How to Choose an Anchor for What You Are Hanging

Why this matters

Most anchor selections start at the catalog and work backwards, which means the first number anyone sees is a capacity. Capacity is the last thing that should matter, because the character of the load disqualifies whole families before any number is compared. A load that never comes off rules out options a heavier intermittent load would allow. A load hanging overhead carries requirements a wall-mounted load of the same weight does not. Get the sequence right and the catalog page you eventually open has one column on it.

What follows is a sequence, not a ranking. Each step deletes candidates, and running them out of order means deleting candidates on the wrong grounds.

Before you go overhead

• Anything suspended overhead is a fall hazard for whatever is under it. Clear and control the area below before the first hole, and keep it controlled while the load is temporarily supported.
• Support the load independently until the anchors are ready to take it. For an adhesive anchor that means until the manufacturer's full cure time has elapsed at the base material's temperature, not the room's. Nobody works under an equipment load carried by uncured adhesive, and nobody removes the temporary support to "see if it holds."
• Overhead drilling in concrete or masonry is a silica exposure and it is an inhalation route, so a glove and glasses do not address it. Construction work falls under 29 CFR 1926.1153, whose Table 1 pairs handheld drilling with water delivery or a dust collector with HEPA filtration; general industry is 29 CFR 1910.1053, with respirators run under a written program per 29 CFR 1910.134. Face protection per 29 CFR 1910.133 for the spoil that comes straight down.
• Working from a ladder or lift brings its own fall-protection requirements, and reaching overhead with a tool that can bind is where people come off ladders. Use a platform that lets you keep both feet planted and your reach inside the rails.
• Scan before drilling. Conduit, pipe and tendons. Where conduit is plausible and the scan is not clean, isolate the affected circuits under 29 CFR 1910.333(b)(2) and prove dead with the live-dead-live sequence in NFPA 70E-2021, 120.5.

Step 1: Resolve the load at the worst anchor, not the average one

Take the operating weight, not the shipping weight. Then find the anchor that carries the most, which is almost never the total divided by the anchor count. Two things push it off the average: an offset center of gravity, and an anchor pattern that is not symmetric about it.

Skipping this step is how a four-anchor pattern gets sized on a quarter of the load when one corner is carrying a third.

Step 2: Classify the load's duration and character

Four categories, and they route differently:

• Sustained load never comes off. It is the case that matters for adhesive creep and for any preload-dependent family.
• Intermittent or live load comes and goes. Higher peaks are tolerable here than in a sustained case.
• Cyclic or vibratory load comes from equipment that runs. It relaxes preload over time, which is what a sibling article on vibration covers, and it favors families that do not depend on preload.
• Shock or seismic load arrives fast and reverses. Where the jurisdiction assigns a seismic design category, equipment anchorage is a designed item under ASCE 7 Chapter 13, in the edition incorporated by the building code your authority having jurisdiction has adopted, and it binds through the permit and the engineer of record rather than on its own.

Step 3: Add the loads that are not the equipment

The nameplate weight is the start of the number, not the number:

• Contents. Water, refrigerant charge, product, condensate in a pan.
• Environmental additions for anything outdoors: ice and snow accumulation, wind uplift on a housing, drifted load against a wall.
• Service loads. Somebody will pull on this to open a panel, or lean on it with a meter in one hand. Shops commonly carry an allowance for that on suspended equipment. Where the anchorage is a designed item, the engineer sets the factor; where you are specifying it yourself on a small unit, state whatever allowance you use on the record so the next person knows the basis.
• Restraint forces. If the equipment is connected to piping or duct that grows and shrinks with temperature and the anchors are the only thing resisting that movement, the anchors are carrying thermal load too. That force is usually larger than people expect, and the fix is usually a flexible connection rather than a bigger anchor.

Step 4: Decide what a single anchor failing does

Redundancy is a design decision, not a fastener property. Ask what happens if one anchor loses capacity: does the load redistribute onto the others, or does the assembly come down? If it comes down, that pattern needs either more anchors than the arithmetic requires or a family with a failure mode that gives warning.

Step 5: Only now, check what the base material can supply

At this point you know the demand and the character. The base material determines which families can meet it, which is a separate question with its own record. A sibling article covers it in full.

Step 6: Size against the right column, and check the interaction

Two errors dominate:

• Reading the wrong table. Solid concrete versus hollow unit, cracked versus uncracked, and the temperature or moisture qualification are separate columns, and the value on the front page of a datasheet is the most favourable one.
• Sizing on tension alone. Tension and shear interact, and an anchor near its tension capacity has little shear left. Use the manufacturer's interaction rule.

Step 7: Write the installation controls onto the work order

An anchor is a designed installation, not a part. The controls that must appear on the work order are the hole diameter and depth, the hole-cleaning procedure and how many cycles of it, the setting method and its verification, the installation torque or setting-tool requirement, and, for adhesive, the cure time at the actual base-material temperature. A correctly selected anchor installed in an uncleaned hole is a wrongly selected anchor.

Step 8: Record the geometry you actually achieved

Effective embedment, edge distance, spacing, base material, and the family and size installed. None of this is visible after the bracket goes on, and all of it is what a later review needs.

Worked example: a suspended unit on four rods

A piece of equipment hangs from a concrete deck on four threaded rods.

Step 1. The nameplate gives a dry weight of 140 lb and an operating weight of 165 lb, so the operating figure is the one that counts. The unit's center of gravity sits toward the connection end, and measuring the pattern against it gives roughly a 60/40 split between the two ends. The heavy pair therefore carries 0.6 x 165 lb, which is 99 lb, split between two rods: 49.5 lb per rod. The naive average of 165 lb across four rods would have been 41.25 lb, so sizing on the average would have under-called the worst anchor by about 17 percent, or put the other way, the worst anchor sits 20 percent above the average.

Step 2. This load is 100 percent sustained, overhead, and in pure tension on the rods. It is also cyclic to a degree, because the unit has a fan, though the unit sits on isolators so the anchors see mostly static load with a small alternating component.

Step 3. The operating weight already includes the fluid. The shop's standing allowance for service load on suspended equipment is 50 percent, giving 49.5 lb x 1.5, which is 74.25 lb of design tension per rod at the worst position. The piping connection is flexible, so no restraint force lands on the anchors.

Step 4. Four rods with a stiff frame means load redistributes if one loses capacity, so this is not a case that needs extra redundancy beyond correct sizing.

Step 5 and 6. The deck is poured concrete, which is where the geometry work in the sibling article on load transfer applies: the demand of 74.25 lb per anchor gets compared against the manufacturer's value for the actual embedment, edge distance and spacing achieved, in the cracked-concrete column, because a deck in flexure is a tension zone and cracked is the default assumption. Compare it against the right kind of value as well as the right column. The 74.25 lb is an unfactored service load: if the report publishes a design strength for strength design, the load is factored before the comparison; if it publishes an allowable, compare service against allowable; if it publishes an ultimate, the report's own factor of safety applies first. Write down which of the three you used.

The step that decides the family. The load case line reads "sustained, overhead, tension." Under ACI 318 Chapter 17, in the edition the adopted building code incorporates, an adhesive anchor installed horizontally or upwardly inclined to carry sustained tension requires a certified installer and continuous inspection. That is not a reason to avoid adhesive, but it is a real requirement with a real schedule cost, and it is why this shop specifies an undercut or a screw anchor qualified for cracked concrete on suspended equipment as a standing practice. Where adhesive genuinely is the right answer, they price the certification and the inspection into the job rather than discovering it at the inspection.

The failure mode if the sequence is skipped. Somebody sizes on 165 lb over four rods, gets 41.25 lb, finds an anchor rated well above that, installs it, and the unit hangs there for two years. The under-called corner is at 49.5 lb rather than 41.25 lb of dead load, and every time a tech opens the access panel and leans on the frame it goes higher. Sustained tension in an adhesive that was never qualified for it creeps. The first sign is a rod that has gone slightly slack against its nut while the diagonal one is tight, which reads as a leveling problem and gets adjusted rather than investigated.

How to verify you got this right

• Check the installation, not the selection. Measure the achieved embedment before the bracket goes on, and check the hole was cleaned to the stated procedure and count. These are the two controls that most often silently fail, and neither is inspectable afterwards.
• Torque-check to the anchor manufacturer's installation value, not to the rod's. For a torque-controlled expansion anchor the installation torque is what sets the expansion, so it is a functional value rather than a preload target.
• For adhesive, do not load until the full cure time has passed at the measured base-material temperature. Read the deck's temperature with a surface instrument. Keep the temporary support in place and the area below clear until then.
• Come back after the first month of operation and check for movement, with the equipment stopped and isolated under 29 CFR 1910.147. Mark each rod and nut with a witness line at handover so this check takes a glance rather than a wrench.
• A proof test, where the job calls for one, keeps everyone out of the load path. The test rig stores energy in a loaded rod; nobody stands in line with it, and the reaction frame bears outside the anchor's own failure cone.

References

• ACI 318 Chapter 17, Anchoring to Concrete, and ASCE 7 Chapter 13 for equipment anchorage in seismic design categories, each in the edition incorporated by the building code your authority having jurisdiction has adopted, binding through the permit and the engineer of record
• Anchor manufacturer installation instructions and evaluation reports for capacity by condition, hole cleaning, installation torque and cure time, which own every specific number
• 29 CFR 1926.1153 and 29 CFR 1910.1053 for respirable crystalline silica; 29 CFR 1910.134 for respiratory protection; 29 CFR 1910.133 for eye and face protection; 29 CFR 1910.147 for isolation during re-inspection
• See related: The Anchor Types and What Each Needs From the Base Material; How an Anchor Transfers Load Into Concrete; Why Things Vibrate Loose