# Is it good to have a higher slew rate? What are some of the effects due to slew rate limitations?

This question is in relation to the working of an op-amp. I know what slew rate is, but I am not entirely sure whether is it good to have a high slew rate.

As far as my understanding goes, a higher slew rate should allow the op-amp to give the output as per the input without time delay. Is my understanding correct?

What are the effects due to slew rate limitations?

• Slew rate does not depend on frequency alone. It depends on voltage as well. – Plutonium smuggler Nov 13 '14 at 4:38
• Yes, a high slewrate can more faithfully reproduce the signal. Higher slewrate components costs more so one tries to get away with as little as possible. High slewrate components will also require more supply current and larger capacitors and cause noise in surrounding components. – KalleMP Nov 13 '14 at 8:38
• High slew rate is not always a good thing. By fourier a perfect square wave requires gain at all frequencies - clearly impossible - but high slew rate amplifiers will have more high frequency gain which may cause stability problems and drive the need for careful layout. – N.G. near Aug 10 '16 at 11:44

Slew rate limits will cause distortion at high output frequency and amplitude. If your amp is slew rate limited, putting in a periodic waveform (sine, square, etc.) will result in something that looks a bit like a sawtooth wave. This can create frequency harmonics that are not present in the original signal, especially when the source signal is a pure sinewave. Generally, you need to have a high enough slew rate for the highest frequency and output voltage that your circuit needs to support.

The slew rate is another term for slope. The maximum slope of a sinewave is equal to its amplitude times its angular frequency (derivative of $A \sin(2 \pi f t)$ at the zero crossing at $t = 0$ is $2 \pi f A$). So a 1 MHz signal at 1 V amplitude (2 V peak to peak) has a maximum slope of $2\pi \times 1 \text{V} \times 1 \text{MHz} = 6.28 \text{V} / \mu \text{s}$. If your amp has a slew rate of less than $6.28 \text{V} / \mu \text{s}$, then you will get a triangle wave if you try to get it to output a 1 MHz 1 V sinewave. Note that the slew rate has to do with the output voltage of the op amp, not the gain. That being said, it usually affects high gain circuits more because the signals tend to be larger.

In op-amps, slew rate and bandwidth tend to be linked - high speed op amps tend to have fast slew rates, otherwise they wouldn't be very useful. Fast slew rates will allow an op amp to overshoot or ring at a larger output swing than an op amp with equivalent bandwidth but a slower slew rate. Slower slew rates can help limit overshoot and ringing in many cases. Another thing to consider is the power supply - the output current has to come from somewhere. Very fast slew rate op amps require a very low impedance power supply. This may require placing multiple capacitors of different values very close to the op amp - generally a combination of large, bulk capacitance and small, high frequency bypass caps .

Slew rate limitations can be helpful for reducing the harmonic content of digital signals. Some devices have a propensity for producing very fast edges (e.g. FPGAs) which, while necessary for high bandwidth communications, can cause problems with lower speed communications. Fast edges can couple to adjacent traces and can cause crosstalk and intersymbol interference. Limiting the slew rate can mitigate this. Transmitting serial data over a limited bandwidth (e.g. for an RF link) also takes advantage of slew rate limiting to confine the bandwidth of the signal.

• Yes, a limited slewrate will cause distortion at high amplitudes and freq but it would I think be more characterised by lowpass filtering rather than harmonic generation which would need even higher slewrates to become significant. Slewrate limiting is used to prevent harmonics in communications lines. – KalleMP Nov 13 '14 at 8:41
• Well, it depends on what signal you're talking about. A perfect sinewave has no harmonics, but a sinewave sent through a slew rate limited amp will gain some harmonics due to the distortion. In communications systems, the starting wave is a square wave that can have harmonics up to incredibly high frequencies due to the slope of the edges. Slew rate limiting makes it look more like a sinewave and as a result attenuates some of those harmonics. – alex.forencich Nov 13 '14 at 9:34

There are several problems that may stem from having "too much" slew rate:

• Slew rate correlates loosely with op-amp bandwidth, so using an op-amp with a much higher slew rate than is actually required means you're making your circuit sensitive to things it doesn't need to be sensitive to.

• An op-amp with a high slew rate is more likely to be susceptible to ringing. You might have to compensate the circuit to fix this.

• Really fast op-amps often don't like being run at unity gain.

Some op-amp datasheets will come right out and tell you this. An example is the OPA227 and OPA228. The OPA228 is about 4× faster, but is only stable in gains of 5 or above. The OPA227 has a phase lead cap inside that limits its bandwidth, allowing it to be unity-gain stable.

Sometimes the op-amp datasheet hides this fact, such as with the AD8397. Its datasheet tells you it's "unity gain stable" on page 1, but then you dig into the details and find the first graph on page 9 which shows peaking in the bandwidth vs gain curve at unity gain. This effectively amounts to positive feedback, which means all you need is a stimulus at the peaking frequency to have a good chance at creating an oscillator. You can end up with a circuit that works fine on your workbench but fails elsewhere due to a different RFI environment.

You mostly care about the slew rate, when the output voltage is large. (Several volts) At lower amplitudes you'll care more about the GBW product. Some opamps will quote the full power bandwidth, the BW at maximum output amplitude. In general this will be determined by the slew rate.

Op amps can be used for many purposes. Generally, one will want to have a slew rate which is fast enough that the op amp will never be slew-rate limited while processing a "continuous" AC signal. On the other hand, if an op amp will be used to process a discontinuous signal which represents a number of DC levels in sequence, the output of the op amp will be sampled some time after the input changes, a slew rate which is slow but still sufficient for the output to reach the required level before it is sampled may reduce the likelihood of overshoot compared with a faster slew rate.

Another way of looking at things is to say that if the input to an op amp will "naturally" be free of transitions that would be sharper than what is wanted or needed in the output, then one should use an op amp whose slew rate is at least as fast as the maximum slew rate that will be commanded from the input. If, however, the input may contain extremely sharp transitions and the output doesn't need to reproduce them, using a slew-rate-limited amplifier may, "for free", help to reduce the sharpness of the transitions in the output and the ringing or other nastiness that such sharpness might cause. Use of an op amp with a higher slew rate than needed may reduce the amount of help that provides.