My Take On...
Controller Area Network - The Unsolved Mystery

For those who are not familiar with Controller Area Network (CAN), CAN is a very simple, yet powerful and extremely reliable technology that allows a fast serial communication between microprocessor systems. The keyword here is “Distributed Control”, which is not only supported by CAN, but many other field-bus systems. However, CAN stands out in terms of reliability and virtual real-time capabilities. 

The technology was originally developed in Germany for use in automobiles (a joint development between Mercedes-Benz, Bosch, Philips and Intel), but due to all the ingenious features that CAN provides it found its way into any kind of industrial application where microprocessors need to communicate with each other or the outside world. Another - actually unintended - side effect is that CAN, like any other networking technology, reduces the need of extensive wiring. The cost savings (purchasing, installation and maintenance) can be quite dramatic. 

So, what is the “unsolved mystery”?

CAN was developed by engineers of the Robert Bosch Company in Germany. In order to design your own CAN controller you need to purchase a license from Bosch, which also entitles you to the – most probably – only complete documentation of CAN. Don’t get me wrong, you do not need to develop your own CAN controller; CAN chips are available from almost all major semiconductor manufacturers and you don’t need to pay any licenses to use them. The point is that you cannot buy literature that describes the complete CAN standard, unless you buy the Bosch license for numerous thousands of Dollars. 

I do have all the “major” works on CAN in my book shelf, some of them in German, some of them translated from German into English - or shall I say Germlish? I am not trying to fall back into that typical German feature of finding everything that is wrong in this world – life’s a constant misery, death is inevitable - but some of my fellow Germans wrote major publications in a way where only the writer – not the more important entity, the reader – would understand these sentences that may span over several pages. Whatever the reasons may be, lack of journalistic skills, over-proportional ego or the attempt to maneuver around a topic that was not understood, the result is not good. 

The translation process must have been pure and brutal punishment. Sometimes the engineers themselves were involved and, maybe because they got bored, tried to add some cute sentences which produced results like “the CPU has to meat the deadline” – this is not a typo on my part; it really happened. 

My point is the apparent lack of those German features that I was exposed to before I came to the United States: Precision and thoroughness. The CAN technology was first introduced in 1986 and after all those years nobody found it necessary to create an official, professional and, after all, complete document that explains all the ingenious features in all details, accompanied by pictures and examples. Well, there are official documents, even an ISO specification that can be purchased for roughly $5 per page, but they all leave more than ample room for interpretation on most topics. Especially the ISO specification is a document in the style of a novel I just wrote. It's called “The Duel”. Here’s the story: “He saw him. He shot him. He left. The End.” Just like that; German efficiency at its best.

After touring around the United States and Canada for several years, conducting numerous technical sessions on CAN and CANopen, I thought it was time to sit down and write a book on CAN. Well, I did and it took me a year of intense research to find all the little undocumented details, but I also learned that I am not done yet. The book was published in October of 2005 and since then I found more information, most of them through mouth-to-mouth propaganda and I am using this web site to publish every bit of information I find.

So, now I call myself the CAN Man, because I can.