The Level 2 method for ophthalmic instruments: symptom, test eye, swap and isolate
Turn a clinician's complaint into a reproducible test on a test eye or target, place the fault in the instrument's block map, separate suspects by swapping and isolating one thing at a time, and know which adjustments are yours and which go back to the manufacturer.
Where Level 1 ends
Ophthalmic Equipment Service — Level 1 gave you the clinic, the instruments and their main assemblies, the light, laser and electrical safety basics, the cleaning rules, the test eyes and the first fault-finding habits. Lesson 6 of that course ended at a clear boundary: the cheap checks are done, the lamp is the right type, the windows are clean, the cables are seated, and the instrument still does not work properly. Level 2 starts exactly there.
The faults you meet now are the ones that survived the first visit. A refractometer that reads out of range on its own test eye after a repair. An OCT whose scans are weak on every patient although the lens looks clean. A microscope head with no power on one rail. A laser that fires but whose spot looks wrong. The method does not change: facts first, block next, cheap and likely causes before expensive ones, one change at a time. What changes is the depth: you now measure supplies, read logs, compare with a twin instrument and decide where your work ends.
Two other courses carry part of the load. The Electronics and Power Supplies course covers supply and board-level work in the workshop; Electrical Safety for Medical Equipment covers the tests after every intervention. This course refers to them rather than repeating them.
From the clinician's words to a test you can repeat
Ophthalmic complaints arrive in clinical language: 'the readings are strange', 'the picture is dark', 'it does not focus on the patient', 'the laser is weak'. Your first job is to translate that into a technical statement and a test. Ask which patient types, which eye, which mode, since when, after what. A refractometer that 'reads strange' only on small pupils or on patients with cataract may be behaving correctly; one that reads strange on its own test eye is not.
The test eye or test target is the ophthalmic engineer's phantom. Refractometers and keratometers are checked on the manufacturer's model eyes or reference spheres, tonometers on their verification devices, biometers and OCT units on the manufacturer's test eyes or targets, cameras and microscopes on resolution charts. A known target removes the patient from the equation: if the instrument is wrong on the target, the instrument is wrong. If it is right on the target and wrong on patients, look at the patient interface: alignment, fixation, chin rest, tear film, operator technique.
A second, independent instrument is the other free reference. Many clinics have two refractometers or a manual keratometer next to an automatic one. If two instruments disagree on the same patient, measure the test eye on both before you decide which one is wrong. Write down the values; a number with a date is evidence, a remembered impression is not.
- Photograph or export the error screen and any log before rebooting.
- Record mode, eye, patient type and the exact value or image the user calls wrong.
- Measure the manufacturer's test eye or target and write the result next to the expected value.
- Ask what changed: a move, a power cut, cleaning, a new PC program, earlier repair work.
The block map: where in the chain is the fault?
Almost every ophthalmic instrument is built from the same blocks: mains supply and internal power supplies; a light source with its driver (halogen, LED, infrared LED, superluminescent diode, laser); the optical path with lenses, mirrors, prisms, filters and windows; a sensor or camera; mechanics and motors (base, joystick, chin rest, focus and scan drives); a controller or PC with software and calibration data; and an output (display, printer, export, network). A treatment laser adds a laser source, a delivery system, an aiming beam and a safety chain.
Place the symptom on that map. No power at all or a dead rail is supply. A dark or uneven image with a working camera is light source or optics. A good image that the software will not measure points to alignment, calibration data or software. Wrong values on the test eye with a clean optical path point to the measuring module, its drive or its calibration. Correct measurements that never reach the printer or the database are output and data. This split is cheap and fast, and it decides which spares you carry on the second visit.
Remember that the block that reports a problem is not always the block that failed. A controller that cannot read a camera may be starved of 5 V. A software message about a motor may be a dusty position sensor. A mainboard that 'does not boot' may be fed by a failing supply. Downstream blocks notice upstream faults first.
Swap and isolate, one change at a time
Swapping is the strongest test you have, and eye clinics often give you the chance: two identical slit lamps, two cameras on one instrument, two refractometers of the same model, a spare power brick. Swap one item, test, and put it back if nothing changed. Swap only like for like: same model, same revision, same connector and the same supply rating. Never swap a module that carries unique calibration data, such as a measuring head, a camera calibrated to its optics or a PC disk with factory files, unless the manufacturer's procedure says it can be moved; a module that works in the other instrument may still measure wrongly there.
Isolation is the other half. Disconnect what is not needed for the test, with the instrument off: the printer, the foot pedal, the second display, the network, the second program on the shared PC. If the fault disappears, add things back one at a time. A great many intermittent faults on PC-based ophthalmic systems are conflicts between two applications or two devices on one USB bus, and isolation finds them in minutes.
Be careful that your own test does not create the symptom. A wrong light level on a camera, a test eye placed in the wrong holder, a manual exposure or gain set by hand, a cover left off so that room light reaches a sensor: all of these produce convincing faults on a healthy instrument. If a result looks impossible, check your test set-up before the instrument.
Measure the supply before condemning the board
The most expensive mistake at Level 2 is replacing a mainboard or a PC module that was only badly fed. Before you order a board, measure the supply rails it receives, at the board connector and under load, and compare them with the values printed on the board, the supply label or the manufacturer's documentation. Look for ripple and dips with an oscilloscope when the fault is intermittent. Check the mains input and the instrument's own fuse and filter as well.
Old instruments have ageing electrolytic capacitors in linear and switch-mode supplies. A supply can show the right voltage on a meter at idle and collapse when the motors or the lamp start. A dried capacitor in a standby circuit can make an instrument start only sometimes. The Electronics and Power Supplies course explains how to test such supplies safely on the bench; at this level the point is to suspect them early.
If a supply has failed, ask why before you power the repaired unit into the instrument. A supply that went over-voltage can take the boards behind it with it, and some of those boards carry firmware or calibration you cannot obtain. Where possible, test a repaired supply on a dummy load first.
What you may adjust and what goes back to the manufacturer
Your side of the line, when the manufacturer's documentation allows it: user-level and service-level cleaning of the surfaces you may touch, lamp and fuse replacement with the specified types, external adjustments described in the user or service documentation, mechanical repairs of bases, joysticks and chin rests, cable and connector repairs with correct parts, PC hardware with backups, and verification on the manufacturer's test eyes.
The manufacturer's or authorised side: internal optical alignment of measuring modules, calibration that needs factory targets or software you are not authorised to use, laser resonator, energy and aiming alignment, firmware and software installation beyond the documented user update, licences and options, programmed consumables such as lamp modules with a memory chip, and any repair of a module whose calibration lives inside it. You do not use passwords, service codes or tools you were not given, and you do not reset counters or edit chips to make a part look original.
When you reach the line, stop and hand over a package: the symptom in technical terms, the test eye results, your measurements, what you swapped and excluded, the logs, and what you suspect. That package often saves the manufacturer a visit and the hospital a wrong part. Lesson 8 returns to who decides when the instrument goes back to patients.
A femtosecond laser written off for a mainboard
A Ziemer femtosecond laser first lost its touchscreen and then stopped booting altogether. The authorised service suspected the mainboard and proposed ordering one from abroad at the hospital's cost, with no guarantee that it would cure the fault. Faced with that bill and uncertainty, the hospital wrote the laser off. When independent engineers later discussed the unit, the cause they named was the power supply, together with a low-cost multifunction module, rather than the mainboard; it was never confirmed on the bench. If they were right, a failing supply produced the symptoms of a dead computer: an unstable screen first, then no start-up.
Before a mainboard is condemned, measure the rails it receives under load. A diagnosis that starts with the most expensive part can end with a written-off instrument.
Two identical instruments, one bad main board
An engineer repaired an ophthalmic system made of two identical units. After he replaced failed supply capacitors and cleared a motor error, one display still flickered and showed stripes. Colleagues suggested the LVDS display driver on the main board or noise on the supply rails, and advised checking the rails, clock and sync lines with an oscilloscope and trying another screen. Because the system had a twin, he swapped the main boards between the two units. The stripes followed the board, which settled the question: the fault lived on that main board and not in the display, cable or supply.
A twin instrument is a test bench. Swap one like-for-like part, see whether the fault follows it, and you have localised it without guessing.
The test that made the fault
An engineer took apart an intraoral X-ray sensor (an RX2 visiograph) that produced smeared images, and suspected the cable near the sensor because moving it changed the picture. Continuity checks found every conductor intact, and a light test gave strange results. In the end it turned out that much of what he saw came from his own test: the exposure was controlled by hand and the sensor was overexposed, which produced artefacts that looked like a fault. The genuine defect was repaired along the way, but the misleading symptoms had been created by the test set-up.
Check your test conditions before the device. A wrong light level, gain or exposure on a healthy sensor produces very convincing faults; the same applies to ophthalmic cameras.
On-site checklist
- Translate the complaint into mode, eye, value and condition.
- Measure the manufacturer's test eye or target and record the result.
- Place the fault on the block map before opening anything.
- Measure supply rails under load before ordering boards.
- Swap only like for like, never modules with unique calibration unless documented.
- Isolate accessories and second programs one at a time.
- Stop at your authorised level and hand over a written package.