Guides

Building a strapping chart for the tank you actually have

A nameplate lies more often than you'd think. Here's how to tape, measure, and prove a depth-to-volume chart for a real tank instead of trusting the drawing — including the five-row worked example that shows exactly why the middle isn't halfway.

By D. Shin · Published September 3, 2026

Sooner or later you inherit a tank with no chart. Maybe the original paperwork went missing with the last owner, maybe it's a tank someone welded up in the yard with no drawing at all, or maybe you just don't trust the laminated sheet taped to the side of it anymore — a couple of dents and thirty years of scale build-up will do that. Whatever the reason, the fix is the same: strap it yourself. That means taking your own dimensions off the physical vessel, running them through the same geometry every dip chart on this site is built on, and — the step people skip — checking the result against something the tank itself tells you, not just your arithmetic.

This covers the manual method: a tape, a few careful measurements, and a hand (or spreadsheet, or the strapping chart calculator) doing the segment math. It's the right scope for a shop, farm, or process tank where "accurate enough to run the tank by" is the bar. For a custody-transfer or legal-for-trade tank — one where the number on the chart settles who pays whom — the formal procedure is API MPMS Chapter 2, Tank Calibration, and it wants an accredited surveyor with calibrated equipment and a documented chain of measurements. Treat what follows as how that process works under the hood, and as entirely adequate for everything short of it.

Wrap the tape, don't span a ruler

The instinct is to measure diameter directly — hold a rule or a set of calipers across the top of the shell and read it off. Resist it. Wrap a tape around the full circumference instead, and get the diameter from D = C ÷ π.

The reason isn't convenience, it's accuracy. No real tank is a perfect circle: rolling tolerance, a weld seam that pulls the plate in slightly, decades of sitting on saddles that have let it settle into a faint oval. A single across-the-top measurement catches whatever diameter happens to lie under that one line — which on an out-of-round shell can read noticeably fat or thin depending on exactly where you put the rule. A tape wrapped all the way around averages over the whole shape. Circumference divided by π gives you the diameter of a perfect circle with the same cross-sectional area as your actual, slightly imperfect one — which is precisely the number the segment-area formula wants. It's also just easier: plenty of tanks are insulated, jacketed, sitting low on their saddles, or simply too big to reach across, and every one of them you can still walk a tape around.

Pull the tape to a consistent, firm tension every time — a loose wrap reads long, a hard yank on a fabric tape stretches it and reads short — and take it at two or three points along the shell's length if the tank is long enough for that to matter, since a shell can belly or taper slightly end to end. Average the readings. If the numbers spread by more than a millimetre or two per metre of circumference, that's telling you something about the tank's condition worth noting on the chart, not just noise to throw away.

Getting from the outside strap to the volume that matters

The tape reads the outside circumference. Liquid lives inside the shell, so before that number goes anywhere near a volume formula it needs the wall thickness subtracted — twice, once for each side of the diameter:

Inside diameter = (C ÷ π) − 2t

where t is the shell plate thickness. Get it off the fabrication drawing if one exists, or a calibrated wall-thickness (ultrasonic) gauge if it doesn't — a corroded tank's plate is often thinner than the number stamped on it decades ago, and using the original spec instead of the current reading will quietly overstate every row in the chart. If the tank is insulated or clad, the strap is reading circumference outside the cladding too, and that layer needs subtracting as well before you get anywhere near the steel.

Set the datum before you touch a formula

Every row in the chart is a depth measured from somewhere, and that somewhere has to be a fixed, physical point — not "the bottom, roughly." Pick the lowest interior point the liquid can reach, and mark it permanently: a punch mark, a small welded pin, or an engraved tag near the gauge point, so the next person to dip the tank — or to re-strap it in ten years — is indexing to the exact same spot you did. If there's a drain sump or a low pocket the outlet can't actually draw from, decide whether that volume counts as usable and set the datum accordingly; a chart that includes dead volume the pump can't reach will always read a little optimistic near empty.

Strap the tank level, or record its tilt if it isn't — most horizontal tanks are laid with a slight intentional slope to drain toward the outlet. A chart built assuming level, read on a tank that's a degree or two off, will be systematically wrong at every depth, not just near empty. For more on what tilt and datum choice do to a reading once the chart exists, see reading a tank dip chart without fooling yourself — it covers the innage/ullage side of this same problem.

Flat heads keep the arithmetic honest — dished ones don't

On a horizontal cylindrical tank, the two ends are either flat plates welded square across the shell, or a shaped head — torispherical, elliptical, or hemispherical — that bulges outward past where the straight shell ends. This matters more than it sounds like it should, because it decides how much extra work the volume math needs.

With flat heads, there's no extra work. The tank is a plain cylinder, full stop, and the volume at any depth is exactly the circular-segment area at that depth times the tan-to-tan shell length — the same segment formula used everywhere on this site (see the cylinder volume calculator for the geometry on its own, or the tank volume calculator for the full tank). With dished heads, each end adds its own bulge term on top of that — a capped-ellipsoid volume that depends on how far the head bulges past the tangent line, which you either pull off the fabrication drawing or measure directly with a straightedge and a tape. Skip that term on a dished-head tank and you'll under-read total capacity by anywhere from a couple of percent to well over ten, depending on how deep the heads are — a bias that doesn't average out, because it's the same error at every single depth.

That's exactly why the worked example below uses flat heads: it's the case where you can verify every number by hand with nothing but a trig table, and it isolates the part of this process that's genuinely new — the measuring and the verification — from head-shape math that's covered in full elsewhere on this site.

A worked chart: 1.82 m tank, flat heads

Take a small horizontal diesel tank with no surviving chart. Three wraps of a steel tape around the shell average out to a circumference of 5.752 m. The plate is 4.8 mm mild steel, confirmed with an ultrasonic thickness gauge rather than trusted from memory. The tan-to-tan length — inside face to inside face, since the heads are flat and flush with the shell ends — is 3.658 m.

• Outside diameter: 5.752 ÷ π = 1.831 m
• Inside (wetted) diameter: 1.831 − (2 × 0.0048) = 1.8214 m, so radius r = 0.9107 m
• Shell length, tan to tan: L = 3.658 m

The wetted cross-section at depth h is a circular segment, and its area is:

A(h) = r² · cos⁻¹((r − h)/r) − (r − h) · √(2rh − h²)

with the inverse cosine in radians, and volume is A(h) × L. Working that through at five depths spaced a quarter-diameter apart — the same close-form geometry the strapping chart calculator uses, evaluated by hand here so every number is checkable:

Notice what the fill-percent column is doing: the first quarter of the depth only buys 19.6% of the volume, but the second quarter buys another 30.4% — the tank fills fastest right through the middle and slowest near the bottom and top, because that's where a horizontal circle is narrowest. This is also a genuinely useful landmark for spot-checking any horizontal cylindrical tank, flat heads or not: the halfway depth — the exact centreline — is always precisely 50% of capacity, regardless of diameter or length, because the top half and bottom half of a circle are mirror images of each other by definition. If your rough chart says anything other than 50% at dead centre, go back and find the arithmetic error before you trust another row.

A real field chart wouldn't stop at five rows — you'd want a table you can actually read a dipstick against, typically every 25 mm or every inch. At a 25 mm increment this same tank works out to roughly 73 rows from empty to full. Rather than grind through that by hand, plug the same three numbers — 1.821 m diameter, 3.658 m length, flat heads — into the strapping chart calculator and it'll generate every row and hand you a CSV to print. The five-row version above exists to show you the shape of the curve and to make every intermediate value checkable by hand; the full-resolution one is what actually goes on the tank.

Picking the increment

Match the step to how finely you can actually read the dip. A dipstick scribed in whole centimetres gains nothing from a chart tabulated to the millimetre — you're just adding rows nobody can resolve by eye, and a longer table is a table people read the wrong line on. A 1 cm or 1 in step covers most field gauges; go finer only if the gauge itself — a sight glass with fine graduations, an electronic level sensor — can actually support it. If you're printing a laminated copy for the tank, keep the underlying calculation (the raw inputs and the un-rounded numbers) on file separately, so a future correction to the wall thickness or length doesn't mean re-climbing the tank to re-measure everything from zero.

Don't trust the arithmetic until a meter proves it

Every number so far has come from a formula fed by two field measurements. That's exactly the kind of chain — one bad tape reading, one wrong thickness, an internal baffle or heating coil nobody accounted for — that produces a chart which is confidently, consistently wrong. The fix is to check it against something the tank tells you directly: fill it by a known amount and see if the dip agrees.

On the example tank above, that means running exactly 4,766 L through a calibrated positive-displacement meter into the empty tank, letting the surface settle, and dipping. The chart predicts 0.911 m — dead centre, for the reason above. If the tape reads close to that, within roughly the meter's own uncertainty (a decent field meter is good to somewhere around ±0.5%), the chart and the physical tank agree and you can trust the rest of the table built from the same geometry. If the dip comes back noticeably off — say 0.93 m or 0.89 m instead of 0.911 m — something upstream is wrong: a misread strap measurement, a wall thickness that's thinner than assumed from corrosion, an internal fitting eating into the cross-section, or a tank that isn't quite as level as you thought. A single metered check at the halfway point is worth more than any amount of double-checking your own multiplication, because it tests the model against the real vessel instead of against itself.

If you don't have a calibrated meter handy, a weighed batch of water works the same way for a tank you can get on a scale, or a scale of known-volume containers pumped in one at a time for a smaller vessel — the point isn't the specific method, it's putting a known quantity in and seeing whether the chart's prediction and the tape agree before the chart goes into service.

Keeping the chart honest over the tank's life

A strapping chart describes the tank as it was on the day you measured it, and tanks don't stay that way. Build a habit of re-strapping — or at minimum, re-checking against a metered fill — whenever any of these happen:

• A repair, patch, or internal modification — a new heating coil, an agitator, a re-profiled bottom — changes the wetted volume at every depth below the change.
• Sediment, wax, or scale has visibly built up on the bottom; the chart's zero no longer matches the liquid's actual zero, and it will read high near empty until you correct for it or clean it out.
• The tank has settled or been re-leveled, changing the tilt the chart assumed.
• A routine metered check starts drifting from the chart by more than the meter's own uncertainty — treat that drift as the tank telling you something changed, not as noise to average away.
• Corrosion has measurably thinned the shell since the last thickness reading, which quietly grows the inside diameter — the opposite direction of every other item on this list, and easy to miss because it makes the tank hold slightly more, not less.

Date every chart, note who took the measurements, and give each revision a number. When a chart is superseded, pull the old copy off the tank rather than leaving it to sit next to the new one — a stale chart that's still legible and still taped up is one of the most common causes of a reading that mysteriously doesn't add up, and it costs nothing to prevent beyond remembering to take it down.