 has better 2nd harmonic behavior between 100 and 1000 Hz than ist larger cousin! On the other hand, distortion measurements on the W22 were probably made in a sealed box which lowers excursion and hence distortion.
W22EX001 vs. L22RN4X/P:
Magnet weight is 0.64 kg which is not much compared to the Scan 21W8555. Nominal gap geometry is identical, but the L22 has the longer coil (more travel but less efficiency). Flux in gap is 1.0 vs. 0.9T. If this is not a rounding issue, this is surprising, because if the W22 has any copper inside the gap, its flux should be lower. Inductance is 0.5 vs. 3.5 mH, clearly a sign that the L22 does not have the copper rings. Cone masse is 29 vs. 43 g, clearly a plus for the magesium cone.
Response data are a little hard to compare because the W22 was measured in a test box and the L22 on an infinite baffle. The first peak is a 5 kHz for the W22 vs. 4 kHz for the L22. The response droop begins at 2 kHz for the W vs. 1 kHz for the L, probably a result of the inductance (the only way to cure this in the L22 is to use a current drive amplifier).
=> Advantages of the W22 are clearly there, but still I would not want to use it out to 2 kHz.
Big 18: W18EX001 vs. L18RNX/P:
Both share the big magnet used also in the W22/L22. Nominal gap geometry is identical, so is the flux of 1.0T. The W18 has the shorter coil and hence higher efficiency and Bxl. Inductance is 0.4 bs. 1.0 T, not big, but clearly due to the copper rings.
Cone mass is 15.5 vs. 14 g, i.e. in this case the magesium cone + coil is heavier. Response is again hard to compare because of box vs. Baffle. Both can equally well be used out to 2 or 2.5 kHz. The W18 sustains response in the 3-4 kHz region a little better, a result of the lower inductance. The first peak occurs at 4.5 kHz ans is very strong, subsequent peaks are very weak, but this probably does not matter. The L18 has a micro-peak at 4.5 kHz, the first major peak occuring at 7 kHz and being 5 dB lower than the 4.5 kHz peak of the W18. There are several almost equally high peaks beyond 7 kHz, but again, this is probably not as important.
=> The W18 might well be the lower distortion driver in terms of motor nonlinearity, but the L18 might be easier to use.
Small 18: W18E001 vs. L18RCY/P:
Both have the smaller 0.42 kg magnet. Due to its small VC diameter, the L18 has a high flux density of 1.25 T. The inductance of 0.9T is surprisingly high considering that the pole piece should be saturated. The W18E seems to have the larger VC diameter of 39 mm used by the W18EX and the W22. This results in a lower flux of 0.88 T. It is not clear whether these figures are correct, because the data sheet contains other inconsistencies. The distortion plot is identical (down to the measurement date and time) to the plot found in the EX data sheet.
Response results are similar to the big 18 findings.
=> Seas should have a look at the W18E data sheet.
Very interseting! L26RFX/P seems like a driver for me... I would think it is better than the Peerless CSX 10". The Peerless has a higher inductance of 2.9mH, does that suggest that its pole piece isn't saturated/no shortcut rings = higher distorsion? The XLS 10" has a high inductance of 4.3mH, and yet the distorsion measurements I've seen has showed very low distortion.
Which frequency range are you refering to? A ported box will have significantly less excursion around the tuning frequency.
Which frequency range are you refering to? A ported box will have significantly less excursion around the tuning frequency.
Vg said:
Inductance itself is not an indicator of distortion. The XLS 10/12 and CSX 10/8/7 drivers are all examples of drivers with relatively high (and in some cases, just plain high) inductance. The CSX10 would be a tough driver to beat. It's a very low distortion unit, I have a hard time believing that the L26 would be that much better distortion wise.
Here are some IMD easurements on the CSX 10 that show very low distortion. I don't have an L26 to play with so unfortunately no direct comparison could be made. The test was set up so the driver would be playing at 96dB@1M in Half space for the 150Hz tone. The 135Hz and 165Hz tones are 6dB down. Anything besides 135, 150, and 165 is ambient noise or distortion. (There isn't much).
Scott
Attachments
capslock said:
Wow, that was a very thorough post. I don't have time to comment on everything, but I would like to comment on the W26/L26 comparison. I've seen both drivers at CES, and I seem to remember the L26 having a curved cone profile while the W26 had a straight cone profiel. I could be wrong, that was almost a year ago now. The reason that I remember it is that the L26 had a cool shimmering effect because of it and looked great. The W26 did too...but I remember noting it as a difference in my mind. The L26 is certainly the better performance for the $ buy between the two, but I wouldn't be so quick to compare the cones on weight alone. I think they are very different.
I would also bet that the W26 has much better intermodulation performance than the L26. The entire W line is stellar in that regard and the "conventional" Seas motors on the whole don't keep up. They are still very good, but they aren't the same.
Scott
Adressing 3 questions/comments:
- The W17-W22 FR were measured in a rather small sealed box. I suspect the distortion measurements were also measured in the same box. My point just is that any kind of enclosure will limit excursion, so boxed distortion measurements should not directly be compared to free air measurements.
- Low inductance indicates that measures have been taken to minimize Le variation and flux modulation. It is also possible to get a rather linear Le vs. excursion by using a T-shaped pole piece that has an extension into the space below the dust cap, even if this increases the total inductance. I suspect Peerless did just that. And the XLS drivers have a pretty large VC diameter, meaning lots of iron inside the loop, hence high, but not necessarily nonlinear inductance. The only thing that cannot be eliminated this way is flux modulation, but this is probably not a major contributor here.
- Excel vs. standard line motors: I have no doubt that the Excel motors are different and better. But comparing the motor data and impedance curves from the W26 and L26 data sheets, I suspect very much that just in this one case the same motor was actually used.
- The W17-W22 FR were measured in a rather small sealed box. I suspect the distortion measurements were also measured in the same box. My point just is that any kind of enclosure will limit excursion, so boxed distortion measurements should not directly be compared to free air measurements.
- Low inductance indicates that measures have been taken to minimize Le variation and flux modulation. It is also possible to get a rather linear Le vs. excursion by using a T-shaped pole piece that has an extension into the space below the dust cap, even if this increases the total inductance. I suspect Peerless did just that. And the XLS drivers have a pretty large VC diameter, meaning lots of iron inside the loop, hence high, but not necessarily nonlinear inductance. The only thing that cannot be eliminated this way is flux modulation, but this is probably not a major contributor here.
- Excel vs. standard line motors: I have no doubt that the Excel motors are different and better. But comparing the motor data and impedance curves from the W26 and L26 data sheets, I suspect very much that just in this one case the same motor was actually used.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.