NE5532 Output Current and Load Drive: What the Dual Op Amp Can Actually Deliver

The NE5532 is a dual bipolar-operational amplifier from Texas Instruments, internally compensated for unity-gain stability, with a typical unity-gain bandwidth of 12 MHz, a slew rate of 5 V/µs, and equivalent input noise of 5 nV/√Hz typ at 1 kHz. It runs on supplies from ±5 V to ±15 V, is characterized at ±15 V and 25 °C, and is built for low-noise signal conditioning and audio-grade gain, buffer, and balanced-line-driver duty — including driving 600 Ω-class loads, which its large-signal gain specs explicitly cover. Each package contains two amplifiers, needs no external compensation, and powers on as soon as the rails are present.

Here's the myth worth killing first: "the NE5532 datasheet numbers everyone has memorized are still correct." They aren't. The current datasheet (Rev. K) revised several headline specs — unity-gain bandwidth changed from 10 to 12 MHz, slew rate from 9 to 5 V/µs, and typical supply current from 8 to 6 mA — and the revision history also records the absolute-maximum supply-voltage values changing from 22 V to 18 V, even though the CAUTION text in the same document still warns about supplies outside ±22 V. Worse, the document contradicts itself internally: the prose in the Feature Description still says "10 MHz" and "9 V/ms," while the normative spec tables say 12 MHz and 5 V/µs. Trust the tables, not the prose, and design to the table values. If you're quoting the NE5532 output current from memory, note that the relevant verified figure here is the output short-circuit current of 38 mA typ — a protection limit, not a load-current spec.

What the NE5532 is actually for

Read the datasheet's own framing and a clear profile emerges. It is a general-purpose high-performance dual op amp whose defining features are very low noise, high output-drive capability, high unity-gain and maximum-output-swing bandwidths, low distortion, and high slew rate. The intended roles that fall out of that:

• Low-noise gain stages and audio front ends. The 5 nV/√Hz typ voltage noise at 1 kHz (8 nV/√Hz typ at 30 Hz) is the headline, and the datasheet specifies maximum limits for equivalent input noise voltage — a guarantee, not a marketing adjective.
• Buffers. The part is internally compensated for unity-gain operation, so gain = 1 needs no external compensation components.
• Balanced/differential line drivers. The only worked typical application uses both halves to produce VOUT+ and VOUT−, with a differential output range of 2 × VREF and a common-mode level of VREF/2. That's the manufacturer's intended pairing of this dual.
• 600 Ω-class loads. Large-signal gain is specified at RL ≥ 600 Ω (50 V/mV typ) and RL ≥ 2 kΩ (100 V/mV typ) at VO = ±10 V — a load spec most op amps never even state.

Equally important is what it is not for. Nothing in the datasheet targets micropower, rail-to-rail, or precision-instrumentation duty. Total supply current is 6 mA typ / 16 mA max for both halves — this is a bipolar workhorse, not a battery-sipping part. The input is not rail-to-rail, the output is not rail-to-rail, and the NE5532/NE5532A grade is rated only from 0 °C to 70 °C (the SA5532/SA5532A cover −40 °C to 85 °C). If your board can see sub-zero temperatures, the suffix choice is a real design decision, not a datasheet footnote.

Wiring the NE5532 into your design

Integration is straightforward, but a few datasheet-specific rules matter:

• Supply range and characterization point. Recommended operation is ±5 V to ±15 V, but every electrical characteristic is measured at ±15 V, 25 °C. Supply-voltage rejection is only specified over ±9 V to ±15 V (80 dB min / 100 dB typ), so below ±9 V the part runs but the datasheet guarantees no PSRR. Single-supply operation is explicitly supported; a +10 V single rail with VCC− grounded is the symmetric equivalent of ±5 V, sitting exactly at the floor.
• Decoupling. Place a 0.1 µF low-ESR ceramic bypass capacitor between each supply pin and ground, physically close to the device. The datasheet warns that noisy or high-impedance power supplies may cause erratic operation and repeats this placement requirement. In single-supply use, one bypass from V+ is acceptable. Beyond the ceramics, a bulk electrolytic at the board entry is reasonable — bypass sizing isn't driven by the part, which draws only 6 mA typ.
• Impedance levels. This is the easiest thing to get wrong. Input resistance is 30 kΩ min / 300 kΩ typ — unusually low for an op amp — and the BJT input stage draws 200 nA typ / 800 nA max bias current at 25 °C (1 µA max over full temperature), with input offset current of 10 nA typ / 150 nA max. At a 10 kΩ source impedance, bias current alone produces roughly 2 mV of error. Keep source and feedback resistances in the hundreds-of-ohms to low-kΩ range, and match the impedance seen by both inputs so bias currents cancel. The datasheet's own application section says: if noise matters, use resistor values of 6 kΩ or lower so resistor thermal noise stays below the amplifier's noise.
• Common-mode window. Common-mode input range is ±12 V min / ±13 V typ at ±15 V supplies. That's roughly 2–3 V of headroom you must preserve, and it shrinks as rails drop. Do not swing an input near either rail.
• Output loading. Drive loads of 600 Ω or higher. The output is not rail-to-rail: the rated swing figure is ±10 V at the specified loads, so budget headroom accordingly. Output short-circuit current is 38 mA typ — that's the protection limit, and your continuous load current should sit comfortably below it.
• Layout. The datasheet asks for external components placed as close to the device as possible (feedback resistors near the inverting input to minimize parasitic capacitance), short input traces, and a driven low-impedance guard ring around critical traces to reduce leakage. Follow those plus the 0.1 µF placement and the surrounding components will perform as specified.
• Thermals. For the D (SOIC-8) package, RθJA is 97 °C/W. Maximum power dissipation follows the datasheet's formula P_D = (T_J(max) − T_A)/θ_JA with T_J(max) = 150 °C — about 1.29 W at 25 °C ambient. At ±15 V with a 600 Ω load, typical total dissipation is roughly 0.27 W, putting the junction near 51 °C at room ambient — thermal is not your binding constraint on this part.

What to watch out for on the NE5532

• Absolute-maximum supply confusion. The revision history moved supply abs-max values from 22 V to 18 V, while the CAUTION text still references ±22 V as the permanent-damage threshold. Trust the current abs-max table you're designing against, and keep rails well inside it.
• The "A" suffix doesn't buy noise or slew. In this revision's operating characteristics table, the NE5532/SA5532 and NE5532A/SA5532A columns are identical for slew rate and both noise specs. Any A-grade improvement lies elsewhere — don't pay for phantom noise performance.
• Slew, not bandwidth, sets your large-signal ceiling. At 5 V/µs, a ±10 V output swing gives a full-power bandwidth of only about 80 kHz. Above that, distortion comes from slew limiting, not the 12 MHz small-signal bandwidth.
• No THD number exists in this revision. "Low distortion" is a qualitative feature statement; Rev. K removed the distortion-adjacent rows along with the peak-to-peak output swing, output-swing bandwidth, output impedance, and small-signal gain rows. Don't quote a THD+N figure from this datasheet, and don't expect a swing number at ±5 V — the parameter was deleted.
• Protection is built in, but bounded. Input-protection diodes and output short-circuit protection are internal, so brief input overvoltage and a momentarily shorted output are survivable. That is not license for sustained out-of-rail input drive — the datasheet specifies no diode current limit, so add a series input resistor if your source can exceed the rails.
• No enable pin. The part is active whenever rails are present; there is no shutdown and no sequencing requirement. If you need a muted output, that's external circuitry.
• One caveat on scope: this revision's absolute-maximum input-voltage and differential-input values were not confirmed against the spec tables, so treat any specific input abs-max number you find elsewhere as unverified for this document.

Testing the NE5532 on your board

• Supply check first. Confirm both rails are within ±5 V to ±15 V and measure total quiescent current with inputs at ground and no load: expect 6 mA typ; 16 mA is the max. A dead-short or oscillating part often shows up here first.
• DC operating point. With both inputs tied to mid-supply through matched low-value resistors, output offset should reflect the 0.5 mV typ / 4 mV max input offset voltage times your noise gain. If offset scales with source impedance rather than gain, bias current (200 nA typ) times an unbalanced impedance is your culprit.
• Small-signal bandwidth. Drive a small signal (well under the slew limit), measure closed-loop bandwidth, and sanity-check it against 12 MHz unity-gain bandwidth divided by your noise gain.
• Large-signal slew. Apply a square wave large enough to slam the output toward ±10 V and measure the edge slope; it should land near 5 V/µs. Ringing or a ramp much slower than expected indicates slew limiting, instability, or an overloaded output.
• Load test. Verify behavior into your real load, and if it's a line-driver application, confirm it stays within the RL ≥ 600 Ω region the gain specs assume. Check output swing against the ±10 V window rather than the rails.
• Noise floor. With input shorted through a low-value resistor, the measured output noise divided by your gain should be consistent with 5 nV/√Hz at 1 kHz. Excess noise usually means feedback resistors above ~6 kΩ are dominating — the resistor, not the amp.
• Thermal sanity. After warm-up at full drive into 600 Ω, case temperature should correspond to roughly 0.27 W of dissipation through 97 °C/W (SOIC-8). Significantly more indicates oscillation or a shorted load.

Troubleshooting the NE5532

• Output stuck near a rail. Check the common-mode window first. An input driven within 2–3 V of either rail at ±15 V supplies violates the ±12 V min input range and the amp stops behaving linearly. Then verify the feedback path is actually closed — a missing or cold solder joint on the feedback resistor leaves an open-loop comparator.
• Oscillation. Revisit decoupling: the 0.1 µF ceramics must sit at the supply pins, not inches away, and the datasheet explicitly blames noisy or high-impedance supplies for erratic operation. Also check for capacitive loads and long unterminated output traces; keep feedback resistors physically near the inverting input to minimize parasitic capacitance.
• Excessive DC offset. Compute bias current (200 nA typ / 800 nA max) times the source impedance seen at each input. Unbalanced impedances convert matching into offset. Add a compensation resistor equal to the parallel combination of source and feedback resistance at the non-inverting input, or lower the impedance values.
• Distortion at high frequency and large swing. That's slew-rate limiting at 5 V/µs — reduce the output amplitude or the frequency. A ±10 V swing is only clean up to roughly 80 kHz of full-power bandwidth.
• Gain lower than expected into a heavy load. Large-signal gain is 50 V/mV typ at RL ≥ 600 Ω but 100 V/mV typ at RL ≥ 2 kΩ; a load below 600 Ω is outside the specified region entirely, and the 38 mA short-circuit limit will start to clamp swing before the load even reaches a true short.
• Erratic behavior at cold or with low rails. Confirm the temperature suffix (NE = 0–70 °C, SA = −40–85 °C) matches your environment, and remember no PSRR guarantee exists below ±9 V rails — a drooping supply can couple rail noise straight into the signal.
• Part runs hot. Suspect self-oscillation driving dissipation far above the ~0.27 W typical figure, or a load below spec. Scope the output before assuming a thermal problem.

The NE5532 remains a well-behaved, low-noise, unity-gain-stable dual op amp as long as you respect three things: the bipolar input stage's appetite for low source impedances, the non-rail-to-rail input and output windows, and the fact that the numbers you memorized a decade ago may not match the datasheet on your desk today.

Part page: NE5532.