It is 10:40 at night at a tanker depot outside Pune. The dispatcher has just spoken to a driver whose speech was a shade slow and who took two attempts to sign the trip sheet. She thinks he has been drinking. She has no way to prove it, no written rule that says she can hold the vehicle on suspicion alone, and a customer expecting delivery by six.
That is the real problem drink driving detection solves. Not curiosity about whether someone drank, but the gap between a supervisor's suspicion and a defensible decision to stop a vehicle leaving the yard. Close that gap and the dispatcher acts in ninety seconds. Leave it open and she either waves the truck out or grounds a driver on a hunch, and both outcomes eventually cost someone.
Drink driving detection happens at three points, and only one of them is yours
Alcohol on the road gets detected in three places. Each has a different owner and a very different cost.
At the gate or depot, before the vehicle moves. This is screening, done by the employer. It is the cheapest point of detection because nothing has happened yet. A reading here ends in a driver going home, not in a court file.
On the road, by the police. Detection here is enforcement. By this stage the vehicle is loaded, in traffic, and carrying your name on the tailboard.
After a crash. Detection at this point is forensic. Blood samples, hospital records, an insurer's assessor, and questions to the operator about what checks were in place.
Operators only control the first one. Everything written below is about making that first point work well enough that the other two rarely matter.
Behavioural signs are a trigger, not a test
Supervisors are usually asked to spot impairment by eye. Smell of alcohol, unsteady gait, bloodshot eyes, slurred or over-careful speech, irritability, delayed responses to simple instructions.
These signs are useful and should be trained into every shift in-charge. They are also unreliable on their own. Fatigue after a long haul, fever, medication and dehydration all mimic them. A driver who has drunk steadily for years may show almost nothing at a level that still puts him over the legal limit.
So treat behaviour as the trigger for a measurement, never as the finding. The finding has to be a number, because a number is the only thing that survives a driver saying "I had nothing, she just doesn't like me."
How the measurement actually works
Alcohol crosses from blood into the air in the lungs at a predictable ratio. A breath analyser samples deep lung air and reports the alcohol concentration, which is why a breath reading stands in for a blood reading in routine checks.
The sensor matters more than anything else on the spec sheet:
Sensor type How it behaves Best use
Semiconductor Low cost, responds to other vapours, drifts faster Personal awareness, low-volume screening
Fuel cell Responds to ethanol specifically, stable over many tests Workplace, fleet, security gate, police-grade use
Format matters too. Passive or contactless units let a supervisor sweep a queue of drivers in seconds without mouthpieces, which is what you want at a depot gate at shift change. Mouthpiece-based units give a specific, recordable value and are what you use to confirm anyone the passive sweep flags. Most serious operations run both: fast sweep, then confirmation.
The third thing, and the one that quietly destroys cases, is calibration. A device that has not been calibrated in over a year produces readings that a driver's union, a lawyer or a labour officer can dismiss in one sentence. Calibration records are part of the evidence, not a maintenance chore.
What the Indian limit actually says
The prescribed limit under the Motor Vehicles Act is 30 mg of alcohol per 100 ml of blood. That is a blood alcohol concentration, not a quantity of drink. The widely repeated idea that "30 ml of alcohol is allowed" is a misreading, and it has put a lot of drivers on the wrong side of a check post.
Thirty milligrams per 100 ml is a low threshold by international standards. For many people a single large peg reaches or crosses it, depending on body weight, food, and how recently it was consumed. There is no safe arithmetic that lets a driver calculate his way to the wheel. We have covered the limit and its consequences in detail in Drink and Drive in India: The Law, the Limit and Who Carries the Risk.
Also worth noting: many transport operators, mining contractors and hazardous-goods clients set their own limit at zero for on-duty drivers. That is a contractual standard, enforced by the employer, and it is entirely separate from what the police measure on the road.
How the police detect it
Enforcement teams set up checks at junctions, exits from commercial districts, and known routes on weekend nights and around festivals. A driver is asked to blow into a handheld breath analyser. If the reading is over the prescribed limit, the process moves to a formal record, and the driver may be taken for a blood test where the reading is contested or where he refuses to provide breath.
Refusal is not a way out. Declining a lawful test carries consequences of its own, and it is recorded.
For a commercial vehicle, the detection does not stop at the driver. Questions follow about the operator, the trip authorisation and whatever checks existed before the vehicle left. That exposure is set out in Driving Under the Influence of Alcohol: The Law, the Test and the Employer's Exposure.
In-vehicle detection: useful, but not a substitute
There is steady research and product development around alcohol interlocks that lock the ignition until a clean breath sample is given, and around passive systems that sense alcohol in the cabin air or through the driver's skin at the steering wheel.
Interlocks are real and are used in some commercial and school transport fleets. Passive cabin sensing is largely still emerging technology and is not something an Indian fleet can buy and standardise on today. Neither replaces a gate check, because an interlock tells you nothing about the driver who has already left the depot and drinks at a dhaba at two in the afternoon.
Random mid-shift and return-to-depot checks are what close that loop, and they need a handheld device in a supervisor's pocket.
A screening policy that survives a challenge
Detection without a written policy creates disputes instead of settling them. The policy needs to answer, in plain language:
• Who is covered. All drivers, or all staff who operate equipment, including contractors and temporary hands.
• When testing happens. Pre-shift for every driver, random selection during the shift, after any incident, and on reasonable suspicion.
• What the threshold is. Usually zero on duty for drivers.
• What a positive reading triggers. Immediate stand-down from driving, a confirmation test on a calibrated unit, a witnessed record, and a stated disciplinary path.
• How readings are recorded. Date, time, device serial number, reading, tester's name, driver's acknowledgement.
• Who is accountable. Named roles, not "the management".
Apply it to everyone, including the depot manager and the owner's nephew. A policy applied selectively is worse than no policy, because it hands any challenged driver a ready-made argument.
Who carries the consequence
When a loaded vehicle causes harm and the driver is found to have been drinking, the driver faces the charge. The operator faces the insurer, the client, the regulator and, in many contracts, a termination clause. The supervisor who let the vehicle out faces questions she cannot answer without a record.
The scope of that accountability for Indian operators is laid out in Driving Under the Influence: What Indian Operators Are Actually Accountable For, and it is broader than most fleet owners assume.
The practical next step
For a depot or transport yard, start with one fuel cell handheld per dispatch point and one passive unit for sweeping queues at shift change, plus a calibration schedule with dates in a diary. That is enough to move from suspicion to evidence on the same night.
If you are specifying units for a fleet, compare sensor type, test throughput and calibration interval across our range of alcohol testing machines before fixing a quantity, because the right format at the gate is not the right format for a roving supervisor.
The dispatcher in Pune does not need a debate about impairment. She needs a number, a rule, and the authority to act on both.