Multipoint grounding – a post in which I draw a target on myself. I’ve been involved in some really contentious arguments on the topic. I wish I could say they were discussions, but the religious fervor of some single-point grounding advocates precludes that.
I’ve had some constructive conversations on this topic in some of our PROFINET one-day training classes. I remember in St. Louis some students came to the defense of multipoint grounding based on their own experiences. After our Detroit class five years ago I was prompted to blog along these lines:
In polite company, we try not to speak about religion, politics or grounding. In our classes we successfully avoid the first two. For PROFIBUS and for PROFINET the official PI recommendation is to ground shielded cable at both ends and at any intermediate points. This is recommended because it’s desirable to bleed off any noise at the closest possible opportunity. To avoid other problems, the grounding points need to be at ground potential – the same ground potential. My impression is that outside North America they will be. Here, not so much… unless careful measures are taken. We recommend you take those measures.
In polite company we try not to speak about religion, politics, grounding, or shielding. In our classes we successfully avoid the first two. But we have to talk about shielding. We recommend it and cannot understand why anyone would argue with that. Some do so argue though. A shield (properly grounded) will help keep the signal clean in electrically noisy environments… electrically noisy environments like any industrial environment. The signal protocols cannot guard against electrical noise; only a shield can do that. But TCP can initiate retries if noise occurs, the arguers argue. True, but not helpful – it takes too long to retry and if the retries are frequent due to noise… well you get the picture.
Feel free to discuss single-point versus multipoint grounding in the comments, but be nice.
–Carl Henning