If you are absolutely sure that F4EC9E is representing -725850, and that this is not just a typo (0 vs. 8) or a bug in the existing conversion program, it could be that your negative numbers are represented in some non-standard way. There are several ways to represent signed integers (the most popular being the two's complement). However, AFAICT, none of the common ones would map the given bit pattern to the decimal number -725850. So... if you are sure about this, it might help if you'd post the existing C conversion routine (if you have the source (and its size is reasonable)). Also, it would be interesting to know what the value FFFFFF is representing.