Bi-Amplifying (or Bi-Amping)

Andy W — Fri, 07 Sep 2007 20:23:00 GMT

This is the practice of running two amplifiers (and two pairs or wire) to each speaker, one amplifier for the woofer, and one amplifier for the tweeter. Of course you can also tri-amp, quad-amp, etc., depending on the number of drivers employed in your particular system.

First, let’s get something out in the open. There is no such thing as “Watts RMS.” Get it out of your vocabulary. It is a misnomer. They are just “Watts.” When you apply a voltage of 1 Volt and you have a resulting current of 1 Ampere, the result is 1 Watt dissipated into the load. In describing AC signals engineers use the term RMS (root mean square) which describes the voltages and current in terms of an equivalent DC voltage or current. Thus a 1 Volt RMS signal over a given period of time has and equivalent EMF (electromotive force) as 1 Volt DC, and a 1 Amp RMS signal is equivalent number of electrons flowing as a 1 Amp DC signal. If you have a 1V_rms signal delivering 1A_rms current the resulting power is still 1W, not 1W_rms. It doesn’t matter how many people use the term, or how long the terms has been used, it is still incorrect.

It’s like saying the word “jeep” came from slurring the letters “GP” which means “general purpose.” This has been repeated so many times, for so many years, that people think it is true. It was even a question, er, answer on Jeopardy. A) The word “jeep” had been used in the military for a number of vehicles before the “jeep.” B) “GP” does not mean “general purpose”. Rather, “GP” came from the term “GPW” which was Ford designation for the WWII “jeep” that they built under Government contract. Now you know where the “G” came from. The “P” was simply a designation for the 80 inch wheelbase. The “W” indicated that the vehicle was built to the Willy’s design. Ford’s own documents prove this. Interestingly I saw some antique farm tractors at the State Fair from the time period and earlier that were labeled “General Purpose”, so maybe that’s how the whole thing got started.

But I digress…

When it comes to amplifier ratings, I’m in the same camp as Paul Klipsch. Rating amplifiers in Watts is a bad idea. Some have suggested rating them in dBW (dB’s relative to 1 Watt) or in horsepower(!?), or as Mr. Klipsch suggested, Watts peak.

I would rather they be rated in Volts rms and drive capability, either specifying a load impedance or current capacity. Why does this make a difference you ask? Because rating an amplifier at 1000W gives consumers the impression that it is 10 times what a 100W amplifier is. In terms of sound output, that’s not the case. A 1000W amplifier will play 10dB louder than a 100W amplifier, if your speakers can take it (they can’t). So what’s 10dB? It’s only a relative doubling of apparent volume. 90dB SPL is twice as loud (to the human ears) as 80dB SPL, which is twice as loud as 70dB, which is twice as loud as 60dB. So what is the power difference between 60dB and 100dB SPL? 10,000:1. Which is part of the reason Paul Klipsch said, “All the world needs is a good 5 Watt amplifier,” or something like that.

Of course, you would also want a pair of Klipschorns if you only had 5 Watts… and he did.

So what can a 1000W amplifier do that a 100W model can’t do? It can play loud transients cleanly without clipping. And it can destroy your speakers very quickly if you get stupid with the volume knob.

But wouldn’t it be less expensive to have more efficient speakers than to pay for a bigger amplifier? Give yourself a cookie. It was true 60 years ago and it’s still true today*. If you increase the sensitivity of your speakers by 6dB, you can decrease you amplifier needs by a factor of four. Now your 25W amplifier sounds like 100W, and you aren’t in any danger of melting your voice coils. If you have K-horns, then your 25W amplifier sounds like 630W (compared to a speaker with a 90dB sensitivity rating), and a 100W amplifier sounds like 2500W. That’s what it means when Klipsch talks about dynamic range. (* Given the current state of power electronics it’s somewhat less true than it used to be, but there aren’t many 1000W per channel consumer amplifiers that don’t cost at least as much as a pair of K-horns. The more enlightened among us (or cheap, like me) just buy a nice Crown or QSC professional amp and call it a day.)

Anyway, back to my point…

Using voltage to describe amplifier power, you would get 12.5W (into 8 Ohms) for a 10V (rms) amplifier, 50W for a 20V amplifier, 200W for a 40V amplifier, 800W for an 80V amplifier, and so on. Each doubling of voltage represents a 6dB increase in output power. Not exactly what we need but it’s a big improvement.

Another option is to rate the log of power (10W is a “1”, 100W is a “2”, and 1000W is a “3”), but that doesn’t quite work either… I need to figure out how to make 10W a “1”, 100W a “2” and 1000W a “4”.

The reasoning Paul Klipsch had for wanting amplifier ratings in peak power output goes like this… If you need 50W for a 200Hz signal and 50W for a 2kHz signal, the amplifier required to reproduce the signal without clipping is a 200W amplifier.

What?

Doesn’t 50W+50W=100W.

Well yes, but…

50W requires a voltage swing of 20V_rms. For a sine wave, that’s 28.28V peak. To fully reproduce both signals at the same time without clipping requires 40V_rms or 56.57V peak. A 100W amplifier can only deliver 40V peak (28.28V_rms).

Now, back to bi-amping…

There is some real “magic” in bi-amping. If you do it right you can theoretically take two 50W amplifiers and make it sound like a 200W amplifier.

How about a little controversy? Another reason Klipsch puts dual binding posts on some of our speakers is so people can bi-amp their speakers.

Other than the obvious, taking 100 Watts worth of amplifier and making it sound like 200W, why would you want to do this?

A good reason would be that the woofers need a lot more power than tweeters. Five or ten Watts on a tweeter is a whole lot of sound. Twenty Watts is getting up there, and fifty Watts, well that’s just not nice at all. On the other hand the woofer might be happy with 100 Watts. The benefit is that you don’t overpower your tweeter, and when you clip your woofer amp (which happens a lot more than you might think) none of the nasty signal gets to your tweeter. Also, some people believe that amplifier “A” sounds better on woofers whereas the tweeter really needs amplifier “B” to sound its best. Whatever (maybe there’s another blog topic.), it’s one more reason why we have dual binding posts on some of our speakers.

So where’s the controversy? Just taking two stereo amplifiers and hooking them up to your speakers (removing the straps on the input terminals) is called passive bi-amping (or “fool’s bi-amping” for the non-PC) and is very close to being just a waste of money (again many people will disagree). You are sending the entire music signal to both amplifiers and both amplifiers are amplifying the entire signal, but you are allowing the passive crossover network to filter the amplified signals. I can’t think of any benefit in doing this. The tweeter amplifier is still reproducing the bass signals and they are still causing the amplifier to clip. (Perhaps there is a slight benefit by not requiring the current of the bass notes, as this can increase the headroom of the tweeter amplifier – by not drawing down the power supply rail voltage.)

So why is it again that we put multiple binding posts on our speakers? Some people claim it makes a difference, and (my own opinion here) we don’t want that to be the reason they don’t our speakers.

One step better than “passive bi-amping” is to add an active filter ahead of the amplifiers, an octave away from the stock crossover point, and thus gain some tangible usefulness (my opinion) of this arrangement. This wouldn’t affect the stock crossover too much and would relieve the amplifiers of amplifying signals that aren’t being used (especially helpful for the tweeter).

Real bi-amping requires active bi-amping which is that the crossover is done electronically with active filters or a DSP and these signals are sent to the amplifiers. The amplifiers then drive the speakers directly with no passive crossover components at all.

The hard part of doing this is that unless you can accurately reproduce the passive crossover with an active system, your speakers will not sound like they were designed to sound, since when you do an active bi-amp you remove the stock passive crossover network.

So real bi-amping requires dual binding posts, but doesn’t require the passive crossover network (which is usually many times the price of the binding posts)… Hmmmm… so if someone really wanted to bi-amp, why would they want to pay for passive networks? Oh well, some things will never make sense to this engineer.

Many of us already enjoy the benefits of a bi-amped home theater system when we use a powered subwoofer. The low frequencies go to the sub, and the receiver is now free of a large power burden. The woofers in the main speakers can now reproduce the upper bass and mids with lower distortion, and overall the system has a much better dynamic range. It also makes it possible to get great sound from smaller unobtrusive designs.

Where bi-amped (or tri-amped) systems really shine is in powered systems. The power output, frequency response (via EQ), and limiting (and/or compression) can be tailored to the drivers in the system. And the drivers can be tailor made for the frequency spectrum they reproduce, not that they aren’t already, but if you relieve, say a 5” woofer from the demand of reproducing the lowest 2-3 octaves, you can design it for a better midrange sound.

Bi-, tri-, quad-amping is commonplace in professional rigs. And if you hear a true bi-amped system with well designed filters (and limiting/compression when the situation calls for it), you’ll want to do on every speaker.

Shameless plug: Klipsch Bi-amp towers announced. Maybe Andre will talk a little about these puppies in one of his blogs. The comment I heard from Kerry Geist after listening was, “If we can’t sell that…”

One side effect of a passive crossover network is that it absorbs some of the power going to your speakers. In a perfect world inductors and capacitors would not absorb any power. When you make them out of real materials they do. In addition, most of the time tweeters will have more output than woofers, so the network will likely use a padding resistor to match the output. If you need to pad a tweeter by just 3dB, then the resistor will absorb half of the power going to the tweeter. So, if you’re bi-amping, you could cut your tweeter amp power in half, or you could have 3dB more overall output.

That sounds like a Klipsch design philosophy, doesn’t it?

Andy's Blog : bi-amp, bi-amping

Bi-Amplifying (or Bi-Amping)