AP22653A — RLIM ↔ Current Limit

Two-way calculator for the current-limit programming resistor on the Diodes AP22652 / AP22653 / AP22652A / AP22653A precision adjustable current-limited power switches. Type in either box — the other one follows. AP22653AW6-7 = LCSC C1511768 (SOT-26, active-high EN, latch-off on overcurrent); the whole family shares one ILIM electrical spec, so every number here applies to all four.
Diodes DS41186 Rev. 4-2 (Oct 2024) — best-fit ILIMIT equations and the Electrical Characteristics spec table. Both models are implemented; the sheet cross-checks them against each other live.
Solve
kΩ
mA
Which curve is the current on?
Model
They agree to within 0.4% from 10–50 kΩ and diverge above it. See Model cross-check below for exactly where and by how much.
Temperature range
Full range widens the envelope only — typ is unchanged. Factors ×0.9673 on min and ×1.0280 on max, taken from the four temperature-paired rows of the spec table (they agree to ±0.1% across all four).
Standard resistor series
The part C1511768

AP22653A — 3.0–5.5 V, up to 2.1 A continuous, 65 mΩ typ RDS(ON), SOT-26. Current limit adjustable 125 mA–2665 mA via RLIM.

On overcurrent it current-limits, and once the internal 6 ms deglitch expires it latches off until EN is toggled or power is cycled. That is the "A" suffix — plain AP22653 stays in constant-current mode instead.

Decode AP2265 X X: 2 = active-low EN, 3 = active-high EN · blank = constant current, A = latch-off · W6 = SOT26, FDZ = W-DFN2020-6.

Answer
→
Full picture at this RLIM
ILIMIT is the DC trip threshold into a normal load. ISHORT is the much lower current the part settles at with OUT hard shorted to GND — it is not a typo, the foldback is real and it is what a dead short actually draws.
Nearest standard values
Resistor tolerance is on top of the silicon spread shown here — the datasheet's min/max curves exclude it. A 1% resistor moves the whole envelope by about ±1%.
Model cross-check best-fit equations vs. the spec table they came from
Sweep click a row to load it
Notes

Pick R against the right curve. The two failure modes pull in opposite directions. If you size on typ and your part lands on the low tail, a legitimate load nuisance-trips (and on the "A" parts, latches off). If you size on typ and it lands on the high tail, you can pull more than the upstream rail can give. Solve both ends and take the window.

The window can be empty. Silicon spread alone is ±7%, so the ratio between your real load and your hard ceiling has to be at least about 1.15× before any single RLIM satisfies both. Below that no resistor value works and you need a different part or a different budget.

Above 50 kΩ, trust the table, not the equations. The best-fit max curve reads low up there — 125 mA against a spec'd 155 mA at 210 kΩ. Low is the dangerous direction for the "never exceed" case, because it makes R look big enough when it isn't. Switch the model selector to the spec table for anything past 50 kΩ.

Layout. Keep the RLIM trace short — the datasheet calls out parasitics on that node as a direct hit to current-limit accuracy. Use a 1% part; the min/max curves already assume the resistor itself is exact.

Do not leave ILIM open. There is no internal default. Shorting ILIM to GND is a defined state (80 mA typ), but floating it is not.

The 2.1 A package limit is separate. RLIM = 10 kΩ programs 2665 mA typ, which is above the 2.1 A continuous-output rating. The current limit is a fault threshold, not a thermal permission slip.