Inspect the Bars

# Inspect the Bars: The Base Is the Weakest Point Mechanically, Not Just Electrically

Every earlier step that worried about base width was worried about electrical consequences — gain, punch-through, reach-through. This inspection step cares about a completely different property of the same thin region: its reduced cross-section is also the mechanically weakest point in the entire bar, the place a crack started anywhere near the cut edges is most likely to propagate toward and through. A bar can carry two junctions, both correctly formed and correctly doped, and still be structurally compromised by a crack introduced during cutting that an electrical probe would never detect, because the crack doesn't necessarily disrupt either junction's electrical function until mechanical stress later — during lead attachment, encapsulation, or ordinary handling — causes it to propagate the rest of the way through.

## 1. Reduced Cross-Section Concentrates Mechanical Stress, Not Just Electrical Field

$$\sigma_{\text{local}} \propto \frac{F}{A_{\text{base}}}, \qquad A_{\text{base}} \propto W_B$$

The same base width $W_B$ that sets current gain through the exponential relationship established back in step sixteen also sets the local cross-sectional area $A_{\text{base}}$ carrying any mechanical load applied to the bar. A narrower base — exactly the direction gain optimization pushes toward — means a smaller load-bearing area at that specific cross-section, and therefore higher local stress for the same applied force. This is why visual inspection at this step is specifically weighted toward the base region rather than distributed evenly across the whole bar: the base is where germanium's inherent brittleness and the device's own electrical design both conspire to make cracks most likely to start or propagate.

## 2. Real Diagram: Where Inspection Actually Looks

Inspection Targets the Base, Where Cracks Actually Propagate the same narrow region that sets gain also concentrates mechanical stress n region base, W_B n region crack from cut edge concentrates right at the narrow base a crack stopping short of the base may still be an acceptable bar; one reaching it is not

## 3. Why Catching This Now Is Far Cheaper Than Catching It Later

$$\text{Cost(scrap at step } n\text{)} \propto n$$

A bar discarded at this inspection step has only the value of steps one through twenty-three invested in it — real cost, but entirely recoverable material loss, with no lead wires attached, no can sealed, no aging or final test time spent. The same structural defect discovered after lead attachment, encapsulation, or parametric test wastes every additional step's labor and material on top of the original loss, for a defect that this inspection could have caught cheaply if it had been looked for at this point. This is the same escalating-cost-of-late-discovery logic that governs quality gates throughout manufacturing generally, applied here to the single specific failure mode — base-region cracking — that this process's own base-width-minimization goal makes especially likely to occur.

Cost of Scrapping a Defective Bar vs. Process Step the same defect costs more to discover the later it's caught process step at discovery → cost of scrapping the bar step 24, caught here step 36, caught here instead same defect, same bar, very different cost depending on when it's found

## Inspect the Bars's Place in the Process Lineage

Inspecting the bars is step twenty-four of the 1951 grown-junction transistor's full manufacturing sequence — immediately after individual rectangular bars have been cut, and before the supporting lead assembly is prepared. It is the step that checks geometry, cracks, and base-region preservation specifically because the same narrow base that current gain depends on is also the bar's mechanically weakest point, and catching a structural defect here costs a fraction of what catching it after any later step would.

Take inspect the bars further

Ask the copilot about this term, or have our engineers assess it against your process.