When selecting an operational amplifier, one of the more challenging selection criteria is the output current or load capability. Most parameters of an op amp are usually explicitly stated in the data sheet table, in the typical performance plots or in an application notes section. With output current, however, the designer may have to look harder. The designer also may have to combine various parameters to get the part to work as stated in the data sheet.
What makes the analysis more complicated is that various semiconductor manufacturers, or even different devices made by the same manufacturer, may specify output current quite differently from one other. This article explains how to predict output current from the device data sheet by considering an example. This will help designers make sure that the part they pick will have sufficient load drive under all conditions.
Predicting drive capability
Output drive is a function of internal and external settings or conditions. The output-stage bias current, drive level, architecture and process are internal factors. In addition to the internal factors, there are external factors that also influence the drive capability. Some of these factors can be manipulated for optimum output drive, and others are less controllable. External factors include output voltage headroom (voltage difference relative to supply rails) to the respective supply voltage; input overdrive; total supply voltage; dc- vs. ac-coupled load; and junction temperature.
The most common way to specify output drive is with the output short-circuit current parameter. The manufacturer specifies how much current is guaranteed to flow when the output is tied to ground. In a single-supply situation, the output is tied to one-half the supply voltage, called Vs/2.
Two numbers may be given, one for sourcing (usually prefixed "+") and one for sinking (usually prefixed "-"). The user can employ this number to predict the op amp behavior in applications where the voltage swing across the load is low, and therefore the output-stage drivers (internal to the op amp) maintain a large voltage headroom to the respective supply rails (V+ for sourcing current and V- for sinking current).
Consider a case where an op amp is tied to a heavy load driven with a voltage close to ground (or Vs/2 in a single-supply case). If this amplifier stage is then hit with a step change, the amount of current that would be available to the load is what the data sheet calls the "output short-circuit current."
Once the output starts changing in response to the step, two things occur: The voltage headroom on the op amp's output voltage is reduced, and the op amp's input overdrive is reduced. In turn, the available output current is reduced due to the first factor and, depending on the op amp design, there may also be a reduction in output current due to the second factor.
Another more useful method of specifying current capability is with the use of output current vs. output voltage plots. An example, taken from National's LMH6642, is shown in Figure 1. For most devices, there is usually one plot for current sourcing (Figure 1a) and one for current sinking (Figure 1b).
(Click to enlarge image)
Fig. 1: LMH6642 output characteristics for source and sink currents
With this form of plot, designers can assess how much current is available for a given amount of output swing. These plots are generated by the semiconductor manufacturers to show the variable output current capability of the amplifier as a function of output voltage.
The vertical axis is "VOUT from V+" for output-sourcing current and "VOUT from V-" for output-sinking current.
One reason for presenting the data this way, instead of displaying the output voltage relative to ground for both plots, is that these plots can be applied more easily to single- or dual-supply operation. Also, since output current is heavily influenced by how much voltage headroom there is, and less influenced by what the total supply voltage is, this way of presenting the data allows for rough calculation at any supply voltage using the closest possible set of supply voltage curves.
This is true even if the exact set of supply-voltage conditions is not present in the data sheet.
Figure 1 plots can be used to predict the amount of swing possible for a given load. If the axes were linear, one could easily extract this kind of information by superimposing a load line over the Figure 1 characteristics and looking for the intercept point. However, as discussed below, the axes are not linear but are log-log, in order to allow good resolution at low currents and with the output a few mill-volts from the supply rails. This is especially the case when dealing with rail-to-rail output op amps. However, plotting a load line over a log-log plot is difficult, as the load line will not be a simple line.
Predicting swing vs. load
With this challenge, how can designers predict the output swing for a given load? A fairly accurate prediction can be made by going through a couple of iterations of predicting swing, going back and forth between the device capability of the Figure 1 plots and external circuit requirements.
Some examples demonstrate how this is done. Figure 2 shows a LMH6642 op amp driving a load of RL = 100 Ω, tied to Vs/2 (one-half the supply voltage). Assume that the LMH6642 output is biased to Vs/2 (5 V).
Fig. 2: Example of op amp driving a modest load
Can you estimate the maximum output swing possible, using the LMH6642 data of Figure 1? The answer is yes.
To do so, construct a table (Table 1), which starts with an initial guess of the output swing (column 2) and then continues with successive corrections on that initial guess as dictated by column 6, (when columns 3 and 5 are compared).
(Click to enlarge image)
Table 1: Iteration used to predict output swing of Figure 2 (LMH6642)
This continues until the final result is reached at the bottom of the table in column 2, under conditions where the device characteristics coincide with the load requirement, and the final swing estimate is established.
From the Table, the iteration results show that the circuit of Figure 2 can swing up to 8.75 V across a 100 Ω load. This translates into a 7.5-VPP ((8.75-5)V x 2 = 7.5 VPP) waveform across the load.
Note these four points about the Table's construction:
1) For the circuit of Figure 2, output current is sourcing only. Therefore, only Figure 1a was used.
2) In each case, the worst-case temperature was assumed in determining the column 5 value from Figure 1.
3) Column 5 entries are values read off from the Figure 1a plot, with the y axis set to the column 4 value.
4) The final answer, in column 2, which was arrived at with the fourth iteration, is an approximation, since column 3 (87.5 mA) is still lower than column 5 (90 mA). However, the plot resolution does not allow further fine-tuning of the final answer.
By using the techniques and definitions in this article, users will be able to better understand how op amp output characteristics are measured and specified, even if the information is not readily available in the data sheet.
About the author
Hooman Hashemi is an application engineer who joined National Semiconductor Corp. in 1995. He has an MSEE from Santa Clara University (1989) and a BSEE from San Jose State University (1983). He currently works in National's Amplifier group.
