Why not make a PS Audio subwoofer crossover?


#1

This suggestion is inspired by Paul’s recent YouTube post “Why are subwoofers attacked”. I believe one of the reasons that inhibit audiophiles in implementing subs in their system is the awkwardness in incorporating them. Most 2 channel preamps don’t have sub outs so it’s up to us to figure out how to do it. That depends on the sub you are working with, which vary from sub to sub. I think the easiest way to accomplish this, apart from a preamp sub out, would to be with a dedicated all analog sub crossover in between the preamp and the amp, along the lines of the JL Audio CR-1. It’s a great unit, if a little pricey at $3k. Why doesn’t PS Audio make a cheaper, but high quality, analog sub crossover so we can all enjoy the benefits of a sub? I’d prefer to keep the subwoofer out of the speaker output chain. I would also assume there are amplification benefits to this as well since the low frequencies are not amplified. Just a thought. So what do you think Paul?


#2

Another high quality, versatile subwoofer crossover option would be nice.

I would prefer however a unit which does not come between the preamp and the amp. I do not want anything else in the signal chain potentially mucking up the signal to the amp.

A better design is to include both line-level inputs from the preamp (if you do not have a sub out, use a Y-connector) and, even better, speaker-level inputs which take their signal from the amplifier.

Connecting a sub is simplicity itself. Depending on the sub, either use a Y-connector from the preamp, or take the signal from the amplifier. The later is the better approach; it better integrates with the system as the sub crossover is getting the same signal as the speakers.

The hard part is setting up the sub. Physical placement is tricky, and setting phase, level, crossover settings, etc. is difficult. Making it even tougher is that placement and settings depend so much on each other.

But audio is a hobby; it is supposed to waste your time.


#3

Elk - not sure, contrary to Paul’s love of REL (not to suggest you’re always in agreement with Paul). I bought a well reviewed REL years ago, and it did not integrate or engage with my system (to me) in the same way as the JL’s. Sent it back. To each his own. I’ve heard them sound fine in other systems.

How does a sub wired through either a Y cable (!) or speaker-level inputs NOT “muck up the signal to the amp” vs. separate line outs from the pre, or to/from a dedicated external crossover (both of which, it would seem to me, present the amp and pre with nominal signals?) Is the assumption that load/feedback through speaker cables or Y cables is transparent/non-existent?

I’m off for a bit, so if I don’t respond until later, don’t read into it. ; )


#4

+1 Beef. I don’t see how either way wouldn’t affect the signal, but I know far, far less than Elk!

One more thought. Why don’t preamps ever have a sub out with a crossover dial? If there are 2 channel analog pre’s that do this, let me know. Seems like a reasonable thing to include.


#5

It seems reasonable to us here, and more nowadays, but has historically been the province of HT amps. Even now, it’s probably low on most manufacturer’s lists. I was interested to find out there was a “sub” out on the little NAD D3020. No crossover frequency control, and it’s a mini jack, but hey.

I accidentally happen to have had pres, HT amps and an external processor that all have 4 or more outs, either full range or crossed over, so I was not limited to subs that only did it via the amp’s speaker outputs. REL naturally says that’s best, and that makes a lot of folks happy. I do run my deck speakers and sub that way though. Whatever works for you and your gear.


#6
badbeef said

Elk - not sure, contrary to Paul’s love of REL . . . it did not integrate or engage with my system (to me) in the same way as the JL’s. Sent it back.

I do not understand what this relates to. I never mentioned REL.

I wrote it is easy to make the physical connections a sub requires, but finding the best physical location for the sub and the settings correct can be tricky.

How does a sub wired through either a Y cable (!) or speaker-level inputs NOT "muck up the signal to the amp" vs. separate line outs from the pre . . . ?
What do you think happens inside a pre to create a duplicate set of outputs? :)

Running a parallel output to the sub as well as the amp is an easy load for the preamp; the sub’s inputs are very high impedance, the preamp does not know they are even there. They just share the signal and the sub merely sees the voltage from the preamp.

This is the same for the speaker level inputs found on subs; they are very high impedance. Wiring the subs to the amp does not impact the signal to the amp sent to your speakers. The advantage, of course, is that the sub now sees exactly the same signal as your speakers, helping integration.

On the other hand, inserting an additional component in serial between the preamp and amp is inherently going to impact the signal - unless it is literally a straight bit of wire between the amp and preamp - which may essentially be the case with external sub crossovers. I am suspicious however. They often process the signal from the pre to the amp, such as by inserting a high-pass filter, with its concurrent phase issues, etc. resulting in loss of clarity.


#7

Thanks for the explanation Elk! Very well said. One thing I’m struggling with is how the crossover point is handled for the speakers. Clearly the sub will roll off with a low pass filter, usually at 24 dB/octave at some crossover frequency chosen by the user. In most HT setups, this goes hand in hand with a concomitant 24 dB/octave high pass filter at the same frequency, i.e. Linkwitz–Riley filter, in order to have a flat frequency response at the crossover point. In what you are describing, I can’t see how there is a high pass filter to mesh with the sub at any frequency other than at the natural lower frequency range of the speaker. And this is probably not 24 dB/octave high pass roll on. If this is true, this isn’t very flexible as you are restricted to the one frequency. Plus, if the crossover is chosen higher than the lower frequency range of the speaker, you would get a significant “bump” where there is summation of signals. Correct me if I’m wrong.


#8
Elk said On the other hand, inserting an additional component in serial between the preamp and amp is inherently going to impact the signal - unless it is literally a straight bit of wire between the amp and preamp - which may essentially be the case with external sub crossovers. I am suspicious however. They often process the signal from the pre to the amp, such as by inserting a high-pass filter, with its concurrent phase issues, etc. resulting in loss of clarity.
I'm sure what you are saying is true, but my Anthem AVM 60 has no such loss of clarity or phase issues at the sub crossover point. My system is both time and phase aligned at the crossover and has a perfectly flat frequency response below 200 Hz. I doubt I could get that in my room using either technique you describe. BTW, I use a miniDSP calibrated microphone in conjunction with Room EQ Wizard to calibrate and monitor my system. It works well, but YMMV.

#9
amgradmd said Correct me if I'm wrong.
You are correct, but there is more going on. For example, one physically places the sub where it does not add energy where the system has a peak (and ideally the sub helps cancel out the peak), etc. Then one addresses phase and level to maximize this integration.

This is when the sub’s crossover is limited. More sophisticated subs allow one to change crossover frequency, slope, phase, level, etc. Really sophisticated crossovers contain multiple-band parametric equalizers with variable center frequency, Q, and level, and more. They also have calibrated mics with a real-time visual graph readout to help setting them up. Even then it is work.


#10

Understood. It still seems like integrating a sub in this manner is difficult as best or even an impossible task, given the room. In my room, which may be tricker than most - hard to say, I use dual JL Fathoms, one of the most sophisticated subs around. And getting them dialed in, even with DSP, was very tough. For example, I must have experimented with at least 20 different locations of the subs and speakers to minimize bass modes and optimize imaging. Once I found the (relative) best location, I equalized the subs to the sub’s location to the room using the subs DSP with their own calibrated mic. THEN, I initiated the Anthem ARC (Anthem Room Correction) with it’s own Anthem calibrated mic, and chose a crossover. And then I listened. I chose another crossover and listened again. Rinse repeat. Finally, I phase and time aligned the speakers using Room EQ Wizard (yet another calibrated mic!), another dicey process. My point is that at least with this complicated process, I can control the process and optimize the process to a reasonable final result with excellent tools. I shudder to think how difficult it would be to get a decent result without it. Again, my room might be trickier than most!

So even staying in the analog domain and eschewing preamp DSP, you should be able to get a fairly seamless integration with a dedicated crossover unit along the lines of the JL CR-1 with proper sub and speaker placement. More easily than without it, I’d guess. And this is one of those times where the benefits of the significantly better sub integration might outweigh the possible detriment of adding something in the signal path. Again, this is based on my suspicions and some experience mucking about. Maybe you, and others, can really get a sub to sing with your method. I don’t know. What do you think, @adminpaul?


#11

These are precisely the issues.

Keep in mind that integrating a sub into a system so that the sub does not jump out as an add-on is very different than equalizing and otherwise treating the signal to smooth bass response. All systems and rooms have bass issues, whether a sub is used or not.

amgradmd said I must have experimented with at least 20 different locations of the subs and speakers to minimize bass modes and optimize imaging.
This is indeed part of the work it takes to get a sub to really shine. Bass is the holy grail of audio.
. . . the benefits of the significantly better sub integration might outweigh the possible detriment of adding something in the signal path.
It certainly may, depending on the system and one's preferences. I still find the best method is to keep the sub crossover, etc. out of the signal path to the main speakers. This way you get all of the benefits of sophisticated subwoofer control without compromising the rest of the sound.

#12

Elk - "I do not understand what this relates to. I never mentioned REL. "

It was about using speaker-level wiring through/to a sub (which AFAIK is how RELs all work) vs. dedicated multiple line outs on a pre to line ins on both mains and subs (as with JLs). We all like what we like. Not a big deal.

“What do you think happens inside of a pre to make a duplicate set of outputs? :)”

I suspect I know less than you about the subject, but I think it’s a short, hardwire, soldered jumper a couple of inches long - vs. separate externally connected Y cables which add two connections. It just seems to me that if you are opposed to putting a pre or a processor in between your source and amp, you would also be opposed to doubling up cables to a single output, or running additional speaker wires to or through subs, or, for that matter, doing DSP correction. But again, to each his own.

I noodled fairly extensively with various forms of DSP, and except in a HT system, (where it’s very convenient to move the speakers around in time, and I’m not as picky about the impact on the sound) I ended up taking it out. With respect to the time delay aspect of DSP, if your subs are not physically closer to the listening seat than the mains, in order to truly time align them, you are delaying the mains speakers. I don’t want my mains run through DSP. Of course introducing EQ is also altering the timing of the signals as well by altering the phase.

I’m currently only using the DEQX as a “pre” or gain stage. It sounds better with it, than directly wiring the Jr. to the amp - despite adding a box and cables, and allows attaching the phono pre and TV directly (shorter paths/less conversions) rather than through the DAC, etc. But ask me next year - I’ll probably be doing something else ; )

DSP devices can certainly do the things they are intended for, to varying degrees, but I ended up feeling they were also doing things I ultimately couldn’t live with, since taking them out made the sound better. My guess is that they are introducing bad phase and micro-timing stuff, sorta the opposite of what makes the DMP help things sound good.


#13

I am a fan of feeding the subwoofer from the output of the power amplifier. I’ve advocated this for years. There’s also a misconception about this connection method floating around, one that suggests it can somehow tax the power amplifier or somehow change the sound of the main system, which it cannot. Let’s dispel that one quickly.

Since the object of subwoofer integration is to not hear the subwoofer, using the amp to feed the sub’s internal crossover and amp is the best way to achieve that. Whatever sonic characteristic the amplifier displays the sub will mimic it.

REL makes some sort of wireless device they claim is transparent and if it were any other company I might raise a suspicious eyebrow. But John Hunter is one of the best listeners in the business and if John says it’s transparent then I believe him.

I would love to see people take subs more seriously.


#14

Paul - I guess I’ll have to take your word for it, though you don’t mention why it can’t affect it. Do you run separate runs of speaker cable directly from the amp to the RELs (i.e. You “Y” it at the amp, or does the speaker cable go to the subs, then loop to the speakers?


#15
Paul McGowan said

Since the object of subwoofer integration is to not hear the subwoofer, using the amp to feed the sub’s internal crossover and amp is the best way to achieve that. Whatever sonic characteristic the amplifier displays the sub will mimic it.

REL makes some sort of wireless device they claim is transparent and if it were any other company I might raise a suspicious eyebrow. But John Hunter is one of the best listeners in the business and if John says it’s transparent then I believe him.

I would love to see people take subs more seriously.

Thanks, Paul, for your imput. You didn't address my concerns that by using this method for integrating the subs, you can't really tailor the crossover point and/or eliminate the dB hump from summation of signals due to lack of utilizing a Linkwitz–Riley filter at the crossover. I would assume that some subs can use custom cutoff slopes to best integrate with the cutoff of the speakers? If so, I haven't seen this with the JL subs. Their calibration is independent of the speakers so how would they know what the speakers limits are? Sorry, but I'm still a little confused on this.

That wireless REL unit looks interesting. Looking at the back, it only has low level and LFE inputs so in 2 channel you’d have to use a Y-splitter, I assume.

https://rel.net/shop/accessories/longbow/


#16

Beef wants to know why feeding the subwoofer’s inputs from the main power amplifier outputs won’t affect the sound of the main amp and the answer’s easy enough: the main amp’s outputs are designed to drive low impedance loads, often as low as 2Ω, while the high-level subwoofer amplifier input’s quite high, usually around 50kΩ. That tiny load on the output of an amplifier is likely small enough to make it impossible to measure, let alone hear. I’ve made the experiment.

Carrying forward the audio qualities of the power amp into the sub has great value for system integration.

To answer amgradmd’s question, yes, a more complex filtering setup might make for a better match between the high-pass function of the loudspeaker and the low-pass function of the subwoofer, but without actual DSP you’ll likely never make it perfect. The high-pass response of a speaker playing in the room is complex, to say the least. Not only do you have the speaker’s crossover to contend with, there’s the box roll off and then how it couples into the room.

I believe you’re over thinking the problem. In my experience, a reasonably simple low pass function is plenty adequate for mating as long as you keep the subwoofer’s highest frequency low enough.


#17

It is precisely the complexity you are talking about that makes me think about this but, like you said Paul, maybe I’m over thinking it. (wouldn’t be the first time!) Maybe my OCD about this comes from looking at the lengths guys go to on AVSForums to dial their subs in in the HT world. Take a look at some of those threads and it’s positively crazy-town! Dudes with 8 subs and all kinds of bass traps and processors. No thanks. If I can’t do it with two subs, I’m not interested!


#18
badbeef said

Elk - "I do not understand what this relates to. I never mentioned REL. "

It was about using speaker-level wiring through/to a sub (which AFAIK is how RELs all work) vs. dedicated multiple line outs on a pre to line ins on both mains and subs (as with JLs).


Got it. I prefer using a wire connected to the amplifier to power the sub. That is, the sub wire is not connected to the amp, but tp the amplifier output along with the speaker. It is like bi-wiring, but the sub is in a second cabinet.
"What do you think happens inside of a pre to make a duplicate set of outputs? :)"

I suspect I know less than you about the subject, but I think it’s a short, hardwire, soldered jumper a couple of inches long - vs. separate externally connected Y cables which add two connections.


Yes, what happens in the pre is electrically equivalent to a Y cable. If the added connection is of concern, one can make a soldered Y cable, but I doubt there is loss from a good quality Y cable.
It just seems to me that if you are opposed to putting a pre or a processor in between your source and amp, you would also be opposed to doubling up cables to a single output, or running additional speaker wires to or through subs, or, for that matter, doing DSP correction.
Again, my preference is running a single set of extra cable from the amplifier's output to the sub. This in no way impacts the signal the main speakers receive.