Linkwitz LX521.4 - new build and impressions

[Keegan]

Since I always find such questions exciting, I have simulated it for you - but only with simple means via VCAD (which is sufficiently accurate) and not via a BEM-LEM simulation.

1. Sketch of the baffle and determine baffle diffraction (BD)
I made a rough sketch of the baffle and exported the OB baffle diffraction frequency response.
View attachment 237133

2. Enclosure simulation, import BD, target curve
Next, the Enclosure tool was used to simulate the behavior of the woofer. The BD FR was imported and adjusted to match our target curve with HP-LR4@120Hz.
The target curve in orange, the frequency response of the driver in the baffle in brown and the driver in the infinite baffle in black.
View attachment 237135
With this, we have finished creating our target frequency response for the woofer of the LX521 in the open baffle with HP-LR4@120Hz XO.

3. Evaluation of the driver excursion
So now we can take a look at the driver excursion at 90dB@1m, 100dB@1m and 105dB@1m in the free field.
Since VCAD's tool always refers to the half-space, we must simulate this in the tool with 96dB (brown curve) for 90dB free field (for 100,105dB accordingly).

View attachment 237136 View attachment 237137 View attachment 237138

The Seas woofer U22REX has two distortion peaks - at 100Hz (3% HD2@90dB-2pi) and 700Hz (2% HD2@90dB-2pi)***. However, both consist mainly of HD2, so should hardly be audible directly due to masking, but could still provide some increased IMD with multitone excitation.
Because of the 100Hz distortion peak, the woofer should, for best results, in no case be crossed lower than 120Hz.

At 100dB@1m the woofer comes to its limits (the distortion should already be very high). At the latest at 105dB@1m is the end, because the woofer with over 8mm excursion comes dangerously close to the maximum 10mm voice coil excursion - so that there is a danger that the voice coil hits the back plate of the motor system, which is not a pleasant noise.

***Source: Hobby HiFi 2013-4

Thanks a lot, Ctrl!
Looks a bit better than I thought.

With that (new information for me) in mind, how is it relative to the sub section? I mean the 8" at 120 Hz relative to the subs at let say 35 - 40 Hz, excursion wize?

I'm asking cause I already bought these sub drivers, but have a chance of buying a second hand complete set of these speakers. So if needed/reasonable, I could double up on the subs. In that case I have to put them together horisontally and use them as a TV table.

 

[Joecarrow]

I once used the LX521 full range for movies downmixed to stereo and I found that the woofers (I don’t call them subs) had a very hard time with the LFE. I had one bottom out or clip the 180w amp channel and that was a bit scary. Highpassed at 40hz, I think you’d be able to play *quite* loudly, with the limit being thermal or amplifier based.

I have found that the woofer section can give satisfying output to below 30hz in the right room with sufficient spacing from walls, but below that it’s a losing battle. I don’t think that open baffle below 30hz, or really much below the lowest room mode, is a great idea.

 

[Joecarrow]

I personally would not deviate from something which Siegfried, after decades of experience and design iterations, designed and specified. He is someone who was dissatisfied with the Orions, which were by far the best speakers I ever heard or owned, and was willing to design something as godawful ugly as the 521.4 in pursuit of improvement. He knew that these designs could never be a mass-market product, or even a small-market product. They are ugly, confusing, finicky, and quite demanding on the user.

The one thing I want to tell people who are building this speaker is that they absolutely need to try them out *as designed* before heading off on some quest to improve what is perceived as a probable shortcoming. What you’re saying is totally correct, he was very knowledgeable, thorough in his investigation, and thoughtful in the design. They represent his vision of the speaker very well.

With the very next breath I’ll tell people how supportive he was of DIY and experimentation and going your own way. If folks want to build it their own way that’s fine, as long as they acknowledge what part of it is theirs and what part is his.

 

[AudioJester]

SL seemed to always be improving on his designs. Its sad we never saw what was coming next.

I use Roon/HQP into an oko8 dac these days, having used many other solutions. The ability to "tune" each driver with audiolense provided me with the best results.
I also agree with others that the bass - both lfe and 100-250Hz is a weak point, particlarly at higher volumes. There are ways around this including using room acoustics. But its startling when much lesser speakers have more/better subjective bass - wharfdale/harbeth etc.

Is the diy route still financially viable - drivers/amps/crossover etc. I know the commercial versions dont make much sense in that you can buy a complete high end mesanovic/neumann/genelec etc for less without all the configuration/setup mucking around.

 

[Burning Sounds]

Firstly, let's be clear the bass unit on a LX521 is not a sub. It is designed to be used together as a single unit with the upper baffle. The LX521 was never designed for HT use, either. If you want greater SPL below 30Hz, then add a sealed box sub.

As for a Harbeth having "better" bass than an LX521 you jest (sorry, couldn't resist the pun). I've heard Harbeths many times and IMO they have not the articulation in the bass compared to a LX521. A dipole will never pressurise a room the way a sealed box will, so if that is important don't build a dipole. That is not to say that the LX521 is weak in the bass, it is not. IIRC 96dB@30Hz is about the limit.

If there is a weak spot in the LX521 it is the UM driver, not the LM. I currently have the magnesium LM driver, but never remember the original polypropylene driver being an issue at all.

As for being ugly, it's certainly unconventional looking, but it doesn't bother me one jot. I don't get confusing and finicky either. I've taken mine to quite a few audio meets and as long as the room is big enough they are easy to setup.

 

[ctrl]

With that (new information for me) in mind, how is it relative to the sub section? I mean the 8" at 120 Hz relative to the subs at let say 35 - 40 Hz, excursion wize?

I have to say in advance, before everyone starts hitting me, that the simulation looks pretty bad (in some cases of low frequency roll off), since as usual for loudspeakers it refers to free field conditions.
In the listening room, standing on the floor with a little help from the front wall, the low frequency range will not turn out quite so bad (this also applies to any other speaker, of course).

I don't have the blueprints of the LX521, if anything is dramatically wrong let me know.

UPDATE:
The baffle used for the simulation did not take into account the sound path extension by the H-frame - pointed out in post#171.
Therefore, I completely replaced the old simulations with new ones which take into account the sound path extension through the H-frame and updated the text accordingly.

The procedure is identical to post#159. Therefore, create a very rough sketch of the baffle for the two subwoofers and create the "open baffle" baffle diffraction frequency response. The baffle size of the LX521 subwoofer cabinet is WxH 13''x24'' but to take account for the path extension by the H-frame concept, I increased the baffle size to 13+6=19'' and 24+6=30''.


I equalized the low frequency range to a 40Hz BU2 high-pass - this corresponds to the roll-off of a sealed speaker. That means the f3 is at 40Hz.
Now we can look at the excursion of the subwoofers under free-field conditions. Shown is the 4pi frequency response as brown curve and in orange the BU2@40Hz high pass target function.
In this case, 90dB@1m and 95dB@1m are sufficient for us:

This does not look good, the limit is reached at about 95dB, if the music is listened to does not contain extremely low frequencies - with 40Hz BU2 high-pass FR.

The linear excursion of 14mm is exceeded very quickly @95db@1m, especially at frequencies below 35Hz. Even at moderate sound pressure levels, there is a risk that the maximum excursion of 28mm will be exceeded, when watching movies >95dB@1m (free field!), the subwoofers can sometimes hit the back plate.


Now many LX521 owners will angrily throw their keyboard against the screen and cry out "This little ass**le has no idea, I can hear much louder than with 95dB". As I said, when the speaker is placed on the floor about 6dB more SPL in the low frequency range is possible (and with the right placement, the front wall can also help).
Comparison baffle diffraction free field (orange) and with one reflective surface (blue, e.g. floor).

In order to compare loudspeakers, however, they are measured (and in our case simulated) under free-field conditions.


In the low frequency range under 30Hz the mutual cancellation of the dipole is so extrem that the subwoofers need to be protected from too much excursion, so I put an additional BU4@20Hz filter to protect the driver.


If we add the protection high pass with BU4@20Hz to protect against too much excursion, the whole thing looks much better at 95dB@1m free field:

Since the L26ROY tolerates a maximum voice coil excursion of 28mm, no problems are to be expected at about 95dB@1m even with movies or music with frequencies below 30Hz.


If you are willing to give up low bass and equalize to a high pass filter of LR4@40Hz, i.e. f3 is now at 50Hz, the maximum SPL increases significantly - 95dB@1m and 100dB@1m free field:

I had one bottom out or clip the 180w amp channel and that was a bit scary.

I have found that the woofer section can give satisfying output to below 30hz in the right room with sufficient spacing from walls, but below that it’s a losing battle. I don’t think that open baffle below 30hz, or really much below the lowest room mode, is a great idea.

The simulations seem to match your experience quite well.

Update: By mistake, I subtracted an additional 6dB for the simulation in the free field. I have corrected the images and my statements.

 

[mamsterla]

As a Linkwitz LX521.4 owner, I don't disagree that the dipole sub tails off around 30Hz. There is very little musical info there and for many people, the speaker will suffice. Sigfried was aware, however, and designed a monopole subwoofer called the Thor for folks who wanted the last octave. I personally do listen to some electronic music with information below 40Hz and went with dual Rythmik F15HP subs crossed over at 40Hz to do that lowest octave.

I think you will see that Sigfried understood that he made some trade offs in his design. Nothing you point out is very controversial IMO.

 

[jhenderson0107]

ctrl -

The LX-521 is designed to exploit the 6 dB gain in the bass region due to floor reflections. And, there are two drivers in the base module, providing an additional 6 dB gain. Do your sims above account for both of these effects?

 

[Joecarrow]

I have to say in advance, before everyone starts hitting me, that the simulation looks pretty bad (in some cases of low frequency roll off), since as usual for loudspeakers it refers to free field conditions.
In the listening room, standing on the floor with a little help from the front wall, the low frequency range will not turn out quite so bad (this also applies to any other speaker, of course).

I don't have the blueprints of the LX521, if anything is dramatically wrong let me know.

The procedure is identical to post#159. Therefore, create a very rough sketch of the baffle (with the wavelengths in the low bass this hardly plays a role anymore) for the two subwoofers and create the "open baffle" baffle diffraction frequency response. I equalized the low frequency range to a 40Hz BU2 high-pass - this corresponds to the roll-off of a sealed speaker. That means the f3 is at 40Hz. Set the LR4 low-pass at 120Hz for the woofer XO. Again in orange is the 40Hz BU2 high-pass target curve and brown is the free field LX521 subwoofer FR.
View attachment 237190 View attachment 237191

Now we can look at the excursion of the subwoofers under free-field conditions.
In this case, 85dB@1m and 90dB@1m are sufficient for us:
View attachment 237195 View attachment 237196
This does not look good, the limit is reached at 90dB at the latest, if no music is listened to that contains extremely low frequencies - if an equalization with HP BU2@40Hz is chosen.
The linear excursion of 14mm is exceeded very quickly, especially at frequencies below 40-30Hz. Even at moderate sound pressure levels, there is a risk that the maximum excursion of 28mm will be exceeded. When watching movies @90dB@1m, the subwoofers can sometimes hit the back plate.


Now many LX521 owners will angrily throw their keyboard against the screen and cry out "This little ass**le has no idea, I can hear much louder than with 85dB". As I said, when the speaker is placed on the floor about 6dB more SPL in the low frequency range is possible (and with the right placement, the front wall can also help).
Comparison baffle diffraction free field (orange) and with one reflective surface (blue, e.g. floor).
View attachment 237197
In order to compare loudspeakers, however, they are measured (and in our case simulated) under free-field conditions.


In the low frequency range under 30Hz the mutual cancellation of the dipole is so extrem that the subwoofers need to be protected from too much excursion, so I also put an additional BU4@20Hz filter to protect the driver.
View attachment 237194

If we add the protection high pass with BU4@20Hz to protect against too much excursion, the whole thing looks much better (again 85dB@1m and 90dB@1m):
View attachment 237200 View attachment 237201
Since the L26ROY tolerates a maximum voice coil excursion of 28mm, no problems are to be expected at 90dB@1m even with movies or music with frequencies below 30Hz.


If you are willing to give up low bass and equalize to a high pass filter of LR4@40Hz, i.e. f3 is now at 50Hz, the maximum SPL increases significantly (again 85dB@1m and 90dB@1m):
View attachment 237219 View attachment 237220 View attachment 237221






The simulations seem to match your experience quite well.

I’m not sure if this simulation takes into account that the woofer section is an H-frame, and has a bit more path length distance from front to back than if it was a flat baffle. It won’t change the results drastically, but perhaps a bit.

 

[ctrl]

I think you will see that Sigfried understood that he made some trade offs in his design. Nothing you point out is very controversial IMO.

Never doubted SF. I have only fulfilled @Keegan's wish

I'm glad you see it so rationally - physics can't be tricked.

The LX-521 is designed to exploit the 6 dB gain in the bass region due to floor reflections. And, there are two drivers in the base module, providing an additional 6 dB gain. Do your sims above account for both of these effects?

In fact, there is even a graph in post#168 that shows exactly the difference between free field conditions and the speaker standing on the ground.

Yes, two L26ROY per speaker are simulated.
The displacement volume is still somewhat limited at about 1000cm³.

It is no coincidence that open baffle speakers often use two 15'' or even 18'' drivers. The displacement volume is then about 1.5 - 3 times as large as with two L26 subwoofers.

 

[ctrl]

I’m not sure if this simulation takes into account that the woofer section is an H-frame, and has a bit more path length distance from front to back than if it was a flat baffle. It won’t change the results drastically, but perhaps a bit.

Thanks for pointing that out. Since I don't have the plans, no. Couldn't really estimate from the pictures how much the mean path extension of the H-frame is.

For comparison, in the simulation above I assumed a baffle with WxH 13''x23.6''. If the effective area of the baffle is increased to 19.7''x31.5'', then there is about 3.5dB more SPL available in the low bass region:


Will make a note about this in my post above.

 

[ctrl]

Gave my post#168 a complete "makeover". Since @Joecarrow correctly pointed out to me that I had not considered the path extension by the H-frame concept in my simulations.

When taken into account, each simulation is boosted by about 3dB SPL, which is probably a bit more in line with the experience of LX521 owners.
The maximum SPL in the low bass range is, for two 10'' subwoofers, of course still low, but normal due to the concept.

 

[fluid]

Gave my post#168 a complete "makeover". Since @Joecarrow correctly pointed out to me that I had not considered the path extension by the H-frame concept in my simulations.

When taken into account, each simulation is boosted by about 3dB SPL, which is probably a bit more in line with the experience of LX521 owners.
The maximum SPL in the low bass range is, for two 10'' subwoofers, of course still low, but normal due to the concept.

I made a comparison of my LX521 to my full range line array in volume displacement terms using Linkwitz's spreadsheet a few yeas ago. Dipole A is the LX521 and the Array Monopole B. They become equivalent at about 50Hz below the monopole wins above the dipole wins. Monopole A represents the Seas woofers in a closed box.

I no longer have the LX521 but the bass was one thing that I remember liking about them for music. For home theatre there was Thor to take over the bottom two octaves.
The SEAS cones used to jump all over the place with tests tracks that contained very low bass but it was always the sound of the treble that made me turn the volume down. I could not stand to listen to them loud enough to cause the woofers damage. They have 28mm of travel each way before damage.

The previous Orion system I had could cause SPIKE protection to kick in on the LM4780 chipamp's I used but that was with the DAC at 0dB putting out 2V. The LX521 system I built was with Hypex UcD amps with the two woofers attached to one UcD400, and it was never anywhere near clipping.

 

[Juhazi]

I have never had dipole lowest woofer (but a friend did before changed to sealed), but downfire sealed single L26ROY pair.
Important to remember is that dipole doesn't add room gain like monopoles.

Like #mamsterla said, SL has designed a sealed box subwoofer Thor for bass/movie freaks. Making it play well with LX521.4 is another story...

Here is my recent in-room measurement of my single L26ROY sealed system, with dipole 12" as lower mid and planar dipole UM and T. It sounds clean to me... My avatar shows it's horizontal dispersion. Looking at distortion, it is obvious that I use LR2 slopes.

#Keegan, LX521 is safe choice if you have a large room, or you can put them on the long wall of the room not very close to the wall. Dipole bass needs lots of space...

 

[suttondesign]

My LX521.4 did not do much in my listening room below about, say, 35-40hz. They could not pressurize the space, clearly. I used a good Rythmik sub crossed at about 45hz with a steep slope, and I high-passed the Linkwitz. By comparison, my Dutch 8C whack the shit out of the room cleanly down to 25hz before rolling off but still useful at 20hz. The Dutch sound like speakers. The Linkwitz do not.

 

[Worth Davis]

My lx521 has quad 18s crossed over at 80 in addition to the standard build

I’ve run it with Dirac or accurate as well.

I haven’t found anything in the 15 to 20k range that outclasses it. I think there is room to grow like all purifi amps, but it’s a great system

 

[Digital1955]

These things looks like some movie props from Beetlejuice.

 

[jeffbook]

These things looks like some movie props from Beetlejuice.

Nice troll......

From Siegfried Linkwitz's website:

Frequently Asked Questions​

When a new loudspeaker looks different from what is conventionally expected or what experts in loudspeaker design consider to be reasonable, then this raises questions, expectations and conclusions based on speculation, previous experience, understanding or a personal agenda. I consider the LX521 to be my best design in a long search for an electro-acoustic transducer, which can set up in a domestic size room a realistic seeming illusion of an acoustic scene, while perceptually moving the loudspeakers, the room's resonances, reflections and reverberation beyond one's acoustic horizon. Q2 - I do not like the appearance of the top baffle and want to change it slightly A2 - Don't change it. Try to get used to it. The midrange/tweeter baffle is an essential acoustic design element. It controls the interference between between front and rear radiated sound waves in order to obtain a dipolar radiation pattern over the whole frequency range. Thus baffle dimensions must be measured in fractions of radiated wavelength and how much phase shift is produced. Acoustic performance was his primary design criteria, not appealing to ones aesthetic values.

 

[jhenderson0107]

Build progresses:

Assembled and sanded:

Painted, awaiting drivers and grill cloth installation:

 

[Madjalapeno]

Build progresses:

Assembled and sanded:
View attachment 240621

Painted, awaiting drivers and grill cloth installation:
View attachment 240622

Nice. Any thoughts on cabling yet?

I used Speakon 8-way connectors, and put a junction box for each upper unit. Makes life much easier.