I've been noticing a disturbing trend on site and perhaps this is due to so many junior members but it seems that there are a lot of problems with simulators. Either the simulators work properly and give good results which can't be duplicated in the real world or they give problems which prevent proper operation of relatively simple devices that could be easily tested in a real circuit.
I don't know if the fault is that universities focus a lot on simulators or whether this is because people are afraid of using the real hardware. Here are a few guidelines of when not to use simulators:
Analog simulators (e.g., Spice) are thoroughly tested and have been proven over a lot of time. So simulating analog circuits with Spice is generally safe as long as the device model can be trusted. As always, ensure you verify the simulator device model if it gives strange results that defy common sense.
I don't know if the fault is that universities focus a lot on simulators or whether this is because people are afraid of using the real hardware. Here are a few guidelines of when not to use simulators:
- You have the hardware readily available and it is fairly simple (i.e., a push-button, a 2-line LCD or a I2C memory). These cases actually occurred in several forums.
- Your application is mainly software and you have a debugger available. It is better to test software on real hardware as opposed to simulators unless you're cycle counting to optimise an algorithm. Be wary of simulators because they may not model the processor cycles properly until you have verified with real hardware the simulator predictions.
- If you have a lot of of interrupts or real-world I/O. Simulators usually are very poor in handling these elements.
Analog simulators (e.g., Spice) are thoroughly tested and have been proven over a lot of time. So simulating analog circuits with Spice is generally safe as long as the device model can be trusted. As always, ensure you verify the simulator device model if it gives strange results that defy common sense.
