When an optics experiment goes wrong, the useful question is not simply “What answer should I have obtained?” It is “What part of the method, measurement or model made that answer unreliable?” In a G3 Physics practical, a student may draw neat rays and still use an uncertain pin position, measure from the surface instead of the normal, mix two different trials, or describe a bright patch as a sharp image. This guide trains the more demanding skill: finding the first defect and repairing the investigation without inventing evidence.
The most reliable starting order is: understand the physical claim, identify what can actually be observed, isolate the variable that matters, set up apparatus safely, collect a traceable record, and evaluate only what the record permits. If the ray path is not stable, no amount of confident prose will make the measurement trustworthy. If the observations are sound but the calculation is wrong, repeating the experiment may be unnecessary. Diagnose before you act.
This volume advances beyond Vol 1089’s glass-block ray tracing, the earlier converging-lens screen-test owner and Vol 1096’s integrated optics mini-mock. Those owners teach core phenomena and combined performance. Here the new task is a fault-finding protocol for practical planning and evaluation: distinguish alignment errors, scale errors, model errors, observational limits and genuinely unexpected results. Parents and students moving from PSLE to G3 can start with the Sengkang G3 transition guide.
At a glance: the troubleshooting sequence
- Write the investigative question in one sentence.
- Identify independent, dependent and controlled variables.
- Sketch the intended ray path and mark the normal before measuring.
- Confirm what the equipment can resolve and what counts as an observation.
- Take repeat readings where practicable and record them without premature rounding.
- Separate measurement faults from incorrect physics explanations.
- Explain one feasible improvement and how it reduces the stated fault.
- Report the boundary of the conclusion, not a claim the apparatus cannot support.
What is official for the 2027 SEC
The 2027 G3 Physics subject is K323, not the earlier 6091 code. SEAB’s published K323 syllabus specifies Paper 1 multiple choice (40 marks, one hour, 30%), Paper 2 structured and free response (80 marks, one hour 45 minutes, 50%) and Paper 3 practical (40 marks, one hour 50 minutes, 20%). Paper 3 assesses planning, manipulation and observation, presentation, analysis, conclusions and evaluation. Its official syllabus lists refraction through glass blocks, total internal reflection and the determination of lens image properties among suggested familiar experimental techniques. The actual paper may use particular apparatus and questions; our tasks below are original learning exercises, not a forecast of equipment or an official specimen. Source: SEAB’s 2027 G3 directory and K323 Physics syllabus.
Do not treat a single home ray-tracing session as a substitute for supervised school practical training. Use approved low-power classroom light sources and teacher-supervised apparatus. Never stare into a laser beam, direct one at people or improvise with mains electrical equipment. The purpose of an exercise is accurate reasoning, not a risky demonstration.
A troubleshooting answer begins before the first measurement
Read the command word and isolate its target. ‘Determine’ asks for a result supported by method and data. ‘Explain’ needs a mechanism. ‘Suggest an improvement’ needs a specific weakness plus a practical change. Some students memorise a full practical description and pour it into every answer. This produces irrelevant steps while the actual weakness remains untouched. When a ray trace seems wrong, make a short diagnostic table: observed symptom, plausible cause, test of that cause, correction and evidence that the correction helped. A good diagnosis is testable. Saying ‘human error’ is not a testable diagnosis because it neither identifies the process nor predicts what better practice would change.
Keep the object of study separate from the apparatus
A glass block refracts light; a set square, ruler and pins help a person represent that light path. If a pencil line looks crooked, the medium has not necessarily behaved strangely. Check whether the optical boundary, the normal, the pin line or the ruler position caused the discrepancy. An investigation asks about light, but the data are mediated by equipment. The examiner cannot observe what you intended. They can assess only the measurements and reasoning you provide. Describe which component each suspected fault belongs to. This prevents the common mistake of treating a messy diagram as proof that the law itself has failed.
Specify a single changing quantity
A fair comparison needs a known independent variable. In a block experiment it may be the angle of incidence. Keep the material, block placement and drawing convention sufficiently stable to compare outcomes. In a converging-lens exercise, moving both object and screen without recording either position obscures which change produced a sharp image. Students should write what changes, what is measured and what remains controlled. Controls are not decorations in a method paragraph: each protects a particular comparison. If room lighting affects visibility, control or record it. If another quantity cannot remain fixed, say so and limit the inference accordingly.
Draw the normal at the correct place
Angles of incidence and refraction are measured from the normal, a line perpendicular to the boundary at the point where the ray meets it. A student measuring from the glass surface can produce a mathematically consistent but physically inappropriate table. Before touching the protractor, mark the exact boundary crossing and construct the perpendicular. Explain the convention in words beside the diagram. If two surfaces are parallel, do not reuse a normal at the wrong location merely because the diagram looks symmetrical. A small misplacement at the entry point travels through the entire angle calculation and can create a false trend.
Do not let line thickness become invisible uncertainty
A thick ray or pencil line may span several millimetres on paper. The protractor can appear precise to one degree while the underlying position of the ray is not. Use a sharp pencil, mark the centres of pin positions and extend constructed lines with a ruler. Keep original marks visible long enough to check their origin. When estimating angles from a diagram, state that it is a drawn measurement rather than a direct instrument reading if that matters. Avoid claiming hundredth-degree certainty because a calculator supplies many digits. Precision of arithmetic cannot rescue a poorly resolved physical mark.
Align pins so the geometry carries useful information
In a pin-alignment method, two pins placed very close together make the direction of the ray difficult to determine. A small lateral positional error then produces a large angular error. Place pins at a sensible separation as directed by the teacher, view without avoidable parallax and mark their actual positions before lifting the equipment. If the eye shifts, two pins that seemed coincident can appear separated. Record the line after checking alignment, not while relying on memory. The improvement is not simply ‘be careful’: it is ‘increase pin separation within the available apparatus space to reduce angular sensitivity to a fixed positional error’.
Distinguish a traced ray from a calculated ray
A traced ray is an observation-based reconstruction. A ray drawn only because it matches the expected answer is a theoretical illustration. Both have educational uses, but they must not be presented as the same evidence. Draw predicted paths in one labelled section of your notebook and actual positions in another. If a calculation predicts a refracted angle of about 19 degrees and the trace suggests 25 degrees, do not erase the latter. Check the normal, pins, block outline and protractor reading. The gap is the starting point of a diagnosis. Quietly replacing measurements with theory destroys the most informative part of the investigation.
Handle glass-block displacement sensibly
For a parallel-sided block, the emerging ray can be parallel to the incident ray yet laterally displaced. Students sometimes draw the ray straight through as if its change inside the block disappears entirely. Track both boundaries in sequence. At each boundary ask which medium the light is entering, and judge the bending relative to the normal. A lateral shift in a trace is not necessarily an error. By contrast, a ray appearing to emerge with a different direction may indicate a misdrawn block, incorrect normals or an observational fault. Name the expected geometry before deciding which reading needs checking.
Recognise the difference between a bright image and a sharp image
A patch of light on a screen is not sufficient evidence of a well-focused real image. Look for distinguishable edges and stable features. If the entire screen brightens but a recognisable object does not form, check the object, lens alignment, screen position and room lighting. Do not report ‘image formed at 18 cm’ simply because light was detected there. For a converging lens, the position and character of an image depend on where the object lies relative to the focal region. An experiment may require moving the screen until a recognisable sharp image forms, recording what moved, and acknowledging cases in which a screen image cannot be produced.
Move one component at a time when recovering focus
Troubleshooting becomes difficult when lens, screen and object all move at once. Use a controlled search. Keep the object steady, set the lens perpendicular to the principal axis, and move the screen deliberately while observing image sharpness. If no suitable real image appears, check whether the object lies inside the focal length rather than assuming the screen is defective. A real image can be caught on a screen; a virtual image cannot be captured there merely by moving the screen. A method that records unsuccessful positions as well as the best focus helps explain how the conclusion was reached.
Check axis and orientation before drawing image conclusions
A lens tilted relative to its intended axis may give blurred or distorted results that a student mislabels as a magnification effect. A screen positioned obliquely can also make one edge appear sharper than another. Align the object, lens centre and screen as appropriate, then judge whether the whole feature set can be focused. If an image is inverted, describe that observation rather than trying to turn the apparatus until the image looks upright. Orientation is a property to interpret. Alignment is a condition to control. Treating one as the other mixes experimental procedure with scientific explanation.
Use repeats for a reason, not for appearance
Repeating a measurement can reveal variability, but repeated copies of one wrong reading do not become reliable merely because their mean is stable. Take genuinely fresh alignments or independent readings where possible. Inspect whether the results spread plausibly. A single outlier might come from parallax, a wrong protractor scale or a slipped block, but the data alone do not prove which. Record the candidate cause and test it. If the physical setup changes between readings, a mean may hide more than it reveals. Explain what was repeated and why those repeats are comparable.
Find the first wrong link in a long solution
Long practical responses often contain a chain: observe pin locations, draw the ray, read the angle, calculate a sine ratio, interpret the constant, draw a conclusion. An answer can go wrong at any link. Repair the earliest defect. Recalculating a sine ratio from a mismeasured angle is not enough; return to the drawing. Likewise, redoing every physical measurement is inefficient when only the calculator was in the wrong angle mode. Use a fault tree with branches for observation, geometry, arithmetic, model and expression. This is a more disciplined response than guessing the error that appears nearest the final answer.
Separate systematic bias from random scatter
If every ray angle is displaced in the same direction, suspect a consistent construction or zero-setting bias. If repeated trials vary around a central value, consider alignment and reading variability. These are different problems and demand different repairs. A constant offset will not necessarily disappear by averaging more data. A random alignment fluctuation may be reduced by sensible repeats, although the size of the improvement depends on the conditions. Do not label a discrepancy systematic solely because it is inconvenient. Look for a pattern that supports the classification.
Describe improvements as cause-and-effect mechanisms
‘Use better apparatus’ is too vague. ‘Use a sharper pencil to reduce uncertainty in locating the ray centre’ is a specific improvement tied to the suspected fault. ‘Place pins farther apart’ is useful when close spacing has made direction sensitive to positional error. ‘Repeat with the block outline fixed’ helps establish whether movement during tracing caused the discrepancy. The most persuasive improvement names a controlled feature and a predicted effect. It also recognises limits: a smaller line thickness cannot correct a conceptual mistake about which angle is measured from the normal.
Make conclusions proportional to the measurements
A small classroom sample can support a limited finding, such as that a ray was refracted towards the normal on entering a denser optical medium under the tested conditions. It cannot by itself prove that every possible material behaves identically or that a device works flawlessly in the field. Distinguish direct observations, quantitative inferences and broader explanations. Use expressions such as ‘under these conditions’, ‘within measurement resolution’, or ‘the data are consistent with’ where needed. This is not evasive language. It tells the reader exactly how far the evidence travels.
Write for the marks actually requested
In a structured practical item, a clear labelled sketch, concise method, meaningful table and causal improvement are often more useful than a long essay about all of optics. Show the relevant point of incidence and normal. Identify the instrument and reading convention. State a valid comparison. When asked to evaluate, connect one limitation to its likely effect. A response should be traceable without requiring the marker to infer which ray or which reading you meant. Good examination technique respects the question’s scope while making the central reasoning visible.
Build transfer by changing the fault, not the words
After solving a pin-alignment error, test the same diagnostic habit on a lens problem. After correcting a wrong angle convention, try a case in which calculator mode is wrong but the diagram is sound. The aim is to recognise which type of evidence is defective. A new context is not automatically a new skill: the learner must explain why the repair transfers. Keep a small error ledger of problem, first wrong step, correction and a delayed changed-context retest. This turns each practical fault into a reusable decision rule rather than a memorised paragraph.
Sixteen original diagnostic cases with model reasoning
Try each case before reading its correction. The stated fault is an opportunity to practise diagnosis, not a predicted SEC question. When checking, mark the exact phrase that connects evidence to cause. Write a new response without looking at the model and change the conditions for a delayed retest.
Diagnostic 1: The block shifts after its outline is drawn
Case. The student traces the incident pins, moves the glass block to draw the emergent positions and obtains an implausible ray.
Reasoning and repair. The evidence chain is broken because the recorded boundary no longer matches the apparatus position. Replace the block on its traced outline and complete the sighting before removal. Do not silently shift the drawn border to make the final angles fit. A strong response identifies the point at which the original geometry was lost.
Transfer question. Transfer: a lens holder moves between trials. Which positions need checking before comparing focus distances?
Diagnostic 2: Two pins are nearly touching
Case. A pair of alignment pins is 5 mm apart and the projected direction looks unstable.
Reasoning and repair. The dominant issue is angular sensitivity to small marking error. Increase separation within a safe, workable arrangement and repeat the alignment; explaining only that the pins were ‘not straight’ misses why the measured angle changed. Compare how a 1 mm positional error affects a short versus a longer baseline.
Transfer question. Transfer: identify the equivalent long-baseline principle in graph-gradient work.
Diagnostic 3: The protractor reads from the surface
Case. A learner reports 62 degrees for an angle that is 28 degrees to the normal.
Reasoning and repair. Correct the reference line before recalculating: these values are complementary, not two independent measurements. Draw the normal at the intersection, mark the acute angle to that line and state the convention. A calculator cannot decide which geometry the question meant.
Transfer question. Transfer: with an angle of 41 degrees to the surface, state its incidence angle.
Diagnostic 4: A clean graph has inconsistent points
Case. Five readings sit on a trend but one was copied from another trial.
Reasoning and repair. This is a record-integrity problem, not an invitation to remove an outlier because it looks untidy. Check the raw notebook for the trial identifier. If the source cannot be recovered, label the reading uncertain and do not claim it was verified. Repeating the measurement may restore a defensible point.
Transfer question. Transfer: what would you do if a graph point has no corresponding raw measurement?
Diagnostic 5: Lens image blurs at the edges
Case. A recognisable centre is sharp but the sides are blurred.
Reasoning and repair. Check alignment, screen tilt and lens suitability before declaring a new law of image formation. A tilted screen can put different parts at different effective distances. Keep a record of the adjustment and whether whole-image sharpness improves. ‘Move the screen’ alone may not correct a tilted optical arrangement.
Transfer question. Transfer: name one observation that would distinguish screen tilt from a uniformly misplaced screen.
Diagnostic 6: Bright screen, no recognisable shape
Case. The screen is illuminated but no distinct object features are visible.
Reasoning and repair. Do not classify this as a successful real-image measurement. Verify the object is illuminated appropriately, the lens is in the path and the object distance permits a real image. Search screen position systematically while leaving other controls fixed. Record the observed criterion for focus instead of treating brightness as accuracy.
Transfer question. Transfer: explain why a virtual image cannot be captured on that screen.
Diagnostic 7: The student changes object and screen together
Case. The report gives one sharp screen position but omits the object position.
Reasoning and repair. The observation may be real, but the variable relationship cannot be analysed from incomplete positions. Record distances from a consistent reference, normally the lens’s optical centre in an idealised thin-lens setup, and change one position at a time. Explain why the missing quantity limits comparison even if the image looks convincing.
Transfer question. Transfer: design a two-column record to recover the next set of readings.
Diagnostic 8: Repeated measurements differ by one degree
Case. Four independent traces give 25, 26, 25 and 24 degrees.
Reasoning and repair. The modest spread is consistent with measurement variability, but without equipment resolution and repeat design it should not be called a precise uncertainty estimate. Report the values honestly, inspect the normal construction and use a sensible summary. Do not pick 25 simply because it fits an expected graph.
Transfer question. Transfer: what additional detail would make a numerical uncertainty statement more defensible?
Diagnostic 9: Every reading is too large
Case. Angles from the same protractor appear consistently high across several controlled trials.
Reasoning and repair. Investigate zero alignment, wrong scale selection and normal construction before taking dozens of repeats. A shared bias can survive averaging. Describe how an independent instrument check or alternate reading direction could test that proposed cause. Distinguish a suspected mechanism from an established fact.
Transfer question. Transfer: which is likely to help more here: another ten repeats or a zero check?
Diagnostic 10: The ratio calculation varies unexpectedly
Case. A student’s sin(i)/sin(r) values differ despite neat numerical work.
Reasoning and repair. First audit whether each pair belongs to the same ray trial and whether both angles were measured from their correct normals. Then check degree mode. Only after these procedural checks should you discuss whether the range of values is consistent with measurement limits. Replacing data to force a constant is not scientific work.
Transfer question. Transfer: a calculator is in radians mode. What symptoms appear when degree-labelled data are entered?
Diagnostic 11: Ray arrows point the wrong way
Case. The drawing shows plausible lines but arrows imply light travels from air into the block on the exit side.
Reasoning and repair. The error is representational yet may corrupt the explanation. Correct direction labels and identify both media at each crossing. A ray path without direction is insufficient for deciding whether the ray bends towards or away from the normal. The improvement is to label source, boundary and travel direction before adding angles.
Transfer question. Transfer: describe the bending direction for light entering air from glass at a non-critical subcritical angle.
Diagnostic 12: An experiment claims perfection from three trials
Case. Three successful screen-focus trials are used to claim that the lens is error-free.
Reasoning and repair. The data establish a limited repeatable observation under the conditions tested. They do not test every object distance, lighting arrangement, mounting angle or lens imperfection. Rewrite the conclusion with a stated scope and one further experiment that would challenge it. Precision in language is part of sound scientific analysis.
Transfer question. Transfer: propose a claim that the same three trials could actually justify.
Diagnostic 13: A clear result is rounded too soon
Case. An angle is recorded as 19 degrees although the instrument and marking procedure allowed a more careful reading.
Reasoning and repair. Preserve reasonable working precision where observations justify it, then round final values according to the task’s requirements. Do not invent decimal places unsupported by a protractor. Early rounding can distort derived trigonometric ratios, but falsely precise input is no better. Explain what the measurement can really distinguish.
Transfer question. Transfer: when a computed answer has six decimals, what determines the digits reported?
Diagnostic 14: A guessed improvement does not fix the fault
Case. A student proposes repeating measurements to correct an angle read from the wrong reference line.
Reasoning and repair. Repeats performed with the same wrong convention reproduce the defect. Redraw the normal and teach the reference-angle rule first. An improvement earns its place by breaking the identified causal chain, not by sounding like a standard laboratory phrase.
Transfer question. Transfer: suggest a direct improvement for the wrong-reference problem and a different improvement for parallax.
Diagnostic 15: Theory and observation disagree
Case. The learner’s predicted emergent ray is parallel to the incident ray, but the traced ray is visibly oblique.
Reasoning and repair. Treat the disagreement as diagnostic evidence. Check whether the block faces are actually parallel, whether the block moved, and whether both boundaries were drawn correctly. Record observations and predictions separately. Do not conclude that refraction is impossible or that the prediction automatically wins without checking assumptions.
Transfer question. Transfer: if the actual object has a wedge-shaped cross-section, which assumption of the prediction fails?
Diagnostic 16: The final evaluation is vague
Case. The answer says ‘measurement was not accurate; be more careful’.
Reasoning and repair. Replace this with the specific weak link: pin positions were too close, amplifying direction uncertainty. State the remedy: increase pin spacing and use a sharp marking tool. State why it helps: the same small location error has a smaller effect on the estimated ray direction. A complete explanation ties symptom, mechanism, action and expected change.
Transfer question. Transfer: rewrite a vague evaluation for an out-of-focus screen measurement.
A practical recording template that helps you find faults
For each trial write the intended independent-variable value, apparatus setup identifier, observation with units, method used to determine the value, any unusual event and a concise evaluation. Give drawings stable labels that match the table. Keep first readings separate from corrected calculations. If you later discover a copying error, show the correction and its source. The examiner need not see your entire private ledger, but practising this way teaches you to distinguish what happened from what you now believe should have happened.
A strong revised conclusion answers three questions: What did I observe? What physical relationship is consistent with the evidence? What weakness or untested condition limits the finding? A more ambitious claim needs stronger evidence. This is how experimental accuracy and examination clarity reinforce each other.
Four-week revision plan
Week 1: practise normals, angle readings, ray arrows and legible records until each can be explained without prompts. Week 2: distinguish bright images from sharp images and separate alignment faults from object-position effects. Week 3: complete mixed-fault cards under a modest time limit, writing the first faulty step before any correction. Week 4: revisit six older cases with changed numbers and locations; keep only errors that persist in the ledger. These timings are suggestions for home study, not SEC examination specifications.
Frequently asked questions
Does Paper 3 use this exact optical-bench scenario?
No. These are original practice cases. The official K323 syllabus states the assessment structure and skills, but it does not promise the exact apparatus or question sequence represented here.
Should I always repeat a reading that looks wrong?
First identify the suspected fault. A fresh repeat helps when alignment or temporary disturbance may be responsible; it does not cure the use of a wrong formula or consistently misread reference angle.
What should I do if the result contradicts a textbook diagram?
Check the physical setup, the drawing and calculation in that order. Keep the actual observation and state the assumptions behind the theoretical expectation. Avoid adjusting raw readings to fit the picture.
How do I write a useful improvement in one sentence?
Name the source of uncertainty, propose one feasible change and explain the mechanism of improvement. A precise causal sentence is usually stronger than several general promises to be careful.
Can I claim that repeated values prove the model?
Repeats support reliability under tested conditions. They do not, by themselves, establish that the underlying model applies in every situation. Explain what variables and assumptions the experiment did not test.
What if I am slower than classmates at practical work?
Practise the small decisions that cost time: marking the normal, assigning trial labels, reading instruments and recording immediately. A reliable sequence can be made more efficient; haste without method often forces a complete restart.
How can a parent help at home without building a laboratory?
Ask the student to explain where a reported observation came from and what would count as evidence against it. Paper diagrams, simple safe lens demonstrations and error discussions can develop reasoning; supervised school practicals remain essential.
What comes after this volume?
The next guide examines numerical reporting and decision thresholds in mathematics. The following English guide repairs conditional grammar in evidence reports. Together they prepare students for a cross-subject optics decision task.
Next reading and official sources
For ray fundamentals revisit Vol 1089; for a focused lens image experiment revisit Vol 0476; for integrated optics rehearsal revisit Vol 1096. The official authority for paper structure is SEAB’s G3 Physics K323 syllabus and the SEC overview. These original exercises are educational examples, not official questions or guaranteed scoring rubrics.
