Why a Mobile Crane Is a Different Machine on Soft Ground

Why this matters

A load chart describes a machine sitting level on firm support. Take either of those away and the chart is not a conservative description of the crane any more, it is a description of a different crane. The mechanism is not dramatic: a float settles an inch or two, the machine tilts less than a degree, and the boom quietly reaches further than anyone planned. This card runs one gate, the manufacturer's level tolerance, against the same crane on two different pads, and shows that the finding in the failing case is not the four and a half percent of extra radius. It is that the chart stopped applying at all.

The gate

Level within the tolerance the manufacturer states for that machine. It is commonly stated as within 1 percent, which is about 0.57 degrees, and the number belongs to the manufacturer's operator manual rather than to a rule of thumb, so read it rather than assuming one percent.

It is a pass or fail, not a scale. Outside the tolerance there is no reduced chart to fall back on and no derate factor published anywhere, because the chart was never derived for that condition. A crane 1.5 percent out of level is not operating at a slightly lower capacity. It is operating with no published capacity.

Why a small tilt costs a large radius

The boom is long and the tilt acts at the end of it. Tilt the whole machine toward the load by an angle and the boom's angle to true horizontal drops by that same angle, and the horizontal projection of the boom, which is your radius, grows by the difference in cosine.

Two things follow that people find counterintuitive. First, the effect gets worse at steeper boom angles, because the cosine is changing fastest there. A tilt costs more radius on a short-radius steep pick than on a long-radius flat one. Second, the swing axis has tilted with the machine, so the boom runs downhill for part of every swing and the radius keeps growing as you rotate toward the low side. The radius you measured at the pick is not the radius you get at the set.

The capacity consequence of a longer radius is owned by the radius card in this library, and the short version is that capacity falls at least as fast as radius grows. What this card adds is that on yielding ground the radius change is not something you planned, and it does not stop when you stop.

Settlement is progressive, and setup is not a test

Ground that carried the machine while the outriggers were being set has been loaded for a few minutes at a fraction of the reaction it will see with the boom out and a load on. Fine-grained and recently placed soils keep consolidating under a sustained load, so a float that read solid at setup can be measurably lower ten minutes into a hold.

Three site conditions make this reliably worse and are worth naming, because all three look fine from the cab:

• A frozen crust. It carries at seven in the morning and is gone by early afternoon. The bearing you tested is a temperature, not a soil property.
• Rain during the lift. Placed fill loses strength as it takes water, and a pad that was adequate dry is not the same pad wet.
• A backfilled trench or a vault under one float. This is a void problem rather than a pressure problem, and no mat size answers it. Mat sizing arithmetic belongs to the ground-bearing card and is not repeated here.

The one thing you may not do about it

Do not re-level under a suspended load. Extending or retracting an outrigger, or adding cribbing, while a load is in the air is prohibited by most manufacturers' procedures, and 29 CFR 1926.1417 requires construction operations to comply with those procedures. Retracting a jack under load transfers the machine's weight onto whatever is left, which is the same event as the settlement you were trying to fix, only faster and under your control instead of the soil's.

The sequence is: stop, land the load in a controlled spot, get everyone outside the swing radius, then correct the support. If there is nowhere to land the load, that is a lift planning failure that has already happened, and the time to have solved it was at the plan.

The gate run twice: the same crane, two pads

The machine, the load and the geometry are identical in both cases. 100 ft of main boom at 70 degrees from horizontal, all boom angles in this card measured from horizontal. Outrigger spread across the machine is 20 ft, which is 240 in. The manufacturer's level tolerance for this machine is 1 percent.

Baseline, perfectly level. Radius is the horizontal projection of the boom: 100 ft x cosine 70 degrees = 34.20 ft.

Case A: engineered pad, compacted granular over a prepared subgrade. After the first pick, the mark on the leading float has dropped 0.5 in.

• Tilt: 0.5 in / 240 in = 0.21 percent, which is about 0.12 degrees
• Boom angle to true horizontal: 70 minus 0.12 = 69.88 degrees
• Radius: 100 ft x cosine 69.88 degrees = 34.40 ft
• Radius growth: 0.20 ft
• Gate: 0.21 percent against a 1 percent tolerance, pass

Two tenths of a foot is inside the tape's own honesty and well inside the boom deflection term the radius card already accounts for. The chart still describes this machine. Keep the mark, re-check after the next pick, and carry on.

Case B: unimproved fill, rained on overnight. After the first pick, the mark on the leading float has dropped 4 in.

• Tilt: 4 in / 240 in = 1.67 percent, which is about 0.955 degrees
• Boom angle to true horizontal: 70 minus 0.955 = 69.045 degrees
• Radius: 100 ft x cosine 69.045 degrees = 35.76 ft
• Radius growth: 35.76 minus 34.20 = 1.56 ft, or 4.6 percent
• Gate: 1.67 percent against a 1 percent tolerance, fail

Now read those two results in the right order. The 1.56 ft of radius is the number people reach for, and using the illustrative chart rows in the radius card it would move the lookup from the 35 ft row to the 40 ft row, which on that table is a 15.1 percent loss of capacity. That is a large number and it is not the finding.

The finding is the gate. At 1.67 percent the machine is outside the condition the entire chart was derived under, so there is no 40 ft row to fall back on either. You cannot correct your way back onto a chart you have left. The capacity arithmetic is useful only as an illustration of how expensive a fraction of a degree is; it is not a permission to continue at a reduced figure.

And it is still moving. A 4 in settlement measured after one pick, on saturated fill, is a rate, not a state. The next check is the one that matters, and the crew has no basis for predicting it.

What Case B actually calls for. Land the load, get clear, and go back to the ground. That means re-establishing the support with the party who owns ground conditions on the site, which under 29 CFR 1926.1402 in construction work is the controlling entity, and re-sizing the support with the mat and reaction data the ground-bearing card walks through. It may also mean re-planning at a shorter radius so that the same support carries a lower reaction. What it never means is levelling the machine back up and taking the same pick again on the same pad, because nothing about the soil changed when the bubble came back to centre.

Measuring the thing you are actually gating on

A 1 percent tolerance is roughly half a degree, and half a degree is hard to see. Three practical consequences:

Use the method the manufacturer specifies, which on modern machines is a level indicator with a stated resolution rather than a spirit level laid on the deck. A bubble that looks centred is not evidence at a half-degree threshold.

Check level at more than one swing position. On a tilted base the deck level reading changes as the machine rotates, so a single reading over the rear tells you about one bearing only.

Measure float elevation directly, not just machine level. Put a mark or a level reference at each float at setup and read it again after the first pick. Machine level tells you the sum of four settlements; float marks tell you which corner is moving, which is the one you have to fix. Take those readings from outside the swing radius, with the load landed and the operator told what you are doing, and never from under the carrier or between a float and the machine.

What this looks like when it goes wrong

The failure almost never announces itself as a tilt. It announces itself as a machine that "feels different", a boom that seems to walk during a swing, a load moment indicator that alarms at a pick which was fine an hour earlier, or a load landing a foot outside its marks. Every one of those is the same event described from a different seat.

The tell that separates a settlement problem from an ordinary planning error is that it changed between two identical picks. A lift that was 85 percent of chart in the morning and alarms in the afternoon, with the same load, the same configuration and the same marks, did not get heavier. The machine moved. When a crew's answer to that observation is to swing more slowly, the diagnosis has been skipped, and the next hour is spent on ground that is now known to be moving and still not measured.

References

• 29 CFR 1926.1402, ground conditions, which requires ground to be firm, drained and graded with supporting materials as necessary and assigns the preparation duty to the controlling entity on a construction site
• 29 CFR 1926.1417, operation, which requires compliance with the manufacturer's procedures, including levelling tolerance and any prohibition on adjusting outriggers under load
• ASME B30.5, mobile and locomotive cranes, in the edition adopted by your employer's program, your contract or your authority having jurisdiction
• Manufacturer operator manual and load chart for the specific machine, which own the levelling tolerance, the level indicator method and every capacity figure
• See related: What Outriggers and Ground Bearing Actually Decide; What Radius Does to a Crane Capacity and Why It Surprises People