PAU Technical Drawing II (Galicia) common mistakes
PAU Technical Drawing II (Galicia) · Understand the traps and how to avoid them.
When a dihedral problem gives a plane by its traces, the candidate draws the auxiliary or rabatting rotation about the horizontal trace only, even when the plane is close to vertical and the horizontal trace is nearly parallel to the ground line; the resulting true shape is distorted or the construction does not close.
They memorise 'rotate about the horizontal trace' as an unconditional rule and never check which trace gives a workable angle or which trace intersects the figure. When the horizontal trace is nearly parallel to the ground line, the rotation angle approaches 90° and the rabatted figure runs off the paper, so they force it and lose accuracy.
What to do instead
Before rotating, compare the angles the two traces make with the ground line and choose the rotation whose axis cuts across the figure. If the chosen trace does not intersect the shape, extend the figure or rotate about the other trace. Practise a set of planes at 15°, 45°, 75° and 90° to the vertical plane and record which trace you used and why.
In an isometric or dimetric axonometric, the candidate takes the drawn lengths along the axes as true lengths and marks the dimensions directly, so the solid comes out 1.22 times too large in isometric or inconsistently scaled in dimetric.
They conflate the axonometric scale with the reduction coefficient. The isometric axes are drawn at 120° with a single natural scale, but true lengths must be multiplied by about 0.816 to be plotted; many textbooks draw at natural scale and simply label the drawing as reduced, which the candidate does not distinguish.
What to do instead
At the start of every axonometric problem, write the scale factor you are using in a corner of the sheet: 1 for natural isometric, 0.816 for true-length isometric, and the two foreshortening factors for dimetric. Apply the factor to every measured coordinate before plotting, and check one known dimension at the end by dividing the drawn length by the factor.
When constructing a conic from focus, directrix and eccentricity, the candidate generates too few points near the vertex and near the asymptotes of the hyperbola, then joins them with a freehand or a single spline, producing a curve that crosses the axis or misses the vertex tangent.
They treat the construction as a point-plotting exercise rather than a locus with known tangency conditions. The vertex is a tangency point where the curve is perpendicular to the axis, and the hyperbola's asymptotes constrain the branch shape; without those constraints a smooth spline through sparse points drifts.
What to do instead
Always locate the vertex first by the ratio distance-to-focus over distance-to-directrix, draw the tangent at the vertex perpendicular to the axis, and for the hyperbola draw the asymptotes through the centre before plotting. Add extra points where the curve changes curvature most, and finish with a French curve aligned to at least three consecutive points, checking tangency at the vertex.
On a multi-view drawing, the candidate places hidden features as visible outlines in the view where they are actually behind material, or omits hidden lines entirely in views where the feature is not directly visible.
They reason from the 3D mental model rather than applying the projection rule: a line is visible only if no material lies between it and the observer in that view. Because the 3D model looks 'open', they draw all edges solid.
What to do instead
For each view, list the edges and test visibility by asking whether the edge is on the near face. Use dashed lines for edges hidden behind material, and never draw a hidden line coincident with a visible line. Do a final pass with a coloured pencil marking every dashed line and confirming it corresponds to a real edge in the other views.
When a section is required, the candidate draws the cutting plane line but hatches the whole view or hatches across the visible background, and mixes hatching direction between adjacent parts so two separate solids read as one.
They confuse a section with a shaded view and do not apply the convention that hatching marks only the cut material. They also forget that adjacent parts in an assembly must be hatched in opposite directions or at different spacings.
What to do instead
Hatch only the areas where the cutting plane passes through solid material; leave voids and background unhatched. For assemblies, alternate hatching direction (45° and 135°) or spacing between neighbouring parts. Check that no hatch line runs parallel to a main outline, since that hides the contour.
The candidate dimensions from a scaled drawing by measuring the printed length with a ruler and writing that number, ignoring the stated scale, so a part drawn at 1:2 is dimensioned at half its real size.
They treat the drawing as a picture to be measured rather than a specification. In CAD-produced drawings the model is often at full size but plotted at a scale, and the candidate does not separate model space from paper space.
What to do instead
Always dimension true sizes regardless of drawing scale, and read the value from the model or the given data, not from a ruler on the sheet. If a dimension is missing, compute it from the given dimensions and geometry rather than scaling the print. State the scale in the title block and verify one dimension against it.
In a CAD exercise the candidate draws everything on layer 0 with the default line type, then cannot produce the required line weights, centre lines or hidden lines, and the plotted sheet fails the line-convention requirement.
They use CAD as a drawing board and never set up layers, line types or a title block before drawing. Because CAD displays everything as thin lines on screen, the line-weight problem is invisible until plotting.
What to do instead
Before drawing, create layers for outlines, hidden lines, centre lines, dimensions and text, each with its line type and line weight, and set the current layer appropriately. Set line-type scale so dashes are visible at the plotted scale, and check the plot preview for line weights before submitting.
The candidate crops or trims entities with the wrong boundary, leaving stray segments, or uses object snap loosely so lines that should meet at a corner stop a fraction of a millimetre short or overshoot.
They rely on visual alignment at the current zoom instead of snapping to geometric points. At high zoom the gap is obvious but at the working zoom it is invisible, and the error propagates into dimensions and sections.
What to do instead
Enable endpoint, midpoint, intersection and perpendicular snaps and keep them on. Use trim and extend with explicit cutting edges or boundaries, then zoom to each junction to confirm closure. Run an audit or select-all to check for zero-length or duplicate entities before finishing.
When croquis-drawing an object, the candidate sets the front view arbitrarily and then cannot fit the three views on the sheet, or the views do not align because they were placed by eye rather than projected.
They start drawing before choosing the viewing direction and the paper layout. The front view choice determines which features appear as true shape and how much space the other views need, and without projection lines the alignment drifts.
What to do instead
First choose the front view so the most characteristic contour is true shape and the object fits the format. Lay out the three views with construction lines projected from the front view, respecting first-angle or third-angle placement as required, and only then darken the outlines.
On an angle or tangent construction, the candidate bisects or draws a tangent arc by eye and does not mark the construction arcs, so the examiner cannot verify the method and the point of tangency is slightly off.
They consider construction lines clutter and erase them, or they skip the formal construction because the answer 'looks right'. Tangency requires the centre to lie on the perpendicular to the tangent line at the contact point, which eyeballing cannot guarantee.
What to do instead
Keep construction arcs and lines in a light weight, and determine tangent points by dropping a perpendicular from the centre to the tangent line or by the line-of-centres method for two circles. Mark the tangency points. Erase only the surplus beyond the final outlines, never the construction that proves the method.