I went on a course a while ago in which I learnt that what you seen on your oscilloscope screen in not necessarily representative of the signal you’re trying to measure
Part two of this lesson in real-world electronics is this…..
I’ve got this board where one of the signals coming from a clock buffer is being distorted. Removing the series resistor that sits next to the output of the buffer makes the signal look nice and clean with sharp edges and everything. So it’s not the input to the circuit. I’ve just lifted the legs of the clock pins the trace is leading to and……it still looks the same. Conclusion the trace is causing the problem.
This is where the simulation comes in. From the layout file the PCB guys designed I can extract the data for that trace and use a tool to simulate the trace with lossy microstrips in place of the tracks on the physical board.
The lesson is that you get out what you put into these kind of simulations. Try as we might we can’t get the trace to represent what we see in the real world (not without entering stupid values anyway – the kind of problem we would have noticed already if that is what is happening). I’m stumped. I can’t think of anything big we’ve forgotton but the simulations don’t match real life.
Simulation is obviously a useful tool when you don’t have a physical piece of hardware infront of you for some reason – e.g. it hasn’t been designed yet. But hopefully today has demonstrated that it isn’t the gospel truth – just another useful tool that has to be used carefully – much like a circular saw.
