HN Debrief

New Texas Instruments 5532 chips are not the 5532s we’ve used for decades

  • Hardware
  • Semiconductors
  • Supply Chain
  • Manufacturing

The linked post argues that Texas Instruments’ current NE5532 is no longer the same general-purpose dual op-amp engineers have used for decades, despite keeping the same part number. The concrete changes called out were not cosmetic. Commenters highlighted a lower absolute maximum supply rating, a slew rate drop from 9 V/µs to 5 V/µs, removal of offset trim, and an input-stage change from NPN to PNP. That matters because the 5532 is a classic commodity analog part used all over audio gear, maintenance workflows, and old designs where people reasonably expect a replacement marked “5532” to behave like the familiar device class.

If your products, repairs, or approved vendor lists assume part-number continuity for legacy analog ICs, re-validate them now. Treat product change notices and fresh datasheet comparisons as mandatory, especially for old “jellybean” parts that feel too familiar to re-check.

Discussion mood

Strongly negative toward Texas Instruments. The anger came from the size of the spec changes, the decision to keep the same part number on a legacy commodity analog part, and a broader sense that TI has become harder to trust on documentation and support.

Key insights

  1. 01

    Fab migration is likely driving it

    Moving legacy analog parts onto newer fabs and larger wafers can make old designs impossible to reproduce exactly. That makes this look less like a random blunder and more like a business choice to preserve sales continuity while accepting behavioral drift that analog customers will absolutely notice.

    When a semiconductor vendor is in the middle of a fab transition, expect more than packaging or yield changes. Audit any long-lived analog part in your bill of materials when the manufacturer starts shifting nodes or wafer sizes.

      Attribution:
    • topspin #1
    • rsynnott #1
    • fc417fc802 #1
  2. 02

    Commodity part numbers are weaker than people think

    The assumption that all 5532s are fungible only holds loosely, and mostly outside disciplined production environments. Authorized distributors track manufacturer-specific part numbers, while competent hardware teams qualify approved vendor lists rather than buying a generic label and hoping every source matches closely enough.

    If your documentation still names a generic part family where performance or limits matter, tighten it. Specify approved manufacturer and package combinations, then make procurement enforce that list.

      Attribution:
    • derefr #1
    • buescher #1
    • RetroTechie #1
  3. 03

    Regulated products get hit either way

    Changing a part number can trigger real compliance and paperwork costs for small manufacturers, but silently changing the silicon under the same number creates a worse problem for audited products. A regulator may tolerate an explicit component revision path. It will not be happy if fielded units contain a materially different part that no one formally evaluated.

    For regulated hardware, build a component-change workflow that budgets for requalification before a crisis. The cheap path is not to hope a vendor keeps old semantics forever.

      Attribution:
    • rpaddock #1
    • rsynnott #1
    • vanderZwan #1
  4. 04

    Analog margins were always part of the job

    Older analog practice assumed more spread in semiconductors, power rails, and passive tolerances than many modern digital-heavy engineers are used to. The useful correction here is not that TI is fine. It is that designs riding right at absolute limits or depending on incidental dynamic behavior were always fragile, and this episode exposed that fragility hard.

    Revisit old analog sections that run near maximum ratings or depend on exact speed and stability behavior. Add headroom now instead of treating old field success as proof the margin is still there.

      Attribution:
    • buescher #1
    • ofalkaed #1
    • CamperBob2 #1

Against the grain

  1. 01

    The better fix may be replacing 5532s

    The NE5532 is popular because it is good enough and everywhere, not because it is still state of the art. One commenter argued that newer parts like OPA1612 now beat it on audio performance, which shifts the practical question from preserving nostalgia to deciding whether a redesign should retire the part entirely.

    If you are touching an audio design anyway, compare the cost of requalifying a modern op-amp against the cost of policing legacy 5532 sources forever. The safer long-term path may be a planned redesign.

      Attribution:
    • ycui7 #1
    • snvzz #1

In plain english

NE5532 ↩
A widely used dual low-noise operational amplifier, especially common in audio circuits.
NPN ↩
A type of bipolar transistor structure used in analog and digital circuits, with different electrical behavior from PNP devices.
offset trim ↩
A feature that lets engineers adjust an op-amp to reduce its input offset voltage error.
op-amp ↩
Short for operational amplifier, an analog integrated circuit used to amplify and process electrical signals.
OPA1612 ↩
A newer Texas Instruments dual audio operational amplifier often used as a higher-performance alternative to older parts like the NE5532.
PNP ↩
A type of bipolar transistor structure that behaves differently from NPN devices and can change how an input stage works.
SKU ↩
Stock Keeping Unit, a specific product configuration sold by a company or retailer.
slew rate ↩
How quickly an amplifier’s output voltage can change, which affects how well it handles fast-changing signals.
V/µs ↩
Volts per microsecond, the unit used to measure slew rate.

Reference links

Video analysis

Coverage and background

Related concepts and examples