How to leak-test the analyser's sample path, where the probe tip belongs, why readings need the cover on and what a sudden temperature drop means.
Leak-test the sample path
A leak anywhere between the probe tip and the sensors lets air in and dilutes the sample, so a poorly burning appliance can look clean. Test the path before sampling, using the analyser's specified leak-test procedure:
- Seal the probe tip. On a tight system the pump can draw nothing and the flow falls to zero. Analysers with a built-in leak test report pass or fail.
- If the flow does not stop, look first at the water trap's plug and seals and the hose connections, and put the leak right before testing.
- Leak detection fluid will not find it: the pump pulls the sample under suction, so a leak draws air in rather than letting gas out.
- A fresh-air reading will not find it either: a leak in fresh air draws in more fresh air, so the oxygen looks normal.
Use the specified sampling position
The appliance's instructions determine the sampling point, insertion depth and sealing arrangements. If they call for a core-flow search, use the temperature indication to locate the hottest part of the stream and secure the probe there. Do not substitute that search for a prescribed fixed position or another sampling method. Keep the test point sealed as instructed so room air is not drawn into the sample.
Take readings with the cover on
The Worcester Greenstar 8000 Life instructions require its combustion chamber cover off for adjustments and on for every CO/CO2 reading; its flue integrity check uses the fitted case. Refit and secure the required covers before measuring. Follow the particular model's procedure: needing the cover off to adjust is not permission to take readings with the combustion circuit open to the room.
A sudden drop in flue temperature
Condensate on the thermocouple can make the indicated flue temperature fall suddenly even while the burner is steady. That does not mean the gas temperature has really fallen. Check the probe and repeat the measurement; treat the temperature and calculated efficiency as unreliable until the cause is resolved. Gas readings that have not changed do not show that nothing else has changed either.