Civil engineering mechanics & structural foundations of Two-Peg Level Collimation.
In civil surveying engineering, optical automatic levels and electronic digital levels establish precise vertical elevations for road grading, drainage pipe invert slopes, foundation benchmarks, and structural settlement monitoring. An automatic level utilizes an optical telescope fitted with crosshairs and an internal wire-hung pendulum compensator (or magnetic damping prism) that automatically levels the line of sight when the circular bubble is roughly centered. However, mechanical shock during transit or thermal expansion can knock the reticle crosshair out of alignment, causing the optical collimation axis (line of sight) to tilt upward or downward at an angle α relative to the true local horizontal plane. When backsight (BS) and foresight (FS) distances are strictly equal, collimation errors cancel out completely: e_BS = e_FS. But on uneven terrain, steep grades, or river crossings where sight lengths are unequal, collimation error induces cumulative elevation errors: e = D · tan(α) ≈ D · α. Developed as the universal calibration procedure, the Two-Peg Test (ISO 17123-2) isolates and quantifies this collimation error factor c = tan(α). Two rigid benchmark pegs (A and B) are driven into firm ground separated by distance D (e.g. 30 to 60 m). The level is set up midway between pegs (Instrument Position 1, where equal sight lengths yield true elevation difference ΔH_true = a_1 - b_1) and then relocated immediately behind Peg A (Instrument Position 2, where sight lengths are heavily unequal, exposing collimation error: c = [(a_2 - b_2) - ΔH_true] / ΔD). In this laboratory investigation, apparent error Δe is monitored across sight distance differentials from 10.0 to 60.0 m to calibrate the collimation inclination factor c = 0.00185 m/m (1.85 mm per meter, or α = 381.6 arcseconds).
Collimation Axis Geometry & The Two-Peg Calibration
Let the line of sight be inclined at an angle $\alpha$ to the true horizontal plane. The collimation error factor is $c = \tan \alpha \approx \alpha$ (in radians). For any staff reading at distance $D$, the observed reading $R_{obs}$ relates to the true horizontal reading $R_{true}$ by:
\[ R_{obs} = R_{true} + c \cdot D \]When the level is placed at midpoint $M$ between pegs $A$ and $B$ (sight distances $D_A = D_B = D_0$):
\[ a_1 = a_{true} + c D_0 \] \[ b_1 = b_{true} + c D_0 \] \[ \Delta H_{true} = a_1 - b_1 = a_{true} - b_{true} \]When the level is moved to Position 2 behind Peg $A$ (distance to $A$ is $D_{near}$, distance to $B$ is $D_{far}$):
\[ a_2 = a_{true} + c D_{near} \] \[ b_2 = b_{true} + c D_{far} \] \[ a_2 - b_2 = (a_{true} - b_{true}) - c (D_{far} - D_{near}) = \Delta H_{true} - c \Delta D \]The apparent collimation elevation error is:
\[ \Delta e = | (a_2 - b_2) - \Delta H_{true} | = c \cdot \Delta D \]Plotting $\Delta e$ against distance differential $\Delta D$ yields linear slope $c = 0.00185\text{ m/m}$.
Angular Reticle Tilt Conversion
The angular tilt of the line of sight in arcseconds is:
\[ \alpha = c \times 206,265'' = 0.00185 \times 206,265'' = 381.6'' \quad (6' 21.6'') \]Viva voce questions & analytical theory
Why does setting the level at the exact midpoint between two pegs eliminate collimation error?
Because the collimation axis tilts at angle α, the line of sight rises (or dips) by error e = D · tan(α); when backsight distance D_BS and foresight distance D_FS are identical, both readings contain the exact same error e, which subtracts to zero in ΔH = (BS - e) - (FS - e) = BS - FS.
What is the physical collimation axis in an optical telescope?
The collimation axis (line of sight) is the imaginary straight line passing through the optical center of the objective lens and the intersection of the reticle crosshairs.
How does earth curvature and atmospheric refraction affect the Two-Peg test?
At the midpoint setup, earth curvature and refraction affect both pegs equally and cancel out; at Position 2 (60 m distance), combined curvature and refraction is tiny (approx. 0.2 mm) but is accounted for when high-precision invar staffs are used.
What is the acceptable tolerance for collimation error in general engineering leveling?
Standard engineering specifications mandate collimation error c ≤ 0.05 mm/m (5 mm per 100 m, or approx. 10 arcseconds); errors exceeding this threshold require physical reticle diaphragm adjustment.
How is the physical reticle adjustment performed after the Two-Peg test?
With the level at Position 2, compute the true correct reading for the far staff: b_correct = a_2 - ΔH_true; using the tommy pin, gently turn the reticle capstan screw until the horizontal crosshair moves to read exactly b_correct on the staff.
Why must a digital level still undergo the Two-Peg test?
Digital levels use electronic bar-code image sensors, but if the optical lens axis tilts relative to the internal CCD linear array, the digital reading carries the identical optical collimation error; digital levels feature automated software routines that calculate c and store an electronic correction factor.
Experimental precautions and laboratory safety
- Hold leveling staffs strictly vertical using the attached bullseye circular bubble during every reading.
- Verify that the internal pendulum compensator is functioning by gently pressing the compensator check button or tapping the tripod.
- Perform the Two-Peg test during overcast conditions or early morning to avoid optical heat shimmer and mirage refraction near ground level.
- Turn reticle adjusting capstan screws with extreme delicacy to avoid stripping threads or shattering the glass reticle.
Theoretical assumptions & boundary condition limits
- The civil engineering physical models apply within the specified calibrated operational window (10.0 to 60.0 m).
- Assumes homogeneous, isotropic continuum material behavior without unmodeled micro-fracture anomalies.
- Continuous sensor readings reflect ordinary least-squares regression across successive observations.
- Extreme inelastic yielding or cyclic degradation requires non-linear constitutive continuum damage mechanics.