ฝึกพูดภาษาอังกฤษด้วยเทคนิค Shadowing จากวิดีโอ: Deep Dive: Investigating Thermal Cycling Failures Cracked MSOP Solder Joints

You know what really gets my gears turning?

These real world puzzles.

And we've got a pretty juicy one for you today popped up on an online forum.

Oh yeah What's the story?

OK, picture this.

You've got an integrated circuit, MSOP package, soldered onto a PCB, standard stuff.

But then it goes through this grueling test, 1,000 cycles.

1,000, wow.

Yeah, extreme temperature swings.

We're talking minus 40 degrees C, way up to 85 C.

Bone chilling to scourging, basically.

Exactly. And the temperature ramps up and down pretty quick, 4 degrees a minute.

And they let it sit at each extreme for like 15 minutes.

Talk about stress.

Yeah, that kind of cycling, that's a serious test.

Puts significant strain on all the materials with the expansion and contraction.

Totally. And the person who posted, they even shared a cross-section image.

You could clearly see it, a crack right there in the solder joint.

OK, so failure analysis time.

Right, but here's the kicker, the part that makes you kind of scratch your head.

Despite this crack, this visible crack, the product, still working.

Still functional.

That is interesting.

Incomplete failure.

Exactly. So the big question we're tackling today is why?

Why did the solder crack, but the whole thing didn't just die?

Good question.

So our mission in this deep dive is to explore the potential reasons behind this MSOP solder joint cracking.

And we're going to really focus in on the role that differences in the coefficient of thermal expansion, CTE, between the materials might play.

It's a fascinating case.

Really highlights that subtle interplay of materials, science, and electronics.

And like you said, the fact it's an MSOP package, that's a bit unusual for this kind of severe cracking, isn't it? Right, that's what I thought, too.

Now, the original post, it gave us a good starting point, but well, it was missing some key details.

Always the way, isn't it?

Yeah. We don't know the exact solder paste used or the specifics of the PCB type, thickness, surface finish, nada.

And the location on the board?

Or if other parts failed?

Nope, none of that.

Or even if other components nearby had similar issues.

It's like trying to solve a puzzle with half the pieces missing.

Those missing details are often so critical.

Solder alloy, for instance, tells you its specific CTE characteristics right away.

Exactly. And just to make sure everyone's on the same page, MSOP.

Mini Small Outline Package.

It's basically a smaller SOP, tighter pin pitch, like 0.5, 0.65 millimeters maybe.

Usually for lower pin count ICs, less than 16 pins typically.

Compact. Yeah. And what's interesting, like we said, is MSOPs aren't usually the first thing you suspect for solder joint cracks.

They're smaller.

Joints tend to be pretty resilient.

Compared to, say, big MLCCs, which hate bending or connectors that see physical stress.

Precisely. So this report of a cracked MSOP definitely got our attention.

It does suggest we need to look beyond maybe the usual suspects.

The small size normally means less leverage for stress.

OK, so the initial buzz on the forum, the first reaction was, CTE mismatch.

Got to be CTE.

The classic answer.

Right. Makes sense on the surface.

Different materials, different expansion rates when things heat up or cool down.

But as you know, just blaming CTE, well, it might be a bit too simple.

It often is.

I mean, CTE differences are pretty much always there in electronics.

The industry spent decades optimizing materials to handle it.

Especially since it was still working, right?

Suggests maybe something else, like mechanical stress, was a bigger player?

Possibly, or at least a significant contributor.

It's rarely just CTE unless the mismatch is huge or other factors line up badly.

OK, so for anyone listening who's maybe not deep into CTE, let's break it down simply.

Imagine two materials stuck together.

Temp goes up, they grow.

Temp goes down, they shrink.

But if they grow and shrink at different rates, different CTEs, that puts stress right where they meet.

Yeah, like two people holding hands, one taking giant steps, the other tiny ones.

There's going to be tension.

Perfect analogy.

So in our MSOP case, there are three main spots where the CTE mismatches could matter.

First, between the solder itself and the IC lead frame, usually copper.

Right, the lead frame.

That's the metal pathway from the tiny chip inside out to the PCB.

Copper is a standard.

Good conductivity.

Exactly. Now, the forum user didn't say, but we can probably assume it's SAC 305 solder.

It's super common, lead-free.

Good assumption, industry standard pretty much.

OK, so SEC 305 has a CTE in the ballpark of, what, 22 to 25 ppm degree C, parts per million per degree Celsius.

Around there, yeah.

And the lead frame, copper, often tin-plated for solderability.

Copper's CTE is lower, maybe around 16.5 ppm degrees.

So there's a difference.

The solder expands and contracts more than the copper lead.

Right, and that difference creates sheer stress at that interface, especially when it gets cold, right?

The solder shrinks more.

Exactly. That low-temp part of the cycle is often where the strain really racks up.

Over many cycles, that repeated stress that leads to fatigue cracks forming.

Makes sense, like bending a paper clip back and forth eventually snap.

And we should probably remember, these solders, like SAC 305, they aren't just tin, silver, copper.

They have tiny amounts of other stuff, too.

These elements, yeah.

They fine-tune the property.

Yeah, I saw a blog post about it.

100

A bit of silver for toughness, fatigue resistance, a bit of copper for stiffness, strength.

101

It's a real balancing act.

102

Absolutely.

103

Those craze additions can really impact long-term reliability under thermal and mechanical loads.

104

OK, so that's solder to lead frame, interface number two.

105

Solder to the IC package body itself, that black plastic mold compound.

106

Ah, yes, the mold compound.

107

What's the CTE story there?

108

Generally, mold compounds are lower, maybe 10 to 20 ppm degrees, so usually lower than the solder.

109

OK. And its main job is protection, right?

110

Protect the chips, support the leads.

111

But its own expansion and contraction, well, it can push sideways on the solder joints, lateral forces.

112

Ah, so it could concentrate stress maybe at the edges where the solder curves up the fillet.

113

Potentially, yeah.

114

Yeah. Especially at the heel toe of the fillet.

115

Yeah. But you know, these mold compounds, they've been used forever.

116

Their properties are pretty stable, well known.

117

OK. Often the lowest CTE in the whole stack up, actually.

118

OK, good point.

119

That brings us to number three.

120

And this is often the big one, right?

121

Solder to the PCB itself.

122

Assuming standard FR4.

123

Yeah, this is usually the most critical interface for a CTE mismatch.

124

FR4, it's tricky.

125

Oh, so? Well, it expands differently depending on the direction.

126

Yeah. In the x and y plane, you know, along the surface of the board, the CTE is maybe 14 to 17 ppm degrees C, not too far off from copper or solder.

127

OK. But in the z-axis, through the thickness, below its glass transition temp Tg, it's already much higher, like 70 to 120 ppm degrees.

128

Whoa, OK, that's a jump.

129

And then when you go above Tg, which definitely happens during reflow soldering, that z-axis CTE, the alpha 2 CTE, it can skyrocket, 190, 200, even 300 plus ppm degrees.

130

Jeez. But the source mentioned cracking is less likely during reflow because the solder's melted, right?

131

Correct. Solder's liquid, it just flows.

132

The stress builds up later during operation, during those repeated temperature cycles.

133

Even if the product only operates, say, below Tg.

134

Right, even below Tg, that repeated cycling causes expansion and contraction.

135

And that big z-axis movement, that up and down stretching and squashing of the board thickness, that pulls and pushes on the solder joints connecting the components.

136

Tearing forces, basically, leading to fatigue over time.

137

Exactly. Even if the operating temps aren't super high, the board remembers that high temp reflow state.

138

And cycling stresses it relative to that.

139

So the XY expansion across the board surface is more uniform, maybe less direct stress on one tiny MSOP joint, unless you have really uneven heating.

140

Generally, yes.

141

The glass fibers in the FR4 do a decent job controlling XY expansion.

142

It's that z-axis movement through the layers that's harder to constrain and often causes more trouble for the joints.

143

OK, let's sum up the CTE likelihood, then.

144

You generally need, what, a difference bigger than 10 or 15 ppm / ˚C for really significant stress and cracking from CTE alone?

145

That's a reasonable rule of thumb, yeah.

146

So looking at the numbers, solder to copper lead frame, maybe 6 to 9 ppm difference.

147

Solder to mold compound, maybe 2 to 15 ppm difference.

148

Solder to FR4 XY, maybe 5 to 11 ppm difference.

149

They're mostly kind of borderline or within that range, not screamingly high mismatches, assuming standard materials like SAC305.

150

If they used, say, a brittle, low temp solder, that's a different story.

151

Right. So CTE is definitely a factor adding stress, but maybe not the sole smoking gun here, especially since it still worked.

152

The community thought CTE first, but maybe the evidence points elsewhere.

153

It seems plausible.

154

That functional aspect really pushes us to consider other things more strongly, like mechanical stress.

155

OK, so if it's not just CTE, how do we figure out what really happened?

156

Time to put on our detective hats, I guess.

157

Indeed. When you face a failure like this, you need a systematic approach.

158

Experience helps, sure, but structure is key.

159

OK, walk us through it.

160

Step one. Step one, information gathering.

161

Always. You can't solve anything without the facts.

162

We use the 5W1H framework, who, what, when, where, why, how.

163

We need more on the product's application, the exact test profile, manufacturing details, all of them.

164

More info equals better chance of finding the root cause.

165

Makes sense.

166

The forum post gave us the what crack joint and some when we're 1,000 cycles, but the why and how are big question marks, plus those missing details like solder type, PCB specs.

167

Exactly. So after gathering what we can.

168

Step two. Get the actual failed sample and ideally related devices, too.

169

Ah, the physical evidence.

170

Yeah. Photos are OK.

171

Descriptions are helpful, but nothing beats having the actual cracked board.

172

Avoid misinterpretation and grab any bits that fell off.

173

Got it. Handle with care.

174

OK, samples in the lab.

175

What's next under the microscope?

176

Step three.

177

Check solderability and crucially, inspect the IMC, the intermetallic compound.

178

IMC. OK, remind us what that is again.

179

It's that thin layer that forms right where the solder meets the metal pad on the PCB and where it meets the component lead.

180

It forms during soldering when heat makes the metals react.

181

And looking at it tells you.

182

It tells you about the initial soldering quality.

183

Was the reflow profile right?

184

Any process control issues, oxidation on the parts, a good IMC is key.

185

So what does a good IMC look like?

186

Uniform continuous layer, both at the PCB pad and the component lead.

187

If it's patchy or way too thick or has voids, that suggests problems in the initial soldering.

188

But you usually need to cross section to see it properly, like the forum user did.

189

Right. And based on that image they shared, what did the IMC look like in this case?

190

The description said the IMC down at the PCB pad looked good, uniform.

191

So probably means the reflow process and the board finish were OK.

192

OK. What about at the lead itself?

193

Harder to tell precisely because of the crack right there.

194

But apparently, there was evidence of IMC at the base of the lead.

195

And the growth on the sides of the lead was normal, so just as likely OK there too initially.

196

So initial soldering probably wasn't the root cause.

197

But you said IMC, even if formed well, can be a weak point.

198

Yes, absolutely.

199

It's essential for the bond.

200

It is the bond, metallurgically speaking, like mortar in a brick wall.

201

But it's also typically more brittle than the bulk solder alloy itself.

202

So when the joint is stressed, mechanically or thermally, that brittle IMC layer is often where cracks like to start.

203

It's the weakest link in the chain, mechanically speaking.

204

Gotcha. OK, so even with good IMC, stress finds the weak spot.

205

What's step four in our investigation?

206

Step four, examine the crack morphology, the shape, the path it took.

207

What did the photo suggest?

208

It looked like the crack probably started right at that IMC interface at the bottom of the MSOP lead.

209

And interestingly, the complete break, the full fracture, seemed to be only at the bottom.

210

The cracks on the sides hadn't gone all the way through, which, as we said, explains why it still worked.

211

Current could still sneak through the partly cracked sides.

212

Makes perfect sense.

213

And reinforces the idea that the IMC was the point of failure initiation.

214

Exactly. And ideally, you'd want to look at other leads on that same MSOP, see if they show similar early signs of cracking.

215

Is it just one pin or systemic?

216

Right, get the bigger picture.

217

Now, switching gears slightly, that number, the 1,000 cycles, the source material pointed out that's pretty extreme for typical commercial stuff, yeah.

218

Oh, absolutely.

219

Standard reliability test, maybe simulating shipping, you might do 5, 10 cycles, maybe 100 for some things, 1,000 cycles like that.

220

That's more like component qualification or high-rel stuff, automotive, aerospace, military.

221

So it makes you wonder about the product's intended use.

222

Why such intense testing?

223

It does. And it really makes you think maybe something beyond just inherent material properties was at play.

224

Was the board itself flexing a lot under those conditions?

225

Good question.

226

And the source author also wondered why they did a destructive cross-section after finding it still worked.

227

Maybe just routine inspection found it.

228

That's plausible.

229

Sometimes non-destructive tests, like X-ray, might misfind cracks like that.

230

Cross-sectioning gives you the definitive answer.

231

OK, so let's recap.

232

CTE mismatch, probably a factor, but maybe not the main driver alone.

233

Initial solder quality IMC looked OK based on the limited info.

234

Seems that way.

235

So where does that leave us?

236

The source material strongly suggests mechanical or bending stress as the most likely culprit.

237

Why that? Well, if the usual suspects aren't clearly guilty, you look for other forces.

238

And PCB warpage during temp cycling is a huge one.

239

Warpage? How does that happen?

240

Multilayer boards.

241

They're complex structures.

242

Copper layers, insulation, via maze, they don't all expand and contract perfectly evenly, especially with big temperature swings.

243

This unevenness causes the board to, well, warp, bend, flex, twist slightly.

244

And that slight bending stresses the components sitting on top.

245

Exactly. The components are rigid.

246

The board beneath them is flexing.

247

That force gets transmitted straight into the solder joints, holding them down.

248

And over 1,000 cycles of that bending back and forth, that's classic fatigue.

249

Causes cracks, often with those telltale branch patterns as the crack finds the easiest path.

250

And thinner boards would warp more, right?

251

Less rigid?

252

Generally, yes.

253

And location matters, too.

254

Components far from mounting points, far from where the PCB is screwed down to the chassis, they often see more stress.

255

Because the board flexes more out in the middle.

256

Right. Or if the chassis itself deforms with temperature, it can pull or push on the PCB.

257

And those components out on their own become the stress relief points.

258

Their solder joints take the hit.

259

OK, that sounds like a very plausible explanation for this case.

260

So if PCB bending mechanical stress is the likely villain, what can engineers do to stop it?

261

Countermeasures.

262

Several options.

263

First, design phase.

264

You can use CAD simulation, finite element analysis, FEA.

265

To predict the stress.

266

Exactly. Simulate the temperature cycles on the whole assembly, PCB, and chassis.

267

See where it deforms most.

268

Then you can modify the design.

269

Maybe change screw locations, add more mounting points, maybe move sensitive components like this MSOP to a less flexy part of the board, or add stiffeners to the chassis.

270

Or even choose a more rigid PCB material, though that costs more.

271

Makes sense.

272

Address the root cause in the design.

273

What about manufacturing process fixes?

274

The source mentioned underfill.

275

Yes, underfill.

276

Basically, squirting epoxy under the component between it and the PCB.

277

To reinforce the joints.

278

Right. It adds mechanical support, helps distribute the stress away from just the solder.

279

It can definitely help.

280

But as the source wisely pointed out, it often just delays the inevitable if the underlying warpage is still happening.

281

It treats the symptom, not the disease.

282

Eventually, the underfill itself might crack under enough stress.

283

So it might get you through warranty, but maybe not solve the fundamental problem.

284

It can be like that sometimes.

285

It's a valid technique, often used.

286

But it's best when combined with addressing the root cause of the stress.

287

So the best approach is always figure out why it's cracking the root cause, verify it, and then fix that fundamental issue.

288

Absolutely.

289

That's the path to a truly robust and reliable solution.

290

Well, this has been a really fascinating deep dive.

291

It's amazing how complex even a simple solder joint failure can be, so many interacting factors.

292

It really is.

293

And we have to remember, our whole discussion here is based on assumptions, given the limited info from that forum post.

294

There could easily be other factors we haven't even touched on.

295

Absolutely true.

296

And to you, listening if you have other ideas, alternative takes on this cracked MSOP, we genuinely love to hear them.

297

Solving these engineering puzzles really benefits from different perspectives.

298

Definitely.

299

And hey, if you enjoy digging into electronics, manufacturing, reliability, stuff like this, make sure you subscribe to our channel.

300

Please do. And if you found this chat valuable, maybe give us a like, subscribe, share it around, a five-star review on your podcast platform.

301

That always helps, too.

302

Helps others find us.

303

Yep. You can find links to anything related we mentioned down in the show notes or episode info.

304

Thanks so much for tuning in for this deep dive.

305

Yeah, thanks, everyone.

306

Until next time, keep asking questions, keep learning.

เทคนิค Shadowing คืออะไร?

Shadowing เป็นเทคนิคการเรียนรู้ภาษาที่ได้รับการรับรองทางวิทยาศาสตร์ พัฒนาขึ้นสำหรับการฝึกนักแปลมืออาชีพ วิธีการนี้เรียบง่ายแต่ทรงพลัง: คุณฟังเสียงภาษาอังกฤษจากเจ้าของภาษาและพูดตามทันที — เหมือนเงาที่ตามผู้พูดด้วยช่วงเวลาห่าง 1-2 วินาที การวิจัยแสดงว่าเทคนิคนี้ปรับปรุงความแม่นยำในการออกเสียง ทำนองเสียง จังหวะ การเชื่อมเสียง การฟังเข้าใจ และความคล่องแคล่วในการพูดได้อย่างมีนัยสำคัญ

ฝึกพูดภาษาอังกฤษด้วยเทคนิค Shadowing จากวิดีโอ: Deep Dive: Investigating Thermal Cycling Failures Cracked MSOP Solder Joints

เกี่ยวกับบทเรียนนี้

เทคนิค Shadowing คืออะไร?