Learn from a transient sensor disturbance

SSLV-D1 launched on August 7, 2022 but did not deliver its payloads into a sustainable orbit. ISRO’s investigation linked a separation disturbance to temporary accelerometer saturation and fault-detection logic that selected a salvage mode. The sensors later functioned normally, but their readings were no longer used by that selected mode.

The account describes open-loop steering and bypass of the VTM in the implemented salvage sequence. The fallback executed its programmed response without achieving the intended mission outcome. A response can behave as written and still be unsuitable for the real combination of conditions.

A saturated sensor is outside its usable measurement range during that event; it is not necessarily permanently broken. In an original diagnostic example, three agreeing sensors encounter a brief vibration pulse, disagree during the pulse and then agree again. The event history matters when deciding what information can be trusted afterward.

A longer observation window can help distinguish a temporary disturbance from lasting degradation, but waiting can also delay response to a real persistent fault. There is no universal rule to “always wait longer.” The system needs requirements for detection speed, false alarms, available backup measurements and the consequences of each decision.

Follow a disturbance through the system

ISRO’s corrective plan combined changes to separation hardware and structural response with revised detection logic, navigation fallback and use of terminal propulsion. The lesson is broader than replacing a sensor: the disturbance, the measurement, the decision and the response interact.

For an original test plan, include short pulses, persistent biased readings and loss of a backup signal. Check both whether the fault is identified and whether the selected response still meets its stated objective. Record the configuration and limitations so a successful software test is not mistaken for qualification of every physical separation condition.