Clarification on Gain Bandwidth vs Slew Rate

Thread Starter

Swarly88

Joined Nov 12, 2018
25
Greetings,

I am finding it difficult to answers regarding Gain Bandwidth. Specifically, WHY an unity gain amplifier, for example, will begin to drive its output to zero at higher frequencies. The math is straight-foreword, but the specifics as to why are either not covered, or I'll get conflicting answers (ie capacitance at high frequencies behave more like inductors/shorts / less voltage across load resistors which means less voltage read on a scope / etc). Additionally, the behavior of gain bandwidth vs slew rate sound very similar: the performance of the amplifier is limit by the frequency that you input to it. Would anyone be able to provide some clarification or point me in the direction of some documentation that I could read up on?
 

Papabravo

Joined Feb 24, 2006
12,500
In both cases the output cannot "follow" the input. The logical conclusion of increasing the input frequency to very high values is that the amplifier can only respond at some set of lower frequencies. This lack of ability to respond means there is a maximum frequency, beyond which we no longer have an amplifier; the circuit has become something else. In this realm concepts like gain and slew rate lose their connection to the behavior of the output with respect to the input. It happens for a variety of reasons which you have already alluded to all happening at the same time. It is also similar to the behavior of transmission lines where the distributed nature of properties like capacitance and inductance change with increasing frequency.

If you have access to a VNA, you can sweep devices and circuits to see what happens as you increase the frequency to nosebleed levels. It is truly mind bending.
 

Thread Starter

Swarly88

Joined Nov 12, 2018
25
We performed a lab where we built a unity gain filter and measured the output at 5k Hz, 50k Hz, 500k Hz, and 5 MHz. Around 500k Hz is where I first noticed the output was shifting more than 180 degrees out of phase of the input, with the 5 MHz signal ultimately driving the output to zero. Part of our write up includes talking about gain bandwidth and slew rate but those concepts were starting to blur together for me.
 

crutschow

Joined Mar 14, 2008
23,506
If you are referring to an op amp, the gain-bandwidth product is determined by the internal compensation which typically rolls off the open-loop gain with a single low-frequency pole at 6dB/octave.
The frequency where this rolloff reaches 0dB gain is defined as the gain-bandwidth value.
This is a small-signal value and does not include any large-signal effects, such as slew-rate.

The slew rate (given in maximum V/s of output signal change) is not directly related to this gain-bandwidth, but is determined by how fast the internal modes can follow a large signal change.
Thus the amp is limited by the signal amplitude it can generate at higher frequencies, even if it's below the gain-bandwidth limit.
 

Thread Starter

Swarly88

Joined Nov 12, 2018
25
Is part of your grade dependent on you being able to locate sources that discuss those concepts?
Oh no. I wouldn't have asked in such a way if it was. If anything, this level of detail isn't required and I'm doing it for myself.
 

MrAl

Joined Jun 17, 2014
6,614
Oh no. I wouldn't have asked in such a way if it was. If anything, this level of detail isn't required and I'm doing it for myself.
Hi,

Gain is a more linear concept where it decreases with increasing frequency mostly because of parasitic resistor-capacitor low pass filter actions and of course a low pass filter gain drops off at some higher frequency.

Slew rate is a more non linear concept that comes about because of the inability of a device to respond instantaneously to an input signal. It is relatively constant affecting all frequencies but is more noticeable at higher frequencies. It can be roughly through of as a delay.

So theoretically, combine a low pass filter with a delay network and you have something like a real op amp.
 

danadak

Joined Mar 10, 2018
3,611
Limitation of bandwidth is rooted in electron mobility, how fast electrons can
follow a signals rate of change. This is generally speaking a small signal,
< 100 mV, environment.

upload_2019-2-5_7-36-48.png

https://en.wikipedia.org/wiki/Transistor

Slew rate is largely predicated by available current to charge the internal compensation
capacitor. This is a large signal issue.

upload_2019-2-5_7-41-16.png

https://training.ti.com/ti-precision-labs-op-amps-slew-rate-1

These limitations are physical in nature. Mobility, geometry......

More than you want to know on this topic -

https://ecee.colorado.edu/~bart/book/book/chapter5/ch5_6.htm



Regards, Dana.
 
Last edited:
Top