The concept of an available slew rate in op amps emerged from the earliest developments to describe the maximum large signal transition rate for a step response. This simple concept has become more challenging with today’s modern slew enhanced devices.

While slew rate is one of the more important characteristics in high speed op amps and Fully Differential Amplifiers (FDAs), it has become a relatively confusing specification as different development groups treat it in slightly different ways. There are two very important points to establish as we move to a better understanding with modern high speed slew enhanced amplifiers.

  • Normally, the rated slew rate is an upper limit that the application should strive not to exceed. The application configuration can influence the physical slew rate delivered. Most device data sheets try to report a single number, but this will be only approximately accurate across different applications (and from device to device and across temperature and supply voltage). Given this variability by application and across temperature, most design flows guard band the required slew rate somewhat higher than an exact analysis would suggest. Detailed slew rates by configuration and across temperature and supply are possible, but the industry wide trend is to report a single number–usually different if more than one specification table is provided at different supply voltages.
  • The classical definition of slew rate in high open loop gain op amp type products and still an important distinction, is true slew rate limiting comes along with the feedback loop opening up as the output cannot keep up with the signal demands. Driving larger and larger steps at the output will transition from a linear response shape, to a partially slew limited response, to steps that show a large region of fixed dV/dT on the transition.

 
F1 slew rate op amp(cr) Figure 1: Diagram showing both a linear region and a true slew limited region.  

Slew rate limits can be seen in both step response and large signal bandwidth (LSBW) curves. Having an approximately accurate number will then allow different applications to stay well below slew limited operation if linear operation is desired.

Many modern high speed amplifiers include a slew enhancement feature. A great deal of the last 30 years of high speed op amp development has been to incorporate better and better “slew rate on demand” type circuits that lie dormant until the need arises. The trick is to transition into and out of the slew rate enhancing regions with minimal disturbance in the waveforms. Slew enhancement offers improved power efficiency while still offering higher LSBW. Slew enhanced devices still reach a slew rate limit–just much higher than the quiescent power might suggest. It is important to note that slew enhanced devices do draw higher dynamic supply currents as their boost features are called on to operate.

One of the most successful slew enhancement approaches first appeared in the Current Feedback Amplifier (CFA’s) designs pioneered by Comlinear in the early 1980’s and widely available today from numerous suppliers in more recent devices like the OPA695. Improved slew enhanced Voltage Feedback Amplifiers (VFAs) followed closely with the “high transconductance input stage” featured in parts like the OPA890. More recently, FDAs have captured the fancy of op amp development groups and end system designers alike. These also come in CFA flavors (like the LMH6554) and, more recently, slew enhanced voltage feedback precision FDA’s like the THS4541.

 
Classical slew rate response (cr) Figure 2: Usual slew rate response on a modern precision op amp.  

Nearly all high speed amplifiers are characterised with some degree of peaking in their small signal frequency response shape. This peaking is a direct indication that the closed loop response is operating with complex pole pairs. That small signal frequency response shape imposes a 2nd order step response shape on the output. The output time waveform will attempt to follow that 2nd order shape until the instantaneous (not average) dV/dT exceeds the available slew rate in the device.

Going beyond simple gain stages, applying any op amp to multi-stage low pass active filter design will by definition be imposing 2nd order step responses on the individual stages. So even if a lower speed op amp operating at high gains might have a simple 1st order response, using that stage inside a low pass, DC coupled, active filter design will change the overall output step response back to 2nd order where the analysis here will apply. This linear step response analysis going from the small signal frequency response shape to output step waveform applies to any op amp or FDA either slew boosted or not.

Michael Steffes works as a Senior Member of Technical staff in the definition and development of signal path components supporting precision and high speed data converters for Texas Instruments.

First published by EDN.