Bi-Amping, practical dilemma - Same amp (XPA-5) or separate

[ronkuper]

Hi Guys,

So I've decided to check out passive* bi-amping for my next geek-out-audio mini project and wish to test some questions and assumptions with you  

Relevant existing gear: 
XPA5 (Gen 2)
Quadral M50 Platinum (Bi-wiring ready, specs below ***) 

Current decisions and their supporting assumptions (welcome to object to any of them): 

Decision: Passive bi-amping - Not going to bypass the speakers' passive crossovers 
Supporting assumptions: 

• I don't have the skills to design an active crossover network that would surpass the manufacturer designed passive xo   
• Passive bi-amping does offer advantages - 1. By dedicating an amp channel each for HF+MF and LF (reducing stress on the HF+MF amp)
                                                                                 2. By separating the electrical input (removing the external jumper) of the HF+MF XOs from the LF XOs and so preventing back current EMF from the LF woofer drivers to interfere with the HF+MF drivers
• It is rather straightforward to try in my setup 

The Dilemmas:

1. Should I configure an active (digital) low/high pass filters before the respective amp channels or give both the full signal? 
2. By giving both amps the full signal - Will it still benefit the HF+MF amp channel (less resistance from drivers?) or will it work just as hard on amplifying the full signal regardless of the drivers / passive XOs it is connected to? 
3. With 'full signal' to both - Does it make sense to use a shared PSU amp such as the XPA-5 or will it make it draw twice as much power (4 Amps channels getting the full fronts signal) which might degrade performance overall? 
4. Other drawbacks of using a shared PSU multi-channel amp vs. dedicated amps with regards to passive bi-amping?
5. If using dedicated amps, should they all be the same model? Could there be amp level differences issues otherwise**? 
6. If using an active low/high pass filters before the amps - Should I worry about phase issues (from the active XOs) with the High-Mid section and the LF section of my speakers?  

Anything else to consider? 

Eager to hear what you guys think on these topics. 


Thanks,
Ron



*Sources for passive bi-amping assumptions (out of many on the web) - 

• www.padrick.net/LiveSound/Biwiring/Biwiring.htm
• www.avforums.com/threads/bi-amping-do-i-need-to-disconnect-the-passive-crossover-in-my-speakers.683245/   (Including an interesting take from an engineer at Bowers and Wilkins) 
  

** Practical dilemma behind this question (5 - same model amps): Currently the XPA5 handles fronts, center and surrounds. Options are:
     1. Get another 2 channel amp and use it to bi-amp the fronts with the XPA5 (One taking the mids+highs and one taking the lows)
     2. Get another 2 channel amp the use it to power the surrounds and then use the freed up channels on the XPA5 to bi-amp the fronts. 

*** Specs of the fronts from the manufacturer website:
Nominal/music power (W) 200/300 W
Frequency response (Hz) 23...50.000 Hz
Crossover frequency (Hz) 280/2300 Hz
Efficiency (dB/1W/1m) 90 dB
Impedance (Ohm) 4...8

[Soup]

FWIW, I would choose option 2 since the surrounds would be used least..................

[leonski]

First assumption under passive is wrong. Your amps will continue to be asked to amplify the full bandwidth. Thus negating one of the 'advantages' of what you call 'passive' biamping.

You also underestimate the advantages of a line-level solution. At that point the speakers crossover is elminated from the mix and an amp will be free to 'damp' the speaker as you'd hope. Depending on the crossover, your damping factor might be
a fraction of what you think it 'is'. Best csee without crossover between amp and the driver results in much higher DF. The LF to Mid/Tweet crossover of 280hz is close enough to the 50:50 point that I woulnd't worry.

You also need some more information. Crossover frequency would be helpful. The 50:50 power point for music is around 350hz. This is where a pair of identical amps would ordinarily shine.
90db SENSITIVE speakers (that's NOT a measure of 'efficiency') are fairly good at turning electricity into sound.

Me? I'd go ACTIVE using MiniDSP. But first? Get more information on the existing crossover. Write the manufacturer. Get a schematic or even a good verbal description.

And read THIS, not the stuff from thos other guys. This guy doesn't have a dog in the fight and so tells you what you need to know, not what they'd LIKE you to know.

sound.whsites.net/bi-amp.htm

This is part I of 2 or 3, but covers most of the essentials fairly easily.

[ronkuper]

Hi Leonski,

Thanks for replying!

First - I didn't mean for an Active vs. Passive debate, I understand Active is superior.

I also realize that it is not practical for me to achieve at this time.

On the other side - trying out passive biamping is very practical for me with my current setup, knowledge, budget and available time.
I really wanted to hear you out and discuss the assumptions and dilemmas that I had in order to get the most (if ever so slight) improvement from this relatively small setup change without mistakenly doing something that would potentially degrade the performance.

I got to skim through Rod Elliott's article you linked several times over the years, just to reach the conclusion that I'm not there yet... What changed is that I have lately read new data from various sources which suggest that passive bi-amping (which is straightforward for me to achieve) does have some benefits on its own (not compared to the real deal but still beneficial nonetheless).

1. Are you certain the HF+MF amp channel would work just as hard getting the full signal but connected only to a high-pass passive filter/mids/tweets?  Also couldn't an active XO (digital in JRiver/Roon/etc) help for the amp side of things (see dilemmas 1 and 6 in the original post)?
2. Are you completely dissing the notion of benefiting from removing back current / EMF?


From source1:

The input impedance of the highs section increases as the frequency decreases, so the highs section draws less power from the amp as the frequency decreases. The amp that is driving the highs is now doing a lot less work than when that amp was also driving the lows section of the loudspeaker, where most of the power is used.

From source2:

So when you have a loudspeaker which has a coil of wire moving in a magnetic field pulsing to a signal, that same speaker is acting in reverse (like a microphone) producing its own back current or EMF, this EMF can travel back up through the short 2" jumper links on the terminal tray at the back of the loudspeaker into the mid/high frequency crossover and induce very low level distortion that can blemish or mask the fine details evident in midrange & treble content.

The higher the volume, the greater distance the bass drivers voice coil moves and so the greater the EMF is produced, and the more bass drive units you have, even more back EMF produced.

[leonski]

Just off the top::
The crossover in the speaker introduces additional resistance which makes an amps damping factor of 500 or 1000 or a Million or whatever, simply academic. By the time you add all the additional resistance of crossover and such, you might be down
to a DF in the 20 to 50 range. Which is probably fine. Being RID of the crossover lets the speaker and amp connect without all that additional BS. Damping factor should go up quite a bit.
If you are curious about BACK-EMF, I will suggest a Very Simple test you can perform in just a few minutes with just a paperclip.

And yes, by all means experiment with passive biamp. To Tell The Truth, that's the system I'm using. I have a pair of Parasound A23 which are roughly 1/2 the power of the A21. I would go 'active' and gain about (max possible?) 3db which would give me more 'Apparent' power than a single A21. My Panels cross @600hz so I'm not quite
at the 50:50 point BUT make up for it in part by restricting the LF to the main speakers and letting the SUB do the heavy lifting.

There is ONE other possible solution. A PASSIVE Line Level XO. These are small, easily made, but require some calculation. I noodled thru one for my panels and it could be made in and Altoids Tin. Easily. Capacitors and such are EXTREMELY small values. Downside? Won't drive long interconnects. Difficult to change values once made.

I'd go ahead with the passive. Cheap and fairly easily done. Than you'll KNOW. I don't think there is much DOWNSide. Keep everything connected correctly and in-order. Watch power-up sequence when testing and keep levels down. Ultimately, since you have 3-way speakers, you may put in a 2-way ACTIVE crossover and leave the mid/tweet crossover
in the speaker connected. And what I noted in MY setup was roughly the same maximum loudness potential as my FORMER 'D' amp setup which was a PSAudio integrated with B&O modules.

I think you're doing fine and reading before acting is always a decent idea. Above All? have Fun.

[ronkuper]

Thanks!

Would you be willing to expand a bit about the 50:50 notion? 

[leonski]

The general idea is music takes energy to reproduce, probably in proportion to that which it takes to make it.

Bass drums take a lot of energy to reproduce while a flute or piccolo is ALL very high frequency and doesn't take that much energy to reproduce. In your speakers? It is possible to FRY a tweeter with a blast of ultrasonic or simply turning it up too loud. Your tweeter might
have the same SENSITIVITY as the other drivers, but far Lower Power Handling. Don't conflate the 2.

For 'normal' music, studies show the approximate distribution of power required reaches 50:50 at about 350hz. People use this kind of information when designing an active speaker system where you might have 50 watts to the woofer, 35 to the midrange and 15 to the tweeter.
The not-so-well known Braun Triamp was configured this way. But I don't know the crossover freuqencies.
For the guy biamping his speakers, some choose to use SS for the bass and Tube for the high end. That's how they can get away with 200 watts on the lows and 50 watts on the highs. And STILL have both amps run out of steam at the same time. It depends on crossover frequency.

The 'Normal' (fake word) distribution calls for the 90:10 point to be maybe 5000hz. In ANY event, never go below maybe 10% power, even if you have a Super-Super tweeter crossing a 10,000hz. Or a woofer with a 25hz crossing point.

The presumption of 'normal' distribution which is not always true. A drum solo will be 90% below 1khz and 10% above. While in some cases, you might need far more HF power than 'normal'. Movies? Up For Grabs since nothing in a movie soundtrack is 'normal'.

In MY case? My speakers cross at 600hz. But I restrict the LF to the speaker below maybe 50hz to 60hz. This has the effect of pushing the 50:50 point UP to a higher frequency. So I don't feel too bad having identical amps on top and bottom. And this goes for when I go active, too.

Of more importance is matching amp gain in such a situation. Most EMO amps except some of the older stuff hovers around 28db or 29db which is good for you. Using identical amps renders that consideration a non-starter. That's why, to keep things simple, and cost effective, I put
ONE stereo amp on each 2-way panel. My preamp handles bass management and I low-cut the mains at what I mentioned as 50hz or 60hz.

Does this help?   The power / frequency chart is buried in the ESP article I linked.    

[MusicHead]

[ronkuper]

Thanks a lot!

[Leonard]

I'm doing passive biamping with my XPA-5 Gen 2 amplifier and B&W CM10 S2 speakers and the difference over traditional single-amping is very audible.

The amplifier sits on a low floor rack between the two speakers in a dedicated listening room that has been acoustically treated.

My XSP-1 Gen 2 preamp has two pairs of stereo outputs. I've got bass management is switched off, so full-range signals go to the XPA-5 via XLR cables. Channels 1 and 2 power the left speaker, and channels 4 and 5 power the right. Channel 3 is unused. To be specific, Channels 1 and 5 drive the bass section of their respective speakers. I did it this way so that the heat generated from playing bass loudly stays at the far left and far right of the amplifier instead of accumulating in the middle.

The CM10s have two pairs of speaker inputs with commonly-used metal jumpers. With those removed, two equal-length pairs of identical cables connect the terminals to the XPA-5.

My understanding of speaker impedance, correct me if I'm wrong, is that it rises so high outside its working frequency range that it presents a lighter and lighter load to the amplifier. For instance, my CM10 is a 3-way speaker with three 6-1/2-inch bass drivers in parallel and a midrange-tweeter combo. As far as the channels driving the woofers are concerned, the impedance it "sees" rises and continues to rise so that only a very amount of its output power is wasted above the crossover frequency. Same goes for the channels that drive the midrange-tweeter combo - the impedance "seen" by the amplifier below the crossover frequency goes up and up so that it eventually presents a very tiny load to the amplifier. What I experience daily is that after some loud listening, the channels that drive the bass sections are considerably hotter than those that drive the midrange-tweeter combos.

For me, the benefit to bi-amping with the XPA-5 is headroom. In the XPA-5, Emotiva wisely utilized a beefy-enough power supply so that all 5 channels can produce their rated power into their rated loads at the same time. So, when bi-amped and when playing way bass-heavy songs, channels that drive the woofers have well over a kilowatt of juice to draw from.

The track I use to verify that bi-amping sounds better than single-amping is "Tricycle" by Flim and the BB's. I have the SACD and it is a very dynamic instrumental jazz-rock track with big SPL differences between soft and loud passages. And the loud passages are particularly bass-heavy. Look it up and give it a listen.

When single-amped and at my preferred volume level, compression and distortion made itself known as harshness and a sense of uneasiness when the music got loud. I even tried bi-wiring and heard no improvements. But when I bi-amped, man oh man, the smile just got wider and wider.

It was then that I understood what reviewers mean when they describe an amplifier's sound as "relaxed" and "unforced," even at high volume levels. So go ahead and use your XPA-5 to bi-amp your speakers. Feed your amp full-range signals for now and try an active crossover later, once you figure out how to bypass your speaker's internal passive crossover.

As long as you observe the correct polarity when hooking up the speaker cables, I believe you will enjoy how it sounds.

[leonski]

Louder almost always 'sounds better'.

Level controlled test is the real way to tell.

[Leonard]

Feb 15, 2018 14:31:22 GMT -5 leonski said:

Louder almost always 'sounds better'.

Level controlled test is the real way to tell.

Yes. Which is why I don't touch the volume on my XSP-1 Gen2 preamp when comparing bi-amp to single-amp. I own two SPL meters and a calibrated microphone. I use them to verify loudness to within a fraction of a decibel. In my system, bi-amp does not get louder. It gets better.

[leonski]

Glad you enjoy your biamp experiment.
I run biamp, too, on may panels, but plan to go to an Active crossover. Deleting the speaker crossover might be your next step, too. Depending on crossover frequency, going active for woofer and leaving the mid/tweet crossover in place with a 2nd amp might be an advantage.
Removing out of band signals from the driving amp is nothing but to the good and CAN net you as much as an additional 3db apparent amp power.
All the extra resistance of a speaker crossover LOWERS damping factor and reduces 'efficiency'. Direct connect of your drivers to an amp should have beneficial effects. And given your ability to measure, you should be able to reset the relative levels very easily and be back in business.
Certain active crossovers will also have the ability to implement FIR filters which have NO phase shift thru the passband. That would be neat for my Maggies which have a 2nd order low pass and a 1st order high pass which puts the 2 drivers 90 degrees out of wack.

And not that I'm an expert in this field, I find the MiniDSP products to be very attractive. The 2x8 balanced might work, though the new 2x4HD (2 in / 4 out) will also handle FIR filters. This works for me since I break the output for my sub out of my preamp.
the MiniDSP stuff also has provisions for as many as 10 PEQ bands per channel AND can 'shelve' or assignable slope (filter Q).

[Gary Cook]

My interest in bi amping has waxed and waned for decades. It was really big in the late 80's early 90's when high powered, high damping factor, high sound quality power amplifiers cost a lot of money. So it was cost effective to share the load with bi amping. Prior to that I had confirmed my choice of running 2.1 (ie; with sub) for stereo music, something I have done ever since. Having amplifiers with quoted damping factors under 100 often resulted in single digits in the real world. Consequently removing the requirement to handle the ~20 to ~80 hz frequency range (ie; sub woofer territory) makes/made a huge difference especially to a Class A power amp rated at 25 watts. It was then able to handle the 80hz to 20khz far better.

Obviously now with a pair of monoblock power amps rated at 250/500 watts and with damping factors over 500, handing ~70 to ~20khz is a doddle and I just can't see myself getting interested in bi amping ever again.


Cheers
Gary

[leonski]

Biamping may continue to make sense even with high power 'cheap' watts.
Some think the lower powered amps in a series may sound better.
And REAL damping factor of 500 is probably impossible to achieve with real speaker loads / crossover and wire. Ridding oneself of speaker crosover might actually allow a higher real-world DF. That doesn't apply to panels and 'Stats, I don't think.

One thing that WAS not available until fairlly recently is DSP and the incredible flexibility they provide. And the Addition of the FIR crossover, not IIR is quite amazing. Imagine a crossover with NO phase shift thru the passband.

Also? You can actually get as much as 3db increase in amp power based on no longer having out of band signal. I think the REALadvantages of Bi / Tri amping are still with us. Some powered speakers are VERY good. I was surprised some time
ago when I heard the Paradigm Atom powered speaker. Compact and decent sounding, a good dorm or small system speaker not needing a lot of other 'stuff'.

[ronkuper]

Such an interesting discussion!

Feb 15, 2018 12:47:06 GMT -5 Leonard said:

I'm doing passive biamping with my XPA-5 Gen 2 amplifier and B&W CM10 S2 speakers and the difference over traditional single-amping is very audible.

Thanks Leonard for sharing, a very similar use case and your impression is very helpful for me!

What do you think about having an active crossover before the amp (I can do it in software such as Roon), without touching the speakers passive XO and with a XO point well below/above the speakers' XO points? 

Potential Pros: 
1. Reducing Intermodulation Distortion* of MF/HF signal by LF signal amplified by the same amp channel
2. Reducing load on the XPA-5 power supply
3. Reducing load on the MF/HF amp channel

Potential Cons:
1. Phase issues (might be prevented if I use FIR filters with linear phase)
2. ?

I don't expect the active XO to interfere with the speakers' passive XO since I am thinking about setting them well below/above their respective roll-off. 


*1.7 on sound.whsites.net/bi-amp.htm

[Boomzilla]

If you got the money, honey, I got the time! Yes, you should hear a difference with passive, vertical bi-amping. No external filters needed, thanks. If this is for HT, the processor will limit the bass before the signal gets to your amps, so no worries there. Since your amp was designed to feed five channels with full range signal simultaneously, the extra current will now be available for the two (right-bass and left-bass) that still need that current. If you have a sub, then the amp will be even less taxed. The R/L high frequency channels will be drawing less than one watt.

Enjoy your rig, ronkuper - I'll bet it sounds awesome!

Boomzilla

[ronkuper]

Thank you kind sir, for your encouraging words!

To further explain, besides power saving, with my suggestion I was hoping to reduce this phenomenon:

1.7 - Intermodulation Distortion
Intermodulation distortion in an amplifier is a form of distortion created when two different frequencies are being amplified simultaneously. The effects of intermod are most noticeable when one of the frequencies is much lower than the other, and the high frequency signal is actually modulated by the low frequency.

Taken from sound.whsites.net/bi-amp.htm.

I was thinking an active crossover might help with that as well while still being a Passive Bi-Amp setup. 

[Boomzilla]

Vertical biamplification alone will NOT affect IM distortion. Why? The same drivers still cover the same frequencies.

Using a subwoofer with a high-pass filter for your main speakers, however, WILL reduce IM distortion (and audibly so). Why? The high-excursion bass range is offloaded to the subwoofer(s), and the woofer cones of the main speakers don't have to travel so far. In general, less cone travel = lower distortion. This is why horn-loaded speakers (that use horns for the BASS range) such as Klipschorns & Klipsch La Scalas have less IM distortion than any other speakers on the market.

You could get less IM distortion by using a bunch of woofers in each speaker, but that's both cost-inefficient and problematic from a wave-launch standpoint. In other words, there's no free lunch...

Cheers - Boom

[leonski]

Feb 26, 2018 16:45:54 GMT -5 ronkuper said:

Such an interesting discussion!

Feb 15, 2018 12:47:06 GMT -5 Leonard said:

I'm doing passive biamping with my XPA-5 Gen 2 amplifier and B&W CM10 S2 speakers and the difference over traditional single-amping is very audible.

Thanks Leonard for sharing, a very similar use case and your impression is very helpful for me!

What do you think about having an active crossover before the amp (I can do it in software such as Roon), without touching the speakers passive XO and with a XO point well below/above the speakers' XO points? 

Potential Pros: 
1. Reducing Intermodulation Distortion* of MF/HF signal by LF signal amplified by the same amp channel
2. Reducing load on the XPA-5 power supply
3. Reducing load on the MF/HF amp channel

Potential Cons:
1. Phase issues (might be prevented if I use FIR filters with linear phase)
2. ?

I don't expect the active XO to interfere with the speakers' passive XO since I am thinking about setting them well below/above their respective roll-off. 


*1.7 on sound.whsites.net/bi-amp.htm

Don't forget that a crossover is NOT a brick wall.    So a Setting of (just an example for easy numbers) of  100hz with a 6db / octave slope results in 6db down @200hz and 12db down @400hz.
Of course, new DSP stuff will allow up to 24db octave and as much as 48db octave slopes.      So, as a practical matter, setting your 'other' crossovers above and below the speakers crossover so as not to 'interfere' may not
actually be possible.   You may also end up creating a 'phase nightmare'.     
An example from my system would be the natural lower limit of my panels is SAID to be mid 30s.   That's fairly optimistic, IMO.    My Sub, OTOH is good to below 20hz, by test and since it is an HT product, I'm guessing it must 
be good to 100hz or higher.     So?   I set my subs crossover to about 40hz to 45hz.   It is fed a full-range signal from my preamp.     I low-cut my main speaker with a 12db /octave slope at about 50hz to 55hz.  I expect the actual
crossover POINT to be about 45hz to 50hz.   The sub crossover is 24db / octave.    So, sub output is falling above crossover while main speakers output is falling Below crossover.   They might be 6db down EACH at the crossover and
so, the idea is they will SUM FLAT.    12db and 24db are 180degrees apart, so the sub's phase switch should let 'em work in phase.     I get good output, no muddiness and it is quite musical when called for.

Another potential CON for the FIR filter is that it is QUITE math intensive and the equipment needed to make it work is somewhat more expensive not to mention calculating all the 'taps' needed to make it work.

The MiniDSP 2x4HD is 205$, which is quite a bump over others in that series.

www.minidsp.com/products/minidsp-in-a-box/minidsp-2x4-hd              I think this one handles FIR filters.   

If you are Dead Set on biamp no reason at all NOT to try, especially given very minimal expense.     Keep the active option open and do some more reading.