Click to view a larger image

A manufacturer of high-speed inkjet printheads was having spurious, destructive failures in their driver chips. They had not been able to observe the failure occurring for two reasons: first, because the failures were infrequent and not related to any particular operating conditions; and second, because the chip's Vss was riding on a 200V, 2 microsecond pulse, preventing them from measuring waveforms accurately.

We first obtained a fully floating, battery powered oscilloscope that allowed us to observe the waveforms with good fidelity. This allowed us to find that one of the inputs was going below ground during the pulse transition, permitting the chip to go into latchup. The source of this voltage was traced to the inductance of a trace on the driver board. The trace was moved to reduce the inductance, and the failure rate dropped.