LM358, TL072, NE5532 Op-Amp Alternatives: A 2026 Sourcing Playbook

Three of the most-installed dual op-amps in the world — the LM358, the TL072, and the NE5532 — are sitting on 26+ week lead times at the authorized channel as of Q1 2026. The TI distributor lead-time advisory from earlier this year put several LM358 grades into the 26-30 week bucket, and onsemi has been quietly trimming older bipolar lines as the analog wafer-fab consolidation continues. Buyers searching for op-amp alternatives have not been this active since the 2021 shortage.

This is the playbook we use at Cosolvic when a customer’s BOM gets stuck on one of these three parts. It covers what each op-amp actually does, which second sources are real substitutes, where the trade-offs hide, and how the gray market for 50-year-old op-amps has gotten dirty enough that we now refuse certain lot codes on sight.

I’m writing this from Shenzhen, where the secondary market for legacy analog is loud, fast, and occasionally dishonest. Sorting it out is our day job.

Why op-amp alternatives suddenly require a real second-source plan

The short version: TI and onsemi are consolidating analog wafer fabs and shutting down older 6-inch lines that produce a lot of the parts engineers reach for without thinking. The LM358 was originally a 1972 Philips design that National Semiconductor productized in 1973. The NE5532 came out of Signetics in the late 1970s. The TL072 is a 1978 TI part. None of them have changed materially in decades, and the wafer processes that build them are old enough that the fab economics no longer work.

What this means in practice: the parts are still in the catalog. The datasheet still exists. The franchise distributor will still take your order. But the quoted lead time on certain package-and-grade combinations — TSSOP-8 LM358BD on extended-temp, for instance — has slipped past 26 weeks, and buyers calling to request expedite are being told there is no expedite. The factory makes a finite quantity per quarter and has already allocated it.

The same story has played out for older Maxim and Linear Technology parts that ADI is rationalizing post-acquisition. We covered the broader pattern in our component lifecycle guide. The op-amp version is just the most visible because these three parts are everywhere.

LM358 alternatives: a clean substitution in most industrial designs

The LM358 is the workhorse. Dual op-amp, single-supply, 3-32V rails, GBW around 1 MHz, input offset around 2 mV typical, common-mode range hitting the negative rail. People drop it into sensor front-ends, simple filters, and comparator-ish duty without thinking too hard. That’s exactly why so many BOMs have one and exactly why the substitution is usually clean.

The strongest direct substitute we keep stock on is JRC’s NJM2904. Same pinout, same single-supply behavior, GBW within 10%, comparable offset specs, and pricing typically 50-70% of TI list. Microchip’s MCP602 / MCP604 is a more modern CMOS alternative with significantly better offset (250 µV) and rail-to-rail output, useful when the design is marginal on input range. ST’s LM358D is a true second-source from a different fab and shows up at 4-6 week lead times even when TI’s version is locked. For low-voltage 3.3V systems, TI’s own LMV358 drops in with a 2.7-5.5V rail.

The Chinese option worth knowing is 3Peak’s TPH3221 family. We’ve qualified these on customer industrial designs and they perform as advertised, but they are not pin-for-pin clones in every package and the offset distribution is wider. Use them when the design has guard-band, not when you’re trimming millivolts.

Decision moment — Engineer. If your LM358 is doing simple gain or filtering on an industrial signal with millivolt-level headroom, the NJM2904 swap is a paperwork exercise. If it’s sitting in a calibrated measurement loop where 1-2 mV of additional offset matters, you need to look at MCP602 or actually upgrade to a precision part. Don’t second-source blindly.

TL072 alternatives: don’t lose the JFET input

The TL072 is not the LM358 with extra GBW. It’s a JFET-input part, which is the entire point. Input bias current sits in the picoamp range instead of the nanoamp range, input impedance is effectively infinite at audio frequencies, and that’s why the part lives in piezo front-ends, electret microphone preamps, and any high-impedance sensor where bipolar input current would load the source down.

If you substitute a bipolar op-amp, your circuit changes character. The DC operating point moves, the noise corner shifts, and a piezo sensor that was happy will start sagging. The substitute has to keep the JFET input.

JRC’s NJM072 is the closest direct equivalent — literally a second-source of the same architecture, with characteristics within a few percent. TI’s OPA1642 is a JFET-input audio op-amp that sits a tier above the TL072 in distortion and noise but costs three to four times as much. Burr-Brown OPA2134 (now under TI) is the audiophile community’s standard upgrade path and works as a drop-in for TL072 in audio gear. TI’s LM4562 is a bipolar audio premium part — it sounds great, but it’s not a JFET substitute, and your front-end behavior will change if the source impedance is high.

External reference: TI’s TL072 product page keeps the current datasheet revisions and the active grade list, which is worth checking before a redesign because TI has been quietly NRND-ing some of the older grades.

NE5532 alternatives: audio designers will hear what changed

The NE5532 is a bipolar dual op-amp from the late 1970s with input noise around 5 nV/√Hz, GBW around 10 MHz, and a slew rate near 9 V/µs. It became the studio-audio default — preamps, EQ stages, headphone amps, the analog stages of every CD player ever shipped. The reason it’s still everywhere in 2026 is that it is genuinely good at what it does and the cost is reasonable.

If you’re swapping it out, audio engineers will tell you they can hear the difference. Sometimes that’s mythology and sometimes it isn’t. The actual measurable differences live in the distortion corner above 10 kHz and the behavior into low-impedance loads — headphone amps especially.

JRC’s NJM2068 is the most common shadow substitute — Japanese pro-audio gear has been using it interchangeably with NE5532 for thirty years. JRC NJM4580 is a step up in linearity and is the Yamaha/Roland default in current production. OPA2134 (JFET, mentioned above) is the audiophile upgrade. TI’s LME49720 is a premium-audio bipolar op-amp that measurably outperforms the NE5532 on every spec but costs accordingly.

The trap on this part is the cross-reference list you’ll find on supplier websites suggesting MIC4427 (Microchip) as a substitute. That part is a MOSFET driver. It’s not an op-amp. Don’t laugh — we’ve seen it on a customer’s spreadsheet, sourced via a non-technical procurement seat. Verify before you order.

External reference: onsemi’s NE5532 product page is the current authoritative source since the part migrated under the onsemi umbrella.

Decision moment — Engineer. If the NE5532 is in a consumer-audio product where the spec sheet says “studio quality” and the actual bottleneck is the speaker driver, NJM2068 is the right swap. If it’s in a measurement instrument or a pro-audio mic preamp where the noise floor is an actual sales feature, you need to think harder — either re-qualify with NJM4580 or budget for an LME49720.

Side-by-side: the eight strongest op-amp alternatives compared

Specs above are typical values from manufacturer datasheets. Always verify against current revisions before committing a design.

Counterfeit risk on 50-year-old op-amps

Long-lifecycle parts in tight supply attract fraud. The LM358, TL072, and NE5532 have all three risk factors stacked: they’re everywhere, they’ve been made by multiple fabs over decades, and the secondary market is full of unverified inventory. We’ve physically inspected enough fraudulent reels of these parts to write a separate guide on it — see how to verify electronic component authenticity for the full method.

Three things get faked on op-amps:

Re-marking. A cheaper or lower-grade part — commercial-temp, slower GBW grade, or a different family entirely — gets sanded and re-laser-marked as the high-spec part. Detection: laser-mark depth wrong under microscope, ink on what should be laser parts, lot codes that don’t match the production date format the original fab uses.

Pulled / refurbished from boards. Used parts desoldered from scrap boards, leads straightened, re-tinned, sometimes re-baked. Detection: lead surface oxidation patterns, solder residue in the package corners, slight bowing of the part body from thermal cycling.

Lot-code inconsistency. A reel marked as a single TI lot code that contains parts from three different actual batches — mixed sources passed off as one. Detection: spot-check 5-10 parts per reel under microscope; look for date-code drift and laser-mark variance.

We test-buy from new suppliers periodically just to keep our own counterfeit detection sharp. The volume of bad LM358 in the Shenzhen secondary market is not theoretical.

How we approach an op-amp alternatives BOM at Cosolvic

If a customer sent us a BOM with all three parts on it tomorrow, here’s the sequence we’d run.

First, separate the parts by application criticality. Industrial-control LM358 with 1 MHz GBW and millivolt headroom — that’s a paperwork swap to NJM2904 and we can quote 4-6 weeks. TL072 in a piezo sensor front-end — needs a JFET substitute, not just any dual op-amp. NE5532 in an audio path that the customer’s audio engineer signed off on — needs a real conversation about which substitute, not a unilateral choice.

Second, check what’s actually available at the secondary level versus what needs to ship from Japan or fab-direct. NJM parts from JRC’s authorized Japan channel typically run 4-6 weeks. NJM parts in the Shenzhen secondary market are sometimes available within 3-5 days but require provenance verification. We don’t quote secondary stock unless we can verify the source chain.

Third, flag the regulatory posture. Anything in a medical, automotive-AECQ, or aerospace-qualified design is not eligible for casual substitution. Those programs have re-qualification cost that can dwarf the part savings, and we’ll tell the customer to pay TI’s lead time rather than chase a cross-reference. The same logic applies to the broader hard-to-find component sourcing playbook we use for obsolete parts generally — the cross-reference is the easy part, qualification is the work.

Have a legacy op-amp BOM you’re trying to source? Send us your BOM at request a quote. We’ll tell you within four hours which lines we have authentic stock for, what’s available within 3-5 days, and which ones genuinely require a redesign instead of a substitution. Every shipment ships with our 100% refund guarantee on authenticity.

For BOMs with Chinese-spec MCUs alongside these analog parts, the STM32 Chinese alternatives playbook covers the same logic on the digital side. External reference: JRC/NJR’s op-amp portfolio page is the authoritative spec source for the NJM2904, NJM072, NJM2068, and NJM4580 cross-reference parts mentioned above.

FAQ

Is the NJM2904 a true drop-in replacement for the LM358?
Functionally yes for most industrial uses — same pinout, same single-supply behavior, GBW within tolerance. Where it diverges: input offset distribution is slightly wider, and the temperature coefficient isn’t identical. For calibrated measurement designs you should re-verify offset and drift. For sensor interface, comparator, or filter duty, it’s a paperwork swap.

Can I substitute a bipolar op-amp for the TL072 in an audio circuit?
Sometimes. If the source impedance is low — under 10 kΩ, e.g., a buffered line input — a bipolar like NE5532 or LM4562 will work and may sound better. If the source is high-impedance — piezo, electret without internal buffer, guitar pickup — the input bias current of a bipolar part will load the source and change the response. Stick with JFET input: NJM072, OPA2134, or OPA1642.

Are 3Peak op-amps reliable enough for production?
For consumer and standard industrial applications, yes — we’ve shipped 3Peak parts on customer designs without field failures. For automotive, medical, or anything safety-critical, no — they don’t carry the qualification documentation those programs need. Use them where price matters and the application has guard-band.

How do I verify an LM358 reel I bought from a non-franchise distributor?
Visual inspection under 40× microscope (laser-mark depth, lead oxidation, package consistency), date-code and lot-code cross-check against the manufacturer’s coding format, and at minimum a parametric spot-check on 5-10 parts (offset, GBW, supply current). For high-volume orders we run X-ray on a sample to confirm die match. The full method is in our authenticity verification guide.

Will TI eventually obsolete the LM358 entirely?
Unlikely in the short term — the part still ships in volumes too large to NRND outright. What’s more likely is that specific grades and packages get NRND while the mainstream SOIC-8 commercial-temp grade stays alive on a longer cycle. Watch TI’s PCN feed for the grade you actually use. The same caution applies to the NE5532 under onsemi.

The legacy op-amp problem in 2026 isn’t “the part is gone.” It’s “the lead time is unacceptable and the secondary market is mixed.” The work is matching the right substitute to the right application criticality — and refusing to ship secondary stock you can’t verify. NJM2904 for industrial LM358, NJM072 for JFET TL072, NJM2068 or NJM4580 for NE5532 audio, with 3Peak as the price-aggressive China option where qualification allows. Everything else is conversation.