Scarlett Zhang
| Picture | Part Number | Manufacturer | Stock | Action |
|---|---|---|---|---|
 | LM1117MPX-3.3/NOPB | Texas Instruments | 8,230 | |
 | NCP1117ST33T3G | onsemi | 4,370 | |
 | LD1117S33TR | STMicroelectronics | 1,460 | |
 | LDL1117S33R | STMicroelectronics | 3,340 |
No. It is a specific MaxLinear 3.3-V regulator. The original manufacturer, package drawing, capacitor, thermal layout and operating sequence must be known before compatibility can be assessed. A generic AMS1117-3.3 label does not supply those details.
The SPX1117 tab is connected to VOUT. For the fixed 3.3-V part, pin 1 is ground, pin 2 is output and pin 3 is input in the datasheet's top view. The tab must not be treated as a ground pad.
SPX1117 Rev. 2.2.0 records corrected dimensions in the recommended land-pattern and stencil drawing. Compare that recommendation with the controlled PCB and assembly process. An older footprint is not automatically defective, but its suitability should be documented.
The 1.1-V figure is typical. The electrical table lists 1.30 V maximum at 800 mA and defines dropout by a one-percent output reduction. A complete headroom budget also includes source tolerance, wiring losses, output requirements and transients.
Category: ComponentSelection&Alternatives
Author: Scarlett Zhang
SPX1117M3-L-3-3/TR is MaxLinear's fixed 3.3-V, 800-mA regulator in SOT-223. It may be a useful candidate for an existing 1117 design, but the shared number and package name do not establish replacement compatibility. Check the output-connected tab, the controlled land pattern, worst-case dropout, thermal dissipation, output capacitor and start-up sequence. The supplied June 2025 datasheet is especially relevant because its revision history records a correction to the recommended land-pattern dimensions.
A board marked “1117-3.3” rarely contains enough information to approve a substitution. The original manufacturer's identity may be missing from a schematic printout, the CAD symbol may show only three electrical pins, and the purchasing system may treat several suffixes as equivalent. Before comparing silicon, recover the original complete ordering code, footprint revision, output capacitor and intended load range. Those details define the circuit that the new regulator must satisfy.
SPX1117M3-L-3-3/TR is a specific fixed-output part. In MaxLinear's ordering table, M3 identifies the SOT-223 package family, the 3-3 portion corresponds to the 3.3-V option, and the complete row specifies tape-and-reel supply. The same document covers adjustable and other fixed-voltage versions. Their common family name does not permit transferring adjustable-pin circuits or another output voltage's test conditions to this exact device. MaxLinear SPX1117 product and document page.
Possibly, after review. MaxLinear specifies at least 2.2 µF for ceramic or tantalum and recommends ESR of 0.5 Ω or less. Check the actual capacitor's effective value under bias and temperature, then verify dynamic behavior on the board.
The datasheet explicitly says SPX1117 is not intended for prebiased-load startup and recommends discharging the output before input power is applied. Identify backfeed from other rails or interfaces and choose a different solution if prebiased startup is required.
A useful first decision is whether this is a component substitution or an opportunity to redesign the rail. If the existing board must remain physically unchanged, capacitor stability, pin mapping and copper area become hard constraints. If a board revision is allowed, a different regulator architecture or larger thermal area may solve a problem that no nominally compatible 1117 can fix. Keeping those paths distinct saves time during approval.
Table 1: The exact SPX1117 option under review
| Item | Supported identity or limit | Why it matters on an existing board |
|---|---|---|
| Ordering code | SPX1117M3-L-3-3/TR | Fixes manufacturer, output option and package/carrier record |
| Nominal output | 3.3 V | Compare against the load's full allowed supply range |
| Rated output current | 800 mA | Thermal conditions still determine usable continuous current |
| Package | Three-lead SOT-223, output-connected tab | Tab copper is electrically part of the output rail |
| Pin assignment, top view | 1: GND; 2: VOUT; 3: VIN | Check symbol, layout orientation and actual board nets |
| Junction range | −40°C to +125°C | Do not substitute ambient or thermal-shutdown temperature |
Source: MaxLinear SPX1117, Rev. 2.2.0, pages 1–3. Compiled by YG GROUP. Limits retain the conditions defined in the electrical table.
The table establishes identity, not a verdict. A three-terminal regulator can be mechanically familiar yet behave differently at light load, during a fast input collapse or with the original output capacitor. The right review starts with the board requirements and then asks whether the exact device meets them across the relevant operating conditions.
MaxLinear's pin-assignment figure places ground on pin 1, output on pin 2 and input on pin 3; the tab is connected to VOUT. On the fixed 3.3-V version, pin 1 is ground, not an adjustable feedback input. The package drawing and pin figure are shown from specified views. A mirrored interpretation can turn a correct datasheet into an incorrect layout, so compare the drawing with the copper rather than relying on the visual appearance of the symbol.
Do not connect the large tab land to ground merely because it is the thermal feature. Its copper area carries the regulated output. That affects clearance, stitching, test access and the placement of adjacent components. A large output copper area may improve heat spreading while also increasing the physical extent of a sensitive supply node. The design should make that tradeoff deliberately.
A continuity review of the released PCB is worthwhile when the board's history is uncertain. The symbol may label a fourth tab pin, combine it with pin 2, or omit it from the drawing while the footprint still connects it. These CAD conventions can all represent the same electrical connection, but they can also hide an old library error. Record how the symbol and footprint implement the tab so the next replacement review does not repeat the investigation.
The supplied datasheet revision history states that Rev. 2.2.0, dated June 10, 2025, corrected dimensions in the recommended SOT-223 land-pattern and stencil figure. It separately notes that the mechanical drawing revision changed for document registration without changing its content. Those are different changes. The useful action is to compare the released footprint and stencil against the current recommendations, not to assume that every drawing changed or that an old board is automatically defective.
The land-pattern drawing is a manufacturing recommendation, and the production assembly process may legitimately use a qualified variation. If that is the case, retain the process evidence and the reason for the variation. The correction still deserves review because an undocumented footprint inherited from an earlier library cannot be assumed equivalent. MaxLinear document access.
Figure 1: Mechanical compatibility has three connected checks: pin mapping, output-connected tab copper and the controlled land-pattern/stencil revision. This editorial schematic is based on SPX1117 Rev. 2.2.0, pages 3 and 9–11; it is not a dimensioned footprint. Diagram: YG GROUP.
The familiar “1.1-V dropout” headline is the typical value at 800 mA. The electrical table gives a maximum of 1.30 V at that load, with the specified full-temperature indication. At 100 mA and 500 mA, the maximum values are 1.20 V and 1.25 V respectively. Those numbers are much more useful for deciding whether an existing supply can maintain regulation than the typical headline alone.
Dropout also has a test definition. MaxLinear defines it as the input-to-output difference that produces a one-percent output decrease relative to the nominal output measured with 1.5 V of headroom. It is therefore not a promise that all accuracy and transient requirements are satisfied exactly at the dropout boundary. If the load has a tight minimum supply limit, preserve margin above that boundary and include regulator tolerance, wiring drop and load transients.
Consider an illustrative 5-V source with a five-percent negative tolerance. Its minimum is 4.75 V before board losses. Relative to a nominal 3.3-V output, that leaves 1.45 V, only 0.15 V above the 1.30-V maximum dropout figure at 800 mA. This arithmetic is not a full pass result: connector loss, source ripple, output tolerance and temperature effects still have to fit. It does show why a board that appears comfortable at a bench supply set to exactly 5 V can become marginal in a product.
Table 2: A headroom review for the 3.3-V rail
| Quantity | Datasheet fact or illustrative assumption | Review implication |
|---|---|---|
| Full-load dropout | 1.10 V typical, 1.30 V maximum at 800 mA | Design from the maximum and its test definition |
| Intermediate-load dropout | 1.25 V maximum at 500 mA | Lower load helps, but use the relevant load envelope |
| Example source minimum | 4.75 V, assuming a 5-V source with −5% tolerance | This is an example, not a regulator specification |
| Example nominal headroom | 4.75 V − 3.3 V = 1.45 V | Only an initial budget before distribution and accuracy effects |
| Input absolute maximum | 20 V | Stress rating; not a sensible default operating target |
| Dropout definition | One-percent output reduction from the stated reference condition | Do not equate dropout with complete regulation compliance |
Source: SPX1117 Rev. 2.2.0, pages 2–3; arithmetic by YG GROUP using explicitly stated source assumptions.
The electrical table contains separate rows for several fixed voltages and an adjustable reference. Its full-range condition cell for the 3.3-V output row is not populated as clearly as the neighboring options in the supplied copy. That is a documentation limitation worth recording. The article does not reconstruct a missing guaranteed input range by copying a row from the 2.5-V or 5-V version. Use the applicable general notes and request clarification when a release depends on that exact corner.
Minimum load deserves attention for the same reason. The relevant output-voltage tests use a 5-mA lower load condition. A system that enters an extremely low-current sleep state should be checked at that state rather than assuming a test beginning at 5 mA guarantees identical behavior down to zero. This does not mean the device cannot operate at light load; it means the quoted test coverage should not be enlarged without evidence.
The quiescent-current figures also caution against treating every 1117 as a modern ultra-low-power regulator. The SPX1117 table lists a typical 5 mA and maximum 10 mA under its stated input conditions. For an always-on battery rail, that can influence the architecture more than a small unit-price difference. For a board drawing hundreds of milliamps continuously, heat and headroom may dominate instead. Choose the relevant constraint rather than ranking parts by one attractive number.
MaxLinear specifies at least 2.2 µF for tantalum or ceramic output capacitors, or 10 µF for aluminum, and recommends ESR of 0.5 Ω or less. It notes that the required value can depend on load and temperature and that larger capacitance, up to 100 µF in the application discussion, can improve load-transient response. These are the requirements to compare with the actual fitted capacitor, not a generic belief that all 1117 devices require the same capacitor technology.
A ceramic capacitor marked 2.2 µF is not automatically 2.2 µF at the operating voltage and temperature. Its effective capacitance depends on its own datasheet, dielectric, dimensions, tolerance and applied bias. The regulator review therefore needs the capacitor's complete ordering code. If the original board used a tantalum part and a later cost reduction introduced a small MLCC, that history is part of the compatibility question even when the regulator itself never changed.
Capacitor placement matters as well. A remote bulk capacitor may help system energy storage while leaving a long inductive path at the regulator. Place and route the local output capacitor so that it actually participates in the regulator's control loop and transient response. An input capacitor similarly needs a short, effective current path. A bench test with temporary wires can create behavior that the intended PCB would not show, or conceal a problem by adding unintended capacitance.
Stability asks whether the feedback loop returns to equilibrium without sustained oscillation under the required conditions. Transient response asks how far and how long the output moves when the load changes. Ripple rejection asks how input disturbances appear at the output. Passing one of these checks does not automatically pass the others. A larger output capacitor may reduce initial droop yet still leave an unsuitable ESR or poorly routed return path.
The supplied transient plots illustrate specific conditions, including a 1.8-V output setup. They are useful examples of behavior and measurement, but they are not measured results for a 3.3-V production board. When evaluating SPX1117M3-L-3-3/TR, define the actual output, capacitance, load step and input source, then measure the board under those conditions. Keep typical manufacturer plots and application-specific measurements clearly separated in the review record.
Table 3: Capacitor and dynamic checks before replacement approval
| Check | SPX1117 evidence or requirement | Board-level evidence to collect |
|---|---|---|
| Capacitance and dielectric | At least 2.2 µF tantalum/ceramic or 10 µF aluminum | Exact capacitor code and effective capacitance at 3.3 V |
| ESR | Recommended value 0.5 Ω or less | Frequency and temperature dependence from capacitor data |
| Local placement | Output network participates in regulation | Layout and return-path inspection |
| Load transient | Datasheet examples have specific conditions | Actual load-step amplitude, edge rate and waveform |
| Light-load state | Quoted output tests begin at 5 mA | Sleep and standby measurements for the real load |
| Input disturbance | Ripple rejection depends on frequency and headroom | Relevant source ripple, dips and start-up sequence |
Source: SPX1117 Rev. 2.2.0, electrical specifications and application information. Compiled by YG GROUP. Board measurements listed here are recommended checks, not results reported by YG GROUP.
For a linear regulator, the first useful loss estimate is the input-to-output voltage difference multiplied by load current. MaxLinear uses this relationship in its thermal discussion. At 5 V input, 3.3 V output and 800 mA, the main pass-device loss is about 1.36 W. This simplified calculation omits the additional contribution associated with operating current and does not include every system loss. Even so, it is large enough to make PCB heat spreading central to the decision.
The supplied thermal experiment demonstrates that the package name does not determine one universal junction-to-ambient resistance. Its board examples range from 46°C/W with substantial copper on both sides to 85°C/W for a smaller board/copper arrangement. Applying those illustrative values to 1.36 W gives temperature rises of roughly 63°C and 116°C respectively. These are calculations using the datasheet's example board values, not predictions for an arbitrary layout.
At an assumed 50°C local ambient, the first simple estimate is about 113°C junction temperature and the second is about 166°C. The latter exceeds the stated 125°C maximum junction operating temperature even before omitted losses are considered. The calculation is intentionally a screening exercise: it identifies when the board needs more copper, less dissipation or a different architecture. It should not be used to claim that a real board has a particular junction temperature without thermal characterization.
A regulator can meet the electrical current rating and still fail the continuous thermal requirement. The SPX1117 protection discussion places thermal shutdown above the normal operating limit. Shutdown is a fault response; cycling into it during ordinary operation is not a thermal design solution. If the device is too hot at the required current, changing only the logo on a similar regulator may leave the underlying dissipation problem unchanged.
The lowest input is often the dropout corner, while the highest input is often the thermal corner. At an illustrative 12-V input and 800-mA load, the main linear loss for a 3.3-V output becomes 6.96 W. That calculation makes clear why the input absolute-maximum rating cannot be read as a promise of full-current operation across the entire voltage span. The required board, enclosure and junction-temperature constraints decide whether an operating point is usable.
A switch-mode preregulator, a lower source voltage or reduced load may be more effective than searching for another 1117. The tradeoff is additional complexity, noise or board area. Conversely, a lightly loaded rail with modest headroom may be an appropriate place for the simple linear solution. The editorial recommendation is to screen dissipation early, before spending effort on a detailed footprint equivalence exercise for an operating point the board cannot cool.
Figure 2: Electrical approval sequence for SPX1117M3-L-3-3/TR. Voltage headroom, capacitor behavior, heat and start-up are separate checks. The thermal relationship shown in the discussion is a first-order calculation, while this figure is an editorial workflow schematic. Diagram: YG GROUP.
The SPX1117 datasheet explicitly says the device is not intended for applications requiring start-up into a prebiased load and recommends discharging the output before applying the input. This is a decisive condition for boards with several power sources. An external programmer, USB-connected peripheral, backup supply or another active rail may keep the output above zero when the regulator starts. The exact mechanism depends on the system; the review should identify it rather than assume a discharged output.
A power-off test is therefore as important as a power-on test. Observe whether the output decays fully, whether another interface feeds the rail and how the sequence behaves during quick restart. If the design requires prebiased start-up, choose a regulator specified for that condition or redesign the power path. A diode added from an old reference circuit is not automatically a complete fix; its effect on discharge, backfeed, voltage loss and fault conditions needs analysis.
The revision history shows that the prebias section was added in an earlier revision and remains in the supplied current document. That is a good example of why a replacement review should use the complete datasheet rather than a first-page feature list. A device can match output voltage, package and nominal current and still be unsuitable because of one operating sequence.
The following parts are specific manufacturer-listed orderable models. They are included because each helps test a different assumption about the existing circuit. The comparison is deliberately limited to supported product-level distinctions; it does not declare any candidate interchangeable on the user's board.
Table 4: Four 3.3-V regulator candidates to compare deliberately
| Orderable model | Manufacturer | Useful comparison | Decisive review boundary |
|---|---|---|---|
| LM1117MPX-3.3/NOPB | Texas Instruments | Familiar 800-mA 1117 in SOT-223 | Exact page specifies 0°C to +125°C; output-capacitor requirements differ |
| NCP1117ST33T3G | onsemi | 3.3-V member of the 1-A NCP1117 family | Compare full electrical, capacitor and package conditions independently |
| LD1117S33TR | STMicroelectronics | Fixed 3.3-V LD1117 in SOT-223 | Similar naming does not establish the same loop or thermal performance |
| LDL1117S33R | STMicroelectronics | 1.2-A family with much lower typical dropout | Different regulator behavior; review its complete operating and capacitor requirements |
Sources: official TI LM1117 ordering page, onsemi NCP1117 product table, ST LD1117 product table and ST LDL1117 product table. Compiled by YG GROUP. These are comparison candidates, not approved replacements.
The TI part illustrates why a family temperature statement must not replace the exact ordering record. TI distinguishes LM1117 from LM1117I, and the selected LM1117MPX-3.3/NOPB record shows the former temperature coverage. If the board's requirement extends below zero, that distinction immediately changes the shortlist. The TI page also calls for a minimum tantalum output capacitor in its product description, which warns against assuming that an SPX1117 ceramic-capacitor implementation transfers unchanged.
The onsemi candidate offers a different current class, but the higher rating is not a free thermal upgrade. At the same input, output and load, the main linear voltage-drop loss remains. Its product family describes fixed-output behavior and protection features that must be compared with the original circuit. A careful reviewer asks which requirement the candidate improves and which existing assumptions need to be retested.
ST's LD1117S33TR is a close functional comparison. LDL1117S33R broadens the discussion toward a newer low-dropout implementation: ST specifies 1.2-A capability and a typical 350-mV dropout at that current for the LDL1117 family. The number is useful for identifying a headroom opportunity, but a typical value at a different load is not a fair worst-case ranking against the SPX1117 maximum. Use each complete electrical specification before making the final choice.
A practical validation plan should reproduce the board's demanding conditions, not just demonstrate that one sample outputs 3.3 V at room temperature. Begin with minimum and maximum input, the actual load envelope, the intended output capacitor and the production layout. Include cold start, warm restart, load changes and the lowest standby current. Add the temperature conditions relevant to the product rather than treating a bench result as coverage of the full junction range.
Measure at the load as well as near the regulator when distribution resistance matters. A correct voltage at pin 2 does not guarantee the remote load sees the same value. The datasheet's layout discussion addresses voltage drop in the adjustable circuit; the general lesson for this fixed-output board is still to account for supply and return paths. Keep the measurement points explicit so two engineers do not report apparently conflicting regulation results from different nodes.
When comparing the original regulator and the proposed SPX1117, keep the source, load, capacitor population and probe arrangement consistent. A long oscilloscope ground lead can add pickup and ringing that do not represent the supply seen by the load. A current step made through long wires can also add inductance. Document the setup and investigate measurement artifacts before interpreting a transient as a regulator defect.
Use the same board revision when the objective is a component-only comparison. If one sample is fitted to a large evaluation board and another to a compact production board, temperature differences include the layout effect. That may still be useful evidence for a redesign, but it cannot isolate device performance. The SPX1117 thermal table makes the importance of copper area explicit, so the comparison should preserve that variable or explain why it changed.
Repeat the sequence after the assembly has reached a stable temperature. A cold, brief load pulse can show acceptable voltage before the device approaches its continuous thermal condition. Conversely, a hot restart can expose a different power-sequencing state because capacitors have not fully discharged. Both observations are useful when attached to their actual initial conditions.
Finally, define what the load is allowed to see. A waveform is only an engineering result when its peak deviation, settling behavior and duration are compared with the load's limits. A visually tidy oscilloscope trace may still violate a narrow supply requirement; a short excursion may be acceptable for another load. This keeps the replacement decision tied to the product rather than to a subjective preference for one trace.
Table 5: Evidence for approving the specific substitution
| Review item | Evidence to retain | Pass decision should mean |
|---|---|---|
| Identity and package | Exact codes, pin maps and controlled drawings | Correct electrical nets and manufacturable footprint |
| Headroom | Worst-case source, load and voltage budget | Regulation requirements met at the required corners |
| Output network | Capacitor code, effective value, ESR and layout | Stable behavior and acceptable transient response |
| Thermal operation | Loss estimate and board measurements | Junction operating requirement met with margin |
| Power sequence | Start-up, shutdown and backfeed observations | No required prebiased-load start-up for SPX1117 |
| Release record | Approved BOM and tested board revision | Purchasing and manufacturing use the verified configuration |
Source: YG GROUP review workflow based on MaxLinear's electrical, capacitor, thermal, prebias and package guidance. The table defines evidence to collect; it does not report completed hardware validation.
SPX1117M3-L-3-3/TR is a reasonable candidate when the existing board provides the right pin mapping, capacitor conditions, voltage headroom, heat spreading and discharged-output start-up. The answer comes from those checks together. The 2025 land-pattern correction and the explicit prebias restriction are particularly useful reasons to reopen the full documentation before approving a familiar-looking 1117 substitution.