The auditor asks a simple question: may I see the calibration certificates for your breath analysers. The admin executive opens the bottom drawer of the store cabinet and finds four devices. Two are the same model bought at different times, one has a cracked mouthpiece port, one nobody recognises. Not a single alcohol level tester in that drawer has been recalibrated since the day it was unboxed, and no certificate exists for any of them.
That drawer is the real exposure in most Indian workplace screening programmes. Not the purchase decision. The purchase was made months ago and it was probably fine. The exposure is that the devices have quietly stopped agreeing with reality, and every reading recorded since then is now hard to defend.
This article is for the person who owns the devices after the box is thrown away: the safety lead, the HR manager, the transport in-charge, the site admin who was handed a unit and told to "do the checks".
An alcohol level tester is a consumable, not a fixed asset
A weighbridge, a fire extinguisher and a breath analyser have one thing in common. All three look fully functional right up to the moment they are needed, and all three fail quietly.
A breath analyser's sensing element reacts chemically with alcohol in the sample. Every test consumes a small part of its working life. Ambient dust, humidity, heat inside a vehicle cabin, and simple age all move the sensor away from the response it was set to at the factory. That movement is drift.
Drift is not dramatic. The unit still switches on, still counts down, still shows a number. It shows the wrong number, in a direction nobody holding it can detect by eye.
The two ways drift hurts you
Reading low. The device under-reports. A worker who should have been stopped walks onto the shop floor, or a driver takes a night run on the highway. Screening happened, records exist, and the programme still failed. The accountability for that outcome sits with the employer, not with the instrument.
Reading high. The device over-reports. A driver is stood down, loses a day's earnings, and disputes it. If you cannot produce a calibration certificate dated within the manufacturer's interval, the dispute is not really about the driver any more. It is about whether your entire testing record means anything.
The second failure is the one that ends programmes. One challenged result with no calibration paperwork behind it, and every previous result at that location becomes arguable.
Calibration intervals: what belongs in your policy
Calibration is the process of exposing the sensor to a known alcohol concentration and adjusting the device's response so it reads that concentration correctly. It is done with a certified reference gas or a wet bath simulator, by the manufacturer or an authorised service centre.
Manufacturers publish an interval. Follow whichever comes first, elapsed time or test count. Broad guidance for planning:
Situation Practical approach
Fuel cell unit, daily entry screening Calibrate on the manufacturer's stated interval, commonly six to twelve months, or at the stated test count
Fuel cell unit, occasional or incident-only use Time interval governs, because test count will never be reached
Semiconductor unit, any use Shorter interval; these sensors drift faster and are less suited to workplace programmes
Device dropped, wet, or left in a hot cabin Verify before next use, regardless of the schedule
Any disputed or contested result Verify immediately and record the outcome
Device returning from a long spell in storage Verify before it goes back into rotation
Set the interval once, in writing, inside the substance policy. "When it looks off" is not an interval. It pushes a technical judgement onto a security guard standing at a gate with a queue behind him.
Verification is not calibration
A verification check confirms the device still reads a known sample correctly. A calibration adjusts it so that it does. Verification is cheap and can be scheduled often; calibration is the corrective action that follows a failed verification or a due date.
Sites that run high test volumes benefit from verifying between calibrations. It catches drift early instead of discovering it during an incident investigation, which is the most expensive possible time to learn that a sensor has moved.
Sensor life: how many tests you actually get
A fuel cell sensor has a working life measured in tests and in years, and both matter. Heavy daily screening at a large plant exhausts test count first. A device kept in a school office or a family car reaches end of life on age, having done a fraction of the tests it was capable of.
Signs the sensor is near the end:
• Readings that vary noticeably between two tests on the same person minutes apart
• Slower response, longer stabilisation, or frequent aborted tests
• Failing verification repeatedly and needing calibration much sooner than the stated interval
• Increasing false triggers in a clean environment
• A unit that no longer holds a stable zero after a rest period
When a sensor reaches end of life, calibration cannot rescue it. The service centre will either replace the sensor module or advise replacing the unit. Budget for this. A screening programme has a running cost, and pretending otherwise is how organisations end up with one dead device and ninety people waiting to be cleared.
What a year without calibration does to a programme
Go back to that drawer twelve months after purchase. Three things have usually happened by then.
First, the readings have stopped being trusted internally. Supervisors know the unit gives odd results, so borderline readings get waved through. The screening has become theatre.
Second, the records have become a liability rather than a defence. A register of results from an uncalibrated instrument documents that you tested, and simultaneously documents that you cannot say what you measured.
Third, nobody owns it. The person who bought the device has moved roles. Purchase, calibration and daily use sit with three different people and no one holds the calendar.
That third failure causes the other two. Assign a name.
Build a calibration register before an audit asks for one
One sheet, kept where audits, client questionnaires and investigations can reach it. For each device:
• Make, model and serial number
• Date placed in service, and location or department
• Date of last calibration, by whom, and certificate reference
• Next calibration due date
• Cumulative test count, if the device reports it
• Verification checks performed since last calibration
• Repairs, drops, sensor replacements
• Current status: in service, awaiting calibration, withdrawn
Tie the due date to a reminder that reaches a person, not a shared inbox. Where a site runs several entry points or a fleet runs many depots, hold spare calibrated units so a device going for service never means a shift screens nobody.
The register does a second job. When a client, insurer or third-party auditor asks how your programme is controlled, a maintained register answers the question in one page. An assurance that "we do check people" does not.
The consumables everyone forgets
Calibration gets attention because it produces a certificate. The rest of the maintenance load produces none.
Mouthpieces. One per person, every time, in any workplace or event setting. Running short is how screening gets skipped. Order them as a standing consumable, not as an emergency purchase.
Batteries. A device with weak cells can abort tests or behave unpredictably. Keep spares wherever testing happens.
Storage. Not the dashboard of a truck in Nagpur in May, and not an unventilated cupboard next to solvents or sanitiser. Heat and vapour exposure both age sensors. Sanitiser is worth naming specifically: alcohol-based hand rub near the intake produces readings that have nothing to do with the person being tested.
Sample discipline. A fifteen-minute wait after any drink, mouthwash or medicated syrup. This is not device maintenance, but it prevents the false positives that get blamed on the device and then used to argue the whole programme is unreliable.
Questions owners actually ask
Can I test for alcohol at home, and how? Yes. A personal fuel cell unit gives a household a usable check before someone drives. Wait at least fifteen minutes after the last drink, rinse nothing with mouthwash, fit a clean mouthpiece and give a steady blow for the full countdown. Then apply the same ownership rules in miniature: the device needs calibration on the manufacturer's schedule, and a unit bought three years ago and never serviced is giving you reassurance rather than information. A family device used a few times a month will reach end of life on age, not on test count. Treat any result near a limit as a reason not to drive, not as permission to try.
Is 40% alcohol level high? That figure is bottle strength, not a breath or blood reading. Forty per cent alcohol by volume describes a spirit. Breath alcohol results are reported in far smaller units, such as milligrams per 100 millilitres of breath or blood, and no functioning tester will ever display 40% as a person's level. If a device shows something wildly implausible, treat it as a fault and pull it from service.
Which alcohol tester is the most accurate? The one that was calibrated on schedule. Sensor technology sets the ceiling, and fuel cell units hold their accuracy far better than semiconductor ones, which is why they are the sensible choice for any workplace, fleet or campus programme. But a well-chosen device a year past its calibration date is less trustworthy than a modest device serviced last month. Accuracy is a maintained state, not a specification printed on a box.
Do we need a printer or data logging? If results are ever likely to be contested, or if a client audit reviews your records, a unit that logs results with date and time removes an argument about what the register says. Handwritten registers are acceptable, but they are only as good as the discipline of the person filling them in.
Recalibrate or replace
Recalibrate when the unit is within its sensor life, passes verification after adjustment, and the physical condition is sound.
Replace when the sensor has reached its rated test count or age, when the housing or mouthpiece port is damaged, when the unit fails verification soon after calibration, or when the model no longer suits the volume you now run. A site that started with twenty tests a week and now screens two hundred has outgrown the device, not broken it.
And retire honestly. An unknown device of unknown history in a store drawer should not go back into use because it still switches on.
The decision in front of you
Pull out every alcohol level tester your organisation owns. Find the serial number, find the last calibration date, and write both down. If you cannot find a date, that device is out of service until it has one.
If that exercise leaves you with more gaps than certificates, the next step is a calibration schedule and a supplier who supports it. AlcoBreath supplies fuel cell breath analysers to Indian workplaces, fleets and institutions along with calibration and consumables, so due dates and mouthpiece supply stop being somebody's side task.
Screening only protects people while the instrument is telling the truth. Keeping it truthful is the part of the job that begins after the purchase order closes.