Blog Signal Chain Basics

Signal Chain Basics #92: The Insidious Claw Curve

Operational amplifier (op amp) datasheets normally have a long list of characterization curves. This list can be overwhelming to new engineers. One of the frequently misunderstood curves is nicknamed the “claw curve” because of its distinctive claw-like shape. (See Figure 1.)

Figure 1

The insidious claw curve.

The insidious claw curve.

The claw curve relates output voltage swing limitation to output current and temperature. The curve itself represents the op amp output voltage when the output is saturated. For example, Figure 2 shows that, at 8 mA, the op amp output swing is limited to about 1.5 V from the positive rail at 125°C and about 0.5 V from the rail at –40°C. Thus, for a 15V supply, the op amp could output 13.5 V at 125°C and 14.5V at –40°C.

Figure 2

Output voltage saturation limit for OPA188, also known as the claw curve.

Output voltage saturation limit for OPA188, also known as the claw curve.

The claw curve shows the output saturation limitation over a wide range of load currents and temperatures. It is important to understand that this is a typical curve, and the actual results will vary from process variation. The specification table will list this limit for a few key loads. (See Table 1.) Notice that the specification table also lists a maximum specification.

Table 1

Output swing from rail for the OPA188 (saturation limit); Vs = ±4V to ±18V.

Output swing from rail for the OPA188 (saturation limit); Vs = ±4V to ±18V.

The output limitations seem pretty simple, so what is so insidious about the claw curve? The problem is that many engineers think that you can get linear swing all the way to the curve. In our example, many engineers assume that you can get a linear output swing of ±13.5 V at 8 mA and 150°C. However, this is not the case. The amplifier starts to become nonlinear as its output approaches the saturation limitation (the claw curve). In many cases, engineers need to know what the linear output range is. To find the linear range, you need to look at the OPEN LOOP GAIN (AOL ) specification. (See Table 2.)

High AOL (AOL > 120 dB) is needed to maintain the best linearity. Notice that the test conditions for the AOL specification limit the output swing to 0.5 V away from the power supply rail. Also, notice that the load resistance is 10 k Ω, and the supply can be as high as ±18 V. This translates to a load current of 1.8 mA. Thus, for a 1.8mA load, you can get linear response up to 0.5 V from the rail, because AOL is greater than 120 dB for this condition. Looking back at the claw curve, it is difficult to read the exact output swing limitation, which is wider than the linear swing range.

Comparing Tables 1 and 2, notice that the room temperature swing to the rail limitation is 250 mV for saturation (Table 1) but 500 mV for linear operation (Table 2). This means that the output can be driven as close as 250 mV from the rail but will be nonlinear from 500 mV to 250 mV from the rail due to decreased AOL in this region.

Table 2

AOL specification for the OPA188 showing output swing condition; Vs = ±4V to ±18V.

AOL specification for the OPA188 showing output swing condition; Vs = ±4V to ±18V.

Thus, it is important to understand the difference between linear output swing and output saturation limitations. The linear output swing is inferred from the AOL specification. The claw curve and the voltage swing from the rail specification, given in the specification table, both represent the output swing limitation due to saturation. The saturation limitation will always be wider than the linear range. While the claw curve provides useful information, make sure that you don't fall victim to the insidious talons of nonlinearity.

Special thanks to Marek Lis for his insight into op amp design.

Please join us next time, when we will discuss the advantages of using a current shunt amplifier in low-side current measurement applications.

For more details on these specifications, download the OPA188 datasheet.

— Arthur Kay is an applications engineering manager at TI, where he specializes in the support of amplifiers, references, and mixed signal devices. He focuses a good deal on industrial applications such as bridge sensor signal conditioning. He has published a book and an article series on amplifier noise. He received his MSEE from Georgia Institute of Technology and his BSEE from Cleveland State University.

1 comment on “Signal Chain Basics #92: The Insidious Claw Curve

  1. onthejob
    August 14, 2014

    Enjoyed this note.  Informative and concise.  Thanks.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.