😕
Surely you should tell the manufacturers, starting from the triassic period of integrated circuits (the triassic part is arguable) that they've been using the term wrong. http://www.ti.com/lit/an/snoa852/snoa852.pdf
P.S. They're all interrelated, albeit overload recovery has a lot more going on than would ever be indicated in slew rate.
"it's how fast an amp recovers From input overload"
I don't think that is quite true, either! "Recovery" is not included.
As said the DUT is driven by a faster square wave that it can handle, the output becomes a ramp as (typically) Cdom is charged by he available current in the bias chain. This ramp slope is measured and becomes max slew rate, V/uS.
So a margin is required to be added to ensure that the amp is never driven into slew rate limit. I believe I am not just splitting hairs.
I don't think that is quite true, either! "Recovery" is not included.
As said the DUT is driven by a faster square wave that it can handle, the output becomes a ramp as (typically) Cdom is charged by he available current in the bias chain. This ramp slope is measured and becomes max slew rate, V/uS.
So a margin is required to be added to ensure that the amp is never driven into slew rate limit. I believe I am not just splitting hairs.
This is a fairly unclear understanding of all the issues. This way of looking at it will not give a clear picture of what happens in the continuum of input error signal vs output. First of all an LTP is never linear there is a transfer function of input voltage vs output current (tanh) which becomes progressively more non-linear as the total bias current is reached at that point it is maxed out. This current and the comp cap. give a maximum rate of change of output (slew rate). The frequency where you just reach this limit at for a full output sine wave is the industry standard definition of full power bandwidth.
Input overload recovery has a very different standard definition which usually is a measure of how badly, for instance, taking one input to the rail breaks internal bias circuits, etc. and the part has to recover. Looking at some data sheets will show pictures of parts going to sleep or going to the other rail for a few microseconds, etc. A square wave applied to a totally well behaved amplifier will make it slew at its slew rate to a final value. No where I have been is this considered input overload recovery.
Sure Scott - I meant commercially available units specifically. No doubt with a new design its possible.
Tons of. LMH5401 has a gain-bandwidth product of 8 GHz, slew rate is 17,500 V/µs and noise is 1.25nV/rtHz at flat band. As a side, distortions (HD2, HD3) are <96dB @100MHz.
Oh, I forgot to mention, it's a VFA. There.
P.S. High voltage (+/-16V) THS4022, 350MHz, 470V/uS, 1.5nV/rtHz. Also VFA.
Last edited:
But that's a differential RF amp for 50 Ohm systems with Ib noise up the wazoo.
THS4022 - Gain of 10 stable.
Last edited:
Are slew rates tested with a square wave? Then the diff input gets overloaded. Look at the voltage across the + and - inputs in the first usecond
And a tanh transfer function is very linear close to zero. And emitter degeneration broaden this linear range.
But neither are linear in the strict sense nor are JFET's, the distinction is important. A bipolar LTP is quite non-linear long before slewrate limit. Daniel posted the classic description from TI this is the way the industry looks at it.
Sorry about my confusion. This is what I learnt in my post graduate EE analog electronics course ( Grey and Meyers textbook) 30 years ago. I guess the definition of slew rate has become analogous to full power bandwidth.
Please read the link I provided. It addresses just that, and as Scott says (and he should know better than, well, pretty much anyone 🙂 ) this is the industry standard.
I know about some of those. They are not designed for audio - I see Scott has already answered.
I do wish you would stop coming back back on me about VFA vs CFA. As I have pointed out numerous times before, I have never been pro either over the other.
Tiresome. But then again this one dimensional approach is to be expected from your sort.
Nice try deflecting. So can you give three clues and three pointers about what is and what should do an opamp "designed for audio"?
I'm sure the semi industry will remain forever indebted to you.
- Status
- Not open for further replies.