The platform: sensors, measuring chamber, fluidics, liquid calibrants, thermostat, electronics and data output

Identify which combiline build is in front of you, trace a sample from the port to the waste bottle, and name the sensor, valve, pump or board responsible for each step.

One chassis, several builds

The Eschweiler combiline is a compact bench analyser built around one housing, one CPU board and one fluidic layout that the factory fits out in different ways. The service manual we work from (version 4.01, for analyser software revision 2.500) covers four builds. The BGA build measures pH, pCO2 and pO2. The ISE build measures electrolytes only: sodium, potassium, ionised calcium, chloride and lithium, depending on which sensors were ordered, together with pH. The BGA+E build combines the blood gas row with electrolytes. The meta build adds glucose and lactate sensors in a separate block with its own temperature control and a second roller pump. A haemoglobin sensor (Hb sensor type II) is an option on the blood gas builds.

Before you touch anything, find out which build you have, because it decides which calibrants are on the reagent wall, which valves exist, which tubing plan applies and which analog board is fitted. Three quick sources tell you. The analog board order number encodes the parameter set: the spare parts list has a separate ESW-ANALOG variant for every combination, for example blood gas only, blood gas with Hb, blood gas with potassium and sodium, and so on, with a separate family for meta builds. The sensor selection page under the standard data menu shows what the software expects to see. And the reagent wall itself shows it: a blood gas build carries BGA calibrant bags, an electrolyte build carries CAL 3 and CAL 4 bottles, and a combined build carries both.

Write down the build, the software revision shown on the power-up screen, the analog board variant and the list of sensors with their dates of installation. A large part of the faults you will meet on this analyser are sensor and consumable questions, and you cannot judge a sensor without knowing how old it is.

The sample path: from port to waste

The operator presents the sample at the sample port, either from a syringe or from a heparinised capillary. The port has an exchangeable insert; the parts list shows a red insert and a white one with a built-in steel cannula, and the red insert is also part of the service set, which tells you it is treated as a wear part. Behind the port sits the sample port valve PV, a pinch valve that is open when unpowered.

From the port the liquid enters the measuring channel inside the thermostat. On its way it passes the sample light barrier, an optical sensor around a short glass capillary. It tells the software whether there is liquid or air in the channel, whether whole blood has arrived (which is also the trigger for an Hb measurement) and when a calibrant has been drawn far enough. The sensors are screwed into the sensor holder in a fixed order that depends on the build; a build with fewer parameters has a shorter row. Take the order from the tubing plan for your build and from the analyser itself, not from memory. On meta builds the liquid continues through the metabolite block with the glucose and lactate sensors.

Movement is done by a peristaltic roller pump, called the reagent transport pump. It runs at a slow and a fast speed, and its tubing is one of the main wear items in the service set. A separate suction pump works together with an intermediate vessel and a dampness absorber for aspiration and emptying, and the spent liquid ends in the waste bottle. Which pump moves which liquid at each step is best traced on the tubing plan of your build. The input valve EV, the suction valve ABS, the ventilation valves VE1 (intermediate vessel) and VE2 (suction pump) and the rinse valve SP are all single-stage pinch valves acting on silicone tubing. SP is closed when unpowered; the others are open when unpowered. That detail matters when you troubleshoot: a pinch valve that has lost power or drive fails to the open position, except SP.

The calibrants do not use pinch valves. They are selected by a block of small SMC solenoid liquid valves mounted behind the sample port: one for each calibrant present on that build (BGA 1 to BGA 4, CAL 3, CAL 4) plus LV, which lets air in. These valves are closed when unpowered. Treat the exact route of each tube as version-specific and keep the tubing plan for your build at hand when you work on the fluidics.

The sensors and what is inside them

All sensors are of the screw-in, rebuildable type. A typical sensor consists of a casing with a screw cap, an inner sensor unit or conducting system, a membrane shell that slides over the tip, a fill solution and two small O-rings. The parts list sells each of these separately, and the sensor lifetime table gives a separate expected life for the inner part and for the membrane. That design is the key to this analyser: much of the routine work is changing membranes and fill solutions, not replacing whole sensors.

The pO2 sensor is an amperometric Clark type. A platinum cathode is held at a negative polarisation voltage generated on the analog board, and the current from oxygen reduction is what the board amplifies. The pCO2 sensor is a Severinghaus type: a pH-sensitive element behind a gas-permeable membrane shell with its own fill solution. The pH sensor is a glass electrode; it has no membrane shell in the parts list, and the consumables include an activating solution for it. The reference sensor closes the circuit for pH and the ion-selective sensors; it has its own reservoir, membrane shell, conducting system and a larger bottle of fill solution than the others.

The electrolyte sensors for potassium, calcium, chloride and lithium are ion-selective electrodes with exchangeable membrane shells. The sodium sensor is built differently: the parts list shows a conducting system and a casing but no separate membrane shell. The glucose and lactate sensors of the meta build are enzyme sensors with short lives counted in days or in number of samples, and they sit in a separate thermostatted block. The Hb sensor is optical: it combines reflection and transmission of infrared light through the whole blood in the channel and needs no calibration, only two sensor factors that are printed on a flag on its cable and entered once from the keyboard.

Calibration with liquids, not gas cylinders

Many blood gas analysers you know from Level 1 and Level 2 calibrate pO2 and pCO2 with gas mixtures from cylinders. The combiline described in this manual does not: the manual lists no gas supply parts, and the manufacturer's product listings say no external gas is needed. In this manual the blood gas calibrants are liquids with known gas values in sealed bags, BGA 1 and BGA 2 on a pure blood gas build, BGA 3 and BGA 4 on combined builds, labelled low and high. The bags hang on a bag adapter with a cannula that pierces the bag and a spring-loaded slide-bar. Two small light barriers inside the adapter detect whether each slide-bar is pressed, so the analyser knows a bag is in place. Electrolytes are calibrated with CAL 3 and CAL 4 from bottles, and the rinse solution is WASH (WASH 1 for BGA, WASH 2 for builds with electrolytes, WASH 2+M for meta). Later Eschweiler material describes a Cal-Pack cartridge system instead, so check which calibrant format the analyser in front of you uses.

Because blood gas values depend on atmospheric pressure, the analog board carries a barometric pressure sensor, and the barometer reading is part of every gas calibration. A wrong barometer reading shifts the calibration of both gas sensors, so it is one of the first things to check when pO2 and pCO2 are both biased in the same direction.

For the engineer, liquid calibration has two practical consequences. First, there is no regulator, gas humidifier or cylinder to service, and no high-pressure gas hazard from the analyser itself. Second, the quality of the gas calibration depends entirely on the calibrant bag or pack: its expiry, how it was stored and how long it has been on the analyser. Treat a bag change as a suspect event whenever blood gas calibration suddenly changes.

Temperature: one thermostat, two circuits

The sensors sit in a thermostat block held at 37 degrees Celsius. Heating is done by heating resistors in the block, regulated by a circuit on the power supply board ESW-MVTC using an NTC sensor in the thermostat. A temperature safety switch in the block cuts the heater if it overheats. The temperature the operator sees on the display comes from a different sensor and a different circuit: a PT100 element read by the analog board. Keep that split in mind. Regulation and indication can disagree, and each has its own adjustment.

A yellow LED below the reagent wall shows the regulation state. It is on continuously when the block is close to 37 degrees, blinks when the block is far from it, faster the larger the deviation, and gives only very short flashes when the NTC is faulty or the safety switch has opened. After switching on, the analyser runs a warm-up phase of about an hour, and functions such as calibration and measurement are locked until it ends. The meta block has its own temperature regulation with a Peltier element, so it can cool as well as heat.

Electronics, printer and data output

The electronics are few and easy to reach. A switching power supply PM110-45-2A gives the logic supply, the plus and minus analog rails and a 24 V rail for the printer, valves, suction pump and heater. The ESW-MVTC board on the inside of the rear wall drives the roller pumps at two speeds, carries the temperature regulation and has trimmers for the pump speeds and the displayed temperature. The ESW-ANALOG board behind the reagent wall holds the sensor amplifiers, the pO2 polarisation and zero circuit, the barometric sensor, the temperature indication circuit, the sample light barrier circuit and, where fitted, the Hb measuring unit.

The CPU board CL Z80ADC carries the processor, the program in EPROM, the RAM, a 12-bit analog-to-digital converter with a multiplexer, opto-isolated driver outputs for valves and motors, a real-time clock and a lithium coin cell that keeps settings and stored parameters. The filling level board CL FÜLLSTAND reads the WASH light barrier, the bag slide-bar barriers and the waste bottle sensor. A keyboard adapter with its own small controller scans the keypad; the LCD has a converter board and a contrast trimmer above the roller pump.

Results go to a built-in thermal printer and, at the same time, out of an RS-232 port that also has an SD card reader for storage. The port sends only. There is no query, no handshake with a laboratory system and no retransmission of stored results from the SD card. A LIS or middleware has to listen and capture each result as it is printed. Each record ends with three codes: the kind of measurement, a mask of sensors that exceeded their permitted measuring time (those values are printed with a question mark) and a mask of sensors that were active. Those masks are useful evidence when you are told that a value 'sometimes does not come through'.

On-site checklist

• Build (BGA, ISE, BGA+E, meta), Hb option and software revision recorded
• Analog board variant and order number noted
• Sensor list with installation dates and membrane change dates recorded
• Calibrant bags and bottles: type, lot and expiry checked against the build
• Tubing plan for this build available
• Temperature LED state observed after warm-up
• Data output path to LIS or middleware understood: one-way RS-232 only
• Age of the CPU lithium cell known from the service record

Check yourself