De-Mystifying Slew Rate in Amplifiers

gregz

gregz

Full Audioholic
In my line of work, circuit response time determines the minimum slew rate I need when choosing an op-amp to use in my design. It's an easily definable term, and one I'm familiar with. It's simply how fast the output voltage can rise or fall.

However, in audio amplifiers, I've seen "don't forget the slew rate is also important" bandied about frequently without any qualification or follow-up. Suddenly, slew rate seems complicated and mysterious, with "more is better" being the only criteria. To me, "More is better" translates to nothing more than "spend as much money as you have in your pocket." Easy for you to say.

I decided to do an idle search for any threads with "slew" in the title, and of course there were none. That situation is now rectified. :)

Allow me to de-mistify this popular term:

If someone makes an amplifier that has an inadequate slew rate, that means that it can't change its speaker output voltage fast enough to keep up with the required output waveform while playing music. The symptom would be that you'd get a triangle shaped wave where a sinewave should be, which would distort the sound (speakers get creative when trying to output triangle waves).

To discuss this, we need a starting point. The audio range commonly reproduced is from 20Hz to 20KHz. Audio CD's cut off at 20Hz, and most speaker systems start dropping out at frequencies not much higher than that.

Just looking from a sine wave point of view, the highest frequency at max power would determine the minimum slew rate:
  • for 200W/ch into 8ohms (40Vrms = 56Vp), minimum slew rate = 6V/us
  • for 100W/ch into 8ohms (28Vrms = 40Vp), minimum slew rate = 5V/us
  • for 100W/ch into 4ohms (20Vrms = 28Vp), minimum slew rate = <4V/us
Ok, the fist thing you notice is that those numbers are awfully small compared to typical amplifier slew rate specs of 50V/us to 120V/us. The next thing you notice is that the slew rate alone is meaningless without the acompanying power specification. The less the power, the less rate is needed.

In fact, if all we were worried about was producing sine waves, then a humble slew rate of 50V/us would be enough for a 200Vp output into 8ohms (2500W). Conversely, it would provide a 100W amp enough output voltage switching speed to produce a 2MHz sinewave! :eek:

I don't know of any audio amplifier that has the bandwidth to output 2MHz; most of the good ones poop out at around 30KHz or 50KHz. Other limiting factors besides slew rate start taking effect at that point.

"That's sinewaves," you say. "What about music?"

Many years ago, Peter Baxandall conducted tests of music waveforms and their slew rate demand on audio amps. His article published in "Wireless World" showed that a range of well recorded LPs produced a waveform with time rate changes no greater than 2KHz. For a 100W amplifier, the needed slew rate would be .5V/us.

With the advent of digital sources, this test has been conducted and re-conducted many times over. Should it be no surprise that music still has 100W amplifier slew rate demands only in the single digits? It shouldn't be; Peter Baxandall's tests may have been conducted in the LP format, but remember that LP needle cartriges are capeable of responses past 30KHz.

Suffice it to say that the slew rate is about as important an amplifier spec as damping factor (see http://www.audioholics.com/techtips/audioprinciples/amplifiers/dampingfactor2.php). Any audio amplifier worth its weight in dirt is capeable of producing 20KHz at least 75% of max power, which is quite a high margin above the slew rate needed to play music.

For more in-depth technical reading on slew rates, here are links to some good articles on slew rate demands of LP's, CD's, and even SACD and DVD-A:

http://stereophile.com/features/404metrics/index1.html

http://www.angelfire.com/ab3/mjramp/golopid5.html
 
WmAx

WmAx

Audioholic Samurai
gregz said:
Suffice it to say that the slew rate is about as important an amplifier spec as damping factor
Ironically, some so-called 'hi-end' amplifier have an output impedance(resulting in very abnormally low damping factor ratio) of sufficient magnitude as to effect audible variations in frequency response on some speakers.

SO of course, in the case of purposefuly bad engineering these parameters are still important! :)

-Chris
 
newsletter

  • RBHsound.com
  • BlueJeansCable.com
  • SVS Sound Subwoofers
  • Experience the Martin Logan Montis
Top