Where horizontal tank dipstick readings go wrong

Why half-depth is not half-full

Stand a cylinder on end and depth tracks volume in a straight line: 50% depth is 50% volume, every inch alike. Lay that same cylinder on its side and the rule breaks, because each inch of depth now spans a different width of liquid. Near the bottom the tank is a narrow lens, so an inch of rise adds almost nothing; the fat middle adds the most it ever will; near the top it narrows again. The volume under a horizontal stick reading follows a circular segment, and the area of that segment is A = r² · acos((r − h) / r) − (r − h) · √(2rh − h²), where r is the tank radius and h is the wetted depth from the bottom. Fill fraction is that area over the full circle, A / (π r²). The trades usually meet this curve as a strapping chart, the per-tank depth-to-volume table that calibration standards like API MPMS §2.2A and ISO 7507 define by physically measuring a tank. Dipfill computes the same depth-to-volume relationship closed-form on every keystroke instead of reading a pre-measured chart.

The table below is the whole argument in five rows. Read depth as a fraction of the tank diameter (a 48 in tank wet to 12 in is 25% depth), then read across to what fraction of the gallons that represents.

The error is symmetric around the centerline and peaks near the quarter and three-quarter marks at roughly 5 points. On a 1,000 gal tank that is about 54 gal of phantom or missing fuel from a reading that was itself dead accurate. The stick was not wrong. The straight-line guess was.

When the flat-end assumption breaks down

Dipfill models the tank as a plain cylinder with flat ends, so capacity is π r² times length and the fill fraction is the segment math above applied uniformly down the barrel. Real tanks often cap off with dished, domed, or hemispherical heads that bulge outward and add volume the cylinder model never sees. Feed Dipfill a 48 in by 96 in tank with dished heads and the readout runs low, because the true tank holds more than the flat-ended one the math assumes. The tool says so directly: every readout carries the line that it assumes flat end caps and that dished or rounded ends may read 5-15% off.

We put that 5-15% band on the readout itself rather than bury it here, because testing kept handing us the same trap: a clean-looking number carrying an invisible asterisk, small enough to look right and large enough to matter on a fuel order.

Is there a fix without re-deriving the heads? Yes, and it is the contractor's move: measure the real full capacity once from the data plate or a known fill, divide Dipfill's full-cylinder capacity into it, and carry that ratio as a correction factor. The non-linearity in the table still holds inside the barrel; the end caps just shift the total.

The diameter trap: what 9 of 10 calculators get wrong

The most common bad input is radius typed into a diameter field. Half the segment-area derivations online are written in r, so a user primed to think in radius meets a tank measured 48 in across the top and guesses which number the box wants. Type 24 when the field wants 48 and every downstream number collapses: capacity falls to a quarter of true, and the stick check rejects honest readings because a 30 in wet mark now exceeds the 24 in "tank."

We spelled out "not the radius" right in the diameter field hint rather than trust the label alone, because "diameter" is exactly the word a radius-primed user will misread. The field is labeled "Tank diameter across," its hint reads "The width of the tank end-to-end, not the radius," and the tool never asks for radius anywhere, so there is no second field to get crossed.

The stick field carries the matching guard. It is labeled "How far up the stick is wet," measured from the bottom, and must be less than or equal to the diameter. Enter a wet mark larger than the tank width and Dipfill refuses with "Stick reading can't be greater than tank diameter" instead of returning a fill above 100%. That single rule catches both the radius-in-the-diameter-field mistake and a mismeasured stick, because both produce the same impossible relationship.

Edge cases at the bottom and the brim

Three field conditions push a reading off even when the tank model is right.

The stick rests on a fitting. Drop a gauge stick down a fill port and it can land on a baffle, a sump, a draw tube, or weld slag instead of the true floor. The wet mark then reflects depth above the obstruction, not above the bottom, and Dipfill reads low because it measures every reading from the actual bottom of the tank. Probe for hard, flat, repeatable contact before you trust the inch mark, and read at the same port every time so the offset, if any, stays constant.

The tank sits tilted. A cylinder pitched even a couple of degrees pools liquid at the low end, so a stick at the high end under-reads and one at the low end over-reads. Dipfill assumes the tank is level, the assumption baked into multiplying one segment area by the full length. On a long tank a small slope moves real gallons, so shim it level or gauge from the center and treat the ends as suspect.

The bottom inch and the brim. Near 0% and near 100% the segment is flattest, so the volume per inch is smallest, which means resolution is worst exactly where you most want it. A 1 in error at 10% depth lands in the steepest, most forgiving part of the curve; the same error against a tank you are about to top off can be the difference between full and an overflow. Round conservatively at the brim. Dipfill displays gallons to one decimal place, but that precision is a property of the arithmetic, not of your tape measure, and the tape is the weaker link.

How to take a reading that the math can trust

Manual stick gauging predates the electronic tank-level sensors that became common around 1990, and the discipline it demands is unchanged. A reading the formula can trust starts with three measurements in the tool's own units. Diameter goes in as inches, the true width across, valid from 1 to 600 in. Length goes in end cap to end cap, inches or feet, up to 500 ft. The stick reading goes in as inches of wetted depth from the bottom. Dipfill parses each as a plain decimal and rejects anything else: a negative number, a stray letter, even scientific notation like 1e5 is dropped rather than guessed at, so a fat-fingered entry shows no result instead of a wrong one.

Take the reading straight down, not canted, because a stick angled across the tank reads a longer wetted length than the true vertical depth and biases every gallon high. Wipe the stick, drop it to firm contact with the bottom, pull it straight, and read the top of the wet line in good light. Use the same port and stick each time so any fixed offset cancels across readings. Confirm the diameter you typed is full width, not radius, then check your stick reading is at or below it; if Dipfill rejects the stick as greater than the diameter, one of those two numbers is wrong, not the tool.

One last caveat for the clipboard: every gallon Dipfill prints assumes flat ends and a level tank, so on a dished-head or visibly pitched tank, treat the readout as a calibrated estimate and reconcile it against a known full capacity before you size the next delivery.

Frequently asked questions

Why isn't a stick that reads half-depth a half-full tank?

On a horizontal cylinder, each inch of depth spans a different width of liquid, so depth and volume only agree at the exact centerline. Near the bottom and top the tank is a narrow lens that holds little, while the fat middle holds the most. The volume under a reading follows a circular segment, not a straight line, so a stick wet to half the diameter is genuinely half full only at that one midpoint.

How do I take a dipstick reading the math can actually trust?

Wipe the stick, drop it straight down to firm contact with the true bottom of the tank, pull it up without canting it, and read the top of the wet line in good light. A stick angled across the tank reads a longer wetted length than the real vertical depth and biases every gallon high. Use the same fill port and the same stick each time so any fixed offset cancels out across readings.

What is the difference between tank diameter and radius, and which does Dipfill want?

Diameter is the full width of the tank measured straight across the end, while radius is half of that distance from the center to the wall. Dipfill asks for diameter, the end-to-end width, never the radius. Typing the radius into the diameter field collapses every result: capacity falls to about a quarter of true and honest stick readings get rejected for exceeding the smaller tank you described.

How accurate is the calculator on tanks with dished or rounded end caps?

Dipfill models the tank as a plain cylinder with flat ends, so capacity is pi times radius squared times length. Real tanks often have dished, domed, or hemispherical heads that bulge outward and add volume the flat model never sees, which makes the readout run low. Expect roughly 5 to 15 percent error on those tanks, and treat every printed gallon figure as a calibrated estimate rather than an exact count.

How can I correct the reading for a tank with rounded heads?

You do not need to re-derive the head geometry. Measure the real full capacity once from the data plate or a known complete fill, then divide Dipfill's full-cylinder capacity into that real number and carry the result as a correction factor. Multiply future readouts by that factor. The non-linear fill curve still holds inside the barrel; the end caps simply shift the total, so a single ratio captures the gap.

What does gallons to fill mean, and how do I use it for an order?

Gallons to fill is the remaining headroom in the tank, the difference between full capacity and the gallons currently present at your stick reading. Size a delivery to that number, not to a guess off the depth percentage. Because fill is non-linear, a tank that looks three-quarters full by depth is closer to eighty percent full by volume, so ordering from depth alone risks an overflow near the brim.

Does Dipfill support metric units like litres and centimetres?

Yes. The output unit switcher converts the same closed-form result between US gallons and litres without changing your inputs, and tank length accepts inches or feet. One US gallon is about 3.785 litres, so a 752 gallon capacity reads as roughly 2,847 litres. The underlying segment math is unit-agnostic; the switcher only changes how the final volume is displayed, so the relative fill percentage stays identical.

Why does Dipfill reject a stick reading that is larger than the tank diameter?

A wetted depth cannot exceed the full height of the tank, which equals the diameter for a cylinder lying on its side. If the stick reading is greater than the diameter, either the diameter was entered as a radius or the stick was mismeasured, and both produce an impossible fill above 100 percent. Dipfill refuses to return a number in that case so a clearly broken input never turns into a confident wrong answer.

Does this calculator work for diesel, heating oil, and propane tanks?

Yes. The math depends only on the tank shape, not on what the tank holds, so the same diameter, length, and stick reading give the right gallons for diesel, heating oil, gasoline, water, or any liquid in a horizontal cylinder. Pressurized propane is the one to watch: Dipfill returns the liquid volume the tank geometry holds, while a propane gauge reads percent and propane is filled to about 80 percent for vapor space, so reconcile against the tank's percent gauge before ordering.

How do I measure my tank's diameter and length for the calculator?

Diameter is the full width across the round end of the tank, wall to wall through the center, in inches. Do not measure to the center and halve it, and do not enter the radius. Length is the straight barrel run from one end cap to the other, which you can enter in inches or feet. If you can only reach the data plate, the stamped nominal capacity is a useful cross-check against the full-cylinder number Dipfill shows.