Depends on how you define "failure". Calibration drift is the most common failure, but since diving doesn't need the accuracy of process control, being 2% off is not a big deal (100 psi per 5000). Problem is that inaccuracy is usually not linear, often being larger at minimum or maximum scale, or can be a fixed amount (needle offset) which makes for huge errors at low pressures.
SPGs should not only be serviced for the moving parts (o-rings, swivels) but also checked for calibration. Several OOA incidents have been attributed to out of calibration SPGs.
Significant calibration drift is present about 6 times per million hours of operation, assuming failure frequencies similar to what is seen in industry for bourdon tube gauges.
Electronic pressure transmitters that have self-diagnostics are usually 10 to 1000 times more reliable than analog instruments (usually because they will tell you when they are having problems), but I have no idea what is being used in dive computers. Even with electronic transmitters, calibration needs to be verified on an annual basis, so if you have an air integrated computer, don't forget to have it checked.
SPGs should not only be serviced for the moving parts (o-rings, swivels) but also checked for calibration. Several OOA incidents have been attributed to out of calibration SPGs.
Significant calibration drift is present about 6 times per million hours of operation, assuming failure frequencies similar to what is seen in industry for bourdon tube gauges.
Electronic pressure transmitters that have self-diagnostics are usually 10 to 1000 times more reliable than analog instruments (usually because they will tell you when they are having problems), but I have no idea what is being used in dive computers. Even with electronic transmitters, calibration needs to be verified on an annual basis, so if you have an air integrated computer, don't forget to have it checked.