</>EmbeddedInterview
← All problems · Signal Processing

Calibration Table Interpolation

Medium

A thermistor is calibrated with a table of (ADC reading, temperature in 0.1 °C) points, sorted by ADC value. Write int16_t adc_to_temp(uint16_t adc) that:

  • finds the two table entries surrounding adc with a binary search,
  • linearly interpolates between them using integer arithmetic only,
  • clamps to the first or last temperature when adc is outside the table.

Write your code on the right and press Run (or Ctrl + Enter). Hidden tests check your functions and print the result below the editor.

Binary-search for the first entry whose adc is ≥ the input. Then interpolate with t0 + (t1 - t0) * (x - x0) / (x1 - x0) using 32-bit signed math.

#include <stdio.h>
#include <stdint.h>

typedef struct { uint16_t adc; int16_t temp; } cal_t;
static const cal_t table[] = {
    {  500,  850 }, { 1000,  600 }, { 1800,  350 }, { 2600,  150 }, { 3400, -100 },
};
#define TABLE_LEN (sizeof table / sizeof table[0])

int16_t adc_to_temp(uint16_t adc)
{
    if (adc <= table[0].adc) return table[0].temp;
    if (adc >= table[TABLE_LEN - 1].adc) return table[TABLE_LEN - 1].temp;

    size_t lo = 1, hi = TABLE_LEN - 1;          /* first index with table[i].adc >= adc */
    while (lo < hi) {
        size_t mid = lo + (hi - lo) / 2;
        if (table[mid].adc < adc) lo = mid + 1;
        else hi = mid;
    }
    const cal_t *a = &table[lo - 1], *b = &table[lo];
    int32_t dt = (int32_t)b->temp - a->temp;
    int32_t dx = (int32_t)b->adc - a->adc;
    return (int16_t)(a->temp + dt * (int32_t)(adc - a->adc) / dx);
}

A piecewise-linear table in flash is how most firmware linearises non-linear sensors (thermistors, flow sensors, battery discharge curves), because it is far cheaper than evaluating the Steinhart-Hart equation with floating point. Binary search keeps lookup O(log n), which matters for large tables; for evenly spaced ADC points you can compute the index directly.

Watch for overflow in dt * (adc - a->adc) when tables have large ranges, and note that integer division truncates towards zero. Add dx / 2 with the right sign for rounding if needed.

solution.c
OutputPress Run to compile and test your code.
Show the hidden tests
/* ---- tests ---- */
static int fails;
#define CHECK(c) do { if (!(c)) { printf("FAIL line %d: %s\n", __LINE__, #c); fails++; } } while (0)
int main(void)
{
    CHECK(adc_to_temp(100) == 850);       /* below table */
    CHECK(adc_to_temp(500) == 850);
    CHECK(adc_to_temp(750) == 725);       /* halfway 850..600 */
    CHECK(adc_to_temp(1800) == 350);
    CHECK(adc_to_temp(2200) == 250);
    CHECK(adc_to_temp(3000) == 25);       /* 150 + (-250 * 400 / 800) */
    CHECK(adc_to_temp(4095) == -100);     /* above table */
    puts(fails ? "Some tests failed" : "All tests passed");
    return fails != 0;
}