What results in good transient response?

[SashaV]

Is it high slew rate, and/or low/no negative feedback, or something else?

[ticknpop]

When i tip the transient washing my windshield at an intersection a dollar and he says thank you I consider that to be good transient response.
I would also consider that to be a high feedback situation.

[Bonsai]

High slew rate, resulting in very fast rise and fall times with no overshoot. You do not need wide open loop bandwidh to accomplish this - you need high LTP tail current with substantial degeneration and a low value Cdom capacitor (here I talk about conventional Lin topology amplifers or fully balanced symmetrical types utilizing conventional Miller compensation). There are many opinions as to what constitutes an acceptable slew rate, but I would venture to say for a 100W amp something equal to or greater than 40V/us is the minimum - a good example is OStrippers class AB amp which you can see at the top of the solid state forum. There are designs out there that do 300V/us, but these are fairly unusual, and typically do not use Miller compensation (See Bob Cordell's Mosfet design for example). I have a design on my website that does around 100V/us - its a 250W symmetrical amplifier.

[DF96]

Yes, high current and low capacitance. Stability helps too.

[jcx]

no commercial music audio recording will have greater than 50 kHz Fc - thats the fastest recording mic sold by Earthworks

1 V / ( 2 * pi * 50 kHz ) = 0.157 V/us

that gives 6.3 V/us as an estimate of what is required for 100W of audio into 8 Ohms (40 Vpk)



testing amps with faster transient can tell you some things, for some compenstion schemes the overshoot, ringing giveas an indication of loop stability


again depending on toplogy, slew rate limit in excees of Audio requirements may give an indication of the level some types of distortion

but it long ago became a numbers race disconnected from real Audio signal requirements or distortion concerns

[Elvee]

Originally Posted by jcx

no commercial music audio recording will have greater than 50 kHz Fc - thats the fastest recording mic sold by Earthworks

1 V / ( 2 * pi * 50 kHz ) = 0.157 V/us

that gives 6.3 V/us as an estimate of what is required for 100W of audio into 8 Ohms (40 Vpk)



testing amps with faster transient can tell you some things, for some compenstion schemes the overshoot, ringing giveas an indication of loop stability


again depending on toplogy, slew rate limit in excees of Audio requirements may give an indication of the level some types of distortion

but it long ago became a numbers race disconnected from real Audio signal requirements or distortion concerns

The voice of reason.... unlikely to be heard, for that very reason (pun!).

In addition, before tackling the large signal response (SR, etc) the small signal response has to be spotless. Including under difficult loads.
Not very glamorous, and therefore often neglected.

[Ron AKA]

The minimum slew rate you need depends on the wattage of the amp, and the impedance of the speaker. More watts needs higher slew rate, and the same with impedance. Higher requires higher slew rate. The article at the link below provides a good explanation. The good news is that most amps have more than enough -- sometimes by a factor of 10.

Slew Rate - Technical Note

It would seem to make sense to have a slew rate that is high enough, but not so high that stability is an issue.

[Bonsai]

Yes JCX, I am aware of those calculations that you show.

However, I don't think designing an amp that just meets the minimum slew rate is necessarily right. After all, a 10 horesepower car is ample to get you from A to B . . .

[Michael Chua]

Originally Posted by Ron AKA

The minimum slew rate you need depends on the wattage of the amp, and the impedance of the speaker.
Slew Rate - Technical Note

From Technical Note.

[jcx]

just tossing some high numbers on the table without giving the actual calculation, requirements, assumptions behind them isn't too useful either

as I said, and as needs emphasizing for non-experts looking at Audiophile press commentary on slew rate - it can be a "numbers race" divorced from signal or circuit requirements

naive op amp rollers may use >1000 V/us CFA in a position designed for a fet input op amp, wax poetic about its improved sonic characteristics - when it is totally inappropriate for their circuit application