What a Tapped Hole in Soft Metal Can Actually Hold

Why this matters

Somebody hangs a bracket off a cast aluminum housing with two good-quality bolts, torques them to the bolt's published figure, and the threads in the housing pull out. The bolts are undamaged. The tech's first read is that the bolts were undersized, so the next attempt uses a higher grade, and the housing strips again faster. Nothing about the fastener was ever the problem. In a soft parent material the joint's capacity belongs to the hole, and the hole's capacity is set almost entirely by how much thread is engaged.

Before you drill or tap anything

Drilling into an assembly creates hazards that the assembly did not have a minute earlier:

• A blind hole in a casting can break through into an oil gallery, a coolant passage, a refrigerant path or a combustion chamber. Confirm from the manufacturer's drawing what is behind the boss before the bit turns. Do not open a new hole in a pressure boundary, a combustion path, or anything serving a relief or protective function without the manufacturer's written authorization.
• Clamp the work. A bit that grabs on breakthrough will spin the part, and a spinning casting takes fingers. Clamp or vise the piece; a hand is not a fixture.
• Chips and cutting-fluid mist go in eyes. Eye and face protection per 29 CFR 1910.133, and where powered tapping throws a mist, local exhaust or a work position out of the plume rather than a glove.
• If the housing is part of energized or pressurized equipment, isolate first: 29 CFR 1910.147 for mechanical and stored energy, 29 CFR 1910.333(b)(2) where an electrical conductor or termination is anywhere in the work zone, with the live-dead-live proving sequence in NFPA 70E-2021, 120.5.

The gate: which set of threads strips first

Every threaded hole is a race between two failure paths. The bolt can break in tension through its own reduced section, or the threads can shear out, and the threads that shear are the weaker of the two sets: the male thread on the fastener or the female thread in the parent.

The stripping area of the female thread is the cylinder of material at the thread's engaged interface, so it scales with engaged length times thread diameter. That is a linear relationship in engagement length, and it is the whole reason engagement is the lever.

For a bolt threaded into a parent of comparable strength, an engagement of roughly one bolt diameter is enough that the bolt breaks before the hole strips, and enough with margin: the convention is set so thread shear capacity exceeds the bolt's tensile capacity rather than merely equalling it. That is where it comes from. It is a statement about a strength ratio near 1, not a universal number, and the margin inside it matters below, because a hole at some percentage of a required engagement develops that percentage of the requirement, and the requirement sits above the bolt's own capacity.

When the parent is weaker, the required engagement scales up roughly with the ratio of strengths. Commonly cited starting conventions: on the order of 1.5 diameters into cast iron, about 2 diameters into common aluminum alloys, and 2.5 to 3 diameters into magnesium and into thermoplastics. Treat those as a place to start a check, not as an answer. The number that governs a specific joint belongs to the fastener or insert manufacturer's data for that alloy, or to the engineer of record.

Why the load does not spread evenly, and why that flips

In a nut and bolt of similar material, engaged thread load is badly non-uniform. The first engaged thread carries a large share, commonly cited as roughly a third, and the share falls off sharply along the engagement. That is why adding engagement past about one and a half diameters in a like-material joint buys you very little: the extra threads are barely loaded.

In a soft parent the same non-uniformity exists at first and then does not persist. The most heavily loaded threads yield locally, the material flows a small amount, and the load redistributes onto threads further down the hole. That local yielding is the reason deep engagement genuinely works in aluminum when it is nearly wasted in steel. It also tells you what you are watching for on a second visit: a soft-parent thread that has taken a slight permanent set does not look damaged, but the fastener will have lost preload, because the joint got shorter by the amount the threads flowed.

The direction is worth holding in mind at both ends. Short engagement in soft metal fails suddenly and completely, pulling a plug of threads out. Generous engagement in soft metal fails gradually as loss of preload over successive heat cycles, long before anything strips.

Two outcomes from the same gate

Take one bolt, 3/8 in nominal with a 16-per-inch thread, and run it against the gate in two housings.

Outcome one: a steel boss, 3/8 in deep. Engagement is one diameter into a parent of comparable strength, so the male thread and the bolt's own section are the weak links, not the hole. The bolt's grade decides the joint's capacity, the manufacturer's torque figure for that bolt and that friction condition applies, and reaching that torque means what it usually means.

Outcome two: a cast aluminum boss, 0.50 in deep. Now run the arithmetic. Take the parent alloy's shear strength as roughly a third of the fastener steel's, which is a defensible working assumption for common cast aluminum against a medium-grade steel fastener and which you should replace with real values from the alloy data when the joint matters. Required engagement to develop the bolt fully is then about three times the steel-in-steel figure: 3 x 0.375 in, which is 1.125 in. Note that this is deliberately more conservative than the 2-diameter rule of thumb quoted earlier for aluminium, which would give 0.75 in. The two disagree because one is a screening rule carrying its own margin and this is a calculation from a strength ratio. Where they disagree the alloy data governs and the rule is the screening check, so run the rule first and do this arithmetic when the rule says the joint is close.

Available engagement is 0.50 in, which is 1.33 bolt diameters and looks generous to the eye. Against the 1.125 in the joint needs, it is 0.50 / 1.125, or 44 percent. Because stripping area is linear in engagement, this hole develops roughly 44 percent of that bolt's tensile capacity before the aluminum threads let go. The bolt's grade is now irrelevant. A higher grade raises the number the hole will never reach.

Carrying that example to a decision

The bracket in question was going to hang on two fasteners into that boss. Two holes at 44 percent each deliver about 89 percent of one fully developed fastener, which is short of a single bolt's capacity, let alone two.

Three routes out, in the order you should test them against the actual casting:

1. Drill deeper. The casting has 0.75 in of sound material before the boss breaks into a passage. That gives 0.75 / 1.125, or 67 percent per hole. Two holes at 67 percent is 133 percent of one bolt, which is real improvement and still not the two-bolt joint that was drawn.
2. Add fasteners. At 44 percent each, three holes give 133 percent of a single fully developed bolt and four give 178 percent. Spreading load across more shallow holes is legitimate, and it is usually the cheapest fix when the casting has face area to spare. It requires that the bracket is stiff enough to actually share load across four points rather than loading two of them.
3. Change what the aluminum sees. A helical thread insert or a solid threaded bushing moves the load in the aluminum out to a larger diameter.

What an insert actually buys, and what it does not

This is where shops over-credit the fix. The tapped hole for a helical coil insert in a 3/8 in thread is on the order of 7/16 in, roughly 17 percent larger in diameter. Because stripping area scales with diameter times length, a 17 percent diameter increase buys about 17 percent more area at the same depth. That alone does not turn 44 percent into 100 percent.

What the insert really buys is different and more useful: a hard steel thread form in place of a cast one, load spread more evenly along the engagement because the coil is compliant, a repairable interface for a hole that will be opened many times, and immunity to the galling that happens when a steel fastener is cycled directly in aluminum. Those are excellent reasons to specify one. Capacity multiplication is not among them.

A solid bushing with a substantially larger outside diameter is the option that does multiply capacity, because you choose its outside diameter. A bushing whose thread into the casting is twice the bolt's diameter roughly doubles the area at the same depth. It costs more material around the hole, which the casting may not have.

Setting a torque the parent can survive

Once the hole is the weak element, the bolt's published torque is the wrong number and using it is how the thread gets pulled on the first assembly rather than in service. Two ways to get the right one:

• The manufacturer's value for that alloy, that thread, and that engagement. This is the answer whenever it exists. Insert manufacturers publish installation torque by parent material; equipment manufacturers publish it by fastener location.
• A destructive test on scrap of the same alloy at the same depth, where no published value exists and the joint is yours to specify. Tap three sample holes, pull each to failure on a torque wrench, record the strip torque, and set the working figure well below the lowest of the three. Do this on a benched sample clamped in a vise, not on the installed equipment, and keep hands clear of the wrench's swing path, because the wrench releases suddenly when the thread goes.

Checking you got this right

• Count turns, not just torque. On installation, the fastener should reach its torque within the turns you expect from the pitch and the joint's crush. A fastener that keeps turning while the reading climbs slowly is already flowing the parent thread. Stop.
• Re-check after the first full heat cycle with the equipment stopped and isolated under 29 CFR 1910.147. Soft-parent joints lose preload here if the threads took a set, and that loss is the early warning that shows up before anything strips.
• Inspect the withdrawn fastener's threads for aluminum transfer. Bright metal smeared in the fastener's thread roots means the parent thread has been galling and shedding, which is engagement loss you cannot see by looking into the hole.
• Write the engagement depth on the job record. The next person to open this joint needs to know they are working against a 0.50 in hole, not a 1.125 in one, and no inspection of the assembled joint will tell them.

References

• Fastener and thread-insert manufacturer data for minimum engagement and installation torque by parent material, which owns any specific value for a given alloy and thread
• 29 CFR 1910.133 for eye and face protection; 29 CFR 1910.147 for mechanical and stored-energy isolation; 29 CFR 1910.333(b)(2) where electrical conductors are in the work zone
• NFPA 70E-2021, 120.5, in the edition adopted by your employer's electrical safety program or your authority having jurisdiction, for the live-dead-live proving sequence
• See related: What Torque Actually Controls and What It Does Not; Why Reusing a Fastener Is Sometimes Fine and Sometimes Not