LM317 Equivalent & Replacement - LM350, LT1084, LM338 Compared

LM317 vs LM350 vs LT1084: Full Comparison

When to Stay With LM317

• Current under 1.5A and you don't care about efficiency: the LM317 is the cheapest option and the most widely second-sourced. If it works, don't over-engineer it.
• Input voltage above 30V: the LM317 handles up to 40V input. The LT108x series tops out at 30V. If you're regulating from a 36V rail, the LM317 is your only option among these three.
• You need through-hole simplicity: LM317 in TO-220 with a clip-on heatsink is the fastest way to get a working adjustable supply on a breadboard.

When to Upgrade to LM350 (3A)

• You need more current and nothing else changes: same pinout, same adjustment formula, same external components. Swap the LM317 for an LM350, upgrade the heatsink, and you've doubled your current.
• But be realistic about heat: dropping 12V to 5V at 3A = 21W of heat. The LM350 doesn't solve the efficiency problem - it just gives you more current headroom before thermal shutdown. If your heatsink was marginal with the LM317, it won't survive with the LM350.

When to Upgrade to LT1084 or LT108x Series

• You need efficiency: the LT1084's 1.3V dropout vs the LM317's 2–3V means roughly half the heat for the same output current. If your input-output differential is tight, this matters.
• You need 5A with manageable thermals: the LT1084 does 5A at low dropout. The LM338 also does 5A but with higher dropout - same current, more heat.
• You need precision: the LT108x reference is trimmed to ±1% vs ±4% for the LM317. For circuits where output voltage accuracy matters, the LT108x is worth the extra cost.
• Remember the output capacitor: LT108x parts require a minimum 10 μF output capacitor for stability. The LM317 works without one (though it's a good idea). If your existing circuit has no output cap, you'll need to add one.

All LM317 Replacements at a Glance

R1 = 120Ω: The Rule That Prevents Half the LM317 Problems on Forums

If you take one thing from this guide, make it this: use 120Ω for R1 (the resistor between OUT and ADJ), not 240Ω, and definitely not 1 kΩ.

The LM317 needs a minimum load current of 5–10 mA to regulate properly. With R1 = 120Ω, the resistor divider draws 1.25V / 120Ω = 10.4 mA - satisfying the minimum load even with no external load connected. With R1 = 240Ω, you get 5.2 mA - marginal, and it degrades at temperature extremes. With R1 = 1 kΩ? 1.25 mA - nowhere near enough. The output drifts high or refuses to regulate at all.

We see this on StackExchange constantly: "My LM317 output is stuck at X volts and won't adjust." Nine times out of ten, R1 is too large. Switch to 120Ω and it works. The 240Ω value in old datasheets was specified for the LM117 (military temp range) - the LM317 needs the lower value. This is the single biggest trap in LM317 circuit design.

Watch Out: Potentiometer Wiper Failure = Full Output Voltage = Fried Downstream Circuit

This one is nasty because the circuit works perfectly - until it doesn't. If you're using a potentiometer as R2 (the variable resistor between ADJ and GND) and the wiper develops a bad contact - from vibration, oxidation, or just age - R2 goes open-circuit. The ADJ pin sees no path to ground. The LM317 responds by driving the output to its maximum possible voltage (close to Vin minus dropout). If your downstream circuit was designed for 5V and suddenly sees 20V, it's dead.

Fix: add a Zener diode clamp between ADJ and GND. Pick a Zener voltage slightly above your intended output. For a 9V max output, use a 10V Zener. If R2 opens, the Zener conducts and limits Vout to VZ + 1.25V. Cheap insurance.

The Counterfeit LM317 Problem

LM317s from AliExpress, Taobao, or uncertified spot-market sellers are a known landmine. The EEVblog forums have multiple threads where users bought batches of LM317s that all failed the same way: regulate fine below 100–200 mA, then hard cut off at higher loads. These are relabeled low-current regulators - "100 mA chips in 1.5A bodies," as one forum member put it. Side-by-side tests with genuine TI parts from authorized distribution showed the fakes failing at less than 10% of rated current.

The quickest verification: load the LM317 to 1A with Vin = 15V, Vout = 5V (10W dissipation, well within spec for a TO-220 with a heatsink). Genuine parts regulate 5V steady. Fakes either fold back to 1.2V or enter thermal shutdown within seconds. Also check: with R1 = 120Ω, R2 = 360Ω (should give ~5V), a fake often outputs a fixed ~1.2V regardless of the divider - the internal reference is damaged or absent.

Common LM317 Failure Modes (And How to Prevent Them)

"Output is stuck at ~1.2V no matter what I do with the resistors"

Two possibilities: (1) You have a counterfeit part with a dead internal reference - swap in a known-genuine LM317. (2) R1 is shorted or the wrong value - measure it in-circuit, verify it's 120–240Ω between OUT and ADJ.

"It works at low current but shuts down above 200 mA"

Counterfeit. Genuine LM317s handle 1.5A all day with adequate heatsinking. If it folds back sharply, you have a relabeled low-current regulator. Replace with a part from authorized distribution.

"It oscillates / gets burning hot even at low load"

Oscillation. Common causes: long wire leads between the LM317 and the output capacitor, missing input bypass capacitor, or R1/R2 placed too far from the ADJ pin. Keep the resistor divider within 5 mm of the ADJ pin. Add a 0.1 μF ceramic across the input. Add a 10 μF electrolytic on the output. If using an ADJ bypass capacitor (for ripple rejection), add a protection diode from ADJ to Vout.

"My potentiometer-adjusted supply just killed my prototype"

Wiper failure → R2 open → Vout maxed out. Add a Zener clamp from ADJ to GND. This is not optional for any design where the downstream circuit can't survive full input voltage.

Where to Buy LM317 Replacements

Contact ICMASS for current pricing on your specific quantity and variant. We stock the full adjustable regulator family - LM317, LM350, LM338, LT1083/84/85/86 - with traceability to TI, ST, ON Semi, and Analog Devices. All parts tested for the 1A load test before shipping.

Frequently Asked Questions About LM317 Replacements

Q1: What is the best direct replacement for LM317?

A: For same specs, lower cost: KA317 (ON Semi/Fairchild). For same current, better efficiency: LT1086 (1.3V dropout vs 2–3V, pin-compatible, only difference is it requires an output capacitor for stability). For more current: LM350 (3A, drop-in replacement, same pinout and adjustment formula).

Q2: Is LM350 a direct drop-in replacement for LM317?

A: Yes. Same TO-220 pinout (ADJ-Vout-Vin, pins 1-2-3), same 1.25V reference, same adjustment formula. Swap the LM317 for an LM350, upgrade the heatsink, and you get 3A instead of 1.5A. No other changes needed.

Q3: What's the difference between LM317 and LT1084?

A: Three things: (1) LT1084 delivers 5A vs LM317's 1.5A. (2) LT1084 has much lower dropout: 1.3V vs 2–3V - roughly half the heat for the same output. (3) LT1084 requires an output capacitor (10 μF min) for stability; LM317 works without one. Also: LT1084 max input is 30V vs LM317's 40V, and LT1084 costs significantly more.

Q4: What resistor value should I use for R1 on LM317?

A: Use 120Ω, not 240Ω. The LM317 needs 5–10 mA minimum load current to regulate. 120Ω guarantees 10.4 mA through the divider - meeting the requirement even with no external load. 240Ω gives only 5.2 mA and can fail at temperature extremes. The 240Ω value from old datasheets applies to the LM117 (military grade), not the LM317. This is the #1 cause of "my LM317 won't regulate" forum posts.

Q5: Can I use LM317 at 3.3V output?

A: Yes. Vout = 1.25 × (1 + R2/R1). With R1 = 120Ω, R2 = 197Ω gives ~3.3V. Use a 200Ω standard value for ~3.33V. The minimum output voltage is 1.25V (R2 = 0). For outputs below 1.25V, you'll need a different topology.

Q6: Why does my LM317 output stay at 1.2V and won't adjust?

A: Either (1) counterfeit part with a dead internal reference - swap with a known-genuine LM317 from authorized distribution, or (2) R1 is shorted or too low - measure between OUT and ADJ, verify 120–240Ω. If R1 is shorted (0Ω), Vout = 1.25V fixed.

Q7: How do I protect my circuit if the adjustment potentiometer fails?

A: Add a Zener diode from ADJ to GND, with VZ slightly above your maximum intended output. If the potentiometer wiper fails open, R2 goes to infinity, but the Zener conducts and limits Vout to VZ + 1.25V. For a 9V design, use a 10V Zener. Without this clamp, wiper failure sends Vout to nearly Vin - destroying your downstream circuit.

Q8: How do I spot a counterfeit LM317?

A: Load test at 1A: set Vin = 15V, Vout = 5V, load with 5Ω/10W resistor. Genuine LM317 regulates 5V steady with a TO-220 on a moderate heatsink. Counterfeits fold back to ~1.2V or enter thermal shutdown within seconds - they're relabeled low-current regulators (EEVblog confirmed). Also: genuine parts draw ~5–10 mA quiescent current. If your "LM317" draws under 1 mA, it's not a real LM317.

Q9: LM317 vs LM338 vs LT1083 - which one for a bench supply?

A: LT1083 if budget allows (7.5A, 1.3V dropout, ±1% reference). LM338 if you need higher input voltage tolerance (40V vs 30V) or lower cost (5A, 2V dropout). LM317 if you only need 1.5A and want the cheapest option. For a serious bench supply: LT1083 + large heatsink + forced air. The lower dropout means less heat to manage at high current.