Active Crossovers (Analog and Digital) and Bi-Amping...

[JonFo]

My Martin Logan SL3's have dynamics! Something they never had before! In fact, that was the only thing I didn't like about them, was the lack of dynamics. ...

Congratulations! see, it wasn't that hard
And yes, once freed of the lossy passive x-over, these speakers (and the amps that drive them) can really shine, and you're not even done with the tuning process yet.

So far, the dbx is just running crossover duty only. The panels are crossed over at 250 Hz with a Linkwitz 48 dB slope, the woofers are also at 250 Hz, but with a Linkwitz 12 dB slope and 16 Hz on the bottom end with a Linkwitz 48 dB slope. The sub is set to 46 Hz with a Linkwitz 24 dB slope.

I'm not quite sure why, but when I initially had the SL3 woofers crossed over at 48 dB @ 250 Hz, there was a big hump in the mid-bass, making everything rather bloated and chesty sounding. Knocking it down to a 12 dB slope eliminated that hump and made everything sound natural. There's more body to vocals like there should be (which is something the SL3's are known to lack a little of). ...

Probably because there is some EQ required, also, I'd try a 24dB L-R low-pass @ 250 on the woofer and see how that works. A first-order (12dB) low-pass means the woofer is playing well into the territory of the panel. That might be the frequency best blend, but THD in that range can take a hit. IIRC, the woofer is not clean past 500Hz. A more modern driver, speced for woofer/mid-bass (40 to 800Hz) would do better, but that's for phase II

Bass content now has snap and texture. Vocals, both male and female sound amazingly real and lifelike. They have life and warmth to them. Treble is smooth and detailed. With no EQ of any kind, they sound like they're being powered by a good tube amp without being rolled off.

The JL sub has completely vanished into thin air. It has become a natural extension of the SL3's. With your eyes closed, you just assume all of the bass you're hearing is coming from the SL3's and nothing more. Then again, with your eyes closed, you really can't tell that any of the sound is coming from the speakers. It's just there in the room with you and that really low bass just shows up when needed.

Listening to Louis Hayes' Serenade for Horace right now, and it's oh so good sounding!

Yep, all that sounds very familiar, and it will just keep getting better as you fine-tune it with the measurement rig.

One immediate recommendation: adjust the delay between panel and woofer, tune this before you do any further EQ and crossover tweaking.

It's best done with REW Impulse response metrics, but you can do a first approximation by using these values that worked on my Monoliths: Set panel delay to 0.26ms.

I hope the tuning continues well this week. Keep us posted.

 

[Chops]

Congratulations! see, it wasn't that hard
And yes, once freed of the lossy passive x-over, these speakers (and the amps that drive them) can really shine, and you're not even done with the tuning process yet.


Probably because there is some EQ required, also, I'd try a 24dB L-R low-pass @ 250 on the woofer and see how that works. A first-order (12dB) low-pass means the woofer is playing well into the territory of the panel. That might be the frequency best blend, but THD in that range can take a hit. IIRC, the woofer is not clean past 500Hz. A more modern driver, speced for woofer/mid-bass (40 to 800Hz) would do better, but that's for phase II



Yep, all that sounds very familiar, and it will just keep getting better as you fine-tune it with the measurement rig.

One immediate recommendation: adjust the delay between panel and woofer, tune this before you do any further EQ and crossover tweaking.

It's best done with REW Impulse response metrics, but you can do a first approximation by using these values that worked on my Monoliths: Set panel delay to 0.26ms.

I hope the tuning continues well this week. Keep us posted.

It was your crossover diagram that I followed to figure things out. Without that, I'd probably still be scratching my head staring at it on the floor. I'm glad there's guys like you around that understand this stuff and share your findings. Bypassing the crossovers is simple. Knowing what wires go to what taps on the transformer and connections beyond that point is another story. Removing the passive crossover networks, the immediate effect noticed is similar to that of doing a full intake and exhaust upgrade on an engine. It's suddenly able to breath much easier and better.

I just realized another part of the problem as to why the bass was a little bloated when crossed at 48 dB... I had the woofers bumped up +1.5 dB when still running the passive networks to help fill in the lower mid-bass. I now have them down to +0.5 dB and at 48 dB slope. It sounds much better now. There's even greater detail now.

Since I have no way of making measurements yet, I didn't mess with the delay. But that will change soon as I have a miniDSP mic on its way. I went ahead and set the panel delay to 0.27ms for two reasons... 1) The Venu360 will do either 0.25 or 0.27, not 0.26. 2) I figured since my woofers look like they sit back a bit further than the ones on your Monoliths, a little more delay probably wouldn't hurt.

According to the Venu360, that's equivalent to almost 3.75". Trying to eyeball the distance without removing the grill over the woofer but using a tape measure along side the cabinet, it looks to be about that. Not sure if I really hear any difference. If I do, it doesn't jump out at me. Probably a good thing. LOL

Thanks for the great tips so far! I really appreciate it!

 

[Chops]

Playing around with crossover frequency some more, I'm now up around 295 Hz with LR 4th order, woofer gain at unity with the panels, and 0.31ms panel delay. There's still a little bit of bloat from the -48 dB slope used on the woofers, but it still sounds better than at -24 dB overall. I'll play around a little with the PEQ on the Venu360 to see if I can get a notch in that "bloat" area. Mind you, this is all being done by ear right now until the miniDSP USB mic shows up. It should be here tomorrow, so I'll have plenty of time this weekend to get some measurements and such.

Also, I've noticed a faint "buzzing" every now and then, both with music and speech. Watching Lewis Black last night and noticed it a few times. Listening to some jazz this morning and noticed it again. So I opened up the RTA on the Venu360 and fired up the sine wave signal generator. The buzz got quite apparent when up around 200 - 215 Hz. At first, I thought it was coming from the middle of the panels, but upon further investigation, it's actually in the middle of the side panels of the bass enclosures, right about where the screws from the internal braces meet the side panels. I'll have to try and address that this weekend as well. I did notice when I was in the left speaker last, that the one brace was split down the middle about 3 inches, exactly where that one screw is. Hmm...

 

[Chops]

So after putzing around with REW for a little bit, I realized that I had the wrong mic cal file loaded, so remeasured with the correct "90 degree" mic calibration file. The titles on each graph tells what's going on. What little EQ I was using via the Venu360 is all disabled. Only thing active is the crossover points and 0.31ms delay on the panels.

 

[Chops]

This one is a bit weird once the sub is added... Looks like the sub killed some of the group delay.

 

[Chops]

Playing around with REW some more and the settings in the dbx Venu360, I started messing around with the driver alignment, this time including the sub into the mix.

After entering my room dimensions, speaker and sub locations as well as my seating location in the "Room Sim" feature of REW, I noticed down in the right window of that screen that it gives you the time alignment. Not knowing if these are correct or accurate, I decided to punch them into the Venu360 anyway.

Originally, I only had the ML's panels delayed by 0.31ms. Well now with the JL added into the mix, it is the starting point of time, so no adjustment made to it. The ML woofers are delayed to 2.00ms, then the panels are delayed to 2.31ms. So...

Sub - 0.0ms
Woofers - 2.0ms
Panels - 2.31ms

I also reactivated the high shelf filter for the panels, starting at 9.34 kHz, increasing to a max of +3 dB to 20 kHz at a rate of 6.6 dB per octave. This is to counteract the natural rolloff of the panels above 10 kHz, just to reintroduce a little air and sparkle.

So anyway, what does all of this time alignment do for the sound?... Well, for starters, bass is a LOT cleaner, tighter, precise and detailed. It's also a little fuller and definitely extends a little deeper. Upper bass and lower midrange is a little fuller and more natural. Vocals are lifted up about a foot or more above where they were, putting them about life sized height. It also reduced that 50 Hz suck-out by about 1.5 dB!

On top of all of that dynamics and sound stage are in spades! One thing that was always a little bit of a letdown with these ML's was the lack of dynamics. I gained a lot of that when I finally bypassed the passive crossovers and went fully active with them. Well, once getting the time alignment tuned in even better, dynamics are crazy good now! I'm quite shocked at the amount of dynamics now!

And as for sound stage, this is also very much improved. Good depth, good height, very good width. It's like using the Carver Sonic Holography!

This just proves how important time alignment is to a system, and what kind of gains can be achieved throughout.

 

[JonFo]

Hi Chops, great to see the progress on this project. Getting more seat-time with REW and making some of the adjustments you've made can sure have an impact. Keep at it, as there is always a bit more to extract from the system.

Good to hear the results are instantly obvious when listening. Time alignment is a critical area to pay attention to, and I find that only good speaker processors (or pre-amps with time-adjsutments) allow one to correctly blend the sub and the mains. Here, you're going the extra step and aligning the panels and woofers in the ESLs, which is the optimum way to do it.

Using impulse responses, you might want to check the sub to woofer phasing and make sure those are optimized, then if there are changes (such as flipping polarity on the woofer), double check panel to woofer phase. Choice of crossover slope and style can also affect phase in the crossover region, so if you chose something other then LR for style, you'll need to check. In some instances, going to a Butterworth will change the phase relationship in beneficial ways, others not so much. Just something to be aware of.

I played around with many settings for weeks on the SL3XC setup, good learning experience.

Congrats on all the great progress and for sharing the results along the journey.

 

[Chops]

Jonathan,

As it sits, I have the woofers in the SL3's inverted via the Venu360. Being inverted, they are in phase with both the panels and sub. I tested them yesterday individually at the crossover point with the panels, and tested them individually with the sub.

I haven't tried any other crossover alignments other than LR. I guess I could play around with BW alignments just see what happens.

It's all great fun with excellent results. That I know!

 

[Chops]

Today I decided to play around with the dbx RTA mic and the AutoEQ "Wizard" on the Venu360. Boy am I glad I did!

Here's the end results right from the RTA mic, via the iPad and the Venu's display. You can't get much better than that. Darn near ruler flat across the board. Of course, I made a few tweaks of my own to get it there, which I'll mention below...

I chose a flat curve and this is what the Venu360 came up with after taking three measurements across all three seating positions on the couch.

The settings the Venu360 made are really good, though it took out a bit too much low end for my liking as well as a little too much upper midrange (#1 and #2).

#1 was pulled down by -7.8 dB... I bumped it up to -2.7 dB.

#2 was pulled down by -4.7 dB, and I bumped it up to -2.6 dB.


Again, these two little tweaks above is what got me to a near perfectly flat room response shown in the RTA.

And just when I thought I had this system dialed in pretty good by ear... Last night I was listening for a short bit through my headphones from the Carver C-1 which is obviously before the dbx. Then when I went back to listening to the speakers, I could quickly tell something was off. Well I can safely say, after running this AutoEQ wizard, it isn't now!

Listening to this right now and man does it sound good!