Slew rate and rise time


Hi, I just wanna ask if slew rate is the same as rise time. Which of the parameters is used to describe how fast an amp responses to input signal? Is there a minimal slew rate or rise time value which is still compatible with good sonics, or are sonics not depending on these values?
dazzdax
Herman makes some good points, but i'd like to take his explanation a bit further.

Think of an electrical circuit as a hose for electrons. Then think of the music as a fire. Different size hoses can deliver different volumes of signal at different pressure levels. Volume and pressure are different things, so don't confuse the two. Fires also burn at different rates in different areas, making some harder to control than others. Music is much like a fire as you have different intensities in different parts of the audible spectrum, all taking place simultaneously.

Think of how fast the signal can make it through the hose ( circuit ) from one end to the other on demand. This is a measure of accelleration and is equivalent to the rise time. In other words, we are clocking how fast the circuit can go from just above a trickle of flow to almost wide open flow. Obviously, when we need water ( electrons ) to put out a fire ( reproduce the signal ), we need to get it there fast. The faster that the circuit can deliver maximum flow, the higher the rise time. The higher the rise time, the more responsive we are to putting out the fire. There's one problem here though.

What happens when we have a hose of a given speed, but different size fires? Hmmm.... Obviously, a garden hose could be used to put out a fire, but whether or not it could keep up with the demand of flow required for a very large fire that changes intensity or direction rapidly is another story. This is where slew rate comes into play.

Slew rate is equivalent to how much flow the hose ( circuit ) is capable of. The bigger the hose ( higher the slew rate ), the more volume we can flow. The more volume that we can flow, the larger the fire ( signal ) that we can put out. After all, it is possible to get water onto a fire in rapid fashion ( fast rise time ), but if the fire ( signal ) is bigger than the volume of water ( limited slew rate ) that we can deliver, that fire ( signal ) goes at least partially unquenched.

The end result would be that our fast response time ( fast rise time ) wouldn't be enough in itself to handle the amount of fire ( signal ) due to a lack of volume of flow ( slew rate ). In effect, we not only need to be able to deliver the signal fast, but we must also be able to deliver the quantity of signal needed as the situation varies.

With all of that in mind, rise time and slew rates are not one in the same and shouldn't be thought of as being directly tied together. While one can derive horsepower figures by looking at the torque rating of a motor at a given rpm, the horsepower and torque curves do not run parallel to each other at any given time. Such is the same with slew rate and rise time. Obviously, rise times and slew rates are two different measures of a circuit i.e. speed and capacity. Having good performance in one area without the other means a limitation in performance somewhere down the line though.

In most cases, a circuit that has a faster rise time will demonstrate a higher slew rate by design. That is, if the engineer / designer is smart enough to take both aspects of circuit performance into account. The faster the rise time and the higher the slew rate, the more responsive the circuit is to any given signal and the less challenged it will be in trying to reproduce that signal. Combining great speed and agility ( fast rise time ) with a capacity for brute force ( high slew rate ) makes for a very well rounded performer.

There are many other spec's that are taken for granted when it comes to circuit speed and finesse that never get mentioned. That is, just as a circuit has a rise time ( how fast it can go up the hill ), there is a fall time to ( how fast it can come down from the peak ). Most circuits offer pretty symmetrical levels of performance here, but not always.

If symmetry is lacking in this area, the peaks will tend to be reproduced TOO avidly, resulting in overshoot. When you get a high peak that wasn't meant to be that sharp, the sound gets "edgy". This has to do with a lack of damping i.e. too slow of a fall or "recovery" time. To continue on with our hose analogy, we can get the water to the fire in both the time and quantity needed, we just couldn't regulate the quantity of flow from the hose before the water itself created further damage. This is sometimes referred to as excessive leading edge energy when one can find a qualified reviewer. They might not understand the what's & why's, but they can hear it taking place : )

Another overlooked spec is Td or the Time Delay of the circuit. Like the hose mentioned above, different circuits have different length paths to them. A shorter hose or circuit can deliver what we need faster with less potential for delay whereas a longer hose or circuit will take longer to deliver the goods. Much like a hose, a longer circuit may end up having more losses along the way with greater variances in flow due to all of the various connections made.

For best results, we want to keep the path short and as simple as is feasible. This results in mininal Time Delay with the least potential for loss or smearing. Most circuits don't do this and the end result is not only a loss in signal, but also timing. The fact that a circuit can have different loss rates at different frequencies along that path can really play games with what we hear. That's because not only can the timing be altered as frequency varies, but the amount of loss can differ as frequency is varied too.

This has to do with dielectric absorption of the parts and why changing passive parts like resistors, capacitors, inductors, wiring, component jacks, etc... can change the sonic presentation that we hear. Not only in terms of tonal balance, but transient response, speed, coherency, spectral purity, etc... These effects also take place in active parts like transistors, diodes, tubes, etc, too. That's why "doping" or the application of chemicals to the external parts of the device can change the electrical and sonic characteristics, etc...

While one could conclude that a longer path is both slower and more lossy ( logical conclusion ), there are also other factors here too. Not only does negative feedback increase the complexity and parts count of the circuit, it also increases the length of the path and the response time of the circuit. The more negative feedback that you have, the more time delay that you'll have. The more time delay that you have, the more loss that you have of various time and frequency related artifacts.

This is why high negative feedback designs sound less "liquid" as they suck the life out of the music. That is, due to the above mentioned non-linear losses with increased parts count and circuit complexity, the timing, harmonic structure and transient response are all mangled to a much higher degree. This makes the music sound lifeless, even though most of the distortion type spec's look great on paper.

This is where the design prowess and listening skills of the manufacturer come into play. It is also the reason why products that are built only to measure well don't sound all that good i.e. the "distortion wars" that resulted in all of the horribly hard, bright and sterile sounding SS gear of the past. That is, the engineers didn't count on all of the negative sonic aspects of electrical error correction and circuit complexity that are involved, nor the non-linearities of the low quality parts that they were using. They assumed that the end would justify the means, which in reality, it really did. Unfortunately, the end sums up the means and the means weren't well thought out or properly executed to begin with. That's why they had to rely on so much negative feedback to get the job done.

This is also why products that can measure similarly don't always sound alike i.e. two different design routes can produce different complexities of circuits with different parts used. As such, the sonics of the circuit are even more important than the measurable accuracy that most engineers base their designs on. This is why many circuits that measure poorly actually sound pretty good. That is, the distortions and non-linearities that they introduce are actually not as detrimental as the means used to achieve some of the better measuring, but poorer sounding gear.

In effect, the "PRAT" or "musicality" of the circuit are directly correlated to all of the above. While there is no real way to quantify the term, "PRAT" is a term that takes into account both the musical accuracy and the electrical accuracy of the circuit. That is, musical notes don't sound like musical notes, whether it is due to non-linear circuit design, using low grade parts, relying on error correction circuitry to compensate for lossy parts and / or non-linear circuitry or any combo of the above. The end result is always audible.

Obviously, there are different levels of PRAT based on how well the circuit does all of the above and preserves the signal being fed into it. With that in mind, once you've heard a system that has even a smidgen of PRAT, you'll know the difference between a "sound system" and a "music reproduction system". One reproduces sounds that emulate music and the other reproduces music. The former is what i refer to as measurable accuracy and the latter is what i refer to as accurate musicality. Sean
>

PS... I did this at 5 in the morning after waking up in the middle of the night. Please cut me some slack if it's a little lacking in flow and / or specificity : )

Sean, excellent "Article." I'm going to keep it for a reference. I have never seen all of this subject matter put together in an understandable format with the relationships such as this. You have again increased my knowledge of what I hear vs design parameters. I knew what I have heard but the why has been in some question.
Bigtee: There have been questions here that i've wanted to respond to in this fashion and i've had private requests to try and explain this too. This specific thread caught me at the right time and in the right mood to attempt this type of "article". I was able to write all of this in about 45 minutes ( including additions and revisions ), so it wasn't as bad as i had thought it was going to be. Obviously, it is not all inclusive, but it should help explain a few things and give people that are interested in learning something to work with.

This might be a good thread to keep going with other technical questions pertaining to spec's, measurements and what we hear. This would give newcomers a point of reference as to what spec's are, how to interpret them, how they interact with each other, etc... Only problem is, getting people to post the pertinent questions & comments within this thread and then keeping those that shouldn't be posting comments & questions in this thread away from it.

I don't have a problem with people presenting alternative points of view, but i do have a problem with those that present information in dogmatic form that they've read elsewhere without having any first hand knowledge on the subject and / or those who refuse to either consider alternative points of view and / or won't learn on their own.

Having said that, most all points of view have some form of validity to them and the key factor is being able to fit all the different pieces that make up the puzzle together in a cohesive manner. Too many people want to grab hold of one aspect of operation and / or theory at the expense of how that part of the equation factors into the sum product. The end result is a partial understanding that is based on fact, which is what makes it believable, but may not be completely applicable when one considers all the factors involved in the grander scheme of things.

This is why i've said that spec's can be used as a very useful tool, but only if the spec's are valid and the type and quantity of spec's are suitable for a meaningful interpretation.

The only variables beyond the spec's would be the actual parts used and the lay-out of the circuit, which can surely present some variables to the sonic outcome of any given product. None the less, and as i've shown on several different occassions, "reasonable guesstimates" of what to expect sonically are possible even with those variables entering into the equation.

Part of this has to do with common sense. If one starts off with a junk platform, even the best parts and excellent execution of the circuit can't make it an awesome performer. On the other hand, even an excellent design that is saddled with poor parts and / or less than optimum execution isn't going to perform the best either. That's why we end up with so many mediocre products i.e. all the factors aren't considered evenly and the end result is a "happy medium" that the bean-counters can live with. Sean
>
Thank you Sean for your descriptive analogy. I'll keep it also. You know, another issue for discussion is also what the relationship is between gain, volume and amount of power that an amp delivers. Let's say an amp is capable of delivering 200wpc into 8 Ohms. Is this with the volume knob wide open or doesn't it relate to the position of the volume knob? But this is as I say another issue to discuss. I'll start a thread about this.
Sean, your analogies are very colorful, one might say even fiery, but I am afraid they are not exactly right. Given the fact that part of what you say is correct and some of it isn't makes it hard to follow.

"Think of how fast the signal can make it through the hose ( circuit ) from one end to the other on demand. This is a measure of accelleration and is equivalent to the rise time." No, how fast it makes through the hose would be propogation delay. The signal could take all day to make it from one end to the other and still have a very fast rise time once it got to the output. The two are unrelated.

"Slew rate is equivalent to how much flow the hose ( circuit ) is capable of." You also refer to it as "volume of water." No, this would be the equivalent of electrical current.

Slew rate, by definition, is simply the maximum rate of change of the output voltage for all possible inputs. In simple terms, how fast the voltage can change.

Mathematically it is the slope of the line on a graph plotting voltage versus time. Take the amount the voltage has changed, divide it by the time it took to change, and you have the slew rate, usually expresed in volts per microsecond.

Since music is made up of sine waves, and higher frequency sine waves have faster rates of change (the fastest being where it crosses zero) at some frequency this rate of change will exceed the slew rate of the amplifier. It will distort this frequency and all of those higher.

Therefore, we can say that the bandwidth of an amplifier is slew rate limited.

Rise time is how long it takes the voltage to get from 10 to 90 percent of it's peak as I stated above. You are correct that there is also a fall time and it can be different.

I admit I may have oversimplified the relationship between risetime and slew rate. Even though you can calculate a rate of change based on the risetime it is not the same as the slew rate. It is possible that a circuit can have a high enough slew rate to handle a signal but still be unable to do so because it is limited by the rise time, and vice versa.

From a practical point of view it makes more sense to talk about rise time with digital circuits since they are always switching between 2 voltage levels. For instance, a zero represented by zero volts and a one represented by 5 volts. The rise time would be how long it takes to get from .5V to 4.5V. The falltime how long it takes to get from 4.5V to .5V. In an analog circuit that has an infinite number of different levels it probably makes more sense to describe it in terms of slew rate.

I could go into a long winded mathematical explanation of the two and show how they are both related to the RC time constants of a circuit, and therefore both are limiting the bandwidth of a circuit, but that is probably more than is needed and it is very difficult to do mathematical equations on this forum.