The Level 2 method for sterilisers: records, alarm history, the curve phase by phase and test loads

Collect cycle records and alarm history, read a pressure-temperature curve phase by phase to place a steriliser fault, choose the test cycle or test load that separates two suspects, and know what you may change and who decides the machine is fit for use.

Where Level 1 ends

In Steriliser Service — Level 1 you learned the types of steriliser, how to make one safe, how steam, water and vacuum move through it, how a cycle is built, what the daily and periodic tests prove, and a basic fault-finding order: supplies first, then reproduce the fault, then open covers. Level 2 keeps that order. What changes is the kind of fault. Here the leak test fails again after a new door gasket, the Bowie-Dick fails once a week and nobody can say why, the load is wet although the drying vacuum looks normal, or the machine stops in the holding phase only in summer.

The expensive mistake at this level is believing the alarm. A steriliser controller can only report what its sensors see, and it reports it in the phase where a limit was crossed, not in the place where the fault lives. 'Vacuum not reached' may be a worn pump, but it may equally be warm seal water, a leaking clamp, a valve that does not close, a pressure transducer that reads wrong or a condenser coil with a pinhole. 'Heating time exceeded' may be a heater, or a generator short of water, or a steam valve that opens half way, or air still in the chamber. Replace the part the alarm named and you may have spent the hospital's money and kept the fault.

Collect the evidence before you touch anything

A modern steriliser keeps more evidence than most engineers use. Large CSSD machines usually store an alarm or event list with time stamps, cycle numbers and the phase in which the alarm occurred, and they print or store a record of every cycle with pressure and temperature at fixed intervals or as a curve. Many store the last tests, a service counter for the pump, door and valves, and sometimes the raw values of every sensor. Benchtop units often keep a smaller list, and older or locally built machines may have nothing but a chart recorder, a dial gauge and the operator's logbook. Whatever exists is your history; save it before anything else.

Work from four sources. The failed cycle itself: programme, phase, the alarm text and the values at that moment. The recent good cycles of the same programme, which are your baseline. The daily and weekly test results for the last weeks: a leak test result that has crept upwards, a Bowie-Dick sheet that shows a faint centre before it fails outright. And the users: what load, what packaging, what changed. A new detergent, a different wrap, a move from cold to hot feed water, work on the building's water or steam plant, another engineer's visit last week - each of these has been the real cause of a 'machine' fault more than once.

Check the controller clock before you trust the order of events. A flat backup battery can reset the date and, on some machines, settings or counters, and it can produce strange alarms of its own on start-up. If events seem to happen in the wrong order, or a machine has 'forgotten' something after a power cut, think of the battery before the board.

• Photograph or export the alarm list and the failed cycle record before any reset.
• Pull two or three good records of the same programme for comparison.
• Read the last weeks of Bowie-Dick and leak test results, not only today's.
• Ask about load, packaging, water, steam, building work and earlier service.
• Note supplies at the moment of the fault: power, water pressure and temperature, steam, compressed air, drain.

Read the curve phase by phase

A cycle record is a pressure and temperature curve, even when it is printed as a column of numbers. Read it the way you would read an ECG: phase by phase, comparing each part with what a healthy machine draws. Level 1 lesson 4 gave you the phases; here is what faults look like in each of them.

In conditioning, look at how long each vacuum pulse takes to reach its set point and how deep it goes. Pulses that take longer every week point at the pump, its seal water or its condenser. A first pulse that is slow while later ones are normal suggests water lying in the chamber or pipework at the start, or a valve that sticks after standing. Pulses that reach their set point quickly but a Bowie-Dick that fails make you suspect the measurement: a pressure transducer that reads lower than the true pressure lets the controller believe it reached a vacuum it never reached.

In heating, look at the come-up time and its shape. A smooth but slow rise points at the steam supply: generator power, water level, a steam valve or a supply pressure that sags. A rise that stalls below the band and then recovers can be air or condensate collecting at the drain sensor, often a blocked drain strainer.

In holding, compare the measured temperature with the saturation temperature for the measured pressure, as you learned at Level 1. Temperature below saturation means air or non-condensable gas, or a temperature sensor reading low. Temperature above saturation means superheat or a pressure sensor reading low. A saw-tooth that hunts across the band points at a steam valve or its control, a slow drift points at supply or sensor. If the machine has two temperature channels, look at whether they disagree and whether the disagreement grows with temperature; a connection that opens when hot shows exactly that.

In drying and air admission, look at how deep the drying vacuum goes and how long the chamber takes to return to atmosphere. A shallow drying vacuum with a good conditioning vacuum suggests steam leaking in through a valve during drying. A slow return to atmosphere points at the air inlet valve or a blocked sterile air filter.

Choose the cheapest test that separates two suspects

At Level 2 you are rarely short of suspects; you are short of evidence. Each test answers one question, so pick the one whose answer removes the most suspects for the least time and money. The routine tests are diagnostic tools as well as release tests. A leak test after a warm-up asks whether the system is tight. A Bowie-Dick asks whether air is removed from a porous load. A hollow-load challenge asks the same for lumens. An empty-chamber test cycle asks whether the machine can run its programme at all. A full production-like load asks whether the machine can cope with real mass and real packaging.

Combine them logically. Leak test passes but Bowie-Dick fails: air is not getting in, so look at air removal, measurement and non-condensable gases. Empty chamber runs well but full loads fail: think of steam capacity, drainage and load. Test cycles pass on site but the fault appears at the customer's busy hours: think of supplies shared with the building. Many machines also offer a manual or service mode in which you can drive valves and the generator one at a time; a generator that holds pressure in manual mode but not in automatic tells you to look at the control loop, its sensor and its wiring rather than the heaters.

Change one thing at a time and keep a written log of each test with its cycle number. When a fault is intermittent, set up to catch it: ask the users to note every occurrence with time, programme and load, leave a data logger or a reference sensor in place if the site allows, and look for what the bad cycles share.

What you may and may not change

A steriliser is a validated process. Its cycle parameters, its sensor calibration and the software that runs it are part of what was proven during validation. That gives you a simple rule: you may repair and replace parts with the specified items and restore the machine to the state in which it was validated, but you do not change the process itself to make a fault disappear. Lengthening a holding time, lowering a vacuum set point, widening an alarm limit, moving a sensor offset so that a reading 'looks right' or switching off an air detector all change the process, and every one of them can turn a visible fault into an invisible non-sterile load.

Some settings are legitimately adjustable during service, such as an altitude or atmospheric pressure setting, the time and date, or a printer option. Even then, record the value before and after, follow the manufacturer's procedure for your authorised level, and ask whether the change could affect a validated cycle. Sensor calibration is adjusted only with a calibrated reference and the manufacturer's method, never against the front panel gauge. Service passwords, software tools and access levels you were not given are not yours to use, and you never copy configuration or calibration files from another machine. When a job reaches that line, stop and hand over a clear package: records, test results, your measurements, what you excluded and what you suspect.

Who releases the steriliser

You do not release loads, and on your own you do not decide that a steriliser may be used for patients' instruments. You make a technical statement: what was found, what was done, which tests were run with which results, and whether the machine is technically fit for use, fit with restrictions or not fit. The decision to put it back into routine use belongs to the people the site has made responsible for decontamination, typically the CSSD manager or the person in charge of the steriliser, often with the infection-control team and, where requalification is needed, with the person who performs it. In Kazakhstan and Russia the sanitary rules also require bacteriological control after repair; agree with the department who arranges it.

Make that division visible. A steriliser that failed a test is labelled out of use until the tests after repair have passed, and your report says plainly what remains open. If the users want to run 'just one urgent load' on a machine that failed its Bowie-Dick, your job is to state in writing that it is not fit for porous loads, not to argue about the theatre list. Lesson 8 returns to return-to-service decisions after major repairs.

On-site checklist

• Save the alarm list, the failed record and two good records of the same programme.
• Check the controller clock and backup battery when events or settings look wrong.
• Measure supplies at the moment of the fault, not only at rest.
• Read the curve phase by phase against the baseline.
• Choose the test that removes the most suspects: leak test, Bowie-Dick, hollow-load test, empty or full load.
• Change one thing at a time and log every test cycle number.
• Never change cycle parameters, alarm limits or sensor offsets to clear a fault.
• Write a technical statement and leave the release decision to the responsible person.

Check yourself