Hi, I have a suspicion about your problem: The DLL you've mentioned: does it in any way communicate with a device? What you said about those ionized molecules supports this. When there's a hardware driver that has been written for internal use, and a DLL that provides an API for it, also meant for internal use, chances are high that these are not as robust as drivers and DLLs meant for the consumer market. That means that it is often easy to write code using such an API or driver that crashes the operating system's kernel (that's the very core of the operating system). It does not mean that the driver and DLL are actually badly designed or written. It only means that they're not supposed to be used by a great number of developers out-of-house. My first guess for a solution: check what you've been writing against the documentation / source code / comments, whatever you have. Second guess: you've actually broken your own code via some flaw in your use of threads that leads to some kind of unexpected behavior that, via the DLL and the driver, crashes the kernel. So, unless the first approach was sufficient, find out whether the behavior you've encountered is reproducible without multiple threads. Cheers from Vienna, The Continent