Slew rate calculations - how much do I need?

[peranders]

I start a new thread which came from here

So, mr. amp_man, how do you calculate the slew rate you will need?

Let's say, 500 watts out, 4 ohms, 50 kHz bandwidth at least. ( I know how to calculate slew rate)

SR= 2*pi*f*SQR(2*P*R) = 2*3.14*50000*SQR(2*4*500) = 19 V/us = absolute minimum. Some overhead is wanted, let's say 25-50 V/us is what you need.

Here is good link for this:
http://hyperphysics.phy-astr.gsu.edu...ic/a741p3.html

amp_man says:

But our Modern DJ's are somewhat more funky towards Scratch effecs and Pitch control, therefore it is necessary to have high Slew Rate in order to have full dynamic transient attack at the ouput.

Somewhere here I got lost. From a CD you can get out 20 kHz and in this bandwidth you have all attacks, transients etc. Absolutely zero above 22 kHz contains music. For a vinylplayer I'll gather that the limit is slighly above. Can a pickup really produce frequencys above 30 kHz?

amp_man says:

High Slew rate increases Damping Factor

From what I know damping factor is output impedance in relation to the speaker impedance. What has slew rate got to do with output impedance?

My conclusion so far is the formula above is valid.

[Nelson Pass]

I believe your numbers are good.

In actual testing of audio, you do not get such high numbers.

The best I have ever seen is 5 V/uS with a vinyl copy of the
Sheffield Drum record just short of clipping a 100 watt amp.
This would translate to about 11-12 V/uS on your larger amp.

Keep in mind that this was the highest I could find at the time,
and .5 V/uS was a far more typical peak, even when you thought
there was a lot of top end. Peter Walker got similar results in
his measurements at Quad.

The question inevitably boils down to how much margin do you
want? Last time I looked John Curl advocated 10 times. For a
complex circuit with ordinary amounts of feedback, I would agree,
but for simple circuits with low feedback, I think the margin can
be less.

[peranders]

I don't have taken into account how much additional speed you must have in order to get distortion figures low. I'll imagine that it's not good to be very near slew rate limiting when it comes to all sorts of distortions.

I'll have serious doubts though that a good PA amp must have SR far beyond 100 V/us. Isn't this asking for trouble in a professional enviroment?

[sam9]

Isn't there also an "internal" slew rate which is a function of tail current and Cdom? I'm real unclear on this. Is this related to above posts or is it an other matter? If related how is it related? Etc, stc, etc (as Yul Brenner once said).

[peranders]

This slew rate you mention is the actual maximum speed of the amp.

[sam9]

I gather, then, that it must always be greater than or equal to what can be directly measured at the output? Other than being an upper limit to what can be measured at the output does it have any additional significance? That is, is there any advantage if it is, for example, twice that achievable at the output versus 1-1/2 times?

[Upupa Epops]

By my opinion we must look at this question globaly. Amp must have also good settling time and fast recovery time - without this is high slew rate for nothing, sound will be not good.

[mikeks]

Originally posted by Nelson Pass
I believe your numbers are good.

In actual testing of audio, you do not get such high numbers.

The best I have ever seen is 5 V/uS with a vinyl copy of the
Sheffield Drum record just short of clipping a 100 watt amp.
This would translate to about 11-12 V/uS on your larger amp.

Keep in mind that this was the highest I could find at the time,
and .5 V/uS was a far more typical peak, even when you thought
there was a lot of top end. Peter Walker got similar results in
his measurements at Quad.

The question inevitably boils down to how much margin do you
want? Last time I looked John Curl advocated 10 times. For a
complex circuit with ordinary amounts of feedback, I would agree,
but for simple circuits with low feedback, I think the margin can
be less.

High slew rates are not required soley for processing audio frequency signals, but to reduce the likelihood of slew overload in the amp. due to RF ingress.....

An amplifier driven to it's slew limit by such spureae is unlikely to simultaneously process audio frequency stimuli......

[UrSv]

Just throwing in some specs to cover the statement:

"The normal Signal transient require a slew rate of no more than 50V/us.
But to Produce fast transient Signal at high Accuracy at especially High Voltage and Current level that is I mean to say that For High Sound Pressure level the high slew rate requirement is evident to produce high voltage transients. Therefore a clear Factor is to increase the normal slew rate to about 3 to 4 times in order to create a signal transient of high speed at high Voltage thereby producing High SPL which is requirement of Large PA's and are catered by Professional equipment only."

Some specs:

#1 Labbgruppen fP6400 at 1200 W stereo in 8 Ohms with a slew rate of 20 V/us
#2 Peavey CS3000H at 700 W stereo in 8 Ohms with a slew rate of ">15 V/us"
#3 Crest Pro 9200 at 1300 W stereo in 8 Ohms with sleq rate ">10 V/us"

Crown, AFAICT, no numbers which speaks against the quote:

"Every high end Pro amp such as Peavey, Crown, Crest, etc Quote the slewrate of their amps so that the customer feels a better possession of technical specs."

With the numbers above I can only conclude that since these are high power amps and the slew rates are much lower than needed according to amp_man_1 then they will all fail to reproduce music with high quality and be unsuitable to Pro use. I never would've guessed given that they are by most people regarded as being pro amps...

[UrSv]

Originally posted by mikeks



High slew rates are not required soley for processing audio frequency signals, but to reduce the likelihood of slew overload in the amp. due to RF ingress.....

An amplifier driven to it's slew limit by such spureae is unlikely to simultaneously process audio frequency stimuli......

Would not most pro amps, and "hifi amps" for that matter, be protected from RF by input filters usually so that this should pose less of a problem?