An emissions analyser draws exhaust through a probe and measures what is in it, and it zeroes itself on clean air before every test, which is a legal requirement. Put the probe in the tailpipe of one car in our yard with its engine off and its key on the bench, and the analyser reads a running engine.
I am an MOT tester at a small garage, and the emissions bench is the instrument I trust most, because it is checked more often than anything else in the building.
How it works.
The analyser zeroes on ambient air before each test, and once a day it is checked against a bottle of calibration gas of known concentration, and both of those are logged, because the test is a legal one and an analyser that drifts is an analyser that fails people who should pass.
The probe goes in the tailpipe with the engine running at idle and then at a fast idle, and the bench reads hydrocarbons, carbon monoxide, carbon dioxide and oxygen, and prints them.
An engine that is off produces nothing. The probe in a cold tailpipe reads ambient air, which is what the zero is.
One car reads a running engine with its engine off.
It is an old estate that belongs to a customer who has it tested with us every year, and it has been coming here longer than I have.
Probe in the tailpipe, engine cold, key on my bench, bonnet up so I can see the engine is still. The bench reads four hundred and twelve parts per million of hydrocarbons, a quarter percent carbon monoxide, and a carbon dioxide figure that is what a warm engine at idle gives.
That is not a leak or a residue. A tailpipe that was running an hour ago will give you a small hydrocarbon reading from what is left in the pipe, and it falls to nothing in a minute as the pump draws clean air through.
This does not fall. It holds, at the readings of an engine idling, for as long as the probe is in the pipe, and the car has been stood in the yard overnight.
What has been checked.
The analyser. Zeroed, calibrated against the bottle, and then swapped for the one from the other bay. Both read it.
The probe and the hose. Replaced.
The car. I have had it up on the ramp with the exhaust off, the whole system, from the manifold to the tailpipe, in three pieces on the floor, and I have put the probe into the tailpipe section lying on the floor, connected to nothing, with the car above it.
Four hundred and twelve.
The cylinder. In case the bench was somehow reading the calibration gas, I ran it with the cylinder disconnected. Same.
The temperature.
An analyser probe has a thermocouple in it, because the test wants the engine hot, and the reading is on the printout.
The tailpipe of that car, engine off, overnight cold, reads sixty degrees at the probe tip, and the tailpipe is cold to the hand at the same moment, and I have had a hand on it while the printout said sixty.
What is done about it.
The car passes. The readings are a pass, comfortably, and with the engine running the readings are exactly the same readings, which is the part I have stopped thinking about: it makes no difference to the bench whether that engine is running.
I have tested it every year for nine years. The certificate is correct because the car is fine, and a tester who put a fail on it because the analyser reads it with the engine off would be a tester who had failed a car for nothing.
What I do.
I test cars and I calibrate the bench.
The one thing I have changed is that I do the emissions test on that car with the customer in the waiting room rather than standing beside me, which is a thing he liked to do.
The reason is the second year, when he was standing beside me and the engine was off and the bench was reading four hundred and twelve, and the probe was in the pipe, and I heard the engine.
Not the car's engine. It was off. I could see the belt not turning. The sound of an engine at idle, from the tailpipe, out of the pipe the probe was in, the quiet thrum an exhaust makes with a warm engine on the other end of it, and the customer heard it too, and he looked at me and said, quite calmly, that it had always done that, and that his father had never been able to get anybody to listen either.