diff --git a/radio/src/Makefile b/radio/src/Makefile
index c80161774..e21c42da3 100644
--- a/radio/src/Makefile
+++ b/radio/src/Makefile
@@ -126,6 +126,13 @@ CORRECT_NEGATIVE_VALUES = YES
 # Values = YES, NO
 ARITHMETIC_OVERFLOW_CHECK = NO
 
+# increases resolution for throttle statistics and throttle counting, in very rare cases.
+# this will be only true if an output channel is used for throttle calculation, and this chanenel has changed limits, e.g. -90% to +80%
+# if you use standard limits for throttle or directly the throttle stick you do not need to bother about this option
+# not valid for ARM based transmitters like Taranis, They do not have this rounding issue at all
+# Values = YES, NO
+ACCURAT_THROTTLE_STATS = NO
+
 # HAPTIC Mods
 # Values = YES, NO 
 HAPTIC = NO
@@ -857,6 +864,9 @@ ifeq ($(ARITHMETIC_OVERFLOW_CHECK), YES)
   CPPDEFS += -DPREVENT_ARITHMETIC_OVERFLOW
 endif
 
+ifeq ($(ACCURAT_THROTTLE_STATS), YES)
+  CPPDEFS += -DACCURAT_THROTTLE_TIMER
+endif
 
 # EXTENSIONS (for Telemetry protocols)
 ifneq ($(EXT), $(filter $(EXT), STD FRSKY ARDUPILOT NMEA JETI MAVLINK TELEMETREZ))
diff --git a/radio/src/opentx.cpp b/radio/src/opentx.cpp
index 0adec523c..19cc7781f 100644
--- a/radio/src/opentx.cpp
+++ b/radio/src/opentx.cpp
@@ -4179,6 +4179,13 @@ void doMixerCalculations()
   if (!tick10ms) return; //make sure the rest happen only every 10ms.
 #endif
 
+#if !defined(CPUM64) && !defined(ACCURAT_THROTTLE_TIMER)
+  //  code cost is about 16 bytes for higher throttle accuracy for timer 
+  //  would not be noticable anyway, because all version up to this change had only 16 steps; 
+  //  now it has already 32  steps; this define would increase to 128 steps
+  #define ACCURAT_THROTTLE_TIMER
+#endif
+
   /* Throttle trace */
   int16_t val;
 
@@ -4199,18 +4206,23 @@ void doMixerCalculations()
     if (lim->symetrical)
       val -= calc1000toRESX(lim->offset);
 #endif
-    // @@@ open.20.fsguruh  optimized calculation; now *8 /8 instead of 10 base; (*16/16 already cause a overrun; unsigned calculation also not possible, because v may be negative)
+
+
     gModelMax -= gModelMin; // we compare difference between Max and Mix for recaling needed; Max and Min are shifted to 0 by default
-    // usually max is 1024 min is -1024 --> max-min = 2048 full range / 128 = max 16 steps
-    
+    // usually max is 1024 min is -1024 --> max-min = 2048 full range
+
+#ifdef ACCURAT_THROTTLE_TIMER
+    if (gModelMax!=0 && gModelMax!=2048) val = (int32_t) (val << 11) / (gModelMax); // rescaling only needed if Min, Max differs
+#else
+    // @@@ open.20.fsguruh  optimized calculation; now *8 /8 instead of 10 base; (*16/16 already cause a overrun; unsigned calculation also not possible, because v may be negative)
+    gModelMax+=255; // force rounding up --> gModelMax is bigger --> val is smaller
     gModelMax >>= (10-2);
 
-    if (gModelMax != 0 && gModelMax != 8) {
+    if (gModelMax!=0 && gModelMax!=8) {
       val = (val << 3) / gModelMax; // rescaling only needed if Min, Max differs
     }
+#endif
     
-    // if (gModelMax!=2048) val = (int32_t) (val << 11) / (gModelMax); // recaling only needed if Min, Max differs
-    // val = val * 10 / (10+(lim->max-lim->min)/20);
     if (val<0) val=0;  // prevent val be negative, which would corrupt throttle trace and timers; could occur if safetyswitch is smaller than limits
   }
   else {
@@ -4221,13 +4233,6 @@ void doMixerCalculations()
 #endif
   }
 
-#if !defined(CPUM64)
-  //  code cost is about 16 bytes for higher throttle accuracy for timer 
-  //  would not be noticable anyway, because all version up to this change had only 16 steps; 
-  //  now it has already 32  steps; this define would increase to 128 steps
-  #define ACCURAT_THROTTLE_TIMER
-#endif
-
 #if defined(ACCURAT_THROTTLE_TIMER)
   val >>= (RESX_SHIFT-6); // calibrate it (resolution increased by factor 4)
 #else