The Level 2 method for defibrillators: logs, readiness, the analyser and the five blocks
Collect self-test results, readiness state and event logs, place a defibrillator fault in power, charge, storage, delivery or monitoring with the cheapest test that separates them, and know what you may and may not change at your level, starting from stored-energy safety.
Where Level 1 stops and Level 2 starts
In Defibrillator and ECG Service — Level 1 you learned the device types, how to make the bench safe, how to test energy, charge time, sync and pacing with an analyser, how to keep batteries and pads in date and how to run a PM. Lesson 6 of that course gave the first-line method: the exact complaint, the event log, the accessories, then the device. Level 2 uses the same method on faults that survived that first round. The self-test passes but the unit 'did not shock' on the ward. It charges slowly only on battery. It reports a charge fault once a week. It delivers the right energy into the analyser but clinicians say pacing does not capture.
At this level you are expected to think in blocks, use the logs the device keeps, choose a test that separates two suspects, and stop at the boundary the manufacturer sets for your authorisation. You are not expected to repair the high-voltage section on the bench. Most defibrillator repairs that work in the field end with a battery, a cable, a set of paddles, a configuration change or a replaced module fitted by the manufacturer's procedure.
Stored energy comes first, every time
Before any method, the rule from Level 1 lesson 2 stands: treat every defibrillator as charged until you have proved otherwise by the manufacturer's procedure. A device that came in because 'it does not shock' is exactly the device in which the switching or dump path may have failed, so the capacitor may still hold its charge. The arithmetic explains the respect: the ZOLL R Series service manual describes a therapy capacitor of 100 microfarads rated at 2800 volts. Half of C times V squared at that rated voltage gives about 400 joules. That is our own upper-bound arithmetic, not a manufacturer figure, and the device does not charge that high in use, but even the energy of one real shock is lethal and it can sit there after the battery is out.
Manufacturers give a defined way to make the device safe before opening it. The Philips Efficia DFM100 service manual, for example, tells you to remove AC and battery, separate the case and discharge the therapy capacitor with a dedicated discharge tool held in place for a stated time; the Nihon Kohden AED-3100 manual warns that the HV capacitor and the pacing converter are both high-voltage units. Use the tool and the steps for that model, never an improvised screwdriver short, and do not trust a single step: if the procedure includes a check that the capacitor is discharged, do it. If you have no procedure, no tool or no training for that model, you do not open the HV section. That is a Level 2 skill too.
Shock testing is only ever into the analyser's load. Never into the air, never with paddles touching each other, never with a person in the circuit, and with the ECG leads disconnected from any patient. The Philips manual adds a point people forget: before an operational check or weekly shock test, confirm the device is not connected to a patient.
Five blocks and the order of a shock
Every defibrillator, from a public-access AED to a monitor-defibrillator with pacing, can be split into five blocks. Power: battery, mains supply, battery charger and power management. Charge: the HV converter that turns a few volts or tens of volts into thousands. Storage: the capacitor, its voltage monitoring, the bleed path and the internal dump or disarm load. Delivery: the switching stage (relays, SCRs or IGBTs depending on design), the waveform control, the impedance measurement, the therapy cable, paddles or pads. Monitoring and control: ECG from pads or leads, rhythm analysis, sync detection, the processor, buttons, display, speaker, printer and logs.
A shock is a sequence through those blocks, and the step where it stops tells you where to look. The device must power up and pass its start-up test; it must recognise a valid cable or pads; it must accept an energy and a charge request; the charger must reach the target voltage in time; the capacitor must hold it; the delivery stage must measure impedance and release the waveform; the device must record what happened. 'Will not charge', 'charges but energy is lost', 'charged but no shock', 'shock but wrong energy' and 'shock delivered but the log is wrong' are five different faults in five different places.
Self-tests, readiness and what they do not prove
Modern devices test themselves on a schedule and show the result on a readiness indicator. Read the schedule in the manual before you rely on it. The Efficia DFM100 runs short hourly tests of power supplies, battery charge, internal communication and temperature, a daily test that adds the clock battery, low-energy internal discharges, pacing, ECG and the printer, and a weekly test with a higher-energy internal discharge through the whole defibrillation circuit. The LIFEPAK 1000 runs a short test at every power-on and automatic tests daily and monthly while off, and stores the results in a test log.
Now read the limits, because the manuals state them honestly. Philips states that its automated tests do not check the therapy cables, paddles, buttons, audio or display; those are covered by the operational check run by a person. An automatic test also cannot see a battery that will collapse under the repeated load of a real resuscitation, a cable that breaks only when bent, or a configuration that leaves sync on after each shock. A green readiness indicator says the device passed what it tested, at the moment it tested it. It does not replace an analyser test, and it says nothing about the pads in the drawer.
Readiness also depends on conditions. An automatic self-test that runs only when a battery is installed, or that does not run if the device was left on, may not have run for weeks. Check the date of the last test in the log, not only its result.
Logs are your first witness
Before you charge anything, collect what the device remembers. Depending on the model that may be an event or code summary with every charge, shock, impedance and error during a use; a test log with self-test results; a hardware and software error log; battery data from a smart pack; and the operational check history. The DFM100 manual, for example, describes exporting hardware and software error logs, readiness and operational check logs and event summaries. Download or print them before you power-cycle, change the battery or update software, because some of these actions overwrite or reset the record. Note the device clock: an event log with a wrong time cannot be matched to the clinical record, and a failing clock battery is a fault in its own right.
Read a log for sequence. A charge timeout that follows a low battery warning by ten minutes is a power story. A shock aborted for high impedance on a device that delivers correctly into the analyser is a pads or patient-contact story. An internal discharge error on the weekly self-test, with normal clinical shocks, points into the storage and dump path. When the log is empty or the device has none, the complaint itself becomes your log: ask exactly which mode, which energy, pads or paddles, battery or mains, and what the screen said.
Tests that separate suspects
Choose the cheapest test that splits the remaining suspects in two. On mains, then on a known-good fully charged battery, then on the customer's battery: charge time across those three separates power from the charge circuit. A known-good therapy cable or pads simulator on the analyser separates the cable from the device. Delivered energy into the 50 ohm load, compared with the selected energy and with the energy the device itself reports, separates storage and delivery errors from measurement errors. The device's own operational check against your analyser check separates a device fault from a test set-up fault. Swapping one thing at a time, and writing down each result, turns an intermittent complaint into a pattern.
A defibrillator analyser is the instrument that makes this possible. Keep its calibration in date, use the adapter the manufacturer names for that cable, and record which analyser and adapter you used. A result from an uncalibrated analyser, or from a home-made load checked with an oscilloscope, can be useful for your own reasoning, but it is not evidence for returning a device to service.
What you may change, and what you may not
Within your authorisation and the manufacturer's documentation you may replace user-replaceable and field-replaceable parts, configure settings the clinical owner agrees to, install software released by the manufacturer for that device, and run the manufacturer's performance verification. You may not bypass a self-test, a pads-detection resistor, an interlock or an energy limit, enter service functions with codes obtained outside the manufacturer's channel, adjust energy calibration without the authorised procedure and equipment, or open and repair the HV section without the procedure and training. When the fault lies beyond your boundary, your job is a clear report of what you tested and excluded, so the next level starts where you stopped.
- Make safe by the manufacturer's procedure; treat the device as charged.
- Save logs, test history and clock state before any reset or update.
- Write the complaint as mode, energy, pads or paddles, battery or mains.
- Place the fault in power, charge, storage, delivery or monitoring.
- Pick the cheapest test that splits two suspects; change one thing at a time.
- Stop at your authorisation boundary and report what you excluded.
| Part | What it does | How it fails | |
|---|---|---|---|
| 01 | Battery and power management | Supplies the charger and the electronics | Collapses under charge load, wrong type, failed smart chip |
| 02 | Mains supply and battery charger | Runs the device and recharges the pack | No charging, overvoltage or overcurrent errors |
| 03 | HV charger (converter) | Raises battery voltage to the capacitor voltage | Slow or failed charge, charge timeout |
| 04 | Therapy capacitor and voltage monitor | Stores the energy and reports its voltage | Charge leaks away, capacity loss, wrong voltage reading |
| 05 | Dump or disarm load | Empties the capacitor when a charge is cancelled | Internal discharge too slow or not complete |
| 06 | Switching stage and waveform control | Connects the capacitor to the patient and shapes the pulse | No shock, aborted shock, wrong energy, relay drive error |
| 07 | Therapy cable, pads and paddles | Carry the shock and the pads ECG | Broken conductor, missing ID, worn paddles, expired pads |
| 08 | ECG, analysis and control | Rhythm analysis, sync, buttons, display, logs | Noisy ECG, missed R-waves, dead button, lost log |
It failed once on the ward and passed every test afterwards
A hospital defibrillator failed to work once in an emergency. Back in the workshop it passed its self-test, charged to 360 J and disarmed itself after the configured delay, and on its battery it delivered seven maximum-energy shocks before the engineer stopped. Colleagues who knew the model said it was 'temperamental about batteries', and that the ward did not always keep it charged. The exact cause was never proven. The recommendations were practical: a quarterly test with discharge into a load, replacement of the battery when it could no longer give the number of maximum shocks the instructions require, an energy check by the metrology service, and a clear rule on the ward for keeping the unit on charge.
A passed self-test after a failure in use proves little. Rebuild the event from the log and the user, test the battery under repeated load, and fix the routine that let the device go to a code half-ready.
Trust the self-test or bring an analyser?
Engineers looking after Primedic Delfi defibrillators discussed whether the internal self-test is enough or an external tester is needed. On that device the capacitor voltage is monitored in hardware and the self-test works through the energy levels in turn, so it is a reasonable check of function. The consensus was that it remains a basic check: it does not measure what actually leaves the device into a known load. Some engineers had built their own loads and watched the pulse on an oscilloscope, which helped their understanding, while a calibrated analyser remained the tool for a documented check.
Use the self-test as evidence of readiness, not of performance. A documented energy result needs a calibrated analyser and the manufacturer's adapter.
An RTC error at self-test
An Axion DKI-N-10 defibrillator reported an RTC error during its self-test, which most likely meant a flat battery for the real-time clock on the board. The engineer planned to fit a new cell. A colleague advised caution before any soldering on an undocumented board: read and save the contents of the memory chip with a programmer first, and study the board, because a slip could lose the device's stored settings. The plan was to fit the new cell on the correct pins without lifting the original memory contents. The case was not fully closed.
A clock error is a real fault: logs with wrong times are worthless. On undocumented boards, preserve the stored data before you touch anything.
On-site checklist
- Follow the model's make-safe procedure before any internal work; use the specified discharge tool.
- Export logs, test history and battery data before power-cycling or updating.
- Check the date of the last automatic self-test, not only its result.
- Test charge time on mains, on a known-good battery and on the customer's battery.
- Shock only into a calibrated analyser with the manufacturer's adapter.
- Record analyser, adapter, battery and cable used for each result.
- Write down what you excluded and where your authorisation ended.