One of the questions asked by test designers is: Do I really care about power consumption while the device is in test mode, on the tester?The answer is yes - power consumption of the device on the tester is also a concern. Although the device usually runs at a low frequency on the tester, the toggling rate of the device on the tester is usually much higher that of normal device operation. Therefore, power consumption of the device on the tester can often be even higher than that of normal operation. Because of this, if the power consumption during testing is not managed well, the device may overheat resulting in abnormal behavior, or in extreme cases, even melt down on the tester.Questions? Comments? Please feel free to post!
I don't know about abnormal behaviour caused by device overheating, but a more likely scenario is that the power-grid and power-pin numbers were based arround a [b]functional[/b] maximum power consuption rather than the peak power drawn during scan-shift (which is the worst part of a test). This can mean a greater than expected IR drop causing unreliable tests. I'm well aware of a number of devices (across the industry, not just nxp) that can only be reliably tested at Vdd-max.So what solutions are there? What do you do to ensure that your device is testable?