Lesson D6 named the trap most schemes fall into: a condensate line sized as if it carried water. This lesson does the arithmetic that lesson promised. It is short, it uses one steam-table lookup you already know from lesson A8, and it ends in a calculator — because the difference between DN15 and DN40 on the same 760 kg/h duty is the difference between a return system that works and one that hammers, floods and gets its healthy traps condemned.
Three fluids, one pipe
Follow a condensate line through one day and it carries three different fluids. At start-up, cold condensate — pure water, no flash, the highest mass flow of the day (warm-up loads run two to three times running load, lesson A10). For the running day, a two-phase mixture — condensate that crossed its trap and flashed. And after a pump, pressurised water again, single-phase, because a pump moves liquid without letting it flash. Three fluids, three sizing rules. The mistake is using the water rule for the two-phase stretch.
Why the flash owns the pipe
Run the practice plant's jacketed vessel: 760 kg/h of condensate leaves the jacket at 3.0 kg/cm²g and crosses the trap into a return line at 0.5 kg/cm²g. Lesson A8's flash fraction: sensible heat above, minus sensible heat below, divided by latent heat at the lower pressure — (603 − 467) ÷ 2226 ≈ 6.1% by mass. Forty-six kilograms per hour of flash steam. It sounds negligible beside 714 kg/h of water.
Now convert both to volume, which is what a pipe actually carries. At 0.5 kg/cm²g the specific volume of steam is about 1.16 m³/kg: 46 kg/h × 1.16 ≈ 54 m³/h of vapour. The 714 kg/h of water occupies 0.7 m³/h. The steam side holds almost 99% of the volume. The line is a steam line with some water in the bottom — and every sizing decision follows from that sentence.
The method: size it as a steam line, at the return pressure
The working rule is lesson C2's velocity method applied to the flash:
- Flash fraction from the pressures either side of the trap (lesson A8, or the calculator below).
- Flash volume = flash mass × specific volume at the return pressure — the lower the return pressure, the bigger the volume, which is why "generous" vented returns need the most metal.
- Bore from velocity. Keep the flash at 15–20 m/s on short trap-discharge lines, and nearer 10–15 m/s on long common returns — slower than a dry steam main (C2's 25–35), because this steam shares the pipe with slugs of water it can pick up and throw.
The jacket line: 54 m³/h at 15 m/s needs a 36 mm bore → DN40, which then runs a comfortable 11 m/s. The water-only calculation — 0.7 m³/h at 1.5 m/s — answers DN15. Fit the DN15 and the 46 kg/h of flash must squeeze through at over 100 m/s: the line screams, water slugs accelerate to hammer speed (lesson C5), back pressure climbs, and every trap upstream loses the differential it was sized on. The plant then replaces the traps. The pipe was guilty.
Two refinements. Long common returns collect flash from many traps — sum the flash loads of everything connected, but take credit for distance only if the line is insulated (bare returns condense their own flash usefully; insulated ones carry it further — either way, size for the worst section). Falling lines are forgiving; rising lines after a trap are slug-flow by construction — avoid them where layout allows, and where it does not, size one step larger, enter the riser from a small collecting pocket, and remember every metre of lift costs ~0.1 kg/cm² of the trap's differential (lesson D5's rule).
After the pump, it is water again
Downstream of a condensate pump — electric or pressure-powered — the fluid is pressurised liquid and the water rules return: 1–1.5 m/s for continuous duty, up to 2 m/s for short pumped bursts. One warning carries over: a pressure-powered pump discharges in strokes, not a steady stream, so its delivery line sees roughly twice the average flow during each stroke — size on the stroke rate, not the daily average, or the check valve chatters its seat away (a G4 failure in the making).
- Stand by your busiest return line at full load: banging, jumping on its supports, or hissing at the receiver vent = sized for the water, not the flash.
- Before condemning a trap that "stopped working", price the differential: metres of lift × 0.1 kg/cm², plus the return-line back pressure. The line is guilty more often than the trap.
- Walk every tee where a trap discharge meets a common return: entries should sweep into the run from the top. A bottom-opposed tee is two water streams meeting head-on — a hammer generator.
- Check the return line size against the calculator's answer for your biggest trap discharge. One size small is common; two is an incident file.
- A running condensate line is a steam line with a wet floor: ~6–16% flash by mass is ~99% by volume.
- Size the flash at the return pressure: 15–20 m/s on trap discharges, 10–15 on long common returns.
- The practice plant's 760 kg/h jacket return needs DN40. The water arithmetic says DN15. The water arithmetic is wrong.
- Pumped condensate is single-phase again: 1–1.5 m/s, sized on the pump's stroke flow, not the average.
- Every metre of lift after a trap costs ~0.1 kg/cm² of differential — count it before you blame the trap.
The "one to two sizes larger" rule of thumb predates the explanation. Fitters in the early plumbing-trade handbooks passed it on as bare instruction — condensate mains one size up from the water table, no reason offered — because crews that followed it stopped being called back for hammering returns. The flash-volume arithmetic that justifies the rule arrived decades later, when two-phase flow became respectable research. The trade knew the answer long before engineering knew the question; this lesson is the question, finally attached.