Highest crossover point from woofer to midbass in tower speaker for best sound? And distance from floor

[ROOSKIE]

Mike I think you are over thinking the benefit of ultra high power handling and yet seem to be neglecting peaks. Your amp needs to be 300-500watts if you are going clipping free when using 50-100watt continuous power delivery as a reference.
Your drivers can't handle that power level but you will still want it if you actually plan to maximize your limits.
Additionally crossover slopes that matter are acoustic, electrical slopes do not matter at all. Measure.
You want the proper acoustical responce, the electrical responce means nothing.
You may also want to blend the drivers more with shallow slopes as some prefer that sound. So listen. Listen and measure.
Anyway I feel the main benefit of higher crossover isn't exactly lower harmonic distortion or higher power handling (though that is a benefit)rather it is much lower IMD and doppler distortions thru reduced cone motion.
Less motion is better.

 

[Tangband]

Coincidentally, the speakers I have are the ATC SCM20 Pro PSL Mk2. The tweeter is the S-Spec, (same as the one used in the most recent 50/100/110), and the midbass is underhung with 12mm travel and a 3" voice coil which uses the dome of the midrange of those 100s in the centre as a dustcap/radiator. The crossover between it and the tweeter is third order (2.1kHz) as well (which purely coincidentally is my favourite slope for mid to tweeter, but not something I ever considered before for mid to bass (esp at such low frequencies), mostly because of component size

I'll be going active with this though, so third order between the two will be easy.

You say up at ear level... I can definitely try that.

To start I have two (active) KEF Kube 10b subs I can cross as high as 140Hz with a 24dB slope. I think if I do that and feed them with 3rd order at 90Hz integration would be decent.

About KEF Kube 10b:
- sealed
- 10"
- 24Hz -3dB
- linear movement (estimate) +-12-15mm (24-30)
- max excursion (estimate) +-20-25mm (40-50mm)
- 300W RMS class D

I reinforced them like I do all my speakers. Even my SCM20s benefitted nicely from some high-force cross bracing (much higher fs and much less reactive at it)

My only significant concern with the Kubes is, because their drivers' moving mass isn't 1.3kg and their amps aren't 5000W, the 24Hz -3dB point comes from DSP. This DSP... I don''t know the effect it will have on phase at the top end of frequency response. Hopefully it's minimal. I know that the amplifier's DSP completely ignores the mechanical limits of the driver (lol), so I think it's unlikely I'll run into compression from any cause other than power limit.

Room correction software - do you think it would adjust the signal well enough to compensate for any unwanted effects from DSP?

Just had a brainwave: to separate room from signal, instead of using a microphone for correction, I could use the audio outs for each, combined to one channel digitally, and use that channel instead of the microphone as the source for the correction software.
Worst or best idea ever?
My first thought was to take the speaker outputs with, say, ~0.1 ohm output impedance) from both channels (sub, satellite) each making the same SPL in the room (90dB or whatever), put each output through 100 ohm resistors, then attach the outputs of those resistors together and into a (~10k) line input.

You need to know how much latency there is in your dsp Kef subwoofers . This delay can be critical for integration if you cross at 140 Hz , or not. The subs cant be on the floor if you cross them at 140 Hz .

ThomasA :s suggestion with a 18/30 crossover is a good one, slightly better sounding than 12/24 - if youre gonna cross at 70-80 Hz and your subs are near the floor . But as I and Rooskie wrote earlier, a good threeway with the drivers a bit off the floor and near eachother is a better sounding option with less compromises.

Roomcorrection is the last thing to do , after the correct installation of your speakers using Tune method and after room treatment . I only correct for the 3 fundamental room resonances , meaning wall-wall, floor-roof, wall-wall. All three are below 80 Hz . I correct for half of the value that the microphone is showing. If theres a fundamental room resonance with a +10 dB peak at 44 Hz, I correct this with -5 dB with the correct Q value.

A very good room and perfect installation with a good source signal dont need any room correction .

 

[mike7877]

Random tidbit of information about ATC's stuff in case anyone's interested:

In a video I saw of the testing/QC of the 3 channel amplifier used in the active SCM50/100/150, I saw that THD+n at 2.83V was in the high -80s. Specifically, I remember seeing -87
This specification considered alone isn't great, but, when you take into account the reality of perfect phase from matched levels and slopes, performance is so good that, for better performance, the amplifier you'd need to use in a passive configuration would have to be EXTREMELY good - one of the best.
And that amp, though it would be more transparent in the high end, ultimately there would be decreased coherence and impact in the upper bass. I think that, overall, the included amp would be preferable.
ATC should just increase the quality of the tweeter amplifier and the series would be unbeatable in so many regards.

If I were them, I'd have two crossover configurations for users to switch between: high power mode and low volume mode, where the intended maximum power rating was 50WRMS. Instead of 3.5kHz to the tweeter, it'd be something like 2.1kHz (like my 20s), and instead of 380Hz to the midrange, it'd be something between 240 and 280.
Better dispersion and quality between 1 and 4kHz

 

[mike7877]

Mike I think you are over thinking the benefit of ultra high power handling and yet seem to be neglecting peaks. Your amp needs to be 300-500watts if you are going clipping free when using 50-100watt continuous power delivery as a reference.
Your drivers can't handle that power level but you will still want it if you actually plan to maximize your limits.
Additionally crossover slopes that matter are acoustic, electrical slopes do not matter at all. Measure.
You want the proper acoustical responce, the electrical responce means nothing.
You may also want to blend the drivers more with shallow slopes as some prefer that sound. So listen. Listen and measure.
Anyway I feel the main benefit of higher crossover isn't exactly lower harmonic distortion or higher power handling (though that is a benefit)rather it is much lower IMD and doppler distortions thru reduced cone motion.
Less motion is better.f

What do you mean by overthinking the benefit of higher power handling?

Officially, the 2 ways are rated for 108dB max continuous per pair - 102dB each.
Each subwoofer is rated for 105dB max continuous each - 111dB per pair
On paper they're perfect for each other! Sub is capable of a bit more than the mains? Great!
But I've noticed while listening to loud music, I way more often run into mechanical limits limiting volume before clipping even becomes a consideration. Since the drivers are underhung and designed well, they've never bottomed out per se, but the further the woofers are supposed to be moving and aren't...

Obviously the severity is dependent on the genre, but it's always the 2-ways hitting their mechanical limits first.

I went over this in another post, but I'll say it again:
137Hz is the optimal frequency to maximize power handling. A 5" driver at 137Hz will make 108dB (match the subwoofer) when it's moving +-6mm (12 p-p)
What's the significance of 6mm?
The very high power and very well made 5" driver I have has:
- cast basket
- >200oz magnet
- 3" voice coil, underhung
- 20mm gap
- rigid damped cone with 3" dome midrange as the dust cap/radiator

The height of the voice coil is 8mm, which moves accurately within a 20mm homogenous electromagnetic field made by the massive magnet
20mm - 8mm = 12mm of absolutely perfect movement.
2 way efficiency is 85dB/w
This means for 105dB, power input needs to be 100W.

The X3700H I have is a 9-channel power amplifier. When driving just one pair of speakers, the transformer and filter caps are having an easyyy time. Also, it's run in bi-amp mode, meaning even if the two channels driving the tweeters have their supply rails drop significantly, the clarity of sound will remain. Eg. Rails drop from 65V to 54V - tweeter amps be like "who caresss?!? My signal's only 14V peak to peak!"

One channel into 8 ohms is 170W continuous, two channels into 8 ohms is 140W continuous. As 8 ohm speakers are usually (conservatively) around 6 in the 50-150Hz range (where all the power is...) that 140W is 190W continuous. Peak will be higher.

I've also done a listening test before with a 12dB high pass filter at 150Hz, and the extra few dB I gained from not being limited by cone movement were distortion free.

Eventually I'll have my Kinergetics KBA-280 fixed (it's been in the shop 18 months...) which is a class A BEAST.
At 1 ohm it'll put out over 1000W per channel RMS
2 ohm, 510W
4 ohm, 280W
8 ohm, 140W

It takes almost 250W sitting doing nothing so it's not something to leave on in the background lol

 

[ROOSKIE]

What do you mean by overthinking the benefit of higher power handling?

Officially, the 2 ways are rated for 108dB max continuous per pair - 102dB each.
Each subwoofer is rated for 105dB max continuous each - 111dB per pair
On paper they're perfect for each other! Sub is capable of a bit more than the mains? Great!
But I've noticed while listening to loud music, I way more often run into mechanical limits limiting volume before clipping even becomes a consideration. Since the drivers are underhung and designed well, they've never bottomed out per se, but the further the woofers are supposed to be moving and aren't...

Obviously the severity is dependent on the genre, but it's always the 2-ways hitting their mechanical limits first.

I went over this in another post, but I'll say it again:
137Hz is the optimal frequency to maximize power handling. A 5" driver at 137Hz will make 108dB (match the subwoofer) when it's moving +-6mm (12 p-p)
What's the significance of 6mm?
The very high power and very well made 5" driver I have has:
- cast basket
- >200oz magnet
- 3" voice coil, underhung
- 20mm gap
- rigid damped cone with 3" dome midrange as the dust cap/radiator

The height of the voice coil is 8mm, which moves accurately within a 20mm homogenous electromagnetic field made by the massive magnet
20mm - 8mm = 12mm of absolutely perfect movement.
2 way efficiency is 85dB/w
This means for 105dB, power input needs to be 100W.

The X3700H I have is a 9-channel power amplifier. When driving just one pair of speakers, the transformer and filter caps are having an easyyy time. Also, it's run in bi-amp mode, meaning even if the two channels driving the tweeters have their supply rails drop significantly, the clarity of sound will remain. Eg. Rails drop from 65V to 54V - tweeter amps be like "who caresss?!? My signal's only 14V peak to peak!"

One channel into 8 ohms is 170W continuous, two channels into 8 ohms is 140W continuous. As 8 ohm speakers are usually (conservatively) around 6 in the 50-150Hz range (where all the power is...) that 140W is 190W continuous. Peak will be higher.

I've also done a listening test before with a 12dB high pass filter at 150Hz, and the extra few dB I gained from not being limited by cone movement were distortion free.

Eventually I'll have my Kinergetics KBA-280 fixed (it's been in the shop 18 months...) which is a class A BEAST.
At 1 ohm it'll put out over 1000W per channel RMS
2 ohm, 510W
4 ohm, 280W
8 ohm, 140W

It takes almost 250W sitting doing nothing so it's not something to leave on in the background lol

In my view maximizing power handling is not really the point. I would like to maximize performance.

What I mean is that for myself by far the biggest benefit of high passing in general and especially higher up essentially making a 3 way speaker is reducing cone travel.
If you are actually going to drive the 5" so hard that you are hitting 6-10mm in the 125-150hrz zone you ought to consider the use of a larger driver.
Reduced cone travel means less IMD and less doppler effect, vastly more significant in my mind(and experiences) vs HD.
The HD is not really the issue. If you have 20% HD at xmax what does it matter if that xmax is at 60hrz or 140hrz? I am pretty sure 20% HD at 140hrz is going to be more audible than 20% at 60hrz though you will have more ultimate SPL capability.

So while I do take advantage of higher SPL capability when I use a 125-150hrz high pass, what I am mainly doing is reducing cone motion and demands on the driver at regular loudish listening levels.

Weight of a magnet is interesting but what matters is actual motor strength to cone mass. A smaller magnet can have much more motor strength vs a poor large one. You can also have a very rigid and lightweight cone that is very well controlled by a smaller magnet and a heavy cone not well controlled by a huge magnet.
Voice coil size also doesn't really matter that much anymore either as computers have made it possible to keep very pedestrian sized voice coils cool with material choices and good venting.

The ACT woofer tested at ASR had typical HD levels by the way. Nothing special, not bad though. I would take the idea of it having substantial excursion limits with a grain of salt.
Measure your HD and if possible IMD levels. Doppler will be impossible to measure though.
I seriously doubt all the numbers you have listed will pan out. Manufacturer specs could be anything in terms of how much distortion they allow before calling things linear and maxed out.
If you are new to design you have build in your baffle effects and ideally actual measured acoustic driver response, though at 150hrz this might not be very important but the higher you go the more it is. If you are going full active including the tweeter then get ready to learn a bit about getting correct/desired responces.
All of the numbers you have are theoretical, so measure instead.

Acoustic response is what matters, nothing else really. Electrical mumbo jumbo and stock 12db this and 24db that, is not usually parallel with reality.
There is no 'optimal' design, optimal q, optimal this or that or guarantee of perfect phase or anything like that. There are many trade offs and design choices and goals.
Especially if you plan for high SPL playback as your driver will behave very differently in the same enclosure. the T/S parameters will change(well they don't change exactly but they measure differently in different conditions and the driver will see the enclosure differently and with small enclosures filled with pressure and turbulence around a high excursion motion this can be big. Make sure to model with high temperatures in the voice coil. (this is an option in your program of choice)
As well, remember a 10db increase in the midrange is equal to about 5-6db in the bass in terms of a human perceived increase in loudness. So what frequency response in the bass that is subjectively and objectively optimal at 80db might actually be very different at 100db.
Whatever the case, measure your stuff and plan for subjective assessment to override objective ideals as very little objective research has been done above 80-85db averages due to liability.
In any case enjoy the ride, should be fun.

What is the measurement system you use? REW and UMIK, or something else?

 

[Tangband]

Mike 7877:

1. Sheck the real dsp latency in your subwoofer.
5ms latency is 1,7 meters soundtravel .
You may have to delay your front speakers with a minidsp .

2. You cant cross the subs at 140 Hz even with a steep crossover if the subwoofers are near the floor . Its more like 70-80 Hz .
I take for granted you have two stereo subwoofers (?) . You can make two stands for your subwoofers, liftning them 50cm.
This way you can put the subwoofers near each main speakers midbass and you can use a 140 Hz crossover .

ROOSKIE has given you very good advices .

3. If you plan to go 3 way active - take away the passive crossover entirely . Otherwise you might end up with an end result worse than its parts .
Use a protection poly capacitor in series with the tweeter with 22 uF or more.
As ROOSKIE has written, its the acoustical slopes that matters .