LM1117 vs AMS1117: Key Differences Engineers Should Know Before Choosing an LDO

The Real Problem with 1117 Parts

I still see people treating LM1117 and AMS1117 as the same type of regulator. Same footprint, similar price, same job. In reality the difference between LM1117 and AMS1117 is large enough that I stopped treating them as interchangeable years ago.

Who Makes What

LM1117 was originally introduced by Texas Instruments, but today the LM1117 series is available from multiple semiconductor manufacturers. The specifications are usually well documented and the production consistency is easier to verify. AMS1117 started with Advanced Monolithic Systems, but most of what you buy now is made by other factories. Some lots are decent. Others have higher dropout or weaker current limit. If you are writing an LM1117 replacement guide for a production board, this is the first thing that bites you.

Current and Dropout

TI rates the LM1117 at 800 mA. Most AMS1117 parts claim 1 A. In an SOT-223 both are limited by heat well before those numbers. Dropout voltage is usually in a similar range on paper, often around 1.1–1.3V at high load. On the board I have seen some AMS versions start to sag earlier when the input is only 1.3–1.4 V above the output and the load steps. The TI part usually holds better.

Heat and Packages

Both are commonly available in SOT-223 packages, while some versions are also offered in other packages such as TO-252. so the footprint can stay the same. Heat is the real limit. I calculate (Vin – Vout) × Iout before I even pick the part. Around 0.8W or higher continuous power loss in an SOT-223 package usually requires careful thermal design. The LM1117 usually shows predictable thermal protection behavior. Some lower-quality compatible versions may show less predictable behavior near their thermal limits. That behavior has caused me more field returns than I care to admit.

Capacitor Issues

Both need some ESR. A 10 µF or 22 µF tantalum works for either. Some ceramic capacitor combinations may affect stability depending on the specific regulator design., though the exact edge is not the same. I always test the actual part I plan to buy if the design uses ceramics. Guessing here has wasted more time than it should.

My Own View on Selection

For anything that has to stay the same for years, or for industrial work, I use the LM1117 or TI's later TLV1117. The numbers are consistent and the thermal behavior is predictable. For cheap consumer boards where the current stays under 400–500 mA and the input has clear headroom, an AMS1117 is usually fine - as long as I can get a real datasheet and not just a marking.

If I have to second-source, I buy samples from the actual vendors and check load-step response and thermal shutdown myself. The difference between LM1117 and AMS1117 is rarely catastrophic, but it is real. Ignoring it is how you end up chasing intermittent resets that only show up after the board has been running for a while.

That is still how I choose an LDO regulator for PCB design.