A Level Physics 9702 · Exam special · Paper 3 (AS) and Paper 5 (A Level) · ~7 min read
A Level Physics 9702: Paper 3 Practical and Paper 5 Planning Guide
How to score on the practical papers: tables and significant figures, graphs and gradients, uncertainties, the limitations-and-improvements question on Paper 3, and planning, log graphs, error bars and worst-fit lines on Paper 5.
Paper formatsTables & unitsSignificant figuresGraphs & best-fit linesGradient & interceptUncertaintiesLimitations & improvementsPlanning
RSby Riyaz Sir
mentorriyaz.in
mentorriyaz.in
Paper tag = where it is examinedDATA & FORMULAE given in the examC1A1B1 how marks are awardedhighlight = comes up again and againHANDY worth memorisingTest yourself at the end
In 30 seconds
- Paper 3 (AS, 2 hours, 40 marks) has two experiments: Question 1 is usually a graph experiment; Question 2 is a two-reading experiment ending with limitations and improvements.
- Paper 5 (A Level, 1 h 15 min, 30 marks) has a planning question (design an experiment) and an analysis question (log graphs, error bars and uncertainties).
- Most marks come from routine skills: a well-formed table, a well-scaled graph, a correct gradient, and specific (not vague) evaluation.
- Uncertainties: percentage uncertainties add for products and quotients; a power multiplies the percentage uncertainty.
By the end you can
- Record data in tables with correct headings, units and precision.
- Plot graphs, draw best-fit lines and calculate gradients and intercepts correctly.
- Estimate and combine uncertainties.
- Identify specific limitations and improvements.
- Plan an experiment and analyse data with log graphs, error bars and worst-fit lines.
1Paper 3: taking readings
Tables
- Headings: quantity / unit, e.g. "l / cm", "T / s", "1/T² / s⁻²". Units in the body of the table lose marks.
- Raw readings to the precision of the instrument, consistently: a metre rule to the nearest mm (e.g. 45.0 cm, not 45 cm); a stopwatch reading, then an averaged period.
- Calculated values: same number of significant figures as (or one more than) the raw data used.
- Range and number: usually at least six sets of readings, spread over as wide a range as the apparatus allows.
- Repeat timings and average. Time many oscillations (e.g. 20) to reduce the percentage uncertainty from reaction time.
Common mistakeInconsistent decimal places in a column, or a heading like "Time (s) for 10 oscillations" with no clear quantity. Write "10T / s".
Graphs
- Scales: points must fill at least half the grid in both directions. Use 1, 2 or 5 units per 2 cm; awkward scales (3, 6, 7) lose marks.
- Plotting: to within half a small square, with small neat crosses or dots.
- Best-fit line: thin, straight, with points balanced either side along its length. Don't force it through the origin.
- Gradient: use a triangle whose hypotenuse is at least half the length of the line. Read points on the line, not data points, and show the coordinates used.
- Intercept: read directly only if the x-axis starts at 0; otherwise substitute a point into y = mx + c.
Worked example
A best-fit line passes through (0.20, 1.35) and (1.10, 5.85). Find the gradient and y-intercept.
- gradient = (5.85 − 1.35)/(1.10 − 0.20) = 5.0 C1 A1
- c = 1.35 − 5.0 × 0.20 = 0.35 A1
2Uncertainties
Estimating and combining
| Situation | Uncertainty |
|---|---|
| Single reading on a scale | about half the smallest division (often the smallest division if judging is hard) |
| Repeated readings | half the range |
| Sum or difference | add absolute uncertainties |
| Product or quotient | add percentage uncertainties |
| Power, e.g. x² | multiply the percentage uncertainty by the power |
Worked example
The length of a pendulum is (80.0 ± 0.2) cm and the period (1.80 ± 0.02) s. Find g = 4π²l/T² and its absolute uncertainty.
- g = 4π² × 0.800/1.80² = 9.75 m s⁻² C1
- % uncertainty = 0.25% + 2 × 1.11% = 2.5% C1
- g = 9.7 ± 0.2 m s⁻² A1
3Paper 3: evaluating the experiment
Limitations and improvements
- Two readings aren't enough to test a relationship. Improvement: take more readings (for different values) and plot a graph.
- Be specific: say what was hard and why, then match it with a precise fix.
| Limitation (specific) | Improvement (specific) |
|---|---|
| Difficult to judge when the ball first hits the floor by eye | use a video camera with a timer, or a light gate / sensor |
| Difficult to measure the diameter because the wire is thin | use a micrometer screw gauge |
| Large percentage uncertainty in a short time | time more oscillations, or use a light gate with a data logger |
| The bob doesn't swing in a single plane | use a fixed clamp and release gently from a marked point |
| Parallax error reading the rule against the moving object | use a set square, or film against a scale |
| Temperature of the water changes during the reading | stir and use a thermostatically controlled water bath |
Common mistake"Human error", "use more accurate equipment", "repeat readings" with no detail score zero. The 2024 Paper 3 report repeats this every year.
4Paper 5: planning
Designing the experiment
- Variables: state the independent variable (what you change), the dependent variable (what you measure), and the key controlled variables, plus how each is controlled.
- Method: a labelled diagram of the working apparatus; how to change and measure each variable, with named instruments (e.g. "micrometer for the diameter", "oscilloscope for the frequency").
- Analysis: state the graph to plot and how to test the relationship: "the relationship is valid if the graph is a straight line through the origin"; then how to find the constant from the gradient or intercept.
- Safety: a precaution linked to a real hazard in this experiment (e.g. "wear goggles in case the wire snaps"; "use a heat-proof mat for the hot water"). Generic safety scores nothing.
- Additional detail: how to make readings more reliable (e.g. repeat and average, check the zero error, wait for thermal equilibrium).
Analysis with logs
y = kxⁿ ⇒ lg y = n lg x + lg k
gradient n, intercept lg ky = Ae−kx ⇒ ln y = ln A − kx
gradient −k, intercept ln A- Error bars: plot the uncertainty of each point (for logs, the absolute uncertainty in lg y is half of lg(y + Δy) − lg(y − Δy)).
- Worst acceptable line: the steepest or shallowest line still passing through all the error bars. Uncertainty in the gradient = ǀbest gradient − worst gradientǀ.
- Significant figures for logs: lg values get as many decimal places as the raw data has significant figures.
Common mistakePlotting lg y but using y-values to find the intercept, or a worst line that misses an error bar. Show the coordinates you used for both lines.
Linking questionMeasurement, accuracy and precision are in Physical quantities and units; log-linear graphs also appear in Capacitance and Nuclear physics.
Test yourself — cover the answers first
Quick check
- What does a column heading need?
- How do you combine uncertainties for a product?
- How big should the gradient triangle be?
- What graph tests y = kxⁿ?
- Give a specific improvement for timing a falling ball by eye.
Show answers
- quantity / unit
- add percentage uncertainties
- hypotenuse at least half the length of the line
- lg y against lg x
- use a light gate or a video camera with a timer
Stuck on a question?
Share with a friendAsk Riyaz SirSend it to me on WhatsApp. I read every message.
Want regular help? Send this to your parent →