Sean...Of course I am a long time believer in biamping, but I do think that the original reason (and I am talking about 50 years ago) was to avoid intermodulation distortion. Power amps have improved so that this is no longer a significant issue. As you say there are power delivery advantages also, but in this regard amplifiers have also improved about tenfold. (Back then a "big" amp was about 25 watts, and now it would be 250 watts or more).
Which leaves us with the elimination of the passive crossover as the remaining advantage. I cannot understand how many people seem to think that driving two separate full range amps into the passive crossover is "biamping". I have two comments here.
1...It is dangerous to run a tweeter directly from a power amp. Turnon and turnoff can be accompanied by "thumps" that the tweeter won't like, and a loose interconnect can be an instant disaster.
2...Passive crossovers do not absorb power to the degree that you suggest. The simple test for this is to consider how hot they would get if they did absorb a lot of power, particularly since they are usually mounted inside a closed box, and surrounded by insulation. To take the scientific approach: Capacitors (ideal) dissipate zero power. EE101. Real capacitors are near ideal in this regard. Inductors will dissipate power because they cannot be made with zero resistance. The externally mounted 3.5mH inductors of my MG1.6 are 10 AWG air core coils with dc resistance of 0.2 ohms, which is about the same as the original equipment iron core inductors. Since the driver is 4 ohms, 0.2/4 which is 5 percent of the power will end up as heat in the inductor. (But they never feel warm). Resistors are used in crossovers, but not in the signal path except for the tweeter, where some power loss is usually necessary to balance SPL. If the tweeter padding resistor needs to be a large value, the wrong model of tweeter has been used.
What is the basis of your suggestion that lost power will be 50 percent? |
Sean, Good explanation overall but I think that there are a few things that might be mis-represented:
* "As taken from Audio magazine September 1975 in an article entitled "Bi-Amplification - Power vs Program Material", it was demonstrated that an actively crossed two way with 30 watts on the tweeter and 60 watts on the woofer clipped at appr the same point as a 175 wpc amp driving the same speaker full range with a passive crossover. " If I had to guess, I'd say that the active xover must have had gain. So, the 30W & the 60W amps each had help from the active xover. IOW, there were 2 amplifiers in series: the active xover & the 30W & 60W amps. Needless to say, it must have been enough. The 175W amp OTOH did not have that help & performed all the gain from input to output in just 1 chassis. Hopefully that 1975 report makes some mention of this??? So, when you do 30W+60W=90W < 175W I think that one must add the active x-over gain. Just a comment here. I do agree with you on the concept of bi-amping & further that external x-overs are better.
* "This is done by reducing the thermal losses that one would normally incur when running the capacitors, coils, resistors found in a typical passive crossover." Thermal losses in a capacitor & coil????? Physics does not allow this! I think Eldartford's latest post also addresses this. This I think is a fundamental error! Inductors will dissipate some heat as a coil of wire will have some resistance. If one is using a low ESR capacitor, very little heat gets generated in a capacitor. If you have an active xover, touch the C & L right after several hours of play & see how warm or hot it is!! The way that one effectively doubles the power in a biamping situation is NOT by minimizing the thermal losses(!) but it is by splitting the power into atleast 2 chassis: one for the tweeter & one for the woofer. If one adds the power linearly then one will notice that the voltage excursions have doubled. Thus, if you calculate power, which is volts squared divide by impedance, one will notice that the power has quadrupled effectively. This is the theoretical limit which is usually not achieved owing to distortions in the amp which disallows one from running the amp wide open. Thus, the practical increase in effective power is more like 2X, as that 1975 article you refer to showed.
* "In order to maintain control over the driver, the amplifier not only has to generate enough power to overcome that reflected voltage that the driver itself has produced, but it has to have even more power on reserve in order to "muscle" the driver according to the intensity of the music signal that it's being fed." I feel that this is not correctly written. Yes, the woofer does produce back-EMF. Does this back-EMF hit the amp terminals directly? Not at all.....there's the reverse isolation of the x-over ckt (active or passive). The back-EMF is attenuated quite a bit before hit the amp terminals. The amp does not have to generate more power to overcome & "muscle" the woofer. What the amp has to do (if it is designed correctly) is provide a very low output impedance path into which the back-EMF can be channeled to ground potential. This merit of the amp is measured by the "damping factor" of the amp. In an actual system the impedance of the speaker wire can totally ruin the overall damping factor & disallow (an otherwise high damping factor amp) to control/damp the woofer.
* "As such, the "beefier" the drivers that you're using and the louder and lower in frequency that you want them to play, the more power that you have to have to maintain control. If you don't have the power to maintain control, the driver...." I really don't think that power maintains control over a woofer. Power/Current into a low impedance is what creates the woofer excursion BUT it is a high damping factor/very low output impedance that REALLY maintains woofer control. As they say often - power w/o control is largely useless. The control comes from the damping factor taking into consideration the speaker wire impedance & not just the amp.
* "Having said that, it has been my experience that high quality drivers can take GOBS of power beyond their power rating. That is, so long as the power remains clean ( non-clipped ) AND the music that one is listening to is "dynamic" in nature." A comment here - actually clipped power is not really that much of an issue in & off itself for a speaker driver 'cuz the clipped signal does not hit the driver directly. It must go thru a xover (active or passive). Remember that amps usually clip @ the top end i.e. when the volume has been turned up near-max. Remember, also, that a clipped signal is rich in harmonics - high frequency harmonics. These high frequency harmonics will be filtered by the tweeter xover ckt beyond 20KHz. The xover is a filter, they have to be filtered off. Thus the intensity of the clipped signal hitting the tweeter is much reduced. What really kills a tweeter is, as you wrote, duration. The longer one lets the clipped signal sustain, the longer it feeds the tweeter & the VC burns up. I think that it was on the Adair Audio website that I read that VCs are only about 1-2% efficient! I was shocked to read such a low number - I knew that they were largely inefficient but wasn't expecting 1-2%! If that's true, it's easy to see why "Duration" would fry a tweeter.
Anyway, my intention is for this post to be viewed in a constructive way. I do hope that you view it the same. Thanks! |
Capacitors DO lose power. This is called "dielectric absorption". ESR is also a term used to describe the series resistance or "internal losses" of a capacitor. If the power loss is severe, the cap "cooks" itself from the inside out and doesn't last very long or changes value to the point that normal circuit operation becomes harder and harder to achieve.
Inductors DO lose power i.e. it is called series resistance. Besides that, some of the energy is dissipated in the magnetic field that the inductor creates and / or through inductive coupling to other nearby parts of the circuit that they shouldn't be in. This is why proper crossover layout is important. I've seen "high end" crossover circuits that had a tremendous amount of crosstalk taking place due to this coupling. The end result is that tweeters and midranges end up being fed energy that should have gone to the woofer. Why did this take place? The inductors are too close together, placed in the same horizontal or vertical plane, etc... Due to their close proximity, the coils become inductively coupled via the previously mentioned magnetic field. By simply changing the horizontal and / or vertical planes of the inductors, without even moving them further apart, crosstalk can be reduced by at least -40 dB's. Not only does this result in less loss in the circuit, but better sound due to less smearing. Power handling is also improved due to having a more effective frequency dividing network with less stray coupling.
The use of a Zobel network that is tuned to operate within the audible pass-band WILL attenuate energy. The capacitor selects what frequency the Zobel comes into play and the value of the resistor dictates how much energy it will consume. If one does not use a resistor of high enough power handling and / or the value is poorly chosen, you WILL burn up the resistor in the Zobel when "cranking" the volume way up on a steady-state basis.
All of these parts DO rise in temperature as they are used and this is part of what "component settling" is all about. If the average power that they are passing is high enough, you would feel the heat that they were dissipating. Since most music is very low in average power consumption and very high in peak power, the energy lost / heat generated within these parts isn't very high and / or consistent. It is these very short duration, high intensity peaks that get "eaten up" in a passive crossover, hence the increase in dynamics and increased detail that one encounters from going active. All of the aforementioned losses in the passive circuit end up costing resolution while increasing the levels of inductive and capacitive reactance that the amp tries to load into.
"Thus, the practical increase in effective power is more like 2X, as that 1975 article you refer to showed".
Doubling the power ( X2 ) is a 100% increase. Losing half the power ( /2 ) is a 50% loss. We've said the same thing using different wording looking at things from opposite points of view.
As far as the 30 / 60 wpc actively crossed amps clipping at the same appr output level as that of the 175 wpc passively crossed amplifier, i can't help you out there. I referenced this from Vance Dickason's Loudspeaker Design Cookbook and do not have access to the original article. Your own response seems to confirm these figures rather than contradict them.
The comments about the active crossovers increasing the gain of the signal have little to nothing to do with amplifier capacity. The amplifier can only put out so much power prior to clipping regardless of whether it has to amplify itself is providing the gain or whether the drive levels are increased. If anything, increasing the drive levels via the active crossover would have driven the amplifier into clipping sooner and more consistently. This is because the amplifier functions at a constant rate of gain, so long as it is linear in operation. More drive means more clipping.
Damping factor is a hoax and is completely taken out of context. I've explained this in the past several times. Couple this with the fact that your .1 ohm output of the SS amplifier is typically fed into 40 - 120 ohm speaker cable to get to your 4 - 12 ohm speakers and you should begin to understand why i rant about using "properly designed" low impedance speaker cables as much as i do. You guys are missing the boat on this one in tall fashion.
Reflected EMF is a reality that the amps have to deal with. If it were not, microphones would not work or produce voltage and we couldn't use them to capture acoustic signals to make recordings with. Now take the microphones that produce EMF and increase their capture area ( cone size ) and motor structure by a few dozen times and tell me that they don't generate voltages, especially when their excursions are made in great amplitude and speed. Just as feeding voltage into a coil in a magnetic field causes the cone to move, moving the cone that has the coil attached to it that is placed within the magnetic field generates voltage. This is an action / reaction that is unavoidable. Obviously, one has to "out muscle" the other opposing source of voltage / current, otherwise the two would cancel each other out as heat and there would be no acoustic output what so ever.
"The externally mounted 3.5mH inductors of my MG1.6 are 10 AWG air core coils with dc resistance of 0.2 ohms, which is about the same as the original equipment iron core inductors. Since the driver is 4 ohms, 0.2/4 which is 5 percent of the power will end up as heat in the inductor".
Bombaywall, El just answered your "Thermal losses in a capacitor & coil????? Physics does not allow this!" comment. On top of that, he was talking about a 5% thermal loss using a 10 gauge conductor, which has very low series resistance. How much more loss is there in an inductor that is wound using a 16 - 20 gauge conductor as found in most commercially produced loudspeakers?
"It is dangerous to run a tweeter directly from a power amp. Turnon and turnoff can be accompanied by "thumps" that the tweeter won't like, and a loose interconnect can be an instant disaster."
If you look at some of the old Stereo Review, Audio and even Stereophile reviews of speakers, i think that you'll find that many tweeters will easily cope with transient bursts into the 100's if not 1000+ watt range, so long as they are limited in duration. I've seen tests were the tweeters handled more power than the woofers did, so long as they were bandwidth limited. Other than that, the easy way to get around "turn on surges" is not to turn your gear off.
"I think that it was on the Adair Audio website that I read that VCs are only about 1-2% efficient! I was shocked to read such a low number - I knew that they were largely inefficient but wasn't expecting 1-2%! If that's true, it's easy to see why "Duration" would fry a tweeter."
Pretty efficient horn designs only come up around 4% - 5% or so. Almost all of the power generated within an amplifier is dissipated as thermal losses in a loudspeaker. As such, removing even several tenths of one percentage point via getting rid of "lossy" passive components between the amplifier and driver can make a sizeable difference. Increasing power transfer via proper impedance matching can also improve system efficiency AND improve transient characteristics.
Sorry if this jumps around quite a bit, but i'm limited on time. Due to my increased work load ( busiest time of the year ) and changes in my personal schedule due to family health problems, i can't hang out here as much as i'd like. Hope this at least clarifies a few things. Sean >
|
Hey Sean,
Neither Eldartford nor I totally denied that L & C do lose power. Eldarford: "Passive crossovers do not absorb power to the degree that you suggest." & I wrote: "Inductors will dissipate some heat as a coil of wire will have some resistance. If one is using a low ESR capacitor, very little heat gets generated in a capacitor."
However, I think that you are mistaken as to just how much gets dissipated in these Ls & Cs on average. Optimizing the losses in a reactive component is NOT the way biamping achieves its perceived increase in power. Separation of power amplification into multiple chassis & the constructive addition of the music signal is what gives biamping its perceived increase in power.
"The comments about the active crossovers increasing the gain of the signal have little to nothing to do with amplifier capacity." True! However an amplified signal fed into an amp needs less amplification to the speaker => more headroom for dynamics. Such a setup would have more air/the music would flow easier/more transparent. The comparison is between passive & active xover cases.
"Couple this with the fact that your .1 ohm output of the SS amplifier is typically fed into 40 - 120 ohm speaker cable" What???? Who in their right mind would use a 40-120 Ohm speaker cable?? Does a speaker cable with 40-120 Ohms even exist? I think that you would ripe for catching much flack for this one, Sean! :-) I think it's a typo. You really meant milli-Ohm, right??
"Reflected EMF is a reality that the amps have to deal with." Oh, it certainly is. No question about & I never denied it in my post. However, not to the degree that you mentioned in your original post. There IS the reverse isolation of the xover & their IS the output impedance of the amplifier/damping factor as only an active circuit can get rid of the back-EMF
"Damping factor is a hoax and is completely taken out of context." No it certainly is not! Without a damping factor, which is really a glorified name for amplifier output impedance there'd be no "out muscling" the woofer. We can agree to disagree here, which is just fine w/ me. ============================
Anyway, I, too, am in the more power, the better camp as long as the user is careful to not overdrive the speaker. If there are transients, then they should be shorter in duration (than being longer) & if one doesn't know then it's better to listen with a slightly lower level to be conservative. Case-in-point: Last summer I fried my Green Mountain floor stander tweeter after trying to break-in the woofer by playing music too loud for 14 hrs at a stretch! What Roy Johnson told me was that tweeter can handle way more than its 100W RMS rating if I play loud music for about 3-4 minutes (1 track) every 20 minutes or so. He says that he does that all shows regularly & the tweeter survives each time. So, now I know....... More power results in the music flowing better, more air, more transparency, more lower level details. Drivers can handle many more watts of peak power even tho they are rated for lower RMS power. (Usually, the driver ratings are RMS power, which is average program power). So, yes, Fishboat, its reserve power mostly. Additionally, a large amp has more current capacity even at lower power levels, which what gives the music is ease of flow. To get that driver moving, you need current in the voice coil.
|
I think the damping factor issue is directly related to the speaker in question, which may very likely have a very low moving mass and very high magnetic field and very short excursion length(such as my speakers), and thus have a very significant amount of its own internal damping ability.
In cases such as this, high numerical "damping factors" may actually inhibit transient response, and may not be an advantage at all.
As long as the relationship between the amp's output impedance and the speaker's relative impedance at any given frequency(damping factor) remains enough to control the speaker(>5), then it is sufficient in that circumstance. It is when the speakers have poor internal damping characteristics(ie high moving mass/low magnetic field strength/very long excursion length) where problems are more likely to be "fixed" by high electrical damping factor numbers.
Regarding high power being "better" than low power, we all know I'm in the "low power" camp. I think(and I'm pretty sure Sean agrees) that you are much better off to have higher efficiency speakers than to try to overcome low efficiency speakers with brute force, because of the exponential curve of power needed to add a few db to the output.
Also, it has been my experience that high power amps seem to be less "delicate" in nuances and details, due to the beefy construction needed to handle all that power. And in addition, to have power like that, you simply have to go out of Class A operation, which I don't like to do.
And, when you have low efficiency speakers, you have a higher "low-level detail threshold" because it simply takes more power to make the speaker move at all, so that some very low level stuff never even makes it out of the speakers, unless you have it turned up alot.
If the speakers are high efficiency, you can still get pretty loud SPL, with great low level detail, with very low power amps. I do realize that this normally sacrifices some of the very low bass. I think it is not a bad trade-off. |