ICD Risk in PCB Drilling — Causes and Control

The one thing to get right

The defect is thermal, the material layer is only one input to the heat balance, and the bit is the other. Changing to a lubricated entry sheet while running a bit past its hit count does not reduce the risk — it moves it. Fix the bit and the parameters first, then use the material to buy margin.

What an interconnect defect is

An interconnect defect is a separation or a degraded connection at the interface between the plated copper and the material it bonds to, at or near the hole wall. It is not the same as a rough wall or a burr, which are geometric: an interconnect defect is a bond that has been disturbed, and it can leave the hole looking entirely acceptable while the electrical connection is compromised.

That is what makes it expensive. A rough wall or a burr is visible on the bench and can be sorted; an interconnect defect passes a visual inspection and fails at electrical test or in the field, which moves the cost from scrap to a much later — and much more expensive — point in the process. It is also why the control discussion is about prevention rather than detection: by the time it is detectable, the value is already in the panel.

Why drilling is where the risk is created

Drilling is a high-energy process concentrated into a very small volume. The bit removes material by shearing it, and the energy that goes into that work leaves as heat in the hole wall, the chip and the bit itself. In a deep hole or a dense pattern there is little time between hits for that heat to dissipate, so the wall temperature rises through the stack and stays elevated for the duration of the cut.

The consequence is a differential response. Copper conducts heat away quickly and expands relatively little; the resin system around it heats more slowly, expands more, and relaxes. If the excursion is large enough, the bond at the interface is loaded beyond what it can hold and does not fully recover as the material cools. The mechanism is mechanical, but the driver is thermal — which is why every meaningful control acts either on how much heat is generated or on how quickly it is removed.

The variables that raise the risk

These are the inputs that move the heat balance. They interact, so treating any of them in isolation gives a misleading answer.

Variable	How it moves the risk	Who owns it
Bit wear and hit count	A dulled edge cuts less and rubs more, so more of the input energy becomes heat in the wall	Tool management
Stack height and panel thickness	Sets how long the bit stays in the hole and how much heat accumulates before it can leave	Process and stack design
Hole density	Raises the heat load per unit area, so the local rise is greater for the same parameters	Design and program
Spindle speed and feed	Determines heat generated per revolution against heat carried away by the chip	Program parameters
Resin system and its thermal response	Sets how much the material moves for a given temperature excursion	Laminate selection
Entry side heat path	Where the initial spike is taken and how quickly it spreads away from the wall	Entry material and gauge
Debris evacuation	Retained chip carries heat back into the wall instead of removing it	Program and material

Read the last column as ownership of the fix, not of the symptom. Most interconnect-defect investigations that stall are ones where a material change is being asked to compensate for a tool-management problem.

What the material layers can change

The entry and backup layers act on different parts of the same problem, and it is worth being precise about which is which.

• The entry layer takes the first spike. The bit contacts the entry sheet before anything else, so the initial temperature rise happens there. A lubricated or coated entry material is aimed at that interface directly, which is why the entry side is the more direct lever on wall temperature. The coated aluminium entry board explainer covers what the coating changes at mechanism level.
• The backup layer works at the exit. It governs the breakout — burr, wall condition at the exit and debris — rather than the temperature through the cut. Its contribution to interconnect-defect risk is indirect: a cleaner exit means less rework and less handling of a marginal wall.
• Neither acts on the resin. The material's thermal response is set by the laminate, not by the consumables under and over it. A consumable can reduce the excursion; it cannot change what the excursion does to a given resin system.

This is the reporting discipline that keeps an investigation honest: if the laminate and the parameters are unchanged and the bit is fresh, then a difference between two entry materials is attributable to the entry materials. If any of those moved, it is not.

Program practices that lower the risk

The bulk of the risk reduction sits in how the program and the tooling are managed, and none of it is material-specific:

• Replace on hits, not on feel. A hit-count limit is what keeps the cutting edge inside the window where it cuts rather than rubs. The hit-count calculator and the drill wear calculator both work that number from different inputs.
• Match the parameter set to the stack, not to the previous job. Panel thickness and stack height change the dwell time, so a parameter set carried over from a thinner stack is running outside its own window.
• Watch dense patterns separately. Where hole density is high, the same parameters that hold on a sparse pattern will not hold — the heat input per unit area is different. The panel hole density calculator gives that input a number.
• Keep the stack complete and flat. Entry board present, backup board present, stack closed evenly. A missing or mismatched backup layer changes the exit condition, and the operator compensates with parameters that then stress the wall.
• Control storage. Moisture in the board changes the stack's dimensional state, which changes how the stack closes and therefore the dwell and the entry condition. The moisture control guide covers the storage side.

Where the process runs into a thicker stack or a higher layer count, the multilayer drilling material guide extends the same reasoning to stack builds from six to twenty-four layers.

Verifying a change

Verification is where most of these investigations lose their value, because a change that is not isolated is not evidence. Three requirements make a trial mean something:

• One variable. Hold the bit, the parameters, the panel lot and the stack position, and change only the material under test.
• A defined sample. Use the coupon position and the sectioning method your quality system already specifies, so the result is comparable with your own baseline rather than with ours.
• A recorded bit-life position. State where in the bit's life the section was taken. A section from a fresh bit and a section from a worn bit are two different tests, and comparing across them is the most common way a false conclusion is reached.

Where the result has to be quantified before it is acted on, the drilling defect rate calculator converts the observed counts into a rate you can compare against your own baseline. And if the trial produces an exit-side symptom instead of a wall symptom, the diagnosis order is different — work it through the defect sequence rather than through this page.

Questions engineers ask