How to Find Out How Much Makeup Water a System Is Really Taking

Why this matters

Makeup volume is the single most decision-changing number on a closed system and the one least likely to exist. Without it, a plugged exchanger, a blackened loop and a depleted inhibitor reserve are three separate complaints. With it, they are one finding.

There are three ways to get it, and the useful way to choose between them is not by accuracy but by which blind spot you can afford on this system. Each method is genuinely blind to something specific, and picking the one whose blind spot lines up with the actual failure is how a measurement turns into a wrong answer that everybody trusts.

Before you start

Do not isolate a makeup connection on a system whose low-water cutoff, pump seal cooling or expansion control depends on that feed, and do not leave one closed unattended. Closing it is a legitimate, supervised, time-boxed test with someone watching the equipment, not a valve you shut and come back to tomorrow.

Cutting into a line to install a meter means isolating the section, relieving pressure until a gauge you trust reads zero, and draining to a controlled point that can take the volume and the temperature. Confirm the drain before you open anything and stand clear of the discharge as you crack the valve; hot water flashing out of a fitting scalds through clothing.

Where isolating that section requires stopping a pump, lock out the stored energy under 29 CFR 1910.147, and handle the pump's electrical disconnect separately under 29 CFR 1910.333(b)(2), since 1910.147 excludes electric utilization installations at 1910.147(a)(1)(ii)(C).

A makeup connection is a cross-connection to the potable supply, so anything you install or modify there needs the backflow protection the local plumbing code requires for the treated fluid, in place before the line is put back in service. Treat a missing one as a public-health finding to report, not a detail.

Dosing a tracer means handling concentrated product. Take the glove class, eye and face protection and any respiratory control from Section 8 of that product's safety data sheet, which your employer must make available (29 CFR 1910.1200), and never introduce it into a feeder or sump that has held a hypochlorite product alongside anything acidic or ammonia-bearing, because those combinations release gas into the room.

The three methods, and what each cannot see

Pick by the blind spot. On a system where you suspect a hose is being used to top up, a meter is the wrong instrument and will read reassuringly low. On a system that leaks only when hot and pressurized, an isolation test run cold answers a question nobody asked. On a loop where the inhibitor is being consumed by active corrosion rather than merely diluted, a tracer method reads the loss as larger than it is.

Method 1: put a totalizer on the line

The permanent answer, and the one to install on any system you will see again. Read the totalizer at every visit and write it next to the date. Two readings and a subtraction give a rate; four give a trend, which is what actually catches a slow deterioration.

The step people skip is recording the reading even when nothing is wrong. A meter with no history answers nothing on the day you need it, and the first useful number arrives one visit after the first recorded one.

Report the result as system volumes per year, not gallons. That unit travels between a 200-gallon loop and a 5,000-gallon one, and it maps directly onto how fast a treatment reserve is being diluted.

Method 2: isolate, hold conditions, and measure what it takes to restore

For a system with no meter and a suspected active loss.

Close the makeup with the equipment watched, note the pressure and the system temperature, and hold the system at a steady operating temperature for the window. At the end, restore the original pressure using a measured volume from a calibrated container and a hand pump, and record what it took.

The reason temperature discipline is the whole method: water expands roughly 3 percent between a cold fill around 40 F and 180 F, so a 600-gallon water loop moves about 18 gallons in and out of its expansion tank on a single heat-up. That figure is for water. A glycol charge expands appreciably more, on the order of 5 to 6 percent for a mid-forties ethylene mix over the same range, so on a glycol loop take the expansion from the fluid manufacturer's data before you decide what the test can resolve. That is larger than most leaks you are hunting. At chilled-water temperatures the swing is far smaller and the test is correspondingly easier. Run the window at one steady temperature, or the expansion dominates the result and you will measure the thermostat.

The second trap is that pressure is not a linear stand-in for volume, because a diaphragm expansion tank's pressure response depends on its precharge and its acceptance volume. Calibrate it on the spot instead of assuming: add a measured volume, note the pressure rise, and use that as your scale for the rest of the test.

Method 3: dilution of a tracer

For a system with no meter, no convenient isolation point, and a loss you need to size rather than locate.

Use a species that is diluted and not otherwise removed. Conductivity works when the loop is chemically quiet. A dedicated inhibitor reserve works when the supplier confirms the product is not consumed at a meaningful rate. What does not work is any species that precipitates, is consumed by corrosion, or is added by a feeder during the window.

Take the starting reading, take the makeup water's reading, wait a window long enough for the change to exceed your measurement noise, and take the ending reading on the same instrument.

Worked example: sizing a loss by conductivity

A 600-gallon hydronic loop, no makeup meter, suspicion of a slow loss. All three conductivity readings on the same handheld, temperature recorded, loop circulating at a steady operating temperature throughout.

• Loop at the start: 1,400 uS/cm
• Makeup water: 350 uS/cm
• Loop 30 days later: 1,180 uS/cm

Set up the ratio. The loop is being diluted toward the makeup value, not toward zero, so the fraction of the original excess still remaining is (1,180 minus 350) divided by (1,400 minus 350), which is 830 / 1,050 = 0.790.

Convert to volume. For fresh water entering a well-mixed loop continuously, the remaining fraction is e raised to minus the ratio of makeup volume to system volume. The negative natural log of 0.790 is 0.235 system volumes. Multiply by 600 gallons: about 141 gallons over the 30 days, or 4.7 gallons a day.

Annualize and convert to the useful unit. 4.7 x 365 = about 1,716 gallons a year, which against a 600-gallon system is about 2.9 system volumes per year.

Now state the resolution honestly. If that 141 gallons had arrived as a single drain-and-refill rather than a continuous drip, the correct model is the straight-line one: the fraction lost is 1 minus 0.790, or 0.210, giving 126 gallons. The two models bracket the answer between about 126 and 141 gallons, a spread of roughly 11 percent. So this method reliably separates 5 gallons a day from 50. It does not separate 4.7 from 4.2, and quoting it to a tenth would be inventing precision the method does not have.

Check the direction before trusting the result. Conductivity fell, and dilution with lower-conductivity makeup is one thing that makes it fall. Name what would make it rise, because if the reading had gone up the method does not simply return zero - it returns nonsense. A rising loop conductivity means something is adding ionic load: a fresh chemical charge during the window, corrosion products entering solution, or a process leak into the loop. Any of those also masks dilution happening at the same time, so a flat reading is not proof of a tight loop either.

What flips the conclusion: if the makeup water's own conductivity varies seasonally, or the building blends two supplies, the 350 is not a constant and the calculation inherits that error directly, since it sits in both the numerator and the denominator. Take the makeup reading at the start and again at the end, and if they differ meaningfully, the tracer method is not available on that site.

Interpreting 2.9 system volumes a year. Set the action thresholds yourself and write them down so the crew uses the same ones. A workable starting point: flag any closed loop above about 10 percent of system volume per year, measured over a window of at least 30 days at steady operating temperature, and above roughly 50 percent per year stop treating it as a chemistry problem and start treating it as a loss to find. At 2.9 system volumes, which is 290 percent, this loop is far past the second threshold - the finding is the leak, and no dosing decision should be made until it is fixed.

The failure mode: the loop gets a fresh inhibitor charge to bring the reserve back to target, the readings on the way out are excellent, and the same conversation happens next year. Charging a loop that is passing three of itself annually is buying treatment for a drain.

How to verify the number you got

Cross-check with a second method whose blind spot is different. A tracer result of roughly 3 system volumes a year on a 600-gallon loop predicts around 4 to 5 gallons a day, which is about 4.7 gallons in 24 hours, roughly a quarter of the expansion volume on a loop this size. Run the isolation window a full 24 hours rather than a shift and convert the pressure fall to gallons using the tank calibration, because 8 hours of that loss is under 2 gallons and sits inside gauge noise. Two methods agreeing within their stated resolutions is a measurement; one method quoted to three digits is a guess with decimal places.

Then close the loop on it: install the meter, take the first reading, and write it in the equipment file with the date. Every argument this article exists to settle becomes a subtraction after that.

References

• 29 CFR 1910.147 for stored energy isolation on pumps and pressure vessels, and 29 CFR 1910.333(b)(2) for the electrical disconnect, which 1910.147 excludes at 1910.147(a)(1)(ii)(C)
• 29 CFR 1910.1200 (Hazard Communication) and Section 8 of the product safety data sheet for handling any tracer or inhibitor product
• Local plumbing code, which sets the backflow protection required at a makeup connection between a treated system and the potable supply
• Treatment supplier documentation for whether a specific inhibitor is diluted rather than consumed, which the tracer method depends on
• See related: Why Makeup Water Is the Variable That Decides Everything; The Difference Between an Open Loop and a Closed Loop Chemically; What Conductivity and Total Dissolved Solids Actually Measure