The Level 2 method: alarm history, the hydraulic diagram, two halves of the machine and test modes
Turn an alarm or a failed self-test into a located fault by reading the history, tracing the hydraulic diagram, splitting blood side from dialysate side and machine from supply, and using test modes without ever weakening a protective system.
Where Level 2 starts
Dialysis Equipment Service Level 1 gave you the map: the water plant, the hydraulic module, the blood module, the cycler, the protective systems, planned maintenance and the basic alarm families. Level 2 starts when a machine keeps failing after the obvious things have been done. The new concentrate is connected, the couplings have new O-rings, the filters are fresh, and the machine still fails its morning test or still alarms during treatment. At this point guessing becomes expensive. Every wrong part costs money, every extra hour costs a treatment slot, and every reboot that 'fixes' the fault hides the evidence you needed.
The method in this course is the same one taught in the Fault-Finding Method course, applied to a machine that carries a patient's blood on one side and a few hundred litres of prepared fluid on the other. Define the symptom exactly. Decide which half of the machine and which stage of the flow path it belongs to. Test the cheapest and most likely explanation first. Change one thing at a time. Prove the repair with the same test that showed the fault, then with the tests that prove the machine is safe. Before any of this, the patient question from Level 1 still applies: the patient has been disconnected by the clinical staff, the machine is out of use and labelled, and nobody moves it to another station to see whether it works there.
The alarm history is evidence
Most machines keep more than they show on the main screen. There is usually a list of recent alarms with date and time; the Nipro Surdial service documentation, for example, describes a maintenance screen that lists the latest alarms with their date, time and number. Self-tests store which step failed: the Fresenius 4008 records an error number for each step of its T1 test, so a failed test tells you not only 'failed' but which check and which kind of failure. Treatment and cleaning programmes have their own messages. Read all of it before you clear anything, and photograph it if the machine cannot export it.
Read the history as a sequence, not as a pile. Ask which event came first, which events follow from it, and what the machine was doing at that moment: preparation, treatment, bypass, rinse, heat disinfection, chemical disinfection. A conductivity alarm that appears only after a chemical disinfection points somewhere different from one that appears halfway through a session. A string of different valve errors on consecutive days is one problem with several faces, not several problems. Errors that appear on several machines at the same hour belong to the water, the concentrate supply or the mains, not to any one machine.
Check that the time stamps can be trusted. After a long power cut or a board replacement the clock may be wrong, and the order of events on the screen may not be the order in which they happened. Ask the nurses what they saw and did, in what order, and what changed recently: a new concentrate batch, new disposables, work on the water plant, a new cleaning agent for the casing. A five-minute conversation is still the cheapest test in dialysis.
Reading the hydraulic diagram of the machine in front of you
The flow diagram is the most useful page in a dialysis technical manual, but only if it is the diagram of this machine. Machines of the same model are built in different configurations. The 4008 technical manual, for example, gives separate diagrams for basic and advanced hydraulics and shows optional parts for online fluid, dry bicarbonate cartridges and central concentrate delivery. Before you trace anything, confirm the configuration and software version of the machine and pick the matching diagram.
Trace the diagram in each operating mode, because the same valve plays a different role in each. In treatment the bypass valve is closed and the dialyser valves open; in bypass the fluid goes around the dialyser; in rinse and disinfection some valves are forced open, recirculation paths are used and the concentrate inlets draw water or disinfectant. Mark on a copy where the measuring points are: the places where the manufacturer tells you to connect a pressure gauge, a conductivity meter or a cylinder. Then draw your own simplified block diagram on one page: water in, heating, degassing, mixing, measurement, balancing, dialyser, return, drain. When a fault appears, place it on that page first. Is it before the balancing system or after it? Inside the closed loop or outside it? On the side of the fresh fluid or the used fluid? That one question often halves the search.
Two halves and two borders
A haemodialysis machine has two halves that meet at the dialyser: the blood side (blood pump, pressure transducers, air detector, venous clamp, heparin pump) and the dialysate side (hydraulics, balancing, conductivity, temperature, blood leak detector). Many symptoms cross the border. A transmembrane pressure alarm uses the venous pressure from the blood side and the dialysate pressure from the hydraulics. A blood leak alarm sits on the dialysate side but is triggered by what happens in the dialyser. Your first job is to decide which half really owns the fault.
There are simple ways to split them. Run the dialysate side alone in a rinse or test programme with no blood lines fitted. Join the two dialyser couplings with a short hose, as described in Level 1, to take the dialyser and the external lines out of the hydraulic circuit. Test the blood side with a test line and saline, never with a patient. The second border is between the machine and what feeds it: the water loop, the concentrate canisters or central supply, the drain, the mains. Compare with a neighbouring machine on the same loop. If two machines on the same water fail in the same way, look at the water before you open either of them.
Test modes and live values
A self-test proves that the alarm chain works. It does not prove that the measurement is accurate. The 4008 T1 test, for example, checks the conductivity and temperature alarms by shifting the measured signal electronically and confirming that the alarm reacts; the cell and its calibration can still be wrong while the test passes. Level 1 showed a case where exactly that happened. Accuracy is proved only by an independent reference instrument at the dialyser couplings or at the point the manufacturer specifies.
Service and diagnostic modes go further. They show live values of sensors and states of switches, and some let you drive outputs: open a valve, run a pump, switch the heater. The 4008 diagnostics programme reads and writes the inputs and outputs of both processors; the Surdial maintenance screens show live values such as blood leak sensor readings, the pressures of the closed-circuit leak test and the temperatures at several points. Live values let you compare the control sensor with the protective sensor, watch a pressure while a valve switches, or see whether a level switch changes state. Use these modes only with the access and training the manufacturer gives you, only with no patient connected, and leave every output in its normal state when you finish.
The protective systems are never bypassed
Every protective system on a dialysis machine exists because patients were harmed without it: air in the venous line, wrong conductivity, overheated fluid, uncontrolled fluid removal, unseen blood loss through a broken dialyser. At Level 2 you will be tempted, by a busy unit and a fault that is almost found, to make a protective system quiet for a while. Do not. That means no jumpers on a level or door switch, no forcing an output while a patient is connected, no configuration that lets the machine skip its safety self-tests (the 4008 manual itself warns that such a setting allows the automatic tests of the safety systems to be bypassed, and puts the responsibility on whoever asked for it), no widening of alarm limits to make a fault disappear, and no calibration to hide a drift whose cause you have not found.
If a protective system blocks treatment, it has done its job. Your job is to find out why, repair the cause, and prove by test that the protection works again before the machine goes back to a patient.
- Patient disconnected, machine out of use and labelled before you start.
- Read and photograph the whole history before clearing anything.
- Use the diagram that matches the configuration and software.
- Place the fault: blood side or dialysate side, machine or supply.
- Self-test checks alarms; a reference meter checks accuracy.
- Service modes only with authorisation, no patient, outputs restored.
| Part | What it does | How it fails | |
|---|---|---|---|
| 01 | Reference conductivity and temperature meter | Checks the fluid at the dialyser couplings | Out of calibration, wrong cell constant, air in the cell |
| 02 | Undamped pressure gauge | Reads hydraulic pressures at measuring points | Damped gauge hides pulsation; wrong adapter leaks |
| 03 | Graduated cylinder or electronic balance | Measures UF pump and balancing volumes | Read at the wrong temperature or with foam |
| 04 | Pressure reference for blood transducers | Verifies arterial and venous pressure readings | Leaking connection, wet protector filter |
| 05 | Short coupling bypass hose | Takes dialyser and lines out of the circuit | Worn ends leak and create the fault you seek |
| 06 | Disinfectant residue test strips | Confirm the agent is rinsed out | Expired strips, wrong strip for the agent |
| 07 | Electrical safety analyser | Earth, insulation and leakage after repair | Wrong applied-part setting, no previous results to compare |
| 08 | Manufacturer's service software or modes | Live values, outputs, calibration steps | Outputs left forced; used without authorisation |
The motor was wet. The question was why.
On a haemodialysis machine a pump motor lost its seal and water got inside. The discussion among engineers went straight to the motor: could it be opened, its rotor and stator cleaned, its bearings and gaskets replaced, or should it simply be swapped for a new one? One colleague described the rebuild; the owner leaned towards a new motor. Nobody in the discussion established where the water had come from, and the final outcome was not confirmed. That gap is the lesson. A motor in the hydraulic compartment does not fill with water by itself. A leaking fitting, a cracked housing, a seal on the pump head or condensation from above is still there after the motor is changed.
A failed part is evidence. Before ordering it, ask what made it fail and find that cause on the diagram and in the cabinet, or the new part will fail the same way.
Disconnect the sensor before you replace it
A steam sterilizer kept reporting a water quality sensor error. Before ordering anything, the engineer unplugged the sensor and checked that the input now saw an open circuit, effectively an infinitely high resistance, and that the error behaved as expected. That showed the input electronics were sound and pointed at the sensor and the water it sat in. The water tank was dirty and was being filled with poor water. The tank was cleaned, the unit was switched to purchased purified water, and the error did not return. The same logic applies to a conductivity or level sensor in a dialysis machine.
Removing or replacing the signal at the input splits the sensor from the electronics in one minute. Then ask what the fluid did to the sensor.
On-site checklist
- Confirm the patient is disconnected and the machine is labelled out of use.
- Photograph the alarm list, the self-test result and the programme in which the fault occurred.
- Ask what changed: concentrate batch, disposables, water plant work, cleaning agents.
- Find the hydraulic diagram that matches the configuration and software version.
- Split the fault: blood side or dialysate side; machine or water, concentrate, drain, mains.
- Verify accuracy with reference instruments, not with the self-test.
- Leave every service-mode output in its normal state and every protective system active.