As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
A leak in still air produces a plume only a few millimetres wide close to the opening. At one inch per second a probe tip spends roughly four tenths of a second crossing a plume that size, which is about the time a sniffer needs to draw the sample along the probe and develop a full response. Double the speed and the dwell halves; quadruple it and the reading never reaches a level worth noticing before the tip has passed. The rule is not tradition, it is the ratio between how long you are in the gas and how long the instrument takes to answer.
Sweep speed decides how long the probe sits in the plume
Picture the gas leaving a pinhole in a brazed joint with no draught in the room. It emerges, spreads a little and then either sinks or rises according to its density, forming a narrow column close to the metal that widens with distance. Within a centimetre or two of the opening, the region where concentration is high enough for a sniffer to respond strongly is small, often no wider than a fingernail. That region is the target, and the probe tip either passes through it long enough to matter or it does not.
The arithmetic is simple division. One inch per second is about twenty five millimetres per second. Crossing a plume ten millimetres wide at that speed takes four tenths of a second. At fifty millimetres per second it takes two tenths; at a hundred, one tenth; at the speed most people actually move a probe when they are bored or in a hurry, considerably less. The sensor is not given a chance to disagree with the air it is in.
Nothing about this changes with how good the instrument is. A better element lowers the concentration at which it can respond, which widens the usable part of the plume slightly, but it does not alter the fact that the signal takes time to build and the probe is leaving. Speed is a setting on the technician, not on the instrument, and it is the one that most often decides whether a small leak is found.
Response time and dwell time are the two numbers that interact
Two delays sit between the gas arriving at the tip and the display changing. The first is transport: on a sniffer with a long gooseneck the sample has to travel the length of the probe to reach the element, and that takes a fixed time set by the flow rate and the internal volume. The second is the element itself, which reaches a defined fraction of its final reading in a stated time. The two add, and the total is what the dwell has to beat.
| Sweep speed | Time inside a ten millimetre plume | Fraction of a half second response achieved | Practical result |
|---|---|---|---|
| 25 mm per second, about one inch | 0.40 seconds | Most of it | A clear, repeatable indication |
| 50 mm per second | 0.20 seconds | Under half | A twitch that is easy to dismiss |
| 100 mm per second | 0.10 seconds | About a fifth | Usually invisible against the background |
| 200 mm per second | 0.05 seconds | About a tenth | The leak is not detected at all |
Published figures let you do this calculation for a specific instrument. The TOPTES PT520A+ states a reaction to combustible gas leaks in half a second, a detection span of 50 to 10,000 parts per million, a seventeen inch gooseneck, seven coloured bars with a 75 dB buzzer, and a preheat of thirty seconds before use. Half a second against a four tenths of a second dwell is already tight, which is the quantitative case for the slow sweep rather than a stylistic preference for one.
A plume is small, still and easily disturbed
Everything above assumes the plume is there to be found. It very often is not, because the smallest air movement destroys it. A ceiling fan, an open door, a running air handler or a technician walking past dilutes a plume to below the threshold of any instrument, and no sweep speed recovers what is no longer in the air. Shutting the space down and letting it settle for a few minutes is the single highest value step in a leak search, and it is skipped constantly.
Density decides where to look. Halogenated refrigerants are substantially heavier than air, so their plumes sink and pool below the leak, in trays, on ledges and along the bottom of an enclosure. Methane and hydrogen rise. Sweeping a joint along its top when the gas is going down is a thorough search of clean air, and it accounts for a surprising share of leaks declared absent.
There is an opposite error to sweeping too fast, and it is worth naming because the correction for one causes the other. Instruments with automatic zeroing continuously re reference themselves against what they regard as background. Park the probe over a leak and hold it there, and a slowly rising signal can be absorbed into that moving background until the display returns to nothing. The reliable technique is to pause briefly, withdraw into clean air, and return.
What the probe geometry adds to the arithmetic
Probe length buys access and costs time. A long flexible gooseneck reaches behind a condenser coil or under a compressor, which is where joints tend to be, but it also lengthens the path the sample has to travel before it reaches the element. The Aprvtio ALD-200 states a sixteen inch probe, a heated diode element covering A1, A2L and A3 refrigerants along with hydrogen and nitrogen tracer mixtures, sensitivity adjustable to 0.14 ounces per year and a colour display whose beep frequency rises as the leak intensifies. Its transport and response times are not stated in the listing bullets, which is a gap worth filling from the manual before deciding how slowly to sweep it.
Tip distance is the other geometric term. Concentration falls steeply with distance from the opening, so a probe held five millimetres away sits in a far richer sample than one held twenty five millimetres away, and the difference is larger than most sensitivity differences between instruments. Riding the tip along the metal, slowly, with the gooseneck shaped so the tip can stay in contact round a bend, is worth more than an upgrade.
Walking a joint at the right speed
Still the air. Choose the side of the joint the gas will travel to. Keep the tip within a few millimetres of the metal and move it at about an inch a second, which feels absurdly slow for the first minute and then stops feeling like anything. Pause where the instrument reacts, withdraw to clean air, and come back to confirm. Mark what you find rather than trusting memory, because the second pass is where false positives get eliminated.
A thermal camera is a useful complement here precisely because it has no dwell problem: it takes in a whole joint at once and shows temperature rather than gas, so a large escaping charge may appear as local cooling and a blocked or overloaded component as local heat. It will not see a small leak, since there is no thermal signature to see. The pocket class is covered in the FLIR One review and the Pro version review, with the resolution differences that matter set out in the Pro against Topdon comparison. When the joint is behind something solid rather than merely awkward, an inspection camera gets eyes on it before the probe goes anywhere near.
Frequently Asked Questions
Where does the one inch per second figure come from?
From the test procedures used to qualify leak detectors, which fix the probe speed so that results are reproducible between laboratories, and from manufacturer manuals that repeat the figure for field use. It is not a physical constant. It is a speed slow enough that a typical sensor and its sample path can respond fully while the tip is still inside the plume, which makes it a sensible default rather than a rule of nature.
Does a faster responding sensor let me sweep faster?
Partly, and less than the specification suggests. A faster element shortens one of the two delays in the chain, but the sample still has to travel from the tip along the probe to the sensing element, and that transport time does not shrink. Beyond that, sweeping faster also reduces the chance of the tip passing through the narrow part of the plume at all, which is a geometry problem rather than a timing one.
Should I hold the probe still over a suspect joint?
Pause, but do not park. A brief hold lets the reading develop and confirms what the sweep hinted at. Holding for a long time invites the automatic zero function to treat the steady elevated signal as the new background and null it out, after which the instrument shows nothing while sitting directly over a leak. Moving away into clean air and returning is the reliable way to confirm a find.
Where should I sweep relative to the joint?
Below it for refrigerants and other vapours denser than air, since the plume sinks and pools, and above it for methane and hydrogen, which rise. For a joint that could be leaking either kind, work all the way around it. Position matters more than speed in practice, because a probe swept perfectly at the wrong height passes through clean air the whole way.
Does air movement change the right speed?
It changes whether sweeping will work at all. Any draught thins and displaces the plume, so the concentration available at the tip collapses and no sweep speed recovers it. Shut doors, switch off fans and wait for the air to settle before searching. If the space cannot be stilled, a different method is needed, such as pressurising the system with a tracer gas or enclosing the suspect component.