Adding a pre amp

[Abtr]

1 hour ago, audiobomber said:

I don't know where you got your info about crossovers. I learned from Vance Dickason's Loudspeaker Cookbook and Phil Bamberg, the designer of my speakers. Many (most) people make this same mistake when crossing over to a sub. You have to account for the phase of the main speakers, not just the crossover filter. You have added LR4 filters, but you definitely do not have an LR4 crossover. If it was LR4, the mains and subs would be in phase.

 

Just try what I said, block the ports, high pass BW2 @ 70 Hz on the mains, LR4 @ 70 Hz low pass on the sub. You will have a phase perfect crossover. I would like to hear your about your results but if you want to continue we should move to PM, as this is off topic.

Off topic? Have you seen some of the threads here?

 

It's perfectly normal for main speakers and sub(s) to be more or less out of phase due to passive filters and their different position in the room. An active crossover can't change that unless its high and low frequency output is not in phase, but you don't want to go there.. A two-way LR4 is phase-perfect. So I adjusted the phase of the sub to get an optimal crossover and 'flat' SPL/frequency response at the listening position, and I'm happy with it.

[audiobomber]

4 hours ago, Abtr said:

Off topic? Have you seen some of the threads here?

 

It's perfectly normal for main speakers and sub(s) to be more or less out of phase due to passive filters and their different position in the room. An active crossover can't change that unless its high and low frequency output is not in phase, but you don't want to go there.. A two-way LR4 is phase-perfect. So I adjusted the phase of the sub to get an optimal crossover and 'flat' SPL/frequency response at the listening position, and I'm happy with it.

It's a perfect electronic crossover, but what you want is a perfect acoustic crossover. For that you need to take speaker phase into account, otherwise you will have what you have now, lumpy bass that chuffs along behind the main speakers. 

 

Is it difficult to change a crossover slope with your system? Is blocking the ports hard to do? What is preventing you from experimenting? 

[audiobomber]

18 hours ago, Abtr said:

It's perfectly normal for main speakers and sub(s) to be more or less out of phase due to passive filters and their different position in the room. 

As long as the drivers of the main and sub are within one wavelength and connected in proper absolute phase, they will be heard as a point source. At 70Hz, that is 16 feet or 5 meters. 

[Abtr]

17 hours ago, audiobomber said:

It's a perfect electronic crossover, but what you want is a perfect acoustic crossover. For that you need to take speaker phase into account, otherwise you will have what you have now, lumpy bass that chuffs along behind the main speakers.

No 'lumpy bass' here. Sub and main speakers are largely in phase around the crossover frequency (70Hz).

 

17 hours ago, audiobomber said:

Is it difficult to change a crossover slope with your system? Is blocking the ports hard to do? What is preventing you from experimenting? 

As I said, blocking the ports of the KEFs is no problem but I like them better with open ports. I can't change the order or slope of the crossovers (Xkits XOVER-2, see my signature). 

 

4 hours ago, audiobomber said:

As long as the drivers of the main and sub are within one wavelength and connected in proper absolute phase, they will be heard as a point source. At 70Hz, that is 16 feet or 5 meters. 

Would it make a difference if I connected the main speakers 180 degrees out of phase instead of the sub?

[audiobomber]

7 hours ago, Abtr said:

No 'lumpy bass' here. Sub and main speakers are largely in phase around the crossover frequency (70Hz).

 

As I said, blocking the ports of the KEFs is no problem but I like them better with open ports. I can't change the order or slope of the crossovers (Xkits XOVER-2, see my signature). 

You may force the subs and mains into phase at the crossover frequency with a phase control, but they do not stay in phase. The mains are rolling off at 24dB/octave, plus the fourth-order filter, i.e. 48dB/octave high pass. The sub is rolling off at 24dB/octave low pass. At times they are in phase, at times they are in opposite phase, hence lumpy bass (reinforcement and cancellation below and above the crossover frequency).

 

Would it make a difference if I connected the main speakers 180 degrees out of phase instead of the sub?

 

That would not help. Your problem is not phase at the crossover frequency. The goal is to maintain perfect phase across the 1-1/2 or 2 octaves where the mains and sub are both audible. There are only two solutions:

 

1. Raise the crossover frequency to 1.5 or 2 octaves above port resonance, and use fourth-order Linkwitz-Riley filters on mains and sub

 or 

2. Block the port and implement a second-order Butterworth filter on the mains at the -3dB point, with 4th-order Linkwitz-Riley filter on the sub.

 

I'm not surprised you preferred the LS50 with ports unblocked, because you did not implement properly. If you used your existing crossovers, the result was 6th-order high pass and 4th-order low pass. Same problem, phase variation on the different slopes as you get away from the hinge point.

[audiobomber]

On 11/8/2018 at 12:45 PM, audiobomber said:

I replaced the 6KΩ miniDSP digital sub crossover with a 100KΩ Marchand XM44 electronic crossover. Load went from 4.6KΩ to 16.7KΩ

 

Correction: I just re-read the XM44 manual. Input impedance is actually 25KΩ , so the load seen by my DAC is 11K. 

[Abtr]

7 hours ago, audiobomber said:

You may force the subs and mains into phase at the crossover frequency with a phase control, but they do not stay in phase. The mains are rolling off at 24dB/octave, plus the fourth-order filter, i.e. 48dB/octave high pass. The sub is rolling off at 24dB/octave low pass. At times they are in phase, at times they are in opposite phase, hence lumpy bass (reinforcement and cancellation below and above the crossover frequency).

 

Would it make a difference if I connected the main speakers 180 degrees out of phase instead of the sub?

 

That would not help. Your problem is not phase at the crossover frequency. The goal is to maintain perfect phase across the 1-1/2 or 2 octaves where the mains and sub are both audible. There are only two solutions:

 

1. Raise the crossover frequency to 1.5 or 2 octaves above port resonance, and use fourth-order Linkwitz-Riley filters on mains and sub

 or 

2. Block the port and implement a second-order Butterworth filter on the mains at the -3dB point, with 4th-order Linkwitz-Riley filter on the sub.

 

I'm not surprised you preferred the LS50 with ports unblocked, because you did not implement properly. If you used your existing crossovers, the result was 6th-order high pass and 4th-order low pass. Same problem, phase variation on the different slopes as you get away from the hinge point.

 

The roll-off of the KEFs is 0dB between 70 and 50Hz (0.5 octave). So I have a perfect 4th order crossover above 50Hz and the port output is 12dB down at the port resonance frequency (52Hz). Below 50Hz all I care about is a flat SPL/frequency curve which is a function of relative main/sub phase, relative main/sub position and room reflections. You cannot match phase at these low frequencies by using e.g. a BW2 and LR4 crossover.

 

You keep insisting that I have 'lumpy bass'. I have not. Bass is tight and  proportionate. My biggest concern is keeping low frequencies away from the main speakers to reduce intermodulation distortion, especially at higher SPL. In this respect obviously a 4th order crossover does a better job than a 2nd order crossover. 

[audiobomber]

You have slow bass because the sub is acoustically behind the main speakers.

 

Crossing over from a bass driver to a subwoofer is no different from crossing over from a tweeter to a midrange. All the rules for a proper crossover still apply, one of which is to cross at least 1.5 octaves from the speakers' rolloff frequency. I suggest you read The Loudspeaker Cookbook for some enlightenment.

[audiobomber]

5 hours ago, Abtr said:

You cannot match phase at these low frequencies by using e.g. a BW2 and LR4 crossover.

Home theater receivers use a 80, 100 or 120Hz BW2 rolloff for small mains. This is intended to match sealed main speakers to an LR4 subwoofer crossover.

[Abtr]

4 hours ago, audiobomber said:

You have slow bass because the sub is acoustically behind the main speakers.

What does "slow bass" mean? You said earlier: "You have to account for the phase of the main speakers, not just the crossover filter." And: "You may force the subs and mains into phase at the crossover frequency with a phase control, but they do not stay in phase." So how can any crossover to a sub account for the phase of the main speakers if main speakers and sub do not stay in phase? You do not make sense. 

 

4 hours ago, audiobomber said:

Crossing over from a bass driver to a subwoofer is no different from crossing over from a tweeter to a midrange. All the rules for a proper crossover still apply, one of which is to cross at least 1.5 octaves from the speakers' rolloff frequency. I suggest you read The Loudspeaker Cookbook for some enlightenment.

I know what you mean, but this is more critical for a tweeter to midrange crossover than for a bass to sub crossover. Human hearing is much more sensitive for problems in the mid/high spectrum than it is for bass. Anyway, as I explained, I have a perfect 4th order crossover for the main speakers and sub, apart from a narrow band of 48dB/octave main slope at -12dB. I don't think this is audible and it doesn't show in the overall frequency SPL curve at all. Even if it was marginally audible, then using a 2nd order filter for the main speakers to fix it will introduce very audible intermodulation distortion, especially at higher SPL.    

 

4 hours ago, audiobomber said:

Home theater receivers use a 80, 100 or 120Hz BW2 rolloff for small mains. This is intended to match sealed main speakers to an LR4 subwoofer crossover.

I don't know about that. The  built in crossover of my sub is 12dB/octave (for mains and sub).

[audiobomber]

21 hours ago, Abtr said:

What does "slow bass" mean? You said earlier: "You have to account for the phase of the main speakers, not just the crossover filter." And: "You may force the subs and mains into phase at the crossover frequency with a phase control, but they do not stay in phase." So how can any crossover to a sub account for the phase of the main speakers if main speakers and sub do not stay in phase? You do not make sense. 

 

Phase is simply another way to describe timing. If phase is off, the drivers do not start and stop in unison, which sounds like slow bass. The speaker with the steeper slope will be delayed in time the further it gets from the crossover point. As phase changes due to the differing slopes, the bass from the sub will at times augment, and at times nullify the bass from the main speakers.

 

A sealed loudspeaker rolls off at 12dB/octave. Let's say the system is  -3dB @ 70Hz. If you add a BW2 high pass filter, -3dB @ 70Hz, the result will be a perfect LR4 slope, -6dB  @ 70Hz. So you roll off your subs with the matching LR4 @ 70 Hz, low pass. As long as the subs and mains are within one wavelength (16 ft) of one another, they will be in perfect phase at every spot in the room. They will act as a single driver. 

 

A ported loudspeaker rolls off at 24 dB/octave, but a port is a compromise. It's a one-note resonance, and the driver is undamped below resonance. You are better off avoiding the port resonance when using a sub. You can do this by crossing over at 1.5 or 2 octaves above (which is too high with your speakers), or blocking the LS50 ports. You will not hear IM distortion, it is a non-issue. You will hear tighter, more accurate bass than what you have now.

 

The air in the room doesn't care what machinations you went through in the electronics or loudspeakers to generate a soundwave. All that matters is whether the sub and mains are trying to move the air in unison or not. A big difference between you and me is that I have tried both ways. The right way, with symmetrical slopes, is superior.

[Abtr]

22 hours ago, audiobomber said:

Phase is simply another way to describe timing. If phase is off, the drivers do not start and stop in unison, which sounds like slow bass. The speaker with the steeper slope will be delayed in time the further it gets from the crossover point. As phase changes due to the differing slopes, the bass from the sub will at times augment, and at times nullify the bass from the main speakers.

 

A sealed loudspeaker rolls off at 12dB/octave. Let's say the system is  -3dB @ 70Hz. If you add a BW2 high pass filter, -3dB @ 70Hz, the result will be a perfect LR4 slope, -6dB  @ 70Hz. So you roll off your subs with the matching LR4 @ 70 Hz, low pass. As long as the subs and mains are within one wavelength (16 ft) of one another, they will be in perfect phase at every spot in the room. They will act as a single driver. 

 

A ported loudspeaker rolls off at 24 dB/octave, but a port is a compromise. It's a one-note resonance, and the driver is undamped below resonance. You are better off avoiding the port resonance when using a sub. You can do this by crossing over at 1.5 or 2 octaves above (which is too high with your speakers), or blocking the LS50 ports. You will not hear IM distortion, it is a non-issue. You will hear tighter, more accurate bass than what you have now.

 

The air in the room doesn't care what machinations you went through in the electronics or loudspeakers to generate a soundwave. All that matters is whether the sub and mains are trying to move the air in unison or not. A big difference between you and me is that I have tried both ways. The right way, with symmetrical slopes, is superior.

 

A BW2 filter (on the main speakers) and LR4 filter (on the sub) will not be in phase. And I currently have -6dB @ 70Hz, and -18dB (not -12dB as I said earlier) @ 50Hz (the rolloff frequency and port resonance of the KEFs). At 50Hz the sub plays at -3dB. So @50Hz the main speakers play 15dB quieter than the sub, all in accord with 'symmetrical' LR4 slopes.

      

I might change the crossover frequency of my filters to 80Hz instead of 70Hz (the 4th order slopes are fixed). That would add another -5dB to my perfect LR slopes. Do you think that would be audible?

[audiobomber]

12 hours ago, Abtr said:

 

A BW2 filter (on the main speakers) and LR4 filter (on the sub) will not be in phase.

 

Yes, if the main speakers are sealed, they will be perfectly in phase, no phase adjustment required. I ran like this for years. Acoustic phase is what matters, not electrical phase. If you are sufficiently far away from the speaker's natural rolloff, acoustic phase and electrical phase are the same. If you are within 1.5 octaves of port resonance, electrical phase and acoustic phase differ.

 

I might change the crossover frequency of my filters to 80Hz instead of 70Hz (the 4th order slopes are fixed). That would add another -5dB to my perfect LR slopes. Do you think that would be audible?

 

100Hz would be better.

[Abtr]

14 hours ago, audiobomber said:

Yes, if the main speakers are sealed, they will be perfectly in phase, no phase adjustment required. I ran like this for years. Acoustic phase is what matters, not electrical phase. If you are sufficiently far away from the speaker's natural rolloff, acoustic phase and electrical phase are the same. If you are within 1.5 octaves of port resonance, electrical phase and acoustic phase differ.

 

I might change the crossover frequency of my filters to 80Hz instead of 70Hz (the 4th order slopes are fixed). That would add another -5dB to my perfect LR slopes. Do you think that would be audible?

 

100Hz would be better.

 

I ordered a number of 'crossover frequency configuration modules'  from Xkitz (basically 10 resistors with different values per module). I will try 80Hz, 90Hz and 100Hz starting with 100Hz. I'll adjust subwoofer volume and main speaker polarity to minimize anomalies in the overall frequency response. I think a 100Hz x-over will make the location of the sub audible and I worry about what a 100Hz crossover may do to the mid range of the KEFs and I remain very skeptical about the overall bass improvement that you foresee for my system, but I will try..

 

On topic, given that a good crossover for main speakers and sub is essential for good sound, an active crossover with an appropriate output buffer, fed by a DAC with (digital) volume control, will theoretically produce less distortion than a DAC > preamp > active crossover chain (note that I think passive line-level crossovers suck).

[audiobomber]

19 hours ago, Abtr said:

I think a 100Hz x-over will make the location of the sub audible and I worry about what a 100Hz crossover may do to the mid range of the KEFs and I remain very skeptical about the overall bass improvement that you foresee for my system, but I will try..

At 100Hz LR4 (-6dB@100Hz, -30dB@200Hz), you definitely will not be able to localize the sub.

 

Minus 6dB @ 100 Hz is below the midrange. The KEF's midrange should improve vs. a lower crossover, though some subs don't like to work that high, so you'll need to listen for that. Why do you have 8" and 12" subs listed in your signature?

 

Quote

an active crossover with an appropriate output buffer, fed by a DAC with (digital) volume control, will theoretically produce less distortion than a DAC > preamp > active crossover chain (note that I think passive line-level crossovers suck).

 

Completely agree.

[Abtr]

4 hours ago, audiobomber said:

At 100Hz LR4 (-6dB@100Hz, -30dB@200Hz), you definitely will not be able to localize the sub.

Some say that frequencies above 80Hz can be localized based on interaural phase difference:

https://en.wikipedia.org/wiki/Sound_localization#Evaluation_for_low_frequencies

 

4 hours ago, audiobomber said:

Minus 6dB @ 100 Hz is below the midrange. The KEF's midrange should improve vs. a lower crossover, though some subs don't like to work that high, so you'll need to listen for that. Why do you have 8" and 12" subs listed in your signature?

IME, quality bass improves perception of mid range (and even high range). Taking bass away from the KEFs may impact mid range good or bad. We'll see..

 

Oh, and I've used both subs and like both. They have the same (1000W) amp. Currently I prefer the 12" version.

[Abtr]

On 11/15/2018 at 4:34 PM, audiobomber said:

At 100Hz LR4 (-6dB@100Hz, -30dB@200Hz), you definitely will not be able to localize the sub.

 

Minus 6dB @ 100 Hz is below the midrange. The KEF's midrange should improve vs. a lower crossover, though some subs don't like to work that high, so you'll need to listen for that. Why do you have 8" and 12" subs listed in your signature?

...

 

So I tried the 100Hz crossovers. It is definitely audible. My subjective impression is a somewhat better integration of subwoofer and main speakers. The frequency response in my room still measures OK. I went back and forth a few times and I think the 100Hz configuration is a SQ improvement over the 70Hz crossovers.

 

I have to listen to it more with more and different music. Anyway thanks for the advice, the result is very interesting. One of my concerns is indeed that crossing the sub at 100Hz might be out of its comfort zone. Both the 8" and 12" sub have a fixed built-in 70Hz (12db/octave) crossover for line level (RCA) mid/high output, and both subs are speced to max 100Hz. On the other hand, the built-in variable crossover (also 12db/octave) has a maximum crossover setting of 120Hz.   

[audiobomber]

16 minutes ago, Abtr said:

One of my main concerns is indeed that crossing the sub at 100Hz might be out of its comfort zone. Both the 8" and 12" sub have a fixed built-in 70Hz (12db/octave) crossover for line level (RCA) mid/high output, and both subs are speced to max 100Hz. On the other hand, the built-in variable crossover (also 12db/octave) has a maximum crossover setting of 120Hz.   

The built-in crossovers are almost assuredly 70 Hz 12db/octave Butterworth filters, in other words -3dB @ 70 Hz.  Your sub low pass would be 24dB Linkwitz-Riley (Butterworth squared) filters, -6dB @ 100Hz, which is about -3dB @ 80 Hz, so not too far off.

 

At 200 Hz, where it matters more, the 70Hz BW2 slope is -18dB, whereas the 100Hz LR4 slope is -24dB. 

 

PS The preamp is back in my system. It has a more muscular, pacier sound through the preamp vs. the DAC outputs, which give a more ethereal sound.  

[Abtr]

22 hours ago, audiobomber said:

...

PS The preamp is back in my system. It has a more muscular, pacier sound through the preamp vs. the DAC outputs, which give a more ethereal sound.  

Ethereal (light) sound may point to an impedance mismatch, such as occurs IME with passive preamps and passive line level filters. Or possibly there's some other electrical problem.. Any decent active output (buffer) stage should be able to drive a power amp.    

[Kimo]

For all the preamp adverse posting here, I have eliminated mine.  Going straight from the DAC, and yes it sounds better without being compromised on dynamics.

 

Of course, the DAC has a really nice analog volume control.  Probably better than most preamps, so that certainly helps.  Even my vinyl rig sounds better.   

 

 

[audiobomber]

On 11/8/2018 at 3:09 PM, barrows said:

As for subs, I always prefer to drive subs from the amplifiers output, which can be done with most subs (there is always a way, even for subs without high level inputs).  Of course your external cross over needs change this equation a bit, but it probably could still be done with a little work.  Amps care not at all about driving a relatively high impedance sub input circuit.

@barrows, I had not considered using the high level sub inputs previously. I was under the impression that a low level cross was superior. Thanks to your advice, I was able to keep my subs and sell my preamp. My DAC is now connected directly to my power amp.

 

I had tried damn near everything; analog and digital xovers, running mains full range and filtered, passive line-level, single cap on mains, 1st, 2nd and 3rd-order crosses. They were all compromised in some way. Sound quality now, with my monoblocks driving the main speakers and the subs, is the best I've experienced. I'm done playing with crossovers, this sounds awesome. 

[audiobomber]

On 3/28/2019 at 10:49 AM, audiobomber said:

I had tried damn near everything; analog and digital xovers, running mains full range and filtered, passive line-level, single cap on mains, 1st, 2nd and 3rd-order crosses. They were all compromised in some way.

Minor correction, I meant that I tried 2nd, 3rd and 4th-order crossovers.