If you have not asked what is the width of your power trace when you’re designing your PCB, you should start asking now. It’s a mistake that I vowed never to repeat again, as it cost me thousands of dollars in failed production boards and hours of redesign and manufacture.
Let’s use the bottom design as an example. A simple microcontroller driving a TPIC6C595 Darlington chip via a voltage translator. There are 8 super-bright LED that I’m driving at 20mA each by a 5V supply.
I auto-routed with the default design rule of my software, which is 10 mil for all traces. And there appear to be no significant problems.
I then used the same concept to design a medical gas panel. where I drive multiple super bright LED arrays as below. (Note – The LED Arrays are in DIP form with 8 LED per package.)
I had a huge issue with the power tracks heating up at a certain point, even though I’ve routed the power trace with 10 mil copper width. Because I did not take make a proper calculation of PCB trace width the total current passing through.
Here’s a simple calculation for the minimum trace width for a 10°C temperature rise I’ve done with an online PCB trace calculator.
For the first example, with only 8 LEDs at 20mA each, the required trace width is 0.944mil. In my design, the power trace is 10mil, which is way exceeding the requirement.
For the second example, I ended up driving 35 LED Arrays with 50mA each. So that gave a maximum of 1.75A current that could be passing through the power trace leading to the LEDs. The same calculation required a minimum 25.6 mil which is definitely higher than the 10 mil trace on the board.
With the power traces heating up drastically at certain areas, I have no choice but to perform a major revision on the board prior to production.