What a Drainage Fixture Unit Is Actually Counting

Why this matters

A drainage fixture unit is a probability, wearing the costume of a volume. It is not gallons, it is not gallons per minute, and it is not what the fixture discharges. It is an index that already has an assumption about simultaneous use baked into it, which is why a table can add fifty fixtures together and arrive at a pipe smaller than the sum of their discharges would ever justify. Techs who treat the number as a flow rate size fine by accident on large systems and get burned in exactly two places: very small branches, where the smoothing the index assumes has not happened yet, and continuous flows, where it never happens at all. This card is about what the index is counting so you know where its assumption runs out.

What the number is built from

Behind every fixture unit value sit two things multiplied together: how much the fixture discharges when it discharges, and how likely it is to be discharging at the instant the system sees its peak. The second term is the one that does the work.

A water closet dumps its whole tank in a few seconds and then sits idle for hours. A lavatory trickles for a minute. If you sized a stack for every fixture in a hotel discharging at once you would run pipe nobody has ever needed, because the physical event that sizing protects against is not "everything at once," it is "the worst realistic coincidence." Fixture units encode a judgement about that coincidence, calibrated against real building use, and hand it to you as an integer you can add.

That is why the same physical fixture carries different values in different columns of the same table. A lavatory in a private bathroom and the identical lavatory in an airport restroom are the same porcelain and different probabilities. The column you pick is a statement about use frequency, not about the fixture.

The three things the number is not

It is not a flow rate, and there is no fixed conversion. The relationship between a fixture unit total and an actual peak flow is a curve, not a constant, because diversity increases as the count increases. Two hundred fixture units do not present twenty times the flow of ten fixture units. Anyone handing you a single multiplier to turn fixture units into gallons per minute has flattened a curve into a line, and the error is largest exactly where small systems live.

It is not additive across populations with different use patterns. Two hundred fixture units of school restroom and two hundred fixture units of apartment building do not produce the same peak, and neither one produces the same peak as two hundred units of stadium concourse. The table has already made a judgement about your building type when it assigned the values, and stacking totals from two different judgements produces a number that describes nothing.

It is not a measure of the largest single event. The index describes an aggregate. It says nothing about what happens in the four seconds a single water closet discharges into a branch that serves only that water closet. Some checks in a sizing exercise are governed by the aggregate and some are governed by that single event, and knowing which is which is the whole skill.

Where the diversity assumption runs out

Diversity is a large-numbers effect. It needs a population to average over, and a branch serving three fixtures does not have one.

Think about what the load on a two-fixture branch actually looks like over a day. It is zero almost all the time. When it is not zero it is one fixture, and occasionally it is two. There is no smoothing available: the branch either sees nothing or it sees the biggest thing on it. Add the fixture units and you get a small number that a table maps to a small pipe, and the table is not wrong, because the table's minimum sizes for the fixtures themselves are what actually governs at that end.

That is the crossover, and it is worth stating plainly because it reverses which rule you reach for:

• On a small branch, the governing constraint is usually the largest single fixture on it and the geometry around it, not the fixture unit total. The total is small enough that every candidate pipe size passes it.
• On a stack, a building drain or a sewer, the fixture unit total governs, because there are enough fixtures upstream for the diversity assumption to be earned.

A tech who only ever sizes branches concludes that fixture units do not matter. A designer who only ever sizes mains concludes that the single-fixture check is pedantry. Both are generalizing from the end of the system they live on.

Continuous and semi-continuous flow has no diversity at all

A condensate line, a pump discharge, an ice machine, a commercial dishwasher's continuous drain, a floor drain taking a cooling blowdown: these do not have a probability of discharging. Their probability is one. The index's whole mechanism is absent.

So the model codes do not let you count them as fixtures. They convert flow to an equivalent number of fixture units at a stated rate, commonly on the order of two fixture units for each gallon per minute of continuous flow. Take the rate from the plumbing code your jurisdiction has adopted and amended, because the model documents differ and the jurisdictions amend. The reason the conversion rate feels harsh compared to a fixture's value is that it should: a continuous flow contributes its whole load to every peak, and a fixture contributes a fraction of its load to a few of them.

Semi-continuous is the case people get wrong. A pump that runs on a level switch is not continuous, but its discharge is not diverse either, because when it runs it runs at its rated flow and it cares nothing about what else in the building is happening. Treat it by its flow, not by its enclosure.

Worked example: a small mixed-use building

Ground floor is a coffee shop, three floors of apartments above, one stack. The question that arrives from the field is whether an added floor sink for a new espresso machine's drain needs a bigger branch.

Step one: pick the column and print the choice. The apartment fixtures come out of the private-use column. The coffee shop's public restroom fixtures come out of the public-use column. That choice happens before any arithmetic and it changes the totals, so write it on the calculation sheet rather than leaving it implicit.

Step two: take the values from the adopted table, not from memory. In the IPC as published, a 1.6 gallon-per-flush water closet for private use carries 3 drainage fixture units and the same closet in a public installation carries 4; the UPC does not always assign the same fixture the same number, so the governing value is whatever the adopted and amended local table says. the UPC's table does not always assign the same fixture the same number, so the governing value is whatever the adopted and amended local table says. Everything below uses the adopted table's values, whatever they are, and the method does not change if the integers do.

Step three: the new branch, where the total does not govern. The new floor sink serves one espresso machine drain and one dump sink. Fixture unit total on that branch: a small number, call it the sum of two low-value fixtures. Every pipe size that satisfies the minimum trap and branch size for those two fixtures also satisfies that total with room left. So the total is discarded here, and it is discarded deliberately: the branch is governed by the minimum size the largest fixture on it requires, plus the trap arm geometry, not by the sum. That is the correction the general rule demands and it is the line most calculation sheets never show.

Step four: the continuous flow, converted and printed as its own line. The espresso machine discharges a continuous drip during operation, and the shop's ice machine on the same branch discharges a continuous flow while it is making ice. Suppose the two together are stated at 3 gallons per minute of continuous flow by the equipment data. At a conversion rate of two fixture units per gallon per minute, that is 3 x 2 = 6 fixture units of continuous flow, which is more than the two fixtures themselves contributed. The branch total is now dominated by the term that has no diversity in it. That is not a quirk. That is the index working correctly: the continuous flow really is present at every peak.

Step five: the stack, where the total does govern. Now roll the branch into the stack. The stack carries the three apartment floors plus the shop. There, the fixture count is large enough that diversity is earned, the totals are meaningful, and the added 6 units of continuous flow move the stack's total by a small percentage rather than dominating it. Say the stack's existing total was 96 fixture units. Adding 6 puts it at 102, a rise of 6.25 percent, and the question becomes only whether 102 crosses the table's threshold for the next stack size. If the threshold is well above 102, nothing changes. If the existing stack was already sitting three units under a threshold, this small addition crosses it, and that is the finding.

What a person who skipped the distinction concludes. They add all the fixture units, including the converted continuous flow, get a branch total, look it up in the stack table, and size the branch. It comes out adequate, because the branch total is tiny by stack standards, and they never check the minimum size the fixtures themselves require or the trap arm fall. The branch gets built undersized against the rule that actually governed it, and the fault shows up as a fixture that drains acceptably alone and unacceptably when the machine is running. Nothing in the arithmetic was wrong. The wrong table was consulted for that end of the system.

How to verify you got this right

• Say out loud which rule governed each pipe you sized, and for every pipe the answer should be one of: the fixture unit total, the minimum size for the largest fixture on it, or a geometry limit. If you cannot name which, you have not sized it, you have looked something up.
• Check that no continuous flow was counted as a fixture. Walk the equipment list, not the fixture list. Condensate, ice, pumps, blowdown, and water treatment backwash are the usual misses, and they are the ones that carry their full load into every peak.
• Check that you used one column consistently within one occupancy and that a mixed-use building did not silently blend two use patterns into one total.
• Sanity-check the direction. Adding fixtures raises the total and can only raise the required size. If your recalculation returns a smaller pipe than before, you changed a column or a code edition, not a load.

References

• The plumbing code as adopted and amended by the local jurisdiction, for the drainage fixture unit tables, the private and public use columns, and the continuous flow conversion rate; the model documents are the IPC (ICC) and the UPC (IAPMO), neither of which binds on its own
• Equipment manufacturer documentation for the stated continuous discharge rate of any appliance on the branch
• See related: How to Size a Branch From Fixture Units Without Guessing; Why Drainage Pipe Is Sized for Partial Flow and Not Full