Typical Electrical Failure Modes of Refurbished Chips
In the electronic components market, refurbished chips circulate at prices far below those of original manufacturer parts, but their latent electrical faults often only gradually surface after they are installed in a system. The refurbishing process typically involves high-temperature desoldering, chemical cleaning, grinding and re-marking, and other steps. These operations introduce physical and electrical defects into the chip interior, causing the device to "appear normal" during initial testing yet fail frequently in actual operation. Electrical failures in refurbished chips have three major characteristics: concealment, batch occurrence, and gradual progression. Their typical manifestations can be identified from the following dimensions.
I. Open-Circuit and Short-Circuit Failures
The most direct electrical failures during refurbishing are open circuits and short circuits. Mechanical stress applied during grinding and refurbishing can cause the chip body to crack, and after the crack propagates, it leads to open circuits between pins. A failure analysis case showed that the internal chip body of an LB1836 refurbished device cracked due to grinding stress, causing an open circuit between pins and ultimately making the instrument's built-in printer completely inoperable. In addition, an uncontrolled temperature profile during desoldering or the "popcorn effect" occurring in a moisture-absorbed chip at high temperature can cause substrate delamination and pad tearing, manifesting as short circuits between pins or intermittent contact failure.
II. Parameter Drift and Timing Disturbances
Parameter drift is the most concealed and also the most dangerous failure mode of refurbished chips. After long-term use, the transistor threshold voltage increases due to negative bias temperature instability (NBTI) and hot-carrier injection, the MOSFET switching speed decreases, and the chip's overall operating frequency consequently drops. Taking a refurbished SDRAM as an example, its tRCD parameter expands from the original manufacturer specification of 18 ns (±0.3 ns) to 19.2–20.4 ns. Under an 80°C environment and full-load burst-write conditions, some samples exhibited abnormally prolonged tRC, triggering the DRAM controller's retry mechanism and further causing the ARM7TDMI core bus to hang. This narrowing of timing margins may not appear at room temperature and light load, but under high-temperature or full-load conditions it can suddenly cause system crashes.
III. Abnormal Power Consumption and Degraded Thermal Stability
Both the static and dynamic power consumption of refurbished chips may deviate from the original manufacturer's specifications. On the one hand, increased leakage current caused by aging raises standby power consumption; on the other hand, some refurbished devices exhibit local overheating at high temperatures, and the thermal stability of power devices is significantly reduced. In power management chips, the internal reference voltage of refurbished parts may shift, causing output voltage ripple to exceed limits. In actual measurements, the VDDQ ripple of a refurbished SDRAM was observed to exceed ±120 mV, far above the original manufacturer's ±50 mV upper limit.
| Failure Type | Typical Manifestation | Root Cause |
|---|---|---|
| Open Circuit | No continuity between pins; device completely inoperative | Chip cracking and pad tearing caused by grinding stress |
| Short Circuit | Short between adjacent pins or from power to ground | Solder ball detachment and bridging; substrate delamination |
| Parameter Drift | Timing parameters such as tRCD and tRC out of specification; frequency drop | Increased transistor threshold voltage; NBTI aging |
| Abnormal Power Consumption | Increased static leakage current; VDDQ ripple exceeding limits | Gate oxide degradation; internal reference shift |
| Thermal Failure | Abnormal operation at high temperature; degraded heat dissipation | Increased thermal resistance; local overheating |
IV. Bidirectional Anomalies in Leakage Current
The leakage current behavior of refurbished chips is peculiar. The NBTI effect causes the threshold voltage of PMOS transistors to increase, which reduces standby leakage current, but this does not mean improved performance—the threshold voltage shift itself lowers switching speed and narrows timing margins. At the same time, local damage caused by ESD breakdown or gate oxide degradation increases leakage current, manifesting as reduced resistance from power pins to ground and IV characteristic curves deviating from the normal range. Therefore, judging whether a chip is refurbished solely by whether leakage current is high or low is unreliable; a comprehensive evaluation combining threshold voltage, switching speed, and other multiple parameters is required.
Frequently Asked Questions
Q: Can refurbished chips pass conventional electrical testing before installation?
A: In many cases, yes. What refurbishing introduces are "latent defects." The device may perform normally in initial testing at room temperature and light load, but will fail after high temperature, full load, or long-term operation. Data from Integra Technologies' laboratory shows that the measured failure rate of refurbished devices is often 20% or even higher. It is recommended to conduct electrical testing under the design temperature and vibration conditions to effectively expose latent defects.
Q: Does parameter drift mean a refurbished chip is completely unusable?
A: It depends on the application scenario. For non-critical consumer-grade control boards, some refurbished chips that have undergone functional screening may barely be usable. However, in scenarios with strict requirements for timing accuracy and reliability, such as real-time control, data acquisition, medical, or automotive electronics, parameter drift directly threatens system functional safety, and refurbished devices should not be used.
Q: How can "parameter drift" be distinguished from "normal batch variation"?
A: Normal batch variation typically fluctuates within the tolerance range specified in the original manufacturer's datasheet, and the distribution is concentrated and predictable. Parameter drift in refurbished chips manifests as exceeding datasheet tolerances, significantly increased distribution dispersion, and is often related to temperature and usage history—within the same batch, devices with longer usage time show more severe parameter shift.
Conclusion
The electrical failures of refurbished chips are not a single mode, but a composite failure interwoven from open/short circuits, parameter drift, abnormal power consumption, and leakage current changes. Their core hazard is that the faults are often invisible in the initial stage, yet suddenly erupt at the moment when the system most needs reliability. For circuit designs involving safety-critical functions, establishing a multi-stage screening process from incoming IV characteristic testing to high-temperature dynamic aging is far more cost-effective than troubleshooting after the fact.
