MarkDown

Most engineers think improving yield means upgrading the SMT line.

But in this case, the breakthrough came from somewhere else:

👉 BGA rework.

🏭 The Situation

An EMS factory we worked with was facing a common but costly problem:

• Stable SMT placement
• Good incoming material
• But high scrap rates during rework

Their challenges were typical:

Inconsistent BGA rework results
Yield fluctuation between operators
Occasional PCB deformation after repair
Increasing cost from scrapped high-value boards

👉 On paper, everything looked “under control”
👉 In reality, rework was quietly eating their margins

⚠️ The Root Cause (Not What They Expected)

Initially, the team suspected:

❌ Operator skill gap
❌ Material issues
❌ Random process variation

But after reviewing their workflow, the real issue was clear:

Their rework process was not truly controlled.

Specifically:

• Temperature profiles were not optimized for different PCB types
• Heating was uneven across large boards
• Alignment relied heavily on operator experience
• No data tracking or process validation

⚙️ What Changed

Instead of changing the SMT line, they focused on standardizing and stabilizing BGA rework.

🔹 Step 1: Profile Optimization
Built dedicated profiles for different PCB thickness & BGA sizes
Adjusted heating curves to match real thermal behavior

👉 Result: Reduced overheating and cold joints

🔹 Step 2: Controlled Thermal Environment
Implemented multi-zone heating strategy
Ensured uniform PCB preheating

👉 Result: Eliminated board warpage issues

🔹 Step 3: Alignment Standardization
Introduced optical alignment workflow
Reduced dependency on manual judgment

👉 Result: Improved placement accuracy consistency

🔹 Step 4: Data-Based Validation
Recorded temperature curves
Compared actual vs target profiles
Introduced basic CPK tracking

👉 Result: Process became measurable, not subjective

📊 The Results (Within 6–8 Weeks)

After implementing these changes:

• BGA scrap rate reduced by 32%
• Rework success rate significantly improved
• Operator variability minimized
• Overall production stability increased

And most importantly:

👉 They avoided replacing their existing SMT line

💡 Key Insight

This case highlights something many factories overlook:

Improving yield is not always about adding new machines.
Sometimes, it’s about controlling the processes you already have.

🚀 Where Equipment Starts to Matter

Once the process is clearly defined, the next limitation becomes obvious:

👉 Can your current equipment execute that process consistently?

In this case, the factory realized:

• Manual adjustments were limiting repeatability
• Lack of integration made control difficult

This is where integrated BGA rework systems started to show real value:

Consistent thermal execution
Repeatable alignment
Traceable process data

👉 Not just improving rework — but stabilizing it

💬 Final Thought

If your factory is experiencing:

• Unstable rework yield
• High scrap on high-value boards
• Operator-dependent results

It may not be a production issue.

👉 It might be a rework process issue.

🔁 Let’s Talk

Have you ever calculated how much rework is actually costing your operation?

Or is it still a “hidden loss”?