How to Calculate DP Level Range for a Sealed Tank

Back to Blog

Define the Geometry and the Two Legs

Start by naming the fixed dimensions in symbols so the arithmetic stays honest. Let d be the vertical distance from the transmitter's high-side tap down to it (or up, depending on mounting), let H be the process liquid height above the lower tap, and let L be the height of the sealed wet reference leg on the low side. Let SG_p be the specific gravity of the process liquid and SG_f the specific gravity of the fill fluid in the reference leg. The transmitter measures high-side pressure minus low-side pressure, and every term is a density times g times a height.

On a sealed tank the vapour space pressure sits on top of both legs equally, so it cancels in the differential and drops out of the math, which is the whole reason a DP arrangement works on a pressurized vessel. What remains is the process head on the high side against the fixed reference head on the low side. If any of this feels abstract, the concept page on differential-pressure level measurement covers why the wet leg exists and what it compensates for.

Derive the 4 mA and 20 mA Points

The differential the transmitter sees is DP = SG_p x g x H - SG_f x g x L, where the process head grows with level and the reference head is a fixed number set by the filled leg. At minimum level, H = H_min (often zero at the tap), so the differential is DP_min = SG_p x g x H_min - SG_f x g x L. Because the reference leg is always full, DP_min is a negative number: the low side is heavier than the high side when the tank is low. That negative value is your 4 mA point, and mapping 4 mA to a negative differential is called zero suppression or elevation.

At maximum level, H = H_max, so DP_max = SG_p x g x H_max - SG_f x g x L, and that value is your 20 mA point. Notice the reference term appears in both DP_min and DP_max, so it does not affect the span: subtract the two and the wet-leg term cancels, leaving span = SG_p x g x (H_max - H_min). The span depends only on the process liquid and the level swing, while the suppression depends only on the reference leg. Keeping those two effects separate in your head is what makes the calculation reliable. The reference column that creates the suppression is the wet leg.

Work a Symbolic Example End to End

Take a sealed tank with a level swing of H_max - H_min = 3.0 m of process liquid at SG_p = 0.85, and a wet reference leg L = 3.5 m filled with the same fluid at SG_f = 0.85. Working in pressure head expressed as metres of water column, the process head at full is 0.85 x 3.0 = 2.55 mH2O and the reference head is 0.85 x 3.5 = 2.975 mH2O. So DP_min (tank empty) = 0 - 2.975 = -2.975 mH2O, and DP_max (tank full) = 2.55 - 2.975 = -0.425 mH2O. Both endpoints are negative, and the span is DP_max - DP_min = 2.55 mH2O, which is exactly SG_p x (H_max - H_min) as predicted.

So the transmitter is configured 4 mA at -2.975 mH2O and 20 mA at -0.425 mH2O, a fully suppressed range that never goes positive. This is why a sealed-tank DP level almost always needs elevation, and why forgetting it makes the transmitter read backward or saturate. Once you have derived these two numbers symbolically you can enter them directly, then confirm by injecting the calculated pressures, which is the setup half of the wet-leg DP zero trim procedure.

Verify the Numbers Before You Trust Them

Sanity-check the derived endpoints against physical intuition before entering them. The empty-tank differential should be the most negative value, equal to minus the reference-leg head, because the high side sees no process head. The full-tank differential should be less negative, because the process head partly offsets the reference. If your arithmetic gives a positive 4 mA point on a wet-reference tank, you have almost certainly dropped the reference-leg term or a sign, and the transmitter will read wrong.

Then prove it by injection. With the transmitter isolated, apply the calculated DP_min to the transmitter and confirm it drives 4 mA, apply DP_max and confirm 20 mA, and check the midpoint for linearity. This closes the loop between the paper calculation and the real electronics. Recording the calculated endpoints and the as-left injection results on the calibration sheet gives you the reference every later drift check will compare against, and pairs naturally with an as-found and as-left calibration record.

Avoid the Common Mistakes

The recurring errors are all sign and reference errors. Using the process density for the fill fluid when the reference leg holds a different glycol or seal fluid throws the suppression off. Forgetting that the reference leg puts both endpoints negative leads people to configure a positive zero and get a transmitter that reads full when the tank is empty. Mixing pressure units, metres of water against inches of the actual fluid, quietly scales every number. Pick one unit system and one sign convention and hold them through the whole derivation.

The other trap is assuming the reference leg stays at its design height. If the wet leg loses fill or gains condensate the suppression term changes and the whole measurement shifts, which is exactly the kind of slow bias a continuous trend catches. When the level tag is monitored, a reference-leg problem shows as a steady offset against periodic hand dips, letting you re-verify the calculation against reality rather than assuming the paper math still holds.

Frequently Asked Questions