lib/rmodbus/ext.rb

@@ -69,6 +69,21 @@ def from_32i
     self.pack('N*').unpack('n*').each_slice(2).map { |arr| arr.reverse }.flatten
   end
 
+  # As seen here: https://en.wikipedia.org/wiki/IEEE_floating_point
+  def to_ieee754f
+    binary = self.map{|e| e.to_s(2).rjust(16, '0')}.inject{|bin, e| bin + e}
+    negative_multiplier = binary[0] == "1" ? -1.0 : 1.0
+    exponent = binary[1..8].to_i(2) - 127
+    mantissa = "1." + binary[9..-1]
+    number = negative_multiplier * 10 ** exponent * mantissa.to_f
+    splitted = number.to_s.split(".")
+    before_comma = splitted[0].to_i(2)
+    after_comma = splitted[1].split(//).each_with_index.map do |bit, i|
+      bit.to_i * 2 ** (-i-1)
+    end.inject{|memo, e| memo + e}
+    (before_comma + after_comma).to_f
+  end
+
   def pack_to_word
     word = 0
     s = ""

spec/ext_spec.rb

@@ -29,6 +29,11 @@
     [20342, 17344, 20342, 17344].to_32f.size.should == 2
   end
 
+  it "should turn an array of bytes in IEEE754 format into the correct 32b float" do
+    [17264,32768].to_ieee754f.should be_within(0.1).of(240.5)
+    [16029, 2081].to_ieee754f.should be_within(0.1).of(0.3)
+  end
+
   it "should turn an array from 32b ints into 16b ints, big endian" do
     [1136676726].from_32i.should == [20342, 17344]
     [1136676726, 1136676725].from_32i.should == [20342, 17344, 20341, 17344]