When Passing Electrical Test Is Not Enough
A board passes ICT. It passes FCT. It works perfectly at room temperature. Then it's installed in an engine compartment where temperatures swing from -20°C on a winter morning to +105°C after an hour of operation. Six months later: intermittent connection. Cracked BGA ball. Field failure.
The root cause is thermal fatigue — the cumulative damage from thousands of expansion and contraction cycles. Every material has a CTE (Coefficient of Thermal Expansion). The PCB substrate (FR-4: ~14-17 ppm/°C), the silicon die (~2.6 ppm/°C), the copper pad (~17 ppm/°C), and the SAC305 solder (~22 ppm/°C) all expand at different rates. Every temperature change creates shear stress at every solder joint. Over time, the stress accumulates into cracks.
IPC-9701 is the standard that defines how to test whether solder joints will survive this thermal cycling over the product's designed lifetime. It specifies the test method — not the pass/fail criteria (which come from the product specification).
IPC-9701 Test Method Overview
Test Conditions (IPC-9701 Table 4-1)
| Condition | Temperature Range | Dwell Time | Ramp Rate | Typical Application |
|---|---|---|---|---|
| TC1 | 0°C to +100°C | 10 min | ≤20°C/min | Consumer electronics (indoor) |
| TC2 | -25°C to +100°C | 10 min | ≤20°C/min | Industrial controls |
| TC3 | -40°C to +125°C | 15 min | ≤20°C/min | Automotive (under-hood) |
| TC4 | -55°C to +125°C | 15 min | ≤20°C/min | Military / aerospace |
| TC5 | -55°C to +100°C | 15 min | ≤20°C/min | Military (ground) |
The dwell time ensures the test board reaches uniform temperature throughout — not just the surface, but the internal layers and component bodies. Ramp rate is controlled to prevent thermal shock (which produces different failure modes than fatigue).
Number of Cycles
IPC-9701 does not mandate a specific number of cycles. The number depends on the product's expected lifetime and operating environment:
| Application | Typical Cycles Required | Standard | Status at Superb |
|---|---|---|---|
| Consumer electronics | 100-250 | TC1 or TC2 | Optional |
| Industrial controls | 250-500 | TC2 | Optional |
| Automotive (passenger compartment) | 500-1000 | TC2 or TC3 | Available |
| Automotive (under-hood) | 1000-2000 | TC3 | Available |
| Aerospace / Military | 1000-3000 | TC4 or TC5 | Available |
Electrical Monitoring
IPC-9701 specifies continuous or periodic electrical monitoring during cycling. The standard method is a daisy-chain network — all the solder joints on critical components are connected in series. If any one joint cracks, the daisy chain opens, and the event is time-stamped. This tells you not just that a failure occurred, but when — critical for Weibull reliability analysis.
At Superb, thermal cycling test boards are configured with daisy-chain monitoring on critical BGA and QFN components. An event detector logs each open circuit with a timestamp, correlated to the thermal cycle number.
Post-Cycle Failure Analysis
After cycling, all boards (failed and surviving) undergo failure analysis per IPC-9701:
Visual inspection: 10-40× microscope examination of all solder joints. Look for cracks, discoloration, and deformation.
X-ray inspection: For BGA/QFN, check for crack propagation in solder balls. Compare voiding before and after cycling.
Dye penetrant test (destructive): A low-viscosity fluorescent dye is applied to the board. It wicks into any cracks in solder joints. The board is then pulled apart, and the dye-stained fracture surfaces are examined under UV light. This reveals exactly which joints cracked and where in the solder joint the crack occurred (bulk solder, intermetallic layer, or pad-laminate interface).
Cross-sectioning (destructive): Suspect joints are cut, polished, and examined under a metallurgical microscope. This reveals crack morphology, intermetallic thickness, and any void coalescence.
When Thermal Cycling Is Worth It
Thermal cycling adds cost and lead time. A 500-cycle test at TC3 takes approximately 2 weeks of continuous chamber operation, plus analysis time. It's not necessary for every board.
We recommend thermal cycling when:
The product will experience thermal extremes in its operating environment (outdoor, automotive, industrial)
The board has large BGA packages (>20mm body) — these have the highest thermal mismatch stress
The product is safety-critical — failure means injury, mission loss, or regulatory liability
You are qualifying a new design or a new manufacturer — one-time validation
You have experienced field failures traced to solder joint fatigue — diagnostic testing
For low-cost consumer products operated indoors at room temperature, thermal cycling is usually not cost-effective. The benefit of passing a 500-cycle test for a device that will never see 50°C is near zero.
What IPC-9701 Means for Your Boards
Thermal cycling per IPC-9701 is the most direct way to answer the question: "Will this board survive in my product's operating environment?"
A pass means the solder joints — the most common thermal fatigue failure site — have demonstrated reliability under accelerated conditions that simulate years of field use. A fail means a design or process weakness has been found before a single field return occurs.
At Superb, we offer IPC-9701 thermal cycling as an optional QC service. We'll work with you to select the appropriate test condition and cycle count for your application, set up the daisy-chain monitoring, and deliver a full test report with failure analysis of any joints that did not survive.
