3.

Instrumentation and Force Measurement System

3.1 Overview
To quantify the aerodynamic performance of the scaled vehicle model, two dedicated load cells
were integrated into the test rig to measure the two principal aerodynamic forces independently:
drag (the force acting parallel to the free-stream airflow) and lift (the force acting perpendicular to
the free-stream airflow, positive upward). Each force component was captured by a separate
sensor to avoid cross-axis interference and to improve overall measurement accuracy.

The instrumentation chain for each channel consists of a single-point strain-gauge load cell
(rated at 1 kg), a HX711 24-bit analogue-to-digital converter (ADC) amplifier module, and a
common Arduino Uno microcontroller that digitises, processes, and transmits the force readings to
a host computer in real time.

3.2 Scaled Vehicle Model and Measurement Reference Point

The test subject is a 3-D printed sedan body modelled at a scale ratio of 1 : 15.33 relative to the
full-size production vehicle. The model was mounted at the geometric centre of the vehicle body
— both longitudinally (mid-wheelbase) and laterally (vehicle centreline) — as the reference
attachment point for the force transducers. Locating the load cells at the centre of mass minimises
moment arms, reduces bending errors at the sensor connections, and ensures that the measured
forces represent the net aerodynamic resultant acting through the body rather than a localised
reaction load.

3.3 Load Cell Specification

Both the drag-axis and lift-axis sensors are identical single-point aluminium-alloy load cells with a
Wheatstone-bridge strain-gauge configuration. Table 3.1 lists the manufacturer's specifications as
provided on the component datasheet.

Parameter Specification

Rated Load 1000 g (1 kg)

Output ± 0.05 mV/V

Output Sensitivity 1.0 ± 0.1 mV/V

Temperature Zero Drift 0.1% F.S

Temperature Sensitivity 0.05% F.S

Repeatability 0.03% F.S

Page 1 | Aerodynamic Force Measurement – Load Cell Instrumentation

 Parameter Specification

Insulation Resistance ≥ 2000 MΩ

Excitation Voltage 5 – 10 VDC

Dimensions 75 × 12.7 × 12.7 mm

Table 3.1 – Load cell specifications (both drag and lift sensors are identical)

The compact bar form factor (75 × 12.7 × 12.7 mm) allowed both sensors to be accommodated
within the structural envelope of the test rig platform visible in Figure 3.1, with minimal impact on the
internal flow field or model attitude.

3.4 Signal Conditioning – HX711 24-bit ADC

The raw differential output of each load cell (in the mV/V range) is too small to be read reliably by a
standard microcontroller ADC. Each load cell is therefore paired with an HX711 instrumentation-
grade amplifier and 24-bit Σ-Δ analogue-to-digital converter. Key characteristics relevant to this
application are:
• 24-bit resolution over the full measurement range, yielding approximately 16 million counts
per full-scale deflection and enabling the detection of very small aerodynamic loads.
• Selectable gain of 128× (Channel A, used here) or 64×/32×, providing adequate
amplification for the ± 0.05 mV/V rated output of the selected load cell.
• On-chip voltage regulator tolerates the 5 V supply from the Arduino Uno without additional
filtering circuitry.
• Two-wire serial interface (DOUT and SCK) minimises wiring complexity and eliminates the
need for SPI or I²C bus sharing.

Two independent HX711 modules are used — one per load cell — to ensure that the drag and lift
measurement channels do not share any signal path, preventing cross-talk between channels.

3.5 Arduino Uno Microcontroller and System Wiring

A single Arduino Uno (ATmega328P) microcontroller manages both HX711 modules. It polls each
converter sequentially, applies the calibration scale factor, and transmits the computed force values
over USB Serial at 9600 baud for logging on a host PC. The pin assignments used in this work are:
• Drag load cell HX711 — DOUT → Pin 3, SCK → Pin 2
• Lift load cell HX711 — DOUT → Pin 5, SCK → Pin 4
• Power supply — Arduino 5 V and GND rails shared by both HX711 modules

The HX711 Arduino Library (by Bogdan Necula / olkal) was used to abstract the serial
communication protocol and to expose tare and scale functions directly in the measurement code.

Page 2 | Aerodynamic Force Measurement – Load Cell Instrumentation

3.6 Sensor Calibration Procedure
Prior to any aerodynamic testing, each load cell was calibrated individually using a series of
traceable reference masses. Calibration was performed with no airflow to eliminate any
aerodynamic contribution to the baseline reading. The procedure followed these steps:
• With the model removed and no mass applied, the tare function was called to zero the
output of each channel.
• A known reference mass (500 g) was placed at the geometric centre of each load cell beam.
• The raw ADC count was recorded and divided by the known mass (in grams) to obtain the
calibration factor (counts / gram) for that channel.
• The calibration factor was stored in firmware and validated against a second reference mass
before the measurement campaign commenced.

The calibration sketch below implements this procedure and prints the resulting scale factors to the
Serial Monitor for use in the main measurement firmware.

Listing 3.1 – Calibration Sketch (Arduino Uno / HX711)

// ─── Load Cell Calibration Sketch (Arduino Uno + HX711) ───────────────

#include <HX711.h>

// ── Drag cell wiring ──────────────────────────────────────────────────

#define DRAG_DOUT 3
#define DRAG_CLK 2
// ── Lift cell wiring ──────────────────────────────────────────────────
#define LIFT_DOUT 5
#define LIFT_CLK 4

HX711 dragScale;
HX711 liftScale;

void setup() {
[Link](9600);
[Link](DRAG_DOUT, DRAG_CLK);
[Link](LIFT_DOUT, LIFT_CLK);

dragScale.set_scale(); // clear any previous factor

liftScale.set_scale();
[Link](); // zero with no load
[Link]();

[Link]("Place known mass on DRAG cell, then send any character.");

}

void loop() {
if ([Link]() > 0) {
[Link]();

Page 3 | Aerodynamic Force Measurement – Load Cell Instrumentation

 float dragRaw = dragScale.get_units(10);
float liftRaw = liftScale.get_units(10);

float knownMass_g = 500.0; // e.g. 500 g reference mass

float dragFactor = dragRaw / knownMass_g;
float liftFactor = liftRaw / knownMass_g;

[Link]("Drag calibration factor: "); [Link](dragFactor, 4);

[Link]("Lift calibration factor: "); [Link](liftFactor, 4);
[Link]("Copy these values into the main measurement sketch.");
}
}

3.7 Drag and Lift Coefficient Computation

Once the raw gram-force readings are obtained from each calibrated load cell, they are converted
to SI force units (Newtons) and used to compute the non-dimensional aerodynamic coefficients.
The drag coefficient C_D and lift coefficient C_L are defined by the standard aerodynamic relations:
F_D = C_D · q · A and F_L = C_L · q · A
where q = ½ · ρ · V² is the dynamic pressure of the oncoming airflow

In the above expressions, F_D is the measured drag force (N), F_L is the measured lift force (N), ρ
is the air density (1.225 kg/m³ at standard sea-level conditions), V is the free-stream velocity (m/s),
and A is the frontal (reference) area of the scaled model (m²). Because the coefficients are non-
dimensional, they are directly comparable to full-scale vehicle data — provided that the Reynolds
number similarity conditions are noted and reported alongside the results.

The measurement firmware (Listing 3.2) implements this computation on-board the Arduino Uno
and transmits the instantaneous C_D and C_L values alongside the raw forces to the host
computer for logging and post-processing.

Listing 3.2 – Aerodynamic Force and Coefficient Measurement Sketch

// ─── Aerodynamic Force Measurement Sketch ─────────────────────────────

// Outputs drag force (N) and lift force (N) over Serial
#include <HX711.h>

#define DRAG_DOUT 3
#define DRAG_CLK 2
#define LIFT_DOUT 5
#define LIFT_CLK 4

// ── Calibration factors (obtained from calibration sketch) ─────────────

const float DRAG_CAL_FACTOR = 420.50; // replace with your value
const float LIFT_CAL_FACTOR = 418.73; // replace with your value

Page 4 | Aerodynamic Force Measurement – Load Cell Instrumentation

// ── Physical constants ──────────────────────────────────────────────────
const float G = 9.81; // m/s²
const float RHO = 1.225; // air density (kg/m³) at sea level
const float V = 10.0; // tunnel airspeed (m/s) – update as needed

// ── Model & scaling geometry ────────────────────────────────────────────

// Scale ratio: 1 : 15.33 → full-size frontal area estimated from scaled model
const float SCALE = 15.33;
const float A_model = 0.0092; // frontal area of scaled model (m²) – measure
actual
const float A_full = A_model * SCALE * SCALE; // full-scale frontal area

HX711 dragScale;
HX711 liftScale;

void setup() {
[Link](9600);
[Link](DRAG_DOUT, DRAG_CLK);
[Link](LIFT_DOUT, LIFT_CLK);
dragScale.set_scale(DRAG_CAL_FACTOR);
liftScale.set_scale(LIFT_CAL_FACTOR);
[Link]();
[Link]();
[Link]("Speed(m/s), Drag_F(N), Cd, Lift_F(N), Cl");
}

void loop() {
float mass_drag_g = dragScale.get_units(10); // grams
float mass_lift_g = liftScale.get_units(10);

float Fd = (mass_drag_g / 1000.0) * G; // drag force (N)

float Fl = (mass_lift_g / 1000.0) * G; // lift force (N)

float q = 0.5 * RHO * V * V; // dynamic pressure (Pa)

// ── Aerodynamic coefficients ──────────────────────────────────────

// Using SCALED model area (wind-tunnel perspective)
float Cd = Fd / (q * A_model);
float Cl = Fl / (q * A_model);

[Link](V, 2); [Link](", ");

[Link](Fd, 4); [Link](", ");
[Link](Cd, 4); [Link](", ");
[Link](Fl, 4); [Link](", ");
[Link](Cl, 4);

delay(500);
}

Page 5 | Aerodynamic Force Measurement – Load Cell Instrumentation

3.8 Measurement Uncertainty
The principal sources of uncertainty in the force measurement system are summarised below:
• Repeatability of the load cell (0.03% F.S) introduces a maximum force error of ±0.3 g at full
scale, equivalent to ±2.9 mN.
• Temperature sensitivity (0.05% F.S / °C) may contribute a small thermal drift during
prolonged test runs; testing was therefore conducted in a climate-controlled indoor
environment to minimise this effect.
• HX711 quantisation noise — with 24-bit resolution and a 128× gain, the effective noise
floor is well below the minimum aerodynamic force of interest at the tested flow velocities.
• Model mounting compliance — the elastic deflection of the load cell beam under load (<
0.3 mm at full scale) was considered negligible for the model attitudes tested.

Overall, the combined expanded uncertainty in each force measurement channel is estimated at ±
2% of the indicated reading, which is adequate for comparative aerodynamic evaluation at the 1 :
15.33 scale.

3.9 Summary
Two 1-kg single-point load cells, each conditioned by a dedicated HX711 24-bit ADC and read by a
single Arduino Uno, provide independent and simultaneous measurement of the drag and lift forces
acting on the 1 : 15.33 scaled vehicle model. Calibration was performed with traceable reference
masses applied at the geometric centre of the test rig, consistent with the model mounting
philosophy. The firmware computes and logs aerodynamic coefficients in real time, and the
combined measurement uncertainty is estimated at ± 2%, which is sufficient for the comparative
aerodynamic assessment objectives of this study.

Page 6 | Aerodynamic Force Measurement – Load Cell Instrumentation