diff --git a/Makefile.dep b/Makefile.dep
index eb0e4c67af39b73e6205f1936b9305d329ce8ff9..c0b24c5f550cfb900c9dfa24d92ce695b1249230 100644
--- a/Makefile.dep
+++ b/Makefile.dep
@@ -543,6 +543,10 @@ ifneq (,$(filter xtimer,$(USEMODULE)))
USEMODULE += div
endif
+ifneq (,$(filter saul,$(USEMODULE)))
+ USEMODULE += phydat
+endif
+
ifneq (,$(filter saul_reg,$(USEMODULE)))
USEMODULE += saul
endif
diff --git a/sys/include/phydat.h b/sys/include/phydat.h
index df49167eb6efffa83f05d12c6847453aff12e351..0ab28226ff389233161efb6a76829e802293c74a 100644
--- a/sys/include/phydat.h
+++ b/sys/include/phydat.h
@@ -178,6 +178,40 @@ const char *phydat_unit_to_str(uint8_t unit);
*/
char phydat_prefix_from_scale(int8_t scale);
+/**
+ * @brief Scale an integer value to fit into a @ref phydat_t
+ *
+ * Insert @p value at position @p index in the given @p dat while rescaling all
+ * numbers in @p dat->val so that @p value fits inside the limits of the data
+ * type, [@ref PHYDAT_MIN, @ref PHYDAT_MAX], and update the stored scale factor.
+ * The result will be rounded towards zero (the standard C99 integer division
+ * behaviour).
+ * The final parameter @p prescale can be used to chain multiple calls to
+ * this function in order to fit multidimensional values into the same phydat_t.
+ *
+ * The code example below shows how to chain multiple calls via the @p prescale parameter
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ {.c}
+ * long val0 = 100000;
+ * long val1 = 2000000;
+ * long val2 = 30000000;
+ * phydat_t dat;
+ * dat.scale = 0;
+ * phydat_fit(&dat, val0, 0, phydat_fit(&dat, val1, 1, phydat_fit(&dat, val2, 2, 0)));
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ *
+ * The prescale scaling is only applied to @p value, the existing values in
+ * @p dat are only scaled if the prescaled @p value does not fit in phydat_t::val
+ *
+ * @param[in, out] dat the value will be written into this data array
+ * @param[in] value value to rescale
+ * @param[in] index place the value at this position in the phydat_t::val array
+ * @param[in] prescale start by scaling the value by this exponent
+ *
+ * @return scaling offset that was applied
+ */
+uint8_t phydat_fit(phydat_t *dat, long value, unsigned int index, uint8_t prescale);
+
#ifdef __cplusplus
}
#endif
diff --git a/sys/phydat/phydat.c b/sys/phydat/phydat.c
new file mode 100644
index 0000000000000000000000000000000000000000..4b90e4006a99685a21667a8acc4dd79b366db1d8
--- /dev/null
+++ b/sys/phydat/phydat.c
@@ -0,0 +1,47 @@
+/*
+ * Copyright (C) 2018 Eistec AB
+ *
+ * This file is subject to the terms and conditions of the GNU Lesser
+ * General Public License v2.1. See the file LICENSE in the top level
+ * directory for more details.
+ */
+
+/**
+ * @ingroup sys_phydat
+ * @{
+ *
+ * @file
+ * @brief Generic sensor/actuator data handling
+ *
+ * @author Joakim Nohlgård <joakim.nohlgard@eistec.se>
+ *
+ * @}
+ */
+
+#include <stdint.h>
+#include "phydat.h"
+
+#define ENABLE_DEBUG 0
+#include "debug.h"
+
+uint8_t phydat_fit(phydat_t *dat, long value, unsigned int index, uint8_t prescale)
+{
+ assert(index < (sizeof(dat->val) / sizeof(dat->val[0])));
+ uint8_t ret = prescale;
+ while (prescale > 0) {
+ value /= 10;
+ --prescale;
+ }
+ int8_t scale_offset = 0;
+ while ((value > PHYDAT_MAX) || (value < PHYDAT_MIN)) {
+ value /= 10;
+ for (unsigned int k = 0; k < (sizeof(dat->val) / sizeof(dat->val[0])); ++k) {
+ dat->val[k] /= 10;
+ }
+ ++scale_offset;
+ }
+ dat->val[index] = value;
+ dat->scale += scale_offset;
+ ret += scale_offset;
+ return ret;
+}
diff --git a/sys/phydat/phydat_str.c b/sys/phydat/phydat_str.c
index b8d2e2ebac5d2251926269f7a25b678262ababc8..2a73661f70938215b6830030942dacd2d6363e57 100644
--- a/sys/phydat/phydat_str.c
+++ b/sys/phydat/phydat_str.c
@@ -7,7 +7,7 @@
*/
/**
- * @ingroup driver_sensif
+ * @ingroup sys_phydat
* @{
*
* @file
diff --git a/tests/unittests/tests-phydat/Makefile b/tests/unittests/tests-phydat/Makefile
new file mode 100644
index 0000000000000000000000000000000000000000..48422e909a47d7cd428d10fa73825060ccc8d8c2
--- /dev/null
+++ b/tests/unittests/tests-phydat/Makefile
@@ -0,0 +1 @@
+include $(RIOTBASE)/Makefile.base
diff --git a/tests/unittests/tests-phydat/Makefile.include b/tests/unittests/tests-phydat/Makefile.include
new file mode 100644
index 0000000000000000000000000000000000000000..7583f4db7aaa35443a4423dcb3eb8e7219c632b8
--- /dev/null
+++ b/tests/unittests/tests-phydat/Makefile.include
@@ -0,0 +1 @@
+USEMODULE += phydat
diff --git a/tests/unittests/tests-phydat/tests-phydat.c b/tests/unittests/tests-phydat/tests-phydat.c
new file mode 100644
index 0000000000000000000000000000000000000000..af7e7b99a171f2e1c0a9192ed7942fb165369524
--- /dev/null
+++ b/tests/unittests/tests-phydat/tests-phydat.c
@@ -0,0 +1,73 @@
+/*
+ * Copyright (C) 2018 Eistec AB
+ *
+ * This file is subject to the terms and conditions of the GNU Lesser
+ * General Public License v2.1. See the file LICENSE in the top level
+ * directory for more details.
+ */
+
+#include "embUnit.h"
+#include "tests-phydat.h"
+
+#include "phydat.h"
+
+#define ENABLE_DEBUG (0)
+#include "debug.h"
+
+static void test_phydat_fit(void)
+{
+ /* verify that these big numbers are scaled to fit in phydat_t::val which is int16_t */
+ long val0 = 100445;
+ long val1 = 2000954;
+ long val2 = 30000455;
+ long val4 = 1234567;
+ phydat_t dat;
+ dat.scale = -6;
+ dat.unit = UNIT_V;
+ uint8_t res = phydat_fit(&dat, val0, 0, 0);
+ /* Check that the result was rescaled to 10044e-5 */
+ /* The scaled number is rounded toward zero */
+ TEST_ASSERT_EQUAL_INT(1, res);
+ TEST_ASSERT_EQUAL_INT(UNIT_V, dat.unit);
+ TEST_ASSERT_EQUAL_INT(-5, dat.scale);
+ TEST_ASSERT_EQUAL_INT( 10044, dat.val[0]);
+ /* Fit the next value in the phydat vector */
+ res = phydat_fit(&dat, val1, 1, res);
+ TEST_ASSERT_EQUAL_INT(2, res);
+ TEST_ASSERT_EQUAL_INT(UNIT_V, dat.unit);
+ TEST_ASSERT_EQUAL_INT(-4, dat.scale);
+ TEST_ASSERT_EQUAL_INT( 1004, dat.val[0]);
+ TEST_ASSERT_EQUAL_INT( 20009, dat.val[1]);
+ /* Fit the third value in the phydat vector */
+ res = phydat_fit(&dat, val2, 2, res);
+ TEST_ASSERT_EQUAL_INT(3, res);
+ TEST_ASSERT_EQUAL_INT(UNIT_V, dat.unit);
+ TEST_ASSERT_EQUAL_INT(-3, dat.scale);
+ TEST_ASSERT_EQUAL_INT( 100, dat.val[0]);
+ TEST_ASSERT_EQUAL_INT( 2000, dat.val[1]);
+ TEST_ASSERT_EQUAL_INT( 30000, dat.val[2]);
+ /* Overwrite the second value in the phydat vector */
+ res = phydat_fit(&dat, val4, 1, res);
+ TEST_ASSERT_EQUAL_INT(3, res);
+ TEST_ASSERT_EQUAL_INT(UNIT_V, dat.unit);
+ TEST_ASSERT_EQUAL_INT(-3, dat.scale);
+ TEST_ASSERT_EQUAL_INT( 100, dat.val[0]);
+ TEST_ASSERT_EQUAL_INT( 1234, dat.val[1]);
+ TEST_ASSERT_EQUAL_INT( 30000, dat.val[2]);
+}
+
+Test *tests_phydat_tests(void)
+{
+ EMB_UNIT_TESTFIXTURES(fixtures) {
+ new_TestFixture(test_phydat_fit),
+ };
+
+ EMB_UNIT_TESTCALLER(phydat_tests, NULL, NULL, fixtures);
+
+ return (Test *)&phydat_tests;
+}
+
+void tests_phydat(void)
+{
+ TESTS_RUN(tests_phydat_tests());
+}
diff --git a/tests/unittests/tests-phydat/tests-phydat.h b/tests/unittests/tests-phydat/tests-phydat.h
new file mode 100644
index 0000000000000000000000000000000000000000..1697ecba831dfa272d553b07e623d50d8403e24d
--- /dev/null
+++ b/tests/unittests/tests-phydat/tests-phydat.h
@@ -0,0 +1,36 @@
+/*
+ * Copyright (C) 2018 Eistec AB
+ *
+ * This file is subject to the terms and conditions of the GNU Lesser
+ * General Public License v2.1. See the file LICENSE in the top level
+ * directory for more details.
+ */
+
+/**
+ * @addtogroup unittests
+ * @{
+ *
+ * @file
+ * @brief Unittests for the phydat module
+ *
+ * @author Joakim Nohlgård <joakim.nohlgard@eistec.se>
+ */
+#ifndef TESTS_PHYDAT_H
+#define TESTS_PHYDAT_H
+#include "embUnit/embUnit.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+* @brief The entry point of this test suite.
+*/
+void tests_phydat(void);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* TESTS_PHYDAT_H */
+/** @} */