The time has come. After more than 50 years of progress in the field of digital electronics, we have reached a breakthrough moment. I am ready now, today, right here, to declare the start of a new era. Just to make things official, here’s my statement:
Be it known in all corners of the globe, that from this day forward, there is absolutely, completely, totally no longer any excuse whatsoever for system problems, glitches, data errors or other artifacts related to ringing in digital signals.
Before we all throw our pocket protectors into the air and shout "hooray!" let’s take a good close look at the wording of my pronouncement. After all, you may be wondering, has there been any change in the laws of physics? Has ringing just gone away? Can we henceforth just ignore it?
The answer to these questions is, of course, an emphatic "No, no, and no."
I didn’t say ringing would no longer exist. Nor did I imply that it would not occasionally raise its ugly head to munch on a tender signal or two. Far from it. All I said was that there would no longer be any *excuse* for problems caused by ringing. I feel confident making this statement because ringing is a totally preventable system problem.
The most important thing to realize about ringing is that it is a deterministic, predictable, system artifact that can be simulated with incredible accuracy.
In the world of ringing, simulation is the ideal way to prevent problems. Why? Because if a problem can be simulated, it is by definition preventable--if a circuit won’t work, nobody will build it.
Please don’t misinterpret this as a blanket endorsement of simulation for all digital design problems. In some applications, today’s simulation technology just doesn’t work. For example, try simulating the crosstalk on traces that pass over a split plane boundary. In that sort of complex, three-dimensional electromagnetic field application, most simulators don’t have enough muscle to do the job. However, the simulation of ringing on PCB traces, in the presence of a solid ground plane, with known source and load impedances, and with a known risetime, is a well-known, easily calculable problem. If the simulation says a circuit OK, it will probably work.
Computer simulation of ringing, even with all its warts, is far better than the kind of simple rules of thumb I see used in design shops all around the world. Old rules like "3 inch trace stubs are OK", that may have served well in the past, don’t even come close to cutting it with today’s super-fast digital logic. If you really want to know how far you can push a trace, simulate it.
I can’t tell you how many engineers have contacted me with some weird bus configuration, wanting to know if it’s going to work. The Answer: if you have to ask, simulate it.
What if you don't have a simulator? That brings me to my last point, the "no excuse" part.
Simulation technology is widespread, easy to use, and cheap. If, by using a simulator, you can save one design spin on one circuit board, the simulator pays for itself. If you don’t have a simulator, get one. If you already have one but aren’t satisfied with it, here are some sources that may interest you: