hard error

A hard error is a equipment failure caused by permanent physical damage to hardware components, not by transient software glitches or temporary signal noise. Once a hard error occurs, power-cycling the equipment will not restore function. The fault persists because the underlying hardware, whether a circuit trace, semiconductor junction, connector pin, or mechanical component, has been physically altered or destroyed and cannot repair itself.

Hard errors differ fundamentally from soft errors, which are temporary faults caused by electromagnetic interference, supply voltage spikes, or cosmic ray strikes that upset logic states without leaving permanent damage. A soft error typically clears when the equipment restarts or the affected circuit is refreshed. A hard error remains until the damaged component is replaced. This distinction matters greatly in troubleshooting: if rebooting fixes the problem, you have a soft error; if it does not, you are facing hardware replacement or repair work.

Common causes and mechanisms

Hard errors often result from electrical overstress events. Latch-up occurs when parasitic thyristors on a CMOS chip are triggered by excessive current or voltage, causing destructive current surge that melts metal interconnects or semiconductor material. Electrostatic discharge (ESD) can rupture oxide layers in transistor gates. Thermal stress from sustained overheating causes solder joints to crack or wirebonds to fail. Mechanical stress, dropped equipment, bent pins, crushed connectors, creates obvious hard damage. Manufacturing defects like voids in solder or contamination in semiconductor material also manifest as hard errors, often after a short period of operation when localized current density exceeds the integrity of the flaw.

In the field, hard errors demand component replacement rather than firmware updates or recalibration. For critical systems, this means understanding the root cause to prevent recurrence. If a hard error occurred without an obvious external event, the equipment may have been exposed to an abnormal supply voltage, excessive ambient temperature, humidity ingress causing leakage paths, or a design margin insufficient for the actual operating environment. Field replacement of damaged boards or modules is standard practice; repair of failed semiconductors or fine-pitch components is rarely economical.

The term reflects a contrast with earlier computing eras when almost all failures were mechanical. It has persisted even as the mechanisms have evolved from relay and vacuum tube failure to modern solid-state damage modes. Understanding whether a given failure is hard or soft guides the entire service response: soft errors warrant diagnostics and environmental review; hard errors warrant visual inspection for physical damage, component-level testing, and replacement planning.