EPA 608 Leak-Rate Triggers: Using the 20% Commercial Refrigeration Threshold to Route Multi-Site Repair Dispatch Before the 30-Day Repair Clock

This guide shows how to route multi-site commercial refrigeration repair dispatch using the EPA 608 20% annual leak-rate threshold. It delivers the calculation and dispatch rule that starts the 30-day repair clock before vendor urgency takes over.

How the 20% trigger actually works

The trigger that matters under EPA 608 is not a smell, a hiss, or a technician's gut feel — it is a calculation. For each commercial refrigeration appliance, the annualized leak rate is computed as the pounds of refrigerant added over a rolling 12-month period, divided by the nameplate charge in pounds, multiplied by 100. When that figure reaches 20% for a commercial refrigeration circuit, the regulatory clock starts. Everything else — pressure loss, oil residue, vendor urgency — is diagnostic signal, not the trigger.

The clock is per-appliance, not per-site, and this is where multi-site portfolios get into trouble. A supermarket with four racks and a dozen display cases does not have one leak rate; it has one per circuit. A site that looks clean on a portfolio-wide refrigerant purchase total can still have a single circuit that crossed 20% months ago. Conversely, a site that burns through cylinders of make-up refrigerant may have every individual appliance sitting below the threshold because each has a large nameplate charge relative to its losses. Aggregate purchasing data cannot substitute for the per-appliance math.

Once a circuit's calculated rate reaches 20%, the owner or operator must repair the leak within 30 days of discovery, or document a retrofit or retirement plan if repair is not the chosen path. Some appliance categories may fall outside this commercial refrigeration requirement, so classification matters before you treat the threshold as binding. The check is simple: confirm each appliance's end-use category and its coverage under the current EPA 608 rules — at eCFR or with your compliance counsel — before assuming the 30-day clock applies.

The verification method is a records pull, not a site visit. For each appliance, retrieve the nameplate charge, sum the make-up refrigerant logged over the trailing 12 months, and run the division. If your service tickets record refrigerant added per visit per appliance, this takes minutes. If they record only "refrigerant" as a line item without quantities, your first dispatch is not a repair call — it is a records-reconstruction task, because you cannot defensibly route on a threshold you cannot compute.

Practical rule for portfolio managers: maintain a per-appliance leak-rate worksheet, updated as make-up charges are logged, and treat the 20% figure as the dispatch trigger. Route on the calculated number, and let the 30-day repair window — measured from the date the exceedance is discovered and documented, as detailed in the worked example later in this guide — drive urgency. That keeps the decision tied to the regulatory mechanism rather than to whoever shouts loudest on the phone.

What the evidence does and does not show

Before you trust any number in this guide, understand where it came from. The source material behind this piece — vendor and industry pages from xccold.com, coldstarcn.en.made-in-china.com, completerefridge.com.au, thecaterer.com, theengineeringmindset.com, and news.google.com — contains no EPA regulatory text, no vendor service-level agreement, and no leak-rate statistics. That means every figure presented here is method, not measured fact. The regulatory thresholds and repair windows referenced in this guide reflect long-standing EPA 608 requirements, but you must verify them against the current EPA regulations before using them to justify a dispatch decision or a compliance file entry.

What the available sources does show is vendor-side capability, not compliance authority. Xinqichen (xccold.com) presents itself as a Qingdao-based commercial refrigeration manufacturer with more than fifteen years in the industry, holding CE and RoHS certifications, producing open chillers, island freezers, meat and deli showcases, beverage coolers, and cold storage rooms. Coldstar (via made-in-china.com) lists a multi-deck open chiller for supermarket display use. Complete Refrigeration & Air in Perth advertises qualified technicians and emergency service. None of these pages states a leak-rate threshold, a repair deadline, or a contractual response time. A manufacturer's certification mark tells you about equipment construction; it says nothing about whether a circuit in your portfolio has crossed a regulatory trigger.

The practical check for a multi-site operator is therefore a two-column exercise. Column one: regulatory fact, pulled directly from the current EPA 608 rules at eCFR or your compliance counsel — never from a vendor brochure, a trade article, or this guide. Column two: vendor capability, taken from the vendor's own published materials and your signed agreement. Dispatch decisions are defensible only when the trigger in column one is documented independently of the vendor who benefits from the trip.

Watch for the category error the available sources invites. Trade coverage — such as thecaterer.com's report on Foster and Gamko's product range, or news.google.com items on supermarket refrigeration companies and industry partnerships — is marketing-adjacent information. It can inform equipment selection and vendor shortlisting. It cannot establish what the law requires of you or when a repair clock starts. If a vendor tells you a site "needs a tech out today for compliance," ask for the regulation citation and the appliance-level calculation behind the claim, in writing.

The engineering content in the available sources, such as theengineeringmindset.com's explanation of multi-ejector refrigerant behavior, is useful for understanding how circuits work — not for deciding when to send a truck. Keep those knowledge sources separate from your compliance sources, and your dispatch log will hold up to audit even when your vendors' urgency does not match the regulation.

Routing options compared: threshold vs urgency

When you manage refrigeration across multiple sites, you eventually have to pick a dispatch policy, and there are really only three on the table. The first is calculated-threshold routing: you compute each circuit's annualized leak rate from nameplate charge and logged make-up refrigerant, and you dispatch same-week only when that number reaches the 20% level. Below it, the circuit rides the normal maintenance route. The second is sensory or urgency routing: a store manager smells something, a technician sees oil residue or unusual icing, and someone calls a vendor. The third is blanket same-day dispatch for every reported symptom, everywhere, all the time. Only one of these three holds up under scrutiny.

Option A wins because it is the only policy that produces a number. A calculated leak rate is auditable: you can show the nameplate charge, the make-up log, the arithmetic, and the resulting percentage for every circuit in the portfolio. That number maps directly onto the regulatory clock described earlier in this guide, so when a circuit crosses the threshold, your dispatch date and your compliance deadline are on the same page. No other policy gives you that traceability.

Option B loses for a simple reason: none of its inputs are comparable to 20%. Oil residue, odor, and icing are real field observations, but they are not quantities. You cannot put "smells like refrigerant" next to a threshold and know whether you are at 8% or 25% of annual charge loss. A sensory report might be a trivial top-off or a major breach, and the policy gives you no way to tell the difference before the truck rolls. It also gives you nothing to defend yourself with if a regulator or an auditor asks why you dispatched — or why you didn't.

Option C, blanket same-day dispatch, is Option B with the brakes removed. It treats every anecdote as an emergency, which burns vendor trip charges on circuits that a calculation would have left on the routine route. It also inverts the priority problem: if everything is urgent, nothing is, and the circuits that genuinely crossed the threshold compete for attention with every icing complaint on the list.

The practical rule: keep sensory reports in your intake system, but never let them trigger a dispatch on their own. Log the observation, pull the circuit's make-up records, run the calculation, and let the number decide. If the computed rate is at or above 20%, dispatch same-week and the 30-day clock is yours to manage. If it is below, the circuit waits for its scheduled route — and you have the paperwork to prove that decision was the defensible one.

Costs and numbers that matter

Every dispatch decision in a multi-site portfolio is a bet, and the two ways to lose money are not symmetric. Route a truck to a circuit that turns out to be leaking well under the threshold, and you have paid a vendor trip charge for nothing. Route no truck to a circuit that is actually at or above the threshold, and you have started a regulatory clock you did not know was running. The first error costs you a line item on your vendor invoice; the second costs you a compliance failure. That asymmetry is why the calculation, not the phone call, should drive the dispatch.

What does a wasted trip actually cost? Use your own contracted rate card — your service agreement's trip charge, diagnostic labor rate, and any after-hours multiplier. This guide deliberately carries no dollar figures for vendor visits, because those numbers live in your vendor agreement, not in any public source, and quoting a generic figure would just invite you to skip the step of checking your own contract. Pull the rate card once, write the trip charge into your dispatch worksheet, and you have the price of the "dispatch in error" direction pinned down for the whole portfolio.

The other direction is harder to price. A missed repair window on a circuit that verified at or above the threshold is a regulatory exposure under EPA 608, and the penalty schedule is enforced per violation. Before you assign any dollar figure to that risk, verify the current penalty schedule directly in the EPA 608 regulations — enforcement amounts are adjusted over time, so treat any number you last saw years ago as stale. What you can state today, from the regulation's structure, is the shape of the exposure: a missed clock on a qualifying appliance is a documented, citable violation, not a soft cost.

Keep the units straight so the arithmetic stays auditable across sites. All refrigerant quantities in this guide are pounds (lb) of nameplate charge and make-up refrigerant, and all time windows are calendar days. When your logs record make-up in ounces or kilograms, convert to pounds before computing the leak rate — a mixed-unit worksheet is the most common way a portfolio-wide calculation goes quietly wrong.

One practical check closes the loop: before each dispatch cycle, confirm that the rate card figure and the penalty figure in your worksheet are dated. A stale trip charge overstates the cost of caution; a stale penalty figure understates the cost of inaction. Both errors push you toward the wrong routing policy, and both are fixed by a two-line date stamp at the top of the worksheet.

What the evidence does NOT establish

The 20% calculation is only as good as its inputs, and there are three situations where the inputs simply do not exist. First, an appliance with no visible nameplate charge — common on older cases or units where the label has been painted over or removed — gives you no denominator, so no leak rate can be computed and the EPA 608 trigger cannot be established for that circuit. Second, an appliance with no 12-month make-up refrigerant log leaves you with no numerator; a technician's recollection of "we topped it off a couple of times" is not a logged quantity and will not survive scrutiny. Third, an appliance whose full charge falls below the regulatory threshold for its appliance class is outside the leak-rate framework entirely — the rule does not apply, no matter how much refrigerant it consumes.

A fourth break case is subtler and catches portfolios right after service work. A circuit that was just repaired and recharged has a fresh full charge, but if you carry the pre-repair make-up quantities into the rolling window, the old losses will be divided against a charge the appliance no longer holds — falsely pushing the calculated rate over 20%. The fix is procedural: reset the 12-month window at the documented repair date, record the post-repair full charge, and start the make-up log from zero. Without that reset, you will dispatch vendors to circuits that are legally compliant, and you will have no paper trail explaining why.

Where the rule still wins is the ordinary case: a circuit with a visible nameplate charge and a clean 12-month log. There the calculation is trivial arithmetic — pounds added divided by full charge, annualized — and the routing decision follows mechanically. Either the verified rate clears the threshold and the 30-day repair clock starts, or it does not and you keep monitoring. That mechanical quality is the point: it removes the argument with the vendor who insists his "gut says this one's leaking bad" justifies an emergency call.

One limitation must be stated plainly: the manufacturer and industry pages behind this guide — xccold.com, completerefridge.com.au, and thecaterer.com — provide no empirical leak-rate distribution across any portfolio.

Nothing in the source material tells you what fraction of circuits typically exceed 20%, how leak rates cluster by equipment type, or how quickly rates drift after a recharge. Every figure here is regulatory method, not measured field data.

The practical check before any dispatch, then: confirm nameplate charge is legible, confirm the make-up log covers a full 12 months, confirm the appliance class is covered, and confirm no mid-window repair reset is pending. If all four hold, compute and route on the number. If any one fails, the calculation cannot be performed — and the defensible answer is to fix the record-keeping gap first, not to dispatch on urgency.

One circuit, one dispatch decision

Take one hypothetical circuit and run it end to end: an open multi-deck chiller on the sales floor of a supermarket in your multi-site portfolio. Suppose the nameplate shows a full charge of 12.0 lb of R-448A, and your refrigerant log shows 2.6 lb of make-up added to that circuit over the last 12 months. Those inputs — nameplate charge, logged make-up, discovery date, and site/circuit identity — are all you need to turn a vague "the chiller seems low" into a defensible dispatch decision. The figures below are illustrative arithmetic, not measured data from any source in this guide.

Checkpoint 1 is the annualized rate itself. Divide make-up by nameplate: 2.6 ÷ 12.0 = 0.2167, or 21.7% per year. That clears the 20% threshold for commercial refrigeration equipment, so the 30-day repair clock is active and this circuit routes to same-week vendor dispatch — not to the next routine maintenance visit, and not to a "monitor and re-check" queue. If the same math had produced, say, a single-digit result, the correct routing would be continued logging, not a service call. The calculation, not the technician's urgency, makes the call.

Checkpoint 2 is the clock. Count 30 calendar days forward from your documented discovery date to get the repair deadline — for a discovery date early in a 31-day month, that deadline typically lands in the first days of the following month, so compute it rather than estimating. Here is the detail that trips up multi-site operators: the clock starts on the discovery date, not the dispatch date and not the vendor's confirmation date. If your work order system stamps only the dispatch date, you have no record of when the regulatory window actually opened. Before the vendor rolls, make sure the dispatch record shows both dates explicitly.

Copy this worksheet into your dispatch log for every circuit, every quarter:

Two disciplines make the worksheet trustworthy. First, the make-up figure must come from a logged refrigerant ledger — cylinders issued, top-offs recorded per circuit — not from a vendor's verbal estimate, because the division is only as good as the numerator. Second, run the calculation per circuit, not per site: a store with four circuits has four independent leak rates and four independent clocks, and averaging them can hide a single circuit that crosses the threshold. One circuit, one calculation, one dispatch decision — that is the whole method.

Decision rules for multi-site dispatch

Once a circuit's annualized leak rate is verified, the dispatch decision collapses into five if/then rules. Apply them in order, and do not skip a rule because a site manager is impatient or a vendor is already in the parking lot.

Rule 1 — rate at or above 20%, discovery date recorded: dispatch the same week and open a 30-day clock entry in the refrigerant log. The entry should carry the appliance identifier, the discovery date, the computed rate, and the nameplate charge the rate was derived from. The clock runs per appliance, so two failing circuits at one store are two entries, not one.

Rule 2 — rate below 20%: route the work to the normal maintenance schedule. Do not consume emergency vendor capacity. A circuit leaking below the trigger is a planned repair, and treating it as an emergency is how portfolios burn through after-hours rates on jobs that never had a regulatory deadline attached.

Rule 3 — nameplate charge or 12-month make-up quantity missing: do not compute a rate and do not dispatch on a guess. Task the next scheduled site visit with capturing the nameplate and back-filling the log, then re-run the calculation before any dispatch decision. A rate built on an estimated charge is not a defensible trigger.

Rule 4 — rate cannot be computed but the circuit is empty or clearly failing: treat it as a service call under the maintenance contract, not as a 20% trigger event. Log the make-up refrigerant added, because that quantity is what lets you compute the rate on the next pass.

Rule 5 — vendor urgency conflicts with the computed rate: the calculation wins. A vendor's "it needs attention now" is not a regulatory trigger, and none of the vendor materials reviewed for this guide — from xccold.com, completerefridge.com.au, or the trade coverage cited above — states a service-level agreement that would convert urgency into a deadline. Dispatch only when a rule above says to.

Run the rules in sequence at every site review, and the dispatch queue becomes an output of arithmetic rather than of pressure. That is the whole point: the 20% calculation is the only input that starts a clock, so it should be the only input that starts a truck.

What to do next

Frequently Asked Questions

Can I dispatch a repair vendor based on pressure loss or oil residue alone?

No — dispatch decisions must rest on a verified leak-rate calculation, not on refrigerant smell, pressure loss, or vendor urgency.

What inputs do I need to compute the annualized leak rate for a circuit?

You calculate it from the appliance's nameplate charge and the logged make-up refrigerant added to that circuit.

How is the annualized leak rate formula actually structured?

The annualized leak rate equals the pounds of refrigerant added divided by the nameplate charge in pounds, multiplied by 100.

Does the repair clock start when a technician suspects a leak?

No — the regulatory clock starts only when the verified calculation meets the commercial refrigeration threshold for that circuit.

Should I calculate the leak rate per appliance or per site?

The calculation is done for each commercial refrigeration appliance, and each circuit's rate should be verified before dispatching any vendor.

What period of make-up refrigerant additions counts toward the calculation?

The pounds of refrigerant added over a rolling 12-month period are used as the numerator in the leak-rate calculation.

Quick answers

Also worth reading: Geospatial Priority Routing Reduces Dispatch MTTA for HVAC: Geospatial Priority Routing Reduces Dispatch · Heating and Electrical Vendor Response: 1,800 Seconds to Contact—Triage, Not Dispatch: Heating and Electrical Vendor Response: · 4-Hour SLA Response Clauses: Clock-Start Traps & Credit Recovery: 4-Hour SLA Response Clauses: Clock-Start