SL3-eXtreme Center channel speaker – a design and build story

[raphant69]

I'm happy to see you're enjoying so much your system because after all, it is the most important thing in all that we do.
I confirm that I'm waiting for new stats for my Sequel II's.
At this moment they must be on a boat in the middle of the atlantic ocean, between USA and France.
I think that they're gonna improve my listening pleasure a lot.
I was also wondering about changing woofers but don't know if it's worth. Or maybe upgrading with aluminium ones...
Anyway I plan to remove the nextel and make a black gloss piano finish on these boxes. Like you I used to build speakers before I bought the Sequel II's and I know how proud one can be enjoying DIY stuff. What is interresting in ML's for DIYers (but also for fixing) is their modularity. For example, with two DIY bass towers with 6x10" drivers and 4 monolith stats, one (who has a big room!) can easily built a kind of cheap statement that would give great results.
I have a question you might could answer to:
Do you know how could I remove the Nextel from the Sequel II's boxes, as far as it is an uncommon product? (nobody knows about it here in France)

Best regards from France
Raphaël.

 

[JonFo]

I'm happy to see you're enjoying so much your system because after all, it is the most important thing in all that we do.
....
I have a question you might could answer to:
Do you know how could I remove the Nextel from the Sequel II's boxes, as far as it is an uncommon product? (nobody knows about it here in France)

Best regards from France
Raphaël.

Hi Raphaël,

Yes, sitting here enjoying some Jacques Lousier playing Bach on SACD. Very clean transients and great attack on the drums, which attributable to the Monoliths and the Line Array.

To build a mini-statement, I'd just add another set of Extremis Line array to each Monolith and be done. That would be fine in my 16x28 room.

As for how to remove the nexel finish, I’m not sure. My SL3’s had some chips on them that peeled back pretty easily. I’d think prying it off with a putty knife should be possible. However, it might involve a lot of sanding.

Also, the base material is MDF, so work accordingly.

But it should be possible to strip down to primer (which is white) and start back up from there.

BTW- Some of the best things come from France, like my wife ;-)

Salut from someone who lived for a few years in Strasbourg,

 

[JonFo]

Final Front Shots

Some additional shots of the screen going on, as we took measurements both before and after the screen is up.
Most of the measurements are with the screen up, as that’s how it will be used, and we also want to adjust levels and EQ for the screen losses.

Note the Diffusion panels behind the ESL on the front wall. Those help mitigate the rear wall reflection straight back into the ESL. They also smooth the horizontal dispersion characteristics by ‘spreading out’ the rear wave. I’ll point out the specific effect of this in the measurements.

{2008 edit} The front wall is now entirely covered with MiniTraps HF to absorb the rear wave from the three front speakers fully. That makes for a much clearer and focused soundstage {/edit}

Here is one more long shot from the seating position:

 

[JonFo]

Screen Up animation

And here is an animation of the front view. First is the bare system, then the black acoustically transparent masking which is required to cut the light spill through the screen. Otherwise, even with 18% grey walls, the light bounce from behind would affect the on-screen image. Also, that big metalized ESL panel is mighty reflective as you can see in some of the shots, so we need to tame its light reflections.
Finally, the Screen itself is up. It fills the entire space between the monoliths (80” wide plus frame).

 

[JonFo]

General Comments on process

Getting into this process is not for those who are unwilling to invest the time and effort to master tools, techniques and learn.
The results have been amazing, but the price paid is high.

When we fork over hard earned cash to Martin Logan, this is what we are paying for, their expertise and many, many hours of testing, measurement, prototyping and evaluation.
The only reason I went down this path is that they don’t make a speaker to my specs.

Also, while I am a firm believer in the benefits of active crossovers and multi-amping, it also introduces complexity into the equation.
I calculate I have more than 85 controls per channel I can adjust. The sub has a 12 band parametric EQ with 4 main controls per band for another 48. Just for the speakers along the front, we are facing more than 300 adjustable parameters. And the interesting thing is that many of them interact a good bit. Cabling is another huge deal, as between the Meridian processor and the amps there are more than 50 cables (thankfully, most are Balanced interconnects) and 10 devices. So getting them all right is both a science and an art.

I’ve spent more than 30 hours in the tuning process over a period of a few weekends. The wife is sick of test tones, as I’ve run hundreds of those. Frankly, I’d rather hear music myself now.

That said, here goes a raft of posts on measurements, the process and the results.
We will compare the SL3XC against the prior Logos as well.


2008 – 2012 period update: While still complex to tune, the system has vastly simplified with the upgrade to a DBX DriveRack 4800 handling crossover duties for all the front channels (L/C/R/Sub) in one box.
Additionally, the replacement of the Meridian gear with the awesome Denon AVP-A1HDci pre-amp with Audyssey Pro room correction has made the tonal and timing match across the system near perfect. Imaging is on a totally different plane than before, its incredible how these large speakers just disappear and all you hear is a focused soundstage managed by whatever the mixer of the recording intended.

 

[JonFo]

Center Channel Measurements - Distortion

One of the main concerns with my design was to achieve a low distortion, high SPL capable center to keep up with the Monoliths.
Looks like we’ve achieved that, with a 0.6% overall distortion number (including the sub), which for speakers is awesome.

 

[JonFo]

Comparison to Logos

These are probably some of the most telling measurements, as the goal was not only to give massively higher SPL capabilities but to do so with a smoother frequency response at the listening positions.
As you can see in this graph, the red trace of the SL3XC is much smoother than the older Logos (blue).
Keep in mind this is taken at the primary listening position and reflects (no pun intended ) the impact of the room as well. That’s usually the dominant element in Frequency Response (FR).

 

[JonFo]

Logos Waterfall plot

Lot’s of resonances and issues in this plot. Part of the reason (and this applies to any speaker) is that it's behind a screen which reflects some of the energy back and creates standing waves behind the screen. This ‘smearing’ is unavoidable, but the Logos seems particularly affected by it compared to the SL3XC.
The reason the Logos has this bad comb-filtering is the panel is an inch away from a solid back (the box for the woofer), forcing the rear wave to come right back through the ESL, causing this.

 

[JonFo]

SL3XC Waterfall plot

Fewer resonance issues, but they are readily apparent here as well. Room modes again affect the FR at this location. More room treatments to come in the future to tame those.
In general, considering it’s behind a screen, not too bad. Much better than the Logos, that’s for sure.

{2008 edit} fixing the obvious comb filtering and resonance seen in this chart is why the wall behind the fronts is entirely covered with absorption devices. {/edit}

 

[JonFo]

Seating comparisons

An obvious question about the SL3XC vs the Logos is horizontal dispersion and coverage at key listening positions.
Thankfully, my room arrangement (again designed around these ML’s) lends itself well to placing all key seats in the 30 degree field from the ESL element of the SL3XC. It even has better high-end coverage on the left than the Logos did.
The mid-bass difference is ‘extreme’

As we’ll see in the psychoacoustic measurements, that mid-bass hump is actually needed for a balanced presentation.
In listening, it is a critical ingredient for anchoring the center. The SL3XC’s ability to deliver high SPL (a full 105db measured) with all the mid-bass there is why action movies now really ‘rock’.

Left Seat – Logos Vs SL3XC (green)

 

[JonFo]

Right Seat – Logos vs SL3XC

Scale wound up being a little tighter on this plot, so it looks more ragged, but essentially it’s the same as the left seat, and again much smother and extended than the Logos.

Right Seat – Logos vs SL3XC (green)

 

[JonFo]

Logos (blue) Vs SL3XC (red) Impulse

One of the most important and useful measurements that can be reviewed is the Impulse response. This shows the speakers ability to start and stop quickly, and shows the time-domain impact of the electronics (crossovers add group delay) the Speaker (cones are slower than ESL’s) and the room (biggest impact of all).

Here is a comparison between the Logos (blue) and the SL3XC (red).
Note how ragged the logos is. Those are artifacts of the speaker itself (it has a three way passive x-over), the reflections from the screen and other room effects.
The SL3XC in contrast is waaay cleaner, even though it too is behind the screen.

Also the Logos had inverted polarity when this measurement was taken (the blue line goes down at first).

Logos (blue) Vs SL3XC (red) Impulse

 

[JonFo]

Gain and Polarity

The process of tuning the center starts with adjusting gain and polarity.

Gain is pretty straight forward, well, maybe not so in my rig, as there a total of seven separate level controls to set to get this one channel correct

1- The Meridian processor (and the Sony TAP-E5000 pre-amp switcher, for SACD/DVD-A)
2- The line level to Balanced matching transformer (ART Cleanbox)
3- The input to the DBX driveRack
4- The Crossover elements – High pass, mid-band bandpass and sub low-pass
5- The sub output goes to a mixer where each input has a level control (MX882 in mixer mode)
6- The mid and high pass also go through a gain matching mixer/splitter (MX882 in splitter / gain adjustment mode)
7- The sub EQ also has input adjustments

Then there is the issue of adjusting polarity/ phase /delay (all of which are interrelated).
This is one of the most critical elements to get right, as time-coherence is vital to precise imaging and to smooth frequency response as well.

Once again, not so straightforward in my rig. As we need to time-align not only the usual 5.1 set, but since we actively cross-over, there are actually 11 ‘channels’ to get right.
No walk in the park I tell ‘ya….

 

[JonFo]

Latency management

First, we do some math. Not my favorite subject, but hey, it makes things work.

The thing to remember about time-alignment is that some processors will add latency (fancy word for delay) to the chain, so that needs to be factored.

So out comes the spreadsheet to figure what the net delay at the speakers will be and what intermediate delays might be required.
In the spreadsheet below, I used the fact that the BFD EQ adds a full millisecond of latency to sort of set the bar for the rest of the channels.
But since the DBX has 0.6ms of latency, we need to factor that back in to the sub. So in the Meridian we add in 0.6ms of delay by fibbing a bit about the distance measurement of the sub.
After that, all the delay management will be done in the DriveRacks and the DCX2496 used on the rears.
These units let me control delay in increments of 0.05ms. So no sweat dialing in exact values.

The Buttkicker tactile transducer is also time aligned so the shaking is in synch with what we hear. Once again we have to subtract out existing latency of the EQ and the delay processor itself (a Midiverb 4). Something about the term ‘anal-compulsive’ is appropriate here ;-)

Here is the spreadsheet.

 

[JonFo]

Phase and Polarity

Now that gain and latency are handled, time to look carefully at the Phase / Polarity of the channels.

Starting with the Subwoofers, I measured all the channels and examine the impulse responses to ensure that A) we are time aligned, and B) that we are matching polarities.

Here is a shot of the Sub channels and what an out of phase element looks like. The Black trace is the rear surround channel, as you can see, it does not trace the Center and the Monolith.
The red trace, which is the LFE channel, looks ragged because there is no crossover before it, so the test tones are wide range and it is trying to reproduce frequencies above 125hz here. We can still see the polarity is a bit offset from the others. But once the Meridian adds in the missing 0.6ms of delay, it will align.

To fix this, I just clicked off a control in the DCX (which manages the rears) and flipped the polarity. I also added a few hundredths of a ms delay to align with the front / center.

 

[JonFo]

Alignment of phase and time

Here is the center a Monolith and the LFE. Pretty well aligned except for the LFE, which I can’t measure through the Meridian, as it only generates output from the LFE channel when decoding Dolby Digital or DTS formats (It’s a setting, and I have it set up this way, there are other ways which would engage the sub out for music).

While watching the ‘Transporter 2’ movie last night, there was no lack of LFE and the time alignment and phase seems correct with Avia and other Dolby digital test tones.

 

[JonFo]

Psychoacoustic response

In this measurement, we see the individual responses of the SL3XC at the primary three listening positions (roughly 16’ from the speaker).
This Psychoacoustic measurement windows the measurements in such a way as to reflect how we hear the various frequencies. Essentially, the lower end is ‘lower’ in this type of measurement.
It really works, as when I adjust to a flat third-octave response, the bass is lacking.

 

[JonFo]

Averaged Psychoacoustic response

Here is an average of the three listening positions. Not too shabby, roughly fits in a six DB window from 16hz all the way to 14Khz.
Again, this is behind a screen, measured in-room, so it’s really good.
Perfect will require more work

 

[JonFo]

Compare to Monoliths

And here is comparison to the recently refreshed Monoliths (new panels and Peerless woofers).
As you can see, the SL3XC is somewhat smoother than the Monolith. Also you can see the room effects impact the speakers in almost opposite ways at certain frequencies.

The SL3XC is the blue trace, the red trace is the right Monolith.

 

[JonFo]

Averaged Psychoacoustic response w/Monolith

In this trace of the summed Center and Monolith response you can see how much smoother the overall response becomes.

This why I really like listening to the ‘TriField’ mode on my Meridian. With a good center it really reduces room impacts, it also provides an enveloping sound (anchored in the front) that leaves stereo in the dust.