How to Tell Whether a Closed Loop Still Has Its Treatment

Why this matters

The test kit is the cheap and reliable half of this job. The sample is the half that goes wrong. A reading taken from the wrong point, with the pump off, without a flush, three days after a chemical addition, is a genuine measurement of something that is not the loop, and it will be recorded as the loop's condition and acted on. Shops have drained and recharged perfectly healthy systems on the strength of a sample drawn from a boiler drain.

What follows is the record that makes a reading mean something, one field at a time, with what each field is protecting you from.

Before you open a sample point

Read the loop's pressure and temperature gauge first and write both down. Above roughly 140 F the water leaving a cracked valve flashes to steam and scalds; either isolate the section and let it cool or draw through a sample cooler, discharging into a container placed so the stream cannot reach you or anyone walking past. There is no position that is reliably out of the way of an unexpected jet, which is why the temperature check comes before the valve and not after.

If the sample point requires removing a fitting, plug or strainer basket, relieve the loop pressure at that section and isolate the circulator before any fastener moves. That is mechanical isolation and stored energy, which is 29 CFR 1910.147, and the stored energy here is a pressurised column of hot water that does not care that the pump is off.

Where the system you are sampling is a domestic water system or feeds a cooling tower, flushing a long stagnant leg raises an aerosol carrying whatever grew in it, and those sites should be operating under a written water management program of the kind ASHRAE Standard 188 describes. On a sealed hydronic loop that is not the risk; the risks there are temperature and stored pressure, handled above.

Read Section 8 of the safety data sheet for any test reagent that goes on your skin, and use the glove class it names. 29 CFR 1910.1200 is what puts that sheet in your hands.

The record

One page per system per sample. Blanks are not neutral; each one removes a specific conclusion you would otherwise be able to draw.

The four fields that decide whether the number is real

Sample point. A drain at a low point, a boiler drain, a plugged tee on a dead riser: these exchange with the circulating loop by diffusion, not by flow. They hold whatever was in the loop the last time water actually moved through them, plus every solid that has settled into them since. Sample from a valve on the circulator's discharge or another point that is unambiguously in the flow path.

Pump run time. After any partial refill, the loop is not homogeneous. Fresh makeup sits where it entered until circulation mixes it. Run the pump for a spell before drawing, and if you are sampling a system that has been off for the season, treat the first reading as provisional.

Flush volume. Push out several times the dead volume held between the loop and the valve seat, which for a valve on a short nipple is a small fraction of a litre and for a valve at the end of a capped stub can be several. Flush until the discharge runs at loop temperature. Temperature is the tell: cold fluid out of a valve on a hot loop means you are still emptying the fitting.

Days since dosing. A sample drawn from the feeder's own leg an hour after a charge reads the charge. Give the loop time to mix and note the interval on the record either way.

The three fields that turn a reading into a diagnosis

Reserve alone is a number. Reserve read against makeup, conductivity and iron is a finding.

• Makeup and conductivity together tell you whether the inhibitor left with water or was consumed in place. Water loss removes every dissolved species proportionally, so conductivity falls with it; consumption removes only the reactive species and leaves conductivity flat.
• Iron tells you what the metal experienced. Reserve at target with iron climbing means the chemical is in the water and not on the surface, which points at a deposit shielding the metal rather than at a dosing error.
• Work done since last sample is the field that stops a false alarm. Fresh metal exposed by a repair consumes inhibitor hard for the first weeks and then settles.

A record that changed its own answer

First draw. Hydronic heating loop, volume estimated at 250 gallons. Sampled from the boiler drain, pump off, no flush, sample came out at room temperature. Appearance black and turbid with visible fines. Nitrite 180 ppm against a supplier target of 1,200 and a published minimum of 700. pH 7.1 against a package target near 9.

On that page the loop is finished. The tech's next line was going to be drain, flush and recharge, which is most of a day plus the fluid cost.

Second draw, same visit. Circulator run for twenty minutes. Sample valve on the pump discharge, flushed until the discharge came out at loop temperature. Appearance slightly hazy, no settled fines. Nitrite 940 ppm. pH 8.9. Conductivity 3,100. Dissolved iron 0.4 mg/L. Makeup meter up 4 gallons since the previous quarter. Last field on the page: circulator replaced eight weeks ago.

The two draws differ by a factor of more than five on reserve because they sampled two different bodies of water. The first one measured a stub that had been out of the flow path since the last fill, holding old fluid and the solids that had been dropping into it for years. It was an accurate reading of a dead leg.

Now read the corrected page as a whole. Makeup of 4 gallons against 250 is 1.6% of the loop replaced, so dilution can only account for a 1.6% fall in anything. Reserve fell from a target of 1,200 to 940, a drop of 260, which against 1,200 is 21.7%. Almost all of that is consumption rather than dilution, and the last field explains it: a circulator swap eight weeks ago exposed fresh metal and a new gasket face, which is exactly the innocent consumption that settles on its own.

So the action is a top up to target, not a recharge. Deficit is 260 ppm across 250 gallons, and 250 gallons is about 946 litres, so 260 mg/L across 946 litres is about 0.246 kg of inhibitor ion. At a product that is 10% active by weight that is about 2.5 kg of product. Cross check that against the supplier's own dosing table before adding anything; if the table and your arithmetic disagree, one of the two numbers you used is wrong and it is usually the active fraction.

Compare the two paths. The first page pointed at a drain, a flush, a recharge and a day of labour. The second page pointed at a top up and a note to resample at the normal interval. The entire difference was twenty minutes of pump run and a flush.

What an empty record means

A loop with black water, no meter, no product name and no history is untreated until a sample proves otherwise. The absence of complaints is not evidence of protection, because the failure this article is about produces no symptom at all until a pinhole appears.

Start such a system with a full page rather than a reserve reading: fit or read a makeup meter, get a volume by measurement if the loop is being drained anyway, name the fluid from whatever is on the feeder or from a lab identification, and take the first reading as a baseline rather than as a verdict. One reading cannot distinguish a loop that has been at 400 ppm for a decade from one that fell there last month, and those two need different responses.

How to verify you drew a good sample

Three checks, all on site, all before you leave.

The discharge came out at loop temperature, which says you reached circulating fluid rather than the fitting. The appearance is consistent with the last record for that system, and a step change in clarity or colour with no work done in between means either the sample point changed or the loop did. And the reading is plausible against the makeup meter: a large fall in reserve with the meter unmoved is a real finding worth chasing, but a large fall in reserve with a large jump in the meter is arithmetic you can do on the spot, and if the two do not reconcile, re draw before you dose.

References

• 29 CFR 1910.147, The Control of Hazardous Energy, for isolating the circulator and relieving stored pressure before removing any fitting from a pressurised loop
• 29 CFR 1910.1200, Hazard Communication, for the safety data sheet whose Section 8 sets glove class for test reagents and treatment products
• ASHRAE Standard 188, Legionellosis: Risk Management for Building Water Systems, where the system sampled is a domestic water system or serves a cooling tower
• Fluid and inhibitor supplier documentation for target concentration, published minimum, active fraction and approved test method
• See related: Why an Inhibitor Runs Out and What Happens After It Does, which owns the dilution versus consumption separation applied here