This started out as looking for better driver options for the V-baffle sub of the LX521 compared to the Seas L26RO4Y that the sub was designed for.
Here's a list of drivers that I either had sitting here or I could get from stock and at a reasonable price. I threw in the venerable Beyma 15K200 that I also have a couple of. The original housing can accomodate two XXLS12 in the original V-baffle if it is made about 1 cm wider (which is possible if you use thin sides as SL did with his prototype). The 15 inch RSS would probably have to live a solitary life in an S-Baffle (but it has about the same displacement as two 10 inch L26). I think it should be possible to put two 15K200 in a W-baffle and just make it fit.
I entered the parameters in WinISD to see what kind of efficiencies I'd really get, and I plotted the SPL of 1000 l closed boxes as an approximation of the FR in an open baffle.
Here's a list of drivers that I either had sitting here or I could get from stock and at a reasonable price. I threw in the venerable Beyma 15K200 that I also have a couple of. The original housing can accomodate two XXLS12 in the original V-baffle if it is made about 1 cm wider (which is possible if you use thin sides as SL did with his prototype). The 15 inch RSS would probably have to live a solitary life in an S-Baffle (but it has about the same displacement as two 10 inch L26). I think it should be possible to put two 15K200 in a W-baffle and just make it fit.
I entered the parameters in WinISD to see what kind of efficiencies I'd really get, and I plotted the SPL of 1000 l closed boxes as an approximation of the FR in an open baffle.
Attachments
Last edited:
The biggest surprise for me was that the datasheet values are all over the place compared to what WinISD calculates. I was familiar with prosound midbasses that had SPLs based on midband on axis SPL, but I hadn't come across this in sub drivers.
WinISD outputs "USPL" as the SPL at 2.83 V / 1 m and "SPL" as the SPL at 1 W / 1 m, so USPL should be the same as the data sheet value.
Seas quotes an SPL of 86 dB for the L26RO4Y which is close enough to what WinISD computes for 500 Hz (ignoring Le effects, because I didn't enter Le). Their L26ROY is quoted at 1 dB more, but this clearly is not the case at 500 Hz, wehre it is 1 dB less efficient. It is correct relative to its sibling if you look at e.g. 80 Hz, which may be considered the upper range of where the sub might be used. But why would one manufacturer who updated drivers and data sheets within months of each other when production was transferred from Taiwan to Norway use two completely different definitions, i.e. the asymptotic level for one and a realistic one for the range of a sub for the other?
Little surprise then that the numbers given by Tymphany, Scan Speak and Dayton make little sense, too.
(to be continued)
WinISD outputs "USPL" as the SPL at 2.83 V / 1 m and "SPL" as the SPL at 1 W / 1 m, so USPL should be the same as the data sheet value.
Seas quotes an SPL of 86 dB for the L26RO4Y which is close enough to what WinISD computes for 500 Hz (ignoring Le effects, because I didn't enter Le). Their L26ROY is quoted at 1 dB more, but this clearly is not the case at 500 Hz, wehre it is 1 dB less efficient. It is correct relative to its sibling if you look at e.g. 80 Hz, which may be considered the upper range of where the sub might be used. But why would one manufacturer who updated drivers and data sheets within months of each other when production was transferred from Taiwan to Norway use two completely different definitions, i.e. the asymptotic level for one and a realistic one for the range of a sub for the other?
Little surprise then that the numbers given by Tymphany, Scan Speak and Dayton make little sense, too.
(to be continued)
Last edited:
So what does this mean in my application, which uses a 120 Hz LR4?
Let's look at the L26ROY (blue) vs. the original L26RO4Y. Although it is nominally less efficient at HF, it is really 0.5 to 1.5 dB more efficient throughout the range where it would be used. Its Q in the baffle might be very close to the 0.5 preferred by SL
Frequently Asked Questions
so one would have to change the EQ but I don't really see any reason why it should not be used. Why didn't SL specify it? Probably historical reasons (he had the old Taiwanese L26ROY altered).
Let's look at the L26ROY (blue) vs. the original L26RO4Y. Although it is nominally less efficient at HF, it is really 0.5 to 1.5 dB more efficient throughout the range where it would be used. Its Q in the baffle might be very close to the 0.5 preferred by SL
Frequently Asked Questions
so one would have to change the EQ but I don't really see any reason why it should not be used. Why didn't SL specify it? Probably historical reasons (he had the old Taiwanese L26ROY altered).
Datasheet/brochure "numbers" are Marketing dictated.
Part of the problem is that many/most buyers don´t look at curves (or don´t understand them) and prefer to read "the number" and it´s marketing suicide quoting (true) , say, 86dB when others, exact same specs, quote, say, 89 or 92 dB so ......
Consider you gifted
for being able to read proper curves and even better handle design software
Part of the problem is that many/most buyers don´t look at curves (or don´t understand them) and prefer to read "the number" and it´s marketing suicide quoting (true) , say, 86dB when others, exact same specs, quote, say, 89 or 92 dB so ......
Consider you gifted
for being able to read proper curves and even better handle design software Thanks Back to the comparison of the two L26. Apparently, SL didn't like the original L26ROY (D1001) and worked with Seas to improve it for use at high excursion, e.g. they redesigned the suspension, nearly doubling the mechanical resistance at the same time (while keeping the fibreglass former, so it cannot be eddy current losses). Why did he also specifify a four layer VC? We don't even know that he did. It may have been Seas who said we are going to launch a new driver with your items but it needs to be four layer because we have a market niche to fill for small BR. There should be little doubt that SL considered the D1004 better suited for his open baffle use than the D1001 at the time.
When they moved production to Norway, they probably incorporated some of the learnings from the D1004 into the XM001 or they had come to use Klippel more efficiently in the meantime anyway. According to the test by Vance Dickason, motor and suspension linearity are beautiful
Test Bench: The L26ROY 10” Subwoofer from SEAS | audioXpress.
I am not aware of any kind of test of the redesigned L26RO4Y (XM004) but probably has equally good Klippel results. It also has yet increased mechanical resistance compared to its predecessor.
So these days, and assuming one is willing to measure the FR and adjust the EQ, which one is the better driver for the V-baffle? Are the higher efficiency above 120 Hz, the lower Qes and higher Rms of the L26RO4Y ends in their own right for this applications?
Or is the new L26ROY, which still meets SL's criterion of Q < 0.5, but also brings higher efficiency in the range of use, lower Rms and lower Le to the table, the sibling better suited these days?
And once we get to the decision that there is no need to retain the L26RO4Y, the other drivers in my spreadsheet appear even more attractive.
Back to the comparison of the two L26. Apparently, SL didn't like the original L26ROY (D1001) and worked with Seas to improve it for use at high excursion, e.g. they redesigned the suspension, nearly doubling the mechanical resistance at the same time (while keeping the fibreglass former, so it cannot be eddy current losses). Why did he also specifify a four layer VC? We don't even know that he did. It may have been Seas who said we are going to launch a new driver with your items but it needs to be four layer because we have a market niche to fill for small BR. There should be little doubt that SL considered the D1004 better suited for his open baffle use than the D1001 at the time.When they moved production to Norway, they probably incorporated some of the learnings from the D1004 into the XM001 or they had come to use Klippel more efficiently in the meantime anyway. According to the test by Vance Dickason, motor and suspension linearity are beautiful
Test Bench: The L26ROY 10” Subwoofer from SEAS | audioXpress.
I am not aware of any kind of test of the redesigned L26RO4Y (XM004) but probably has equally good Klippel results. It also has yet increased mechanical resistance compared to its predecessor.
So these days, and assuming one is willing to measure the FR and adjust the EQ, which one is the better driver for the V-baffle? Are the higher efficiency above 120 Hz, the lower Qes and higher Rms of the L26RO4Y ends in their own right for this applications?
Or is the new L26ROY, which still meets SL's criterion of Q < 0.5, but also brings higher efficiency in the range of use, lower Rms and lower Le to the table, the sibling better suited these days?
And once we get to the decision that there is no need to retain the L26RO4Y, the other drivers in my spreadsheet appear even more attractive.
Is the efficiency of any of these drivers a concern for you?
All of these drivers in an open-baffle application like the LX521 can easily be driven with amplifiers of nominal capability. In the LX521 it's entirely about power capability to push the driver to Xmax at the bottom of the LX521 range. As you go up in frequency electrical drive to the woofer decreases at a very steep rate, so power requirement reduces greatly.
Dave.
All of these drivers in an open-baffle application like the LX521 can easily be driven with amplifiers of nominal capability. In the LX521 it's entirely about power capability to push the driver to Xmax at the bottom of the LX521 range. As you go up in frequency electrical drive to the woofer decreases at a very steep rate, so power requirement reduces greatly.
Dave.
I am trying to learn if there is anything that the RO4Y does better than any of the other drivers.
I am too well aware how most amplifiers degrade in performance when operated near nominal power, and then there are VC heating effects. Last not least, the RO4Y has a very lossy suspension, which is bound to be non-linear and may also exhibit thermal changes. So if there is no compelling reason to use it, I think I may have better fits on my shelf.
I am too well aware how most amplifiers degrade in performance when operated near nominal power, and then there are VC heating effects. Last not least, the RO4Y has a very lossy suspension, which is bound to be non-linear and may also exhibit thermal changes. So if there is no compelling reason to use it, I think I may have better fits on my shelf.
Ok, and then if it is ok to use a higher efficiency, slightly lower Q driver, the next more sensitive is the 30W4558T that gains another dB but has an even lower impedance, loading the amp more. There are HobbyHifi magazine tests of the old L26ROY and 26W4558T that seems to share its motor with the 30W, and the 26W fared substantially better in terms of LF distortion, but then we know that SL saw the need to improve on the old L26ROY. The 26W and 30W are also supposed to be similar to Peerless 835016 and 017. The 835 16 did not look nearly as beautiful in Vance Dickason's Klippel tests as the new L26, but it had significantly lower HD3. We don't know the box volume in each of these tests, though.
The Peerless XXLS 830845 is a different story. It is an 8 Ohm driver, so it is much easier on many amps. The voltage base graph here does not tell the full story. It is nearly 3 dB more efficient than the L26RO4Y! It is also very quiet in terms of wind noise. Since I have it sitting here, I will probably try it first.
The RSS is impressive in that a single driver has the same displacement as two L26. However, if I used a single driver, sensitivity would take a hit. It would also have to live in an S baffle, though, which is a significant change from the original design. I haven't been able to find how it performs in terms of wind noise, and its mechanical losses are double those of the L26RO4Y, or the same when scaled to Sd but suspension hysteresis might matter most at small displacement. So not a good choice at this point.
The 15K200 looks more attractive, but it would require a W baffle. Wind noise is not a problem but some units have problems with ticking noise from surround or tinsel leads when operated free air. While it is theoretically 10 dB more efficient than L26RO4Y, due to its low Q, it really is only 1.5 dB more efficient than 830845 at 30 Hz so might need current drive and MFB to realize its true potential.
The Peerless XXLS 830845 is a different story. It is an 8 Ohm driver, so it is much easier on many amps. The voltage base graph here does not tell the full story. It is nearly 3 dB more efficient than the L26RO4Y! It is also very quiet in terms of wind noise. Since I have it sitting here, I will probably try it first.
The RSS is impressive in that a single driver has the same displacement as two L26. However, if I used a single driver, sensitivity would take a hit. It would also have to live in an S baffle, though, which is a significant change from the original design. I haven't been able to find how it performs in terms of wind noise, and its mechanical losses are double those of the L26RO4Y, or the same when scaled to Sd but suspension hysteresis might matter most at small displacement. So not a good choice at this point.
The 15K200 looks more attractive, but it would require a W baffle. Wind noise is not a problem but some units have problems with ticking noise from surround or tinsel leads when operated free air. While it is theoretically 10 dB more efficient than L26RO4Y, due to its low Q, it really is only 1.5 dB more efficient than 830845 at 30 Hz so might need current drive and MFB to realize its true potential.
You seem to have a big selection of drivers already. Whatever drivers you already have four of.....those are the ones to consider. Don't purchase any more because these are all quite expensive.
You seem to be in analysis paralysis on this. The LX521 is easy duty for any driver with longish excursion and low aerodynamic noise. It's primarily about the mechanicals in this application. Concerns about voice coil heating are largely irrelevant. The necessary (relative) EQ changes are easily performed and any good amplifier rated at 50-70wpc will be just fine.
Dave.
You seem to be in analysis paralysis on this. The LX521 is easy duty for any driver with longish excursion and low aerodynamic noise. It's primarily about the mechanicals in this application. Concerns about voice coil heating are largely irrelevant. The necessary (relative) EQ changes are easily performed and any good amplifier rated at 50-70wpc will be just fine.
Dave.
Hmm, from an efficiency POV they're all pretty abysmal [scroll down for math]: Thiele Small parameters equations - How each one affects the others
Davey, as long as this paralyzes my buy now - happy index finger, this is fine Seriously, I take it that the super low Qes and high Rms of the L26RO4Y are not ends of their own in this application.
GM, if you include the Beyma in your statement, you can't be serious. The most efficient PA midbass drivers approach 5% efficiency, but have virtually no bass due to extremely low Qes, stiff suspension and lack displacement. The 15K200 at 2.4% is about as good as it gets for a driver with low fs and highish displacement, and even there, you can see that its advantage over the 830845 down at 30 Hz disappears if you normalize to Sd.
Seriously, I take it that the super low Qes and high Rms of the L26RO4Y are not ends of their own in this application.GM, if you include the Beyma in your statement, you can't be serious. The most efficient PA midbass drivers approach 5% efficiency, but have virtually no bass due to extremely low Qes, stiff suspension and lack displacement. The 15K200 at 2.4% is about as good as it gets for a driver with low fs and highish displacement, and even there, you can see that its advantage over the 830845 down at 30 Hz disappears if you normalize to Sd.
The Rms is rather high, but in this application you've removed any box, and almost all acoustic loading. (Relative to a typical application.)
The L26R04Y driver is like many subwoofer drivers nowadays designed with a very strong motor. Most all of these drivers are NOT designed with this sort of usage in mind.
So, the trade-offs for open-baffle usage need to be understood. SL was, of course, aware of all this and considered the L26R04Y the best driver (at the time) for the Orion application.
That's not to say the two-layer VC version won't work fine as well. In fact, that's what I use. The minimal power requirement being even less in that case.
Dave.
The L26R04Y driver is like many subwoofer drivers nowadays designed with a very strong motor. Most all of these drivers are NOT designed with this sort of usage in mind.
So, the trade-offs for open-baffle usage need to be understood. SL was, of course, aware of all this and considered the L26R04Y the best driver (at the time) for the Orion application.
That's not to say the two-layer VC version won't work fine as well. In fact, that's what I use.
The minimal power requirement being even less in that case.Dave.
Yes, of course you need to change the EQ....because it's a different driver.
These dipole woofer projects require four things:
1. 6db/octave dipole correction.
2. Driver Q correction.
3. Cavity resonance correction.
4. Low-pass crossover.
If, for example, you were considering either the L26RO4Y or the L26ROY in the same defined baffle configuration. Items 1, 3, and 4, would probably remain the same, but you do have to adjust for differing Q. (Obviously, that's assuming you're targeting the same acoustic response in either case.)
Dave.
These dipole woofer projects require four things:
1. 6db/octave dipole correction.
2. Driver Q correction.
3. Cavity resonance correction.
4. Low-pass crossover.
If, for example, you were considering either the L26RO4Y or the L26ROY in the same defined baffle configuration. Items 1, 3, and 4, would probably remain the same, but you do have to adjust for differing Q. (Obviously, that's assuming you're targeting the same acoustic response in either case.)
Dave.
