Vacuum Pipe Sizing: Why Does Crane's 100 psi Chart Fail?

The pump gauge reads 25 in. Hg. The gauge at the tool, three hundred feet down the header, reads 12. Nobody changed anything. The pump is fine, the filter is clean, and the line has no leaks worth chasing. The header was sized off a compressed-air pressure-drop table without the density correction, and the mistake only shows up when production loads all the drops at once.

Get it running first: throttle back the number of simultaneous users, or drop a temporary larger hose in parallel with the worst run. Then work the checks below in order, because each one tells you whether the next one is even worth doing.

Check 1 — Confirm whether your flow number is SCFM or ACFM

This is where most vacuum header calculations die. Pumps are rated in ACFM at a stated vacuum level. Process demand is usually stated in SCFM. Pressure-drop tables in Appendix B-14/15 are tabulated in free air (SCFM) at 100 psig and 60°F.

Convert before you touch a table. At 25 in. Hg vacuum the volumetric multiplier is:

ACFM = SCFM x 30 / (30 - 25) = SCFM x 6

So 175 SCFM is roughly 1,050 ACFM at the vacuum end. At 100 psig the same air occupies about 1/7.8 of its free volume. Multiply those two factors and you get about 47 — the same number that appears in the Crane correction ratio, and that is not a coincidence. It is the density ratio between 100 psig air and 2.43 psia air.

Outcome: if your 175 was ACFM, the pump is smaller than you think and the header is smaller still. If it was SCFM, go to Check 2 and apply the ratio.

Check 2 — Apply the Crane pressure ratio, do not skip it

The Appendix B-14/15 note is explicit: for inlet or average pressures other than 100 psi and temperatures other than 60°F, multiply the tabulated drop by

(100 + 14.7) / (P + 14.7)  x  (460 + t) / 520

For 25 in. Hg vacuum, gauge pressure is −12.27 psi, so absolute pressure is 2.43 psia:

114.7 / 2.43 = 47.2   (at 60 F)
114.7 / 2.43 x (460+68)/520 = 47.9   (at 68 F)

The claim that density does not enter the pressure-drop calculation is wrong. Friction loss at a fixed mass flow scales inversely with absolute pressure because the gas is 47 times less dense and therefore moving 47 times faster through the same bore. There is no physical reason the correction would apply to positive pressure and switch off below atmosphere. If you correct a 125 psig header, you correct a vacuum header.

Reading 1.58 psi/100 ft off the chart for 175 SCFM in 1-1/4 in. Sch 40 (1.38 in. ID) and calling that the vacuum loss is the error. Corrected, the same table predicts 1.58 x 47.2 = 74.6 psi per 100 ft.

Check 3 — Is the predicted drop more than 20% of inlet absolute pressure?

74.6 psi of drop from a 2.43 psia source is impossible. That result is not telling you the correction is wrong; it is telling you the ratio method has left its valid range and the pipe is badly undersized.

The reason is acceleration. In a constant-diameter pipe carrying gas, pressure falls along the run, density falls with it, and volumetric flow rises. Mass flow is constant, so velocity must increase toward the outlet. Liquid pressure drop is friction only. Gas pressure drop is friction plus the momentum change from that acceleration, and the acceleration term is non-linear in pressure. Below roughly 20% drop relative to upstream absolute pressure, the acceleration contribution is small and a simple ratio is defensible. Above it, ratios stop working.

For a vacuum header at 2.43 psia, 20% is 0.49 psi — about 1 in. Hg. That is your entire allowable line loss before the shortcut method becomes invalid, which is why vacuum piping runs so much larger than the compressed-air line feeding the same process.

Read the last column as a diagnostic. The "ideal" ratio of 47.2 is the pure density effect. Real isothermal calculations come out above it, and the excess is the acceleration term. At 4 in. the actual ratio is 55.4 — 17% above ideal, because the drop is a meaningful fraction of 2.43 psia. At 6 in. it falls to 48.6, close to ideal, because the fractional drop is small. When your computed ratio sits near 47–49, the Crane shortcut is safe. When it climbs well above, upsize.

Check 4 — Run the velocity number at actual conditions

Velocity is the fastest sanity check on the shelf and it needs no software.

1. Get flow area. For 1.38 in. ID: A = 0.0104 ft².
2. Use ACFM, not SCFM. 175 SCFM at 25 in. Hg is roughly 1,050 ACFM.
3. V = Q / A / 60 = 1050 / 0.0104 / 60 ≈ 1,700 ft/s (higher, near 1,900 ft/s, depending on the standard-condition basis you use).
4. Compare against sonic velocity for air, about 1,126 ft/s at 68°F.

The line is choked at the inlet before it even starts. Running the same arithmetic with 175 treated as ACFM gives 280 ft/s, which looks acceptable — and that is exactly how an undersized vacuum header gets approved. The 280 ft/s figure answers a different question than the one you asked.

One more correction worth internalizing: velocity is not equal at inlet and outlet just because the bore is constant. Constant area with falling density means rising velocity. Size on the outlet (lowest pressure, highest velocity) end of the run.

Why the uncorrected chart value looks plausible

Skipping the correction produces a number in a familiar range — 1.58 psi, or 3.2 in. Hg per 100 ft — and 3 in. Hg of loss per 100 ft sounds like something a vacuum system could tolerate. It is plausible in magnitude while being wrong in derivation, which is the worst combination. Installations built this way run for years without complaint because they are lightly loaded, or the runs are short, or the pump has enough surplus displacement to mask the loss. Load every drop simultaneously and the header becomes the process bottleneck.

The same reasoning explains why the pump's own inlet connection is a bad sizing guide. A pump rated near 195 ACFM may ship with a 2 in. flange. That flange is sized for a few inches of machine porting, not several hundred feet of header with elbows, a filter, a check valve, and branch tees. Equivalent length of fittings matters far more in a vacuum header than in a compressed-air line, because the velocity head those fittings act on is 47 times larger for the same mass flow.

Size it properly: the compressible-flow method

Appendix B-14/15 exists because the manual predates desks with computers on them. Use the isothermal compressible flow equation instead — equations 1-6 and 1-7 — which handles friction and acceleration together and does not care whether you are above or below atmosphere. It fits on one spreadsheet row.

1. Fix the design vacuum at the process point and the vacuum at the pump inlet. The difference is your total allowable loss, including fittings.
2. Total the SCFM demand with a realistic simultaneity factor. Convert to ACFM at the process-end pressure, not at the pump.
3. Pick a trial size that keeps actual velocity in a sane range — well under sonic, and low enough that the fitting losses do not dominate.
4. Add equivalent lengths for every elbow, tee, valve, filter, and separator, then solve equation 1-6 for the pressure at the far end.
5. Verify the computed drop stays under about 20% of the upstream absolute pressure. Above that, go up a size and re-solve rather than scaling the previous answer.
6. Re-check the pump curve at the resulting inlet vacuum. Pump ACFM capacity falls as vacuum deepens, so a header loss of 3 in. Hg forces the pump to a deeper — and less productive — operating point.

For the 175 SCFM case at 25 in. Hg, 1-1/4 in. is not workable over several hundred feet. Expect to land at 2 to 2-1/2 in. minimum, and 4 in. if you want the drop to stay inside the 20% band with fittings included.

Verify on the installed system

1. Fit vacuum gauges at the pump inlet flange and at the furthest process connection. Compound gauges in in. Hg, same make and range on both ends.
2. Run the plant at true worst case — every drop open simultaneously, not one machine at a time.
3. Record both readings. The difference is your real header loss.
4. Convert to absolute and compare against the calculation: 25 in. Hg gauge = 2.43 psia; 20 in. Hg gauge = 4.9 psia. If the measured loss exceeds the computed loss by more than a modest margin, you have unaccounted fittings, a partially closed valve, a loaded filter, or an in-leak on a branch.
5. Blank off branches one at a time and re-read. A branch whose removal recovers several in. Hg is either leaking or undersized on its own.
6. With the header isolated, confirm the pump reaches its rated vacuum at its own flange. If it does not, the problem is the pump or its inlet filter, not the piping.

Stop here and escalate

Stop if the corrected calculation predicts a drop larger than the available absolute pressure, or if computed velocity approaches sonic — those results mean the model has broken down and no amount of re-reading the chart will fix it. Stop also if the process requires deeper vacuum than about 27 in. Hg, if the stream carries condensable vapor or liquid carryover, or if the run mixes long headers with fast-cycling loads, because none of those are single-phase isothermal problems.

Take the layout, the demand list, and the measured gauge readings to the pump manufacturer's application engineering group and to a piping engineer with compressible-flow software. Get the pump curve at your actual inlet vacuum in writing before anyone orders pipe.

FAQ