Bi-Amping Discussion

[audiobomber]

In the comments section of the MartinLogan Motion 40i review, biamping was raised. I am starting this new thread for an in-depth discussion of the topic. I find it endlessly fascinating and far more complex than some would have you believe. The notion that ripping out the passive crossover in your loudspeakers and connecting an active crossover will bring instant nirvana is simply bogus. My moniker on Audio Asylum is Audio Fan. I co-wrote this AA FAQ with another member in 2000:
https://www.audioasylum.com/messages/general/57210/biamplification-basics-an-article 
I am happy to say that I still agree with virtually everything in the FAQ.

Vertical Biamping
"The purpose of vertical biamping is to reduce the current demand on an amp. Because the main current draw is in the low frequencies, each power amp in a vertical setup only drives one woofer. The channel driving the woofer gets to hog most of the current provided by the amp's power supply, since the channel driving the high frequencies doesn't need as much current. This is particularly advantageous if the amp uses a common power supply for both channels. Vertical biamping in this manner requires identical amps for each channel to maintain balance integrity."

 

I have a pair of 50W stereo amps. Driving my 4-ohm 89dB/W/m speakers with a single stereo amp sounds OK. Driving the speakers biwired to a single channel of both amps sounds much better, more detailed, more dynamic, tighter bass. There is a small further improvement when I drive the speakers in vertical biamp mode. I no longer use the 50W stereo amps in my main system because the Meitner 100W monoblocks sound best, so clearly biamping is no panacea.

 

Horizontal Biamping
"In horizontal biamping, one stereo amp is connected to the low-frequency speaker posts on both the left and right channel, and a second stereo amp is connected to the high-frequency posts of the left and right speakers. This arrangement is used for biamping with dissimilar amps.

The purpose of this arrangement can be to maximize the virtues of different amplifiers. If amp X has superb bass and amp Y is a bit soft on bass but has a glorious mid and top then amp X is used for low end reproduction and amp Y for the upper end of the spectrum."

 

I tried horizontal biamping with a 50W Meitner in combination with various other SS amps (50W NAD, 70W Classe, 100W Bryston). While I could easily achieve better bass and treble with these combos, none were as satisfactory as a single stereo amp powering both channels. In all cases, the horizontal configuration resulted in a loss of coherence. The sound was more hi-fi, less musical and quite disappointing. I did not experiment with tube amps, but I find it very hard to believe that a SS/tube amp combo would make me happy, due to the different sound signatures. I don't doubt that some combination of gear and sonic preferences may arrive at a different conclusion on the merits of horizontal biamping. 

 

I have been experimenting with active and passive systems for twenty years. I currently own a Marchand XM44 three-way active analog crossover, a two-way Marchand XM46 passive line level crossover, a miniDSP 2x4 digital crossover and several pairs of FMOD inline crossovers. Even though I have these crossovers, I am running my (sealed) monitors full range with the subs connected in parallel to the main speakers, via the miniDSP (BW2@63Hz Low Pass, with Linkwitz Transform and correction for the main room node (-12dB@42Hz, Q=5)).

 

The main reasons I have settled on the above configuration are listed in the AA FAQ:

1. Quality speakers are designed with a crossover network matched to the particular characteristics of the drivers. Substituting an active crossover defeats the speaker designer's efforts to truly match the drivers, cabinet, and crossover together.

2. Anything that is added to the signal path alters it in some way. Passive components do too (a cable is a passive component, for example), but they may be less likely to add coloration than active components.

3. An active crossover performs a sophisticated function, roughly similar in complexity to a phono stage. An active crossover has the potential to degrade the sound of a system if its quality does not match the rest of the components.

Ultimately arguments for or against either configuration are academic - the proof is in the listening. 
As with all links in the chain, the quality of the active crossover influences the sonic result. A poorly designed active crossover might not yield better results than the internal high quality passive crossovers in a well designed loudspeaker with components carefully matching the drivers.

 

The reason not listed above is that I value simplicity these days, I don't want to go back to tri-amping, though I admit the miniDSP SHD Series processor/streamers are a temptation, LOL!

 

All comments are welcome. 

[Ralf11]

if an engineer rips out the passive crossover in the loudspeakers they make and adding an amp for each driver things, are likely to improve...

 

next level down would be a consumer ripping out (bypassing) the passive crossover in their loudspeakers that are designed for bi-amping...

 

otherwise, not a fan

 

I used to use a 5 channel amp to drive my old maggies, bypassing the Xover in each one - but my new ones have a better Xover, so I sold my old Sunfire amp

[semente]

What's wrong with the KISS principle?

[audiobomber]

17 hours ago, Ralf11 said:

if an engineer rips out the passive crossover in the loudspeakers they make and adding an amp for each driver things, are likely to improve...

 

next level down would be a consumer ripping out (bypassing) the passive crossover in their loudspeakers that are designed for bi-amping...

 

otherwise, not a fan

 

I used to use a 5 channel amp to drive my old maggies, bypassing the Xover in each one - but my new ones have a better Xover, so I sold my old Sunfire amp

Maybe an acoustic engineer. Just being an electrical or mechanical engineer doesn't give you enough knowledge to design a great loudspeaker without specific training. There are plenty of DIY'ers who know more about designing speakers than most engineers will ever know. 

 

Well designed passive crossovers use techniques that most people are unaware of. They think you can just replace a passive crossover with active fourth order slopes at some chosen frequency and that's going to be an improvement. In fact it will likely sound like crap, because of improper driver timing and no baffle step correction. 

[audiobomber]

3 hours ago, semente said:

What's wrong with the KISS principle?

Do you mean why not just stick with a passive loudspeaker and one amp? Yeah, definitely the best option for most people. Experimenting is fun though, a way to get more personalized sound quality, and can provide superior performance if done well. The downside; it can be tortuous.

[photonman]

What about the "dual-mono" single chassis amps out there, do they qualify as biamping too?  Specifically if you have an amp with a seperate power supply and amp board for each channel but only one common mains connection, do you get the same benefits of bi-amping?  Seems you would stay within KISS parameters.

[semente]

1 hour ago, photonman said:

What about the "dual-mono" single chassis amps out there, do they qualify as biamping too?  Specifically if you have an amp with a seperate power supply and amp board for each channel but only one common mains connection, do you get the same benefits of bi-amping?  Seems you would stay within KISS parameters.

 

That's identical to what I am using but mine's even because simpler because it does without the preamp, an integrated amplifier with one single input. (it's not a pure dual-mono because both channels have a separate PSU but share the ground).

[Ralf11]

2 hours ago, photonman said:

What about the "dual-mono" single chassis amps out there, do they qualify as biamping too?  Specifically if you have an amp with a seperate power supply and amp board for each channel but only one common mains connection, do you get the same benefits of bi-amping?  Seems you would stay within KISS parameters.

 

one driver; one amp

 

dula mono could work if you put 2 of them in a stereo system

[Ralf11]

4 hours ago, audiobomber said:

Maybe an acoustic engineer. Just being an electrical or mechanical engineer doesn't give you enough knowledge to design a great loudspeaker without specific training. There are plenty of DIY'ers who know more about designing speakers than most engineers will ever know. 

 

Well designed passive crossovers use techniques that most people are unaware of. They think you can just replace a passive crossover with active fourth order slopes at some chosen frequency and that's going to be an improvement. In fact it will likely sound like crap, because of improper driver timing and no baffle step correction. 

 

did you miss this: "the loudspeakers they make"

[audiobomber]

32 minutes ago, Ralf11 said:

 

did you miss this: "the loudspeakers they make"

Sorry, I did indeed.

[audiobomber]

3 hours ago, photonman said:

What about the "dual-mono" single chassis amps out there, do they qualify as biamping too?  Specifically if you have an amp with a seperate power supply and amp board for each channel but only one common mains connection, do you get the same benefits of bi-amping?  Seems you would stay within KISS parameters.

The amp you describe is a stereo amp. You need two stereo amps for biamping. 

[Kal Rubinson]

1 hour ago, audiobomber said:

You need two stereo amps for biamping.

You need two stereo amps for biamping in stereo.  What he was asking about was something else.

[Abtr]

11 hours ago, audiobomber said:

...

Well designed passive crossovers use techniques that most people are unaware of. They think you can just replace a passive crossover with active fourth order slopes at some chosen frequency and that's going to be an improvement. In fact it will likely sound like crap, because of improper driver timing and no baffle step correction. 

What do you mean by "improper driver timing"? Phase errors or impulse-response errors or both?

[Speedskater]

12 hours ago, Abtr said:

What do you mean by "improper driver timing"? Phase errors or impulse-response errors or both?

There is a lot going on in a passive crossover/speaker combination that doesn't show-up in the simplified schematic. You can't just pop frequency & slope into an active crossover and get the same transfer function.

[audiobomber]

13 hours ago, Abtr said:

What do you mean by "improper driver timing"? Phase errors or impulse-response errors or both?

I was thinking primarily of phase. For example, where the woofer drops by 6dB/octave due to its internal inductance, so a 2nd-order Butterworth is used with 3rd order on the tweeter for a final BW3 transition.  

[audiobomber]

18 hours ago, Kal Rubinson said:

You need two stereo amps for biamping in stereo.  What he was asking about was something else.

I don't understand what you mean. The Bryston 3B-ST for example is a dual mono amp, it has separate power supplies for each channel, but both channels are amplifying bass and treble frequencies. Using a single 3B-ST for stereo does not fit the definition of biamping, where bass and treble are handled by separate channels.

 

I have friend who uses a pair of 3B-ST's in vertical biamp mode with B&W N802 speakers. Biamping is notably better than a single 3B in his system due to double the current capacity and lower IM.

[Abtr]

1 hour ago, audiobomber said:

I was thinking primarily of phase. For example, where the woofer drops by 6dB/octave due to its internal inductance, so a 2nd-order Butterworth is used with 3rd order on the tweeter for a final BW3 transition.  

But with a BW3 crossover HF and LF output will be 270 degrees out of phase at the crossover frequency and it introduces a group delay for the LF output. One could avoid/correct that with a digital crossover and a FIR filter. 

[audiobomber]

36 minutes ago, Abtr said:

But with a BW3 crossover HF and LF output will be 270 degrees out of phase at the crossover frequency and it introduces a group delay for the LF output. One could avoid/correct that with a digital crossover and a FIR filter. 

Yes, digital crossovers have an advantage there. But the point I originally made is that people who think they can simply use active LR4 crossovers and done are highly unlikely to achieve a good result.

[Kal Rubinson]

3 hours ago, audiobomber said:

Using a single 3B-ST for stereo does not fit the definition of biamping, where bass and treble are handled by separate channels.

That's why you need two.

[Abtr]

10 hours ago, audiobomber said:

Yes, digital crossovers have an advantage there. But the point I originally made is that people who think they can simply use active LR4 crossovers and done are highly unlikely to achieve a good result.

Possibly. I successfully use an active LR4 for crossing my main speakers and sub, but I never tried bi-amping a tweeter and a mid/bass driver. I'd probably use something like a miniDSP for that.

[audiobomber]

2 hours ago, Abtr said:

Possibly. I successfully use an active LR4 for crossing my main speakers and sub, but I never tried bi-amping a tweeter and a mid/bass driver. I'd probably use something like a miniDSP for that.

I have one, but there's no way I'm going to run my main speakers through a miniDSP, the quality is just not sufficient. Their SHD Studio looks interesting though.

[Confused]

Devialet have developed an active crossover system, Devialet AXD, as they call it.  I spoke to Devialet's chief designer at dealer event a while back, the way he explained it was that the speaker manufacturer could specify the perfect design for a given speaker's crossover and Devialet could create this virtually within AXD.  The point being that it is not possible to create the perfect crossover using passive electronics, there will always be compromises in the design, but the "virtual" AXD crossover could indeed be perfect.  Devialet have demonstrated this system, but I have not seen much evidence of the system gaining much commercial success, or for it being offered for a significant range of speakers.  It is an interesting concept, although being cynical you could argue that it is a way for Devialet to sell customers far more amplifiers than they really need.  More details per the links below.

 

https://www.devialet.com/media/picture/image/w/h/whitepaper_axd_2018_en.pdf

 

http://designwsound.com/dwsblog/2018/10/devialet-axd/

 

https://hifipig.com/devialet-add-axd-tech-to-their-expert-pro-line-of-amplifiers/

[Kal Rubinson]

3 hours ago, Confused said:

Devialet have developed an active crossover system, Devialet AXD, as they call it.  I spoke to Devialet's chief designer at dealer event a while back, the way he explained it was that the speaker manufacturer could specify the perfect design for a given speaker's crossover and Devialet could create this virtually within AXD.  The point being that it is not possible to create the perfect crossover using passive electronics, there will always be compromises in the design, but the "virtual" AXD crossover could indeed be perfect.  Devialet have demonstrated this system, but I have not seen much evidence of the system gaining much commercial success, or for it being offered for a significant range of speakers.  It is an interesting concept, although being cynical you could argue that it is a way for Devialet to sell customers far more amplifiers than they really need. 

To accomplish this, Devialet needs the cooperation of the speaker designer and, as others have found before them, this is not likely to happen.  It is not in their interest to cooperate. 

[firedog]

Much of this discussion is a good argument for internally powered speakers where each driver has it's own amp, and crossovers are digital and designed for the individual drivers. 

[Abtr]

13 hours ago, audiobomber said:

I have one, but there's no way I'm going to run my main speakers through a miniDSP, the quality is just not sufficient. Their SHD Studio looks interesting though.

MiniDSP SHD Studio indeed seems to be the way to go. You will be stuck with the miniDSP's volume control though, which may or may not be a problem.

 

I suspect that the insufficient sound quality of the miniDSP 2x4 (HD) is mainly due to a lacking  implementation of the analogue output circuitry (power supply/voltage regulator/output stage). When this is done well, it could be used in principle as a quality preamp.

 

For example, I replaced the LM317 based voltage regulator circuit of my active Xkitz LR4 analogue crossovers with an LT3045 based circuit and the SQ improvement is remarkable. Best preamp/buffer I had sofar for my Schiit Vidar stereo power amp and KEF LS50 main speakers + active sub! And it also performs very good with a pair of JBL LSR305 active (internally bi-amped) monitors which rival the passive KEFs for a fraction of the price. The bass response with the JBLs may even be superior. The JBLs have the advantage of an ADC > DSP > DAC induced group delay of about 10ms which is approximately equal to the analogue LR4 induced low pass delay.  So impulse response of the JBLs and subwoofer are more in line and I think this is audible..