diff --git a/src/logic/unload_to_finda.cpp b/src/logic/unload_to_finda.cpp
index 0ea887d..8896a03 100644
--- a/src/logic/unload_to_finda.cpp
+++ b/src/logic/unload_to_finda.cpp
@@ -24,7 +24,7 @@ void UnloadToFinda::Reset(uint8_t maxTries) {
 
 // @@TODO this may end up somewhere else as more code may need to check the distance traveled by the filament
 int32_t CurrentPositionPulley_mm() {
-    return mm::stepsToUnit<mm::P_pos_t>(mm::P_pos_t({ mm::motion.CurPosition(mm::Pulley) }));
+    return mm::axisUnitToTruncatedUnit<config::U_mm>(mm::motion.CurPosition<mm::Pulley>());
 }
 
 bool UnloadToFinda::Step() {
@@ -48,7 +48,7 @@ bool UnloadToFinda::Step() {
         return false;
     case WaitingForFINDA: {
         int32_t currentPulley_mm = CurrentPositionPulley_mm();
-        if ((abs(unloadStart_mm - currentPulley_mm) > config::fsensorUnloadCheckDistance.v) && mfs::fsensor.Pressed()) {
+        if ((abs(unloadStart_mm - currentPulley_mm) > mm::truncatedUnit(config::fsensorUnloadCheckDistance)) && mfs::fsensor.Pressed()) {
             // fsensor didn't trigger within the first fsensorUnloadCheckDistance mm -> stop pulling, something failed, report an error
             // This scenario should not be tried again - repeating it may cause more damage to filament + potentially more collateral damage
             state = FailedFSensor;
diff --git a/src/modules/axisunit.h b/src/modules/axisunit.h
index faa59d7..ba57611 100644
--- a/src/modules/axisunit.h
+++ b/src/modules/axisunit.h
@@ -112,14 +112,30 @@ static constexpr AU unitToAxisUnit(U v) {
     return { (typename AU::type_t)(v.v * axisScale[AU::axis].stepsPerUnit) };
 }
 
-/// Convert an AxisUnit to unit::Unit.
-/// The scaling factor is stored with the pair config::AxisConfig::uSteps and
-/// config::AxisConfig::stepsPerUnit (one per-axis).
-template <typename U, typename AU>
-static constexpr typename U::type_t axisUnitToUnit(AU v) {
+/// Convert an AxisUnit to a physical unit with a truncated integer type (normally int32_t).
+/// Inverse of unitToAxisUnit, with the additional constraint that type casts are
+/// performed earlier so that no floating point computation is required at runtime.
+/// @param U Base unit type (normally U_mm)
+/// @param AU AxisUnit type (implicit)
+/// @param T Resulting integer type
+/// @param v Value to truncate
+/// @param mul Optional pre-multiplier
+/// @returns Truncated unit (v * mul)
+/// @see unitToAxisUnit
+template <typename U, typename AU, typename T = int32_t>
+static constexpr T axisUnitToTruncatedUnit(AU v, long double mul = 1.) {
     static_assert(AU::unit == U::unit, "incorrect unit type conversion");
-    //static_assert(U::base == axisScale[AU::axis].base, "incorrect unit base conversion");
-    return { (typename U::type_t)(v.v / axisScale[AU::axis].stepsPerUnit) };
+    static_assert(U::base == axisScale[AU::axis].base, "incorrect unit base conversion");
+    return { ((T)v.v / (T)(axisScale[AU::axis].stepsPerUnit / mul)) };
+}
+
+/// Truncate an Unit type to an integer (normally int32_t)
+/// @param v Value to truncate
+/// @param mul Optional pre-multiplier
+/// @returns Truncated unit (v * mul)
+template <typename U, typename T = int32_t>
+static constexpr T truncatedUnit(U v, long double mul = 1.) {
+    return (T)(v.v * mul);
 }
 
 /// Convert a unit::Unit to a steps type (pos_t or steps_t).
@@ -129,13 +145,6 @@ static constexpr typename AU::type_t unitToSteps(U v) {
     return unitToAxisUnit<AU>(v).v;
 }
 
-/// Convert a steps type (pos_t or steps_t) to a unit::Unit.
-/// Extract the raw step count from an AxisUnit with type checking.
-template <typename U, typename AU>
-static constexpr typename U::type_t stepsToUnit(AU pos) {
-    return axisUnitToUnit<U, AU>(pos);
-}
-
 // Pulley
 typedef AxisUnit<pos_t, Pulley, Lenght> P_pos_t; ///< Pulley position type (steps)
 typedef AxisUnit<steps_t, Pulley, Speed> P_speed_t; ///< Pulley speed type (steps/s)