Welcome: ADD Components Limited
Language:
∷
∷
∷
Stress is the invisible killer of discrete semiconductor joints. You cannot see it under the microscope. You cannot measure it with a multimeter. But it is there, sitting inside every solder joint, waiting for a thermal cycle or a vibration event to turn a good joint into a crack. The joint looks perfect when it leaves the line. It passes every inspection. Then three months later, the field returns start rolling in, and nobody can figure out why.
The answer is almost always stress. Too much heat during soldering created residual stress in the joint. The component expanded, the board expanded, they expanded at different rates, and when they cooled, the stress got locked into the solder. That locked-in stress does not show up as a visible defect. It shows up as a sudden failure when the board flexes or heats up in service.
Controlling stress during discrete semiconductor soldering is not about slowing down the process. It is about understanding where the stress comes from and eliminating it at the source.
Most people think stress comes from thermal shock — heating the board up fast and cooling it down fast. That is part of it, but it is not the whole story. Stress in a solder joint comes from three sources: thermal mismatch, mechanical constraint, and uneven heating. All three happen during soldering, and all three can be controlled if you know what to look for.
Every material expands when it heats up and contracts when it cools down. The rate of expansion is different for every material. Copper expands at about 17 parts per million per degree Celsius. Silicon expands at about 2.6 parts per million per degree Celsius. The epoxy in a transistor package expands at about 15 to 20 parts per million per degree Celsius depending on the formulation.
When you heat a discrete semiconductor during soldering, the copper lead expands faster than the silicon die. The epoxy package expands at a different rate than both. The solder joint sits in the middle, trying to hold everything together while everything pulls in different directions. When the joint cools, the materials contract at different rates, and the solder gets stretched or compressed. That stretch or compression is residual stress, and it stays in the joint forever.
For power discrete semiconductors with large copper tabs, this mismatch is even worse. The tab is thick copper. The die is tiny silicon. The thermal mass difference is enormous. The tab heats up slowly and cools down slowly, while the leads heat up fast and cool down fast. The solder joint connecting the tab to the board absorbs all of that differential movement, and it cracks eventually.
A discrete semiconductor sits on a PCB that is clamped, screwed, or otherwise held in place. The board cannot expand freely during soldering. The component cannot contract freely during cooling. Everything is constrained, and constraint plus thermal expansion equals stress.
For through-hole discrete parts, the lead goes through the board and is soldered on both sides. The lead is locked in place. When the board heats up during reflow or wave soldering, the board expands but the lead does not. The solder joint on the bottom side gets stretched. When the board cools, the board contracts but the lead does not move with it. The solder joint gets compressed. That stretch-and-compress cycle happens every time the board goes through a thermal cycle, and it accumulates damage with each cycle.
For SMT discrete parts, the component sits on top of the pads. The solder fillet is the only thing holding the component to the board. If the fillet is thin or weak, the component shifts during thermal cycling, and the stress concentrates at the edges of the pad. That is where cracks start.
When one side of a discrete component heats faster than the other, the solder on the hot side melts first and flows toward the cold side. By the time the cold side melts, the solder has already shifted, and the joint is uneven. The hot side has a thick fillet. The cold side has a thin fillet. The stress is concentrated on the cold side, and that is where the joint fails first.
This happens constantly during hand soldering of through-hole parts. The iron touches one lead first, that lead heats up, the solder melts and flows to the other lead. By the time you get to the second lead, the first joint is already solid and the second joint is starved. The stress is locked into the second joint from the moment it solidifies.
If you do nothing else in this entire article, do this: preheat the board before soldering. Preheating reduces thermal shock, equalizes the temperature across the board, and gives the solder joint time to relax before it solidifies. It is the cheapest, easiest, most effective stress control method available, and most shops do not do it properly.
When a cold board hits a 250-degree-Celsius solder wave or a 240-degree-Celsius reflow peak, every material on the board expands rapidly. The copper pads expand. The component leads expand. The silicon die expands. The epoxy package expands. They all expand at different rates, and the solder joint absorbs all of that differential expansion as stress.
When the board is preheated to 100 to 120 degrees Celsius before it hits the solder, the temperature differential is much smaller. The materials expand less, the differential expansion is smaller, and the solder joint absorbs less stress. The joint solidifies with less residual stress, which means it survives more thermal cycles in the field.
The preheat also gives the flux time to activate before the solder melts. Activated flux wets better, which means the solder flows more evenly and creates a more uniform fillet. A uniform fillet distributes stress evenly across the joint instead of concentrating it at one point.
For through-hole discrete parts going through wave soldering, preheat the board to 100 to 120 degrees Celsius. The ramp rate should sit at 1.5 to 2.5 degrees Celsius per second. Too fast and the flux solvents flash off before the solder melts. Too slow and you waste throughput without gaining anything.
For SMT discrete parts going through reflow, the soak zone between 150 and 200 degrees Celsius serves as the preheat. Hold the board in the soak zone for 60 to 120 seconds. This gives the entire board time to reach thermal equilibrium before the reflow zone. The components on the edges of the board have time to catch up to the components in the center, which eliminates the hot-spot-cold-spot problem that creates uneven joints.
For hand soldering, use a heat sink clip on the lead during soldering. The clip pulls heat away from the junction and slows the heating rate. This reduces the thermal shock to the die and lowers the residual stress in the joint.
The peak temperature during soldering is the biggest driver of residual stress. Higher peak temperature means more expansion, more differential expansion, more stress locked into the joint when it cools. The obvious solution is to use the lowest peak temperature that still gives you a good joint.
For tin-lead soldering of discrete SMT parts, the peak should sit at 220 to 235 degrees Celsius. For lead-free, it should sit at 235 to 250 degrees Celsius. Going above these ranges does not improve the joint — it only increases the stress.
For wave soldering of through-hole discrete parts, the wave temperature should sit at 245 to 260 degrees Celsius for tin-lead and 260 to 275 degrees Celsius for lead-free. The board preheat brings the component up to temperature before it hits the wave, so the actual thermal shock is much smaller than the wave temperature suggests.
For hand soldering, keep the iron tip at 340 to 370 degrees Celsius for discrete semiconductors. A 400-degree-Celsius iron does not make a better joint. It makes more stress. The extra heat soaks into the component junction and creates thermal damage that no amount of good soldering can fix.
The longer the solder is molten, the more time the materials have to expand and contract at different rates. The time above liquidus should be as short as possible while still giving you a good joint.
For discrete SMT parts, the time above liquidus should sit between 40 and 70 seconds. Less than 40 seconds and the flux does not fully activate, which creates poor wetting and high stress. More than 70 seconds and the solder over-flows, the intermetallic layer grows too thick, and the joint becomes brittle.
For through-hole discrete parts in wave soldering, the contact time with the wave should stay between 3 and 5 seconds. More than 5 seconds dumps too much heat into the component and creates excessive residual stress. Less than 3 seconds and the joints do not fully wet, which creates stress concentrations at the edges of the pads.
Everyone talks about heating rate. Almost nobody talks about cooling rate. But the cooling rate is just as important for stress control as the heating rate. If you cool the board too fast, the outer layers solidify before the inner layers, and the differential contraction creates stress. If you cool too slow, the intermetallic layer grows too large and the joint becomes brittle.
For most discrete semiconductors, the cooling rate should sit between 3 and 5 degrees Celsius per second. This is fast enough to produce a fine-grain joint with thin intermetallic layers, but slow enough to avoid thermal shock.
Lead-free solder needs a slightly slower cooling rate — 3 to 4 degrees Celsius per second — because lead-free alloys are more prone to thermal shock. Tin-lead can handle 4 to 5 degrees Celsius per second without issues.
For power discrete semiconductors with large copper tabs, slow the cooling rate to 2 to 3 degrees Celsius per second. The tab has a huge thermal mass, and if you cool it too fast, the tab contracts faster than the board. The solder joint between the tab and the board gets stretched, and that stretch becomes residual stress.
Forced air cooling is uneven. The components on the edges cool faster than the ones in the center. The tall components cool slower than the short ones. This creates thermal gradients across the board that cause warping and uneven stress distribution.
Use convection cooling instead of forced air whenever possible. Convection uses a laminar flow of heated gas that cools the board evenly from all sides. The cooling rate is uniform, the thermal gradients are smaller, and the stress is distributed evenly across every joint on the board.
If you do not have a convection system, slow the fan down to the minimum setting that still gets the board to room temperature within a reasonable time. Faster is not better here. Controlled is better.
Thermal stress is not the only kind of stress. Mechanical stress — from vibration, from board flex, from component weight — adds to the thermal stress and pushes the joint over the edge. Controlling mechanical stress during and after soldering is just as important as controlling thermal stress.
This sounds obvious, but it happens constantly. A technician solders one end of a through-hole transistor, then moves the board to solder the other end. The first joint is still liquid. The board flexes as it is moved. The liquid solder joint deforms and creates a stress concentration that never heals.
Hold the board in place until every joint on the component is solid. For tin-lead solder, that means waiting 2 to 3 seconds after the last joint. For lead-free, wait 4 to 6 seconds. Use a fixture or a clamp to hold the board steady. Do not rely on your hands — your hands shake, and that shake is enough to deform a liquid joint.
A solder joint needs time to fully solidify and relax. If you flex the board while the joint is still cooling, you introduce mechanical stress that adds to the thermal stress. The joint solidifies with combined stress, and it fails earlier than it should.
Do not snap the board to test if it is stiff enough. Do not drop the board on the table. Do not apply pressure to the component to check if it is seated properly. Let the board sit flat on a clean surface for at least 30 seconds after soldering. The joint needs that time to relax.
For power discrete semiconductors with heavy heatsinks, the heatsink adds mechanical stress to the tab joint. The weight of the heatsink pulls on the solder joint while it is cooling. Use a clamp or a spring clip to hold the heatsink in place until the joint is fully solid. Remove the clamp only after the solder has completely solidified under pressure.
The leads on a through-hole discrete component act as levers. When you bend the lead to insert it into the board, you are pre-stressing the lead. When you solder it, that pre-stress gets locked into the joint. When the board flexes in service, the pre-stress adds to the service stress, and the joint cracks.
Trim the leads to the minimum length required by the datasheet. Shorter leads mean less lever arm, less pre-stress, and lower residual stress in the joint. Do not leave extra lead length "just in case." That extra length is a stress concentrator waiting to fail.
Bend the leads gently. A sharp bend at the base of the lead creates a stress concentration point that propagates into the solder joint during thermal cycling. Use a smooth, gradual bend with a radius of at least 1.5 millimeters. This distributes the stress along the lead instead of concentrating it at one point.
The pad design determines how stress is distributed across the solder joint. A good pad design spreads the stress evenly. A bad pad design concentrates the stress at one point, and that is where the crack starts.
A small annular ring concentrates stress at the edge of the pad. The solder fillet has to bend sharply around a small copper ring, and that sharp bend is a stress concentrator. Every thermal cycle puts more stress on that bend, and eventually the crack starts there.
Push the annular ring to 0.3 millimeters or more for through-hole discrete semiconductors. The larger ring gives the solder a gentle curve instead of a sharp bend. The stress is distributed across a wider area, and the joint survives more thermal cycles.
For SMT discrete parts, make the pad extension beyond the component body at least 0.2 millimeters on each side. The extra pad area gives the fillet room to spread out, which reduces the stress concentration at the pad edge.
If one pad is larger than the other, the solder joint on the larger pad wets better and is stronger. The joint on the smaller pad wets poorly and is weaker. Under thermal cycling, the weaker joint takes all the stress, and it fails first.
Make both pads identical in size, shape, and thermal relief. For SOT-23 transistors, measure both outer pads with a caliper. They must be within 0.05 millimeters of each other. For through-hole diodes, both pads must be the same size and the same distance from the component body.
Asymmetric pads also cause tombstoning, which is a mechanical stress failure. The component lifts off the board during reflow because one joint wets faster than the other. The stress of lifting cracks the weaker joint, and the component ends up standing on one end like a tombstone. Symmetric pads prevent this entirely.
You cannot see residual stress with the naked eye. But you can see the damage it causes. Stress-related defects have specific signatures that you can look for under magnification.
Look at the junction between the solder fillet and the copper pad. If you see a hairline crack running along the pad edge, that is a stress crack. It started at the point of highest stress concentration — usually the edge of a small annular ring or the corner of a pad — and propagated into the joint.
Pull a cross-section on any discrete semiconductor joint that shows a crack. Look at the fillet shape, the pad size, and the wetting angle. A crack at the pad edge almost always means the annular ring was too small or the pad was too narrow. Fix the pad design and the cracks go away.
A head-in-pillow defect looks like a good joint from the top, but the component lead did not fully wet to the pad. Instead, the lead sits on top of a ball of solder like a head on a pillow. This is a stress defect caused by uneven heating during reflow.
The solder on one side of the pad melted first and formed a ball. The lead arrived after the solder had already solidified into a ball, so it could not wet to the pad. The joint has no mechanical strength and will fail under the slightest vibration.
Prevent head-in-pillow by ensuring the reflow profile heats the board evenly. Use a soak zone long enough to bring the entire board to thermal equilibrium. Check the oven temperature uniformity with a thermocouple grid. If the center of the board is more than 10 degrees Celsius hotter than the edges, the profile is not uniform enough for fine-pitch discrete parts.
If the copper pad lifts off the board around a through-hole lead, that is a stress failure. The solder joint was strong enough to hold the lead, but the stress from thermal cycling cracked the pad-to-board interface instead.
This happens when the annular ring is too small or when the board is too thin. Increase the annular ring to 0.4 millimeters or more. Use a board thickness of at least 1.6 millimeters for boards with through-hole discrete semiconductors. The thicker board absorbs more stress without cracking.
The best stress control technique is not a piece of equipment. It is a habit. Technicians who follow these habits produce boards with lower residual stress and higher field reliability.
Never solder a component with cold flux. Cold flux does not activate, the solder does not wet, and the joint solidifies with maximum stress. Always apply fresh flux to every pad before soldering.
Never hold the iron on a joint longer than necessary. For discrete semiconductors, that means 2 to 3 seconds per lead for hand soldering. Every extra second dumps more heat into the component and creates more stress.
Never skip preheating. Even a simple hot plate under the board for 30 seconds before hand soldering reduces thermal shock dramatically. For wave soldering and reflow, the preheat zone is not optional — it is the most important zone in the entire process.
Always let the joint cool under controlled conditions. No quenching, no blowing, no cold surfaces. Let the solder solidify slowly and evenly, and the stress will be lower.
Inspect the first board of every new lot. Pull cross-sections, check the fillets, measure the pads. Catch stress-related defects early, fix the root cause, and move on. The ten minutes you spend on that first inspection saves you ten hours of field failure analysis later.
Contact: Joanna
Phone: Info@addcomponents.hk
Tel: 852 5334 3091
Email: info@addcomponents.hk
Add: FLAT/RM C -13/F HARVARD ,COMMERCIAL BUILDING 105-111 THOMSON ROAD,WAN CHAI HK