diff --git a/qucs-filter/qf_filter.cpp b/qucs-filter/qf_filter.cpp
index da9c1047..9a38ea35 100644
--- a/qucs-filter/qf_filter.cpp
+++ b/qucs-filter/qf_filter.cpp
@@ -128,7 +128,7 @@ QString filter::num2str(qf_float num) {
       num /= pow(10.0, (qf_float)(3 * expo));
     }
   }
-  QString str = QString::number(num);
+  QString str = QString::number((double)num);
   if (c) {
     str.append(c);
   }
@@ -145,7 +145,7 @@ QString filter::to_qucs() {
   compos += QString("<Pac P1 1 %1 290 18 -26 0 1 \"1\" 1 \"%2 Ohm\" 1 \"0 "
                     "dBm\" 0 \"1 GHz\" 0>\n")
                 .arg(x)
-                .arg(imp_);
+                .arg((double)imp_);
   compos += QString("<GND * 1 %1 320 0 0 0 0>\n").arg(x);
   wires += QString("<0 200 %1 200 \"\" 0 0 0>\n").arg(space / 2);
   wires += QString("<%1 200 %1 260 \"\" 0 0 0>\n").arg(x);
@@ -264,7 +264,7 @@ QString filter::to_qucs() {
   compos += QString("<Pac P2 1 %1 290 18 -26 0 1 \"2\" 1 \"%2 Ohm\" 1 \"0 "
                     "dBm\" 0 \"1 GHz\" 0>\n")
                 .arg(x)
-                .arg(imp_);
+                .arg((double)imp_);
   compos += QString("<GND * 1 %1 320 0 0 0 0>\n").arg(x);
   wires += QString("<%1 200 %1 260 \"\" 0 0 0>\n").arg(x);
 
diff --git a/qucs-filter/qf_math.h b/qucs-filter/qf_math.h
index 1b678d36..18dcc3dd 100644
--- a/qucs-filter/qf_math.h
+++ b/qucs-filter/qf_math.h
@@ -2,7 +2,7 @@
 #define QF_MATH
 
 namespace qf {
-typedef double qf_float;
+typedef long double qf_float;
 static const double pi = 3.1415926535897932384626433832795029; /* pi       */
 static const double one_over_pi = 0.3183098861837906715377675267450287;                       /* 1/pi     */
 static const double ln2 = 0.6931471805599453094172321214581766; /* log_e(2) */