PCB Drilling Defects: Causes and Prevention

The one thing to get right

Change one variable at a time, or you have not changed anything. A defect that disappears after you swapped the backup board, the entry board and the bit in the same afternoon tells you nothing except that you are now running a different process. Isolate the stack variable, keep the machine side fixed, and re-run.

The twelve defect classes

These twelve cover most of what reaches inspection across rigid, multilayer and HDI work. Sorting them by the side of the hole they appear on is the fastest way to narrow the cause, because the side tells you which part of the stack to argue about.

Defect class	Where it shows	First axis to question
Drill breakage	In the stack, at any depth	Bit condition, feed and speed, stack height
Hole damage	Wall and rim	Bit wear, chip evacuation, entry support
Hole-position offset	Across the panel	Machine, program, panel location, stack closure
Hole-size distortion	Diameter, entry to exit	Bit runout, parameter window, heat build-up
Missed holes	Program level	Program, tool change, machine logic
Burr	Exit side	Exit support and bit sharpness
Incomplete drilling	Depth, blind holes	Z-axis setup, stack thickness, plate flatness
Curled debris	Exit side, on removal	Resin build-up and heat at the break-out
Hole blockage	Inside the hole	Debris retention, vacuum and cleaning
Rough hole wall	Wall, both ends	Bit wear, spindle speed, material response
White ring	Hole edge, entry side	Heat and surface-layer response at entry
Nail head	Hole mouth, entry side	Entry-side support and the first revolutions

The table is a routing guide, not a verdict. A rough hole wall can come from bit wear, from spindle speed or from the material's response to heat, and telling those apart takes one controlled change — the drill wear calculator and the multilayer drilling guide both narrow it from different directions.

The order worth checking in

Work the list in this sequence, and resist the temptation to short-circuit it because a spare board is within reach:

• Breakage and hole damage first. These stop the line, and they almost always trace to the bit, the parameters or the equipment rather than to the consumable under the panel. Settle them and you stop chasing a backup-board change for a machine problem.
• Hole-position offset and hole-size distortion second. Position belongs to the machine, the program and how the stack closes; size belongs to bit runout, the parameter window and heat. Both are stack-sensitive, which is why they get misattributed to the board.
• Burr and hole-wall quality third. This is the group where the exit-side material genuinely has leverage, and where changing the backup or entry board is a legitimate lever rather than a displacement activity.

Within each stage, confirm the defect is reproducible before you change anything. A single panel with one rough hole is a sample of one; three panels with the same signature in the same region is a finding. The drill breakage risk calculator gives you a first read on whether the breakage stage is even plausible for the stack you are running.

Symptom, cause, test, action

The twelve classes above sort defects by where they show. The matrix below sorts the recurring symptoms by what to do about them: the likely cause, the one test that separates the candidates, and the action that follows from the result. It is deliberately narrow — one controlled change per row, because changing two variables at once tells you nothing about either.

Symptom	Likely cause	Test that separates candidates	Action if confirmed
Burr height rises late in a bit's life	Bit wear, not the backup grade	Compare burr at low and high hit counts on the same stack	Tighten the replacement point; keep the backup grade unchanged
Burr appears on a fresh bit	Exit support too soft for the stack	Re-run the same panel count on a harder backup grade	Step up the backup family; re-check flatness after conditioning
Nail head grows with hole count	Entry-side heat and friction build-up	Hold bit and parameters; swap the entry grade only	Move to a coated or lubricated entry sheet for that hole range
Position drift follows stack position	Stack closure or registration, not the machine	Re-seat and re-nest the same panel set on the same program	Fix nesting and pinning before touching parameters
Rough wall at both ends	Bit wear or spindle speed	One controlled parameter change at a time	Replace the bit first; adjust speed only once wear is excluded
Incomplete drilling in blind holes	Stack thickness drifted from the programmed depth	Measure the actual stack against the program	Correct the depth setting and confirm backup thickness tolerance
White ring at the hole edge	Entry-side heat response of the surface layer	Same stack, entry material swapped only	Re-qualify with an entry grade that runs cooler at entry

Rows are starting points, not verdicts. Where a row ends in a material change, run it as a controlled trial — same bit, same program, same panel count — and read the result against the wear and hit-count logs before calling it.

Inspection closes the loop. A practical PCB drilling quality inspection routine samples hole wall quality, burr height and positional accuracy at fixed intervals and records each finding on the defect chart used across shifts. Keeping the chart and the parameter log in the same document makes troubleshooting faster: most recurring defect patterns trace back to a parameter drift or a consumable grade change that the log shows immediately.

Which layer owns which symptom

The stack has three functional layers — entry board on top, panel in the middle, backup board underneath — plus the bit and the machine. Attributing a symptom to the right one is most of the diagnosis.

Layer	What it controls	What it does not
Entry board	Bit guidance in the first revolutions, entry-side heat path, nail head and white ring	Exit-side burr, hole position
Backup board	Exit support, exit burr, breakout quality, debris at the exit	Hole position, drill breakage from a worn bit
Panel and stack	Resin behaviour, aspect ratio, heat accumulation	Substituting for a mis-specified bit
Drill bit and machine	Hole position, size, breakage, wall roughness from wear	Being fixed by a consumable change

Where two layers plausibly own the same symptom the test is a swap, not an argument. If the defect follows the stack position with the same bit and the same program, the consumable side owns it — and that is where a material change, such as moving to a lubricated backup board for heat-driven exit defects or to the phenolic PHE grade where positional accuracy is the constraint, starts to be worth running.

Mechanism behind the exit-side defects

Exit-side defects are governed by what happens in the last fraction of the cut. As the bit breaks through, the copper and the resin are no longer backed by sound material, the cutting edge is at its most exposed, and heat that has built up through the cut has nowhere to go. Three things follow:

• Burr. The unsupported exit edge deforms rather than cuts. Support and a sharp bit are the levers; the harder and flatter the backing surface, the less the edge is able to fold over.
• Curled debris and blockage. Resin softened by heat smears instead of clearing as chip. Reducing the heat reaching the break-out is the object, which is a material and parameter question together.
• Breakout roughness. A bit that has already worn through thousands of hits leaves a coarser wall at the exit than at the entry, which is why the same board can look acceptable on one spindle and marginal on another with a duller bit.

That is also why the exit-side material discussion is a heat discussion as much as a hardness discussion. It is the reason a coating that changes how the surface behaves at temperature can matter even when the core density is unchanged. If you want the geometry side of the same problem, the exit burr calculator and the back-drilling principle guide cover the process angle.

A prevention routine that holds

Prevention is mostly about holding the variables still, because a defect that appears after a change to any one of them is a signal rather than noise.

• Log the bit. Record hit count against every defect event. Rough wall and burr events that cluster late in a bit's life are a bit-management problem wearing a material costume. The hit-count calculator puts a number on the replacement point.
• Condition the boards. Moisture changes the core's dimensional state and therefore the flatness of the stack. The moisture control guide covers storage and handling in humid plants.
• Keep the entry and backup boards as separate decisions. Changing both when a symptom appears destroys the attribution you spent the trial to obtain.
• Re-qualify when the smallest hole changes. A grade that held at 0.25 mm is not automatically the grade for 0.10 mm; the controlling mechanism changes with diameter.

For line-level snap and breakage events specifically, the drill bit snapping guide works the same problem from the failure side rather than the inspection side.

And when the question is the parameter side — how spindle speed, feed rate and chip load bound what each layer can do — the PCB drilling process guide covers the machining half of the same stack.

When a complaint arrives

Work the drill bit and the process parameters first, the entry and backup boards second. In most cases the root cause sits in the parameters, the bit or the equipment rather than in the board under the panel — the board is simply the layer where an exit-side symptom is easiest to see, which is why it is the first thing blamed and often the last thing at fault.

Ask for the same three things every time: the smallest hole and the stack, the bit and its hit count, and the parameters the line was running. With those, the defect class narrows the cause to one or two layers before anyone reaches for a spare box of boards. If the answer points at the exit side after that, the backup board selector narrows the grade, and the drilling consumables glossary keeps the terminology aligned between your plant and ours.

Questions engineers ask