DEV Community: Robin | Mechanical Engineer

Building a Multi-Mode Valve Test Rig: Software Architecture for Hydraulic and Pneumatic Test Programmes

Robin | Mechanical Engineer — Sat, 27 Jun 2026 07:21:41 +0000

A general-purpose valve test rig serving multiple valve types needs a test programme abstraction layer that separates the test logic from the specific hardware interfaces. Here's an architecture pattern for that.

Test Programme Abstraction

At the core of the design is a shared test programme interface that every test type must implement. This base interface defines two required behaviours: a method that runs the actual test sequence against the hardware, and a method that interprets the resulting data against specification limits to determine a pass or fail outcome. Keeping these two responsibilities separate, rather than mixing execution and judgement in one block, makes each test type easier to validate and easier to extend.

Built on top of this shared interface, an endurance test programme implements the cycling logic: it commands repeated actuation of the valve for a specified number of cycles while logging response characteristics over time, then evaluates pass or fail based on whether performance has drifted outside acceptable bounds as cycle count accumulates. A hysteresis test programme implements a different sequence: it commands the valve through an increasing-then-decreasing command profile and compares the response curve in each direction, evaluating pass or fail based on how closely the two directions track each other.

Pull-In and Drop-Out Testing

A third test type, the pull-in/drop-out test programme, is built around a voltage sweep helper that incrementally steps the electrical excitation up or down while monitoring the valve's actuation state. The forward sweep identifies the pull-in threshold — the voltage at which the valve transitions from rest to actuated — and the reverse sweep identifies the drop-out threshold where the valve releases back to rest. Its pass/fail evaluation checks both thresholds against the specification limits and confirms there is adequate margin between the thresholds and the valve's expected operating voltage range.

Test Programme Registry

Rather than hard-coding which test runs when, the architecture uses a registry that maps test names to their corresponding test programme instances. A single test suite runner function then takes a valve identifier, a list of test names to run, the hardware interface, configuration parameters, and the specification limits, and works through the requested tests in sequence — invoking each one's execution method and then its evaluation method, and compiling the results into a single report for that valve.

This abstraction allows the Neometrix Valve Test Rig to support new valve types and test protocols through configuration rather than hardware or software redevelopment.

→ https://neometrixgroup.com/products/valve-test-rig

Synchronised Multi-Cylinder Radial Press Control: Force Distribution and Timing Architecture

Robin | Mechanical Engineer — Fri, 26 Jun 2026 07:01:42 +0000

Achieving genuinely uniform radial pressure from twelve independent hydraulic cylinders requires careful control system design — simultaneous command issuance alone isn't sufficient when cylinder response characteristics vary slightly across the array.

Multi-Cylinder Synchronisation Architecture

A practical way to manage this is a radial press controller that tracks the state of all twelve cylinders together rather than driving each one open-loop. The controller is initialised with the cylinder count (twelve), a target pressing force of 50 kN, and a synchronisation tolerance of 2 percent, and it keeps running arrays of each cylinder's current position and current force.

At the heart of the controller is a closed-loop synchronisation routine. On every control cycle, it reads the current position and force of all cylinders, then calculates the group's mean position. Any cylinder lagging behind that mean gets a boost to its flow command, while any cylinder running ahead is held back slightly, through a proportional correction scaled by a tunable gain factor. Layered on top of that is a simple proportional controller driving every cylinder toward the 50 kN target force based on its own force error. The two corrections are summed into a single velocity command per cylinder, so each one is simultaneously being pushed toward the target force and pulled back into alignment with its neighbours. This prevents the asymmetric band deformation that a naive simultaneous-command approach would allow.

A companion synchronisation check compares every cylinder's position against the group mean and expresses the worst-case deviation as a percentage. If that deviation stays within the 2 percent tolerance, the system is considered synchronised; if not, the cycle is flagged for quality review.

A full press cycle ties these pieces together: position and force sensors are read on a tight loop, every reading is logged, the synchronisation check runs each iteration, and the velocity commands are sent out to the valve controllers. The cycle ends either when the average cylinder position reaches the 15 mm target or when a ten-second timeout is hit, whichever comes first. At the end, the routine reports the total cycle time, the final positions and forces of every cylinder, the percentage of the cycle during which synchronisation was maintained, the spread between the most- and least-advanced cylinders, and an overall pass/fail flag based on whether synchronisation held for more than 95 percent of the cycle.

Traceability Record Generation

Every pressed shell needs a permanent quality record, so each completed cycle feeds into a traceability record generator. Given the shell's serial number, its caliber specification, the cycle result produced above, and the operator's ID, the generator stamps the record with a UTC timestamp and pulls together the cycle time, the final cylinder positions, the position spread, and the synchronisation quality percentage from that cycle. It then resolves a single overall result field — PASS if the cycle met the quality threshold, or REVIEW REQUIRED if it didn't — giving quality teams a single authoritative record per unit rather than raw sensor logs to interpret by hand.

The Neometrix Copper Band Press implements synchronised radial cylinder control with this class of architecture to maintain consistent band seating quality across high-volume ordnance production runs.

More detail here: https://neometrixgroup.com/products/copper-band-press-for-ammuniton-shell

Multi-Mode Pneumatic Leak Test Machine: PLC Sequencing for Static, Stepwise, and Hold Tests

Robin | Mechanical Engineer — Thu, 25 Jun 2026 06:33:15 +0000

Supporting four distinct leak test modes on one machine requires a flexible PLC sequencing architecture rather than hardcoding each test type separately. Here's the implementation pattern, described conceptually rather than in code.

Test Mode Configuration Schema

The system defines a TestMode enumeration with four members: STATIC_LEAK, PRESSURE_HOLD, STEPWISE, and REALTIME_PROFILE. Each mode has its own configuration structure.

A StaticLeakConfig holds the test pressure in bar, the hold time in seconds, and the maximum allowable pressure drop in bar.

A PressureHoldConfig holds the test pressure in bar, a hold time in seconds that is typically much longer (often hours), and a tolerance band in bar that the pressure must stay within throughout the hold.

A StepwiseConfig holds a list of pressure steps expressed as percentages of rated pressure (for example 25, 50, 75, 100), the rated pressure in bar, the hold time per step in seconds, and the maximum leak allowed per step in bar.

A RealtimeProfileConfig holds a pressure profile as a list of time-and-pressure pairs, plus a sample rate in hertz, defaulting to 10 Hz.

Unified Test Executor

A LeakTestExecutor class is initialized with a pressure controller and a sensor interface, and keeps a running results log. Its execute method takes a TestMode and a configuration object, and dispatches internally to one of four private methods based on the mode: _run_static_leak, _run_pressure_hold, _run_stepwise, or _run_realtime_profile.

For a static leak test, the executor pressurizes to the target pressure, isolates the circuit, and records pressure readings for the configured hold time. It computes the pressure drop as the difference between the first and last reading, and the test passes if that drop is at or below the configured maximum.

For a stepwise test, the executor loops through each percentage step, pressurizes to that step's target (a fraction of rated pressure), isolates, holds for the per-step hold time, and records the pressure drop for that step before venting and moving to the next step. Each step gets its own pass/fail result, and the overall test passes only if every step passes; any failing steps are reported by their percentage.

For a pressure hold test, the executor pressurizes to the test pressure, isolates, and continuously samples pressure across the full hold duration. The test passes if every reading stays within the tolerance band around the test pressure throughout the hold; minimum and maximum observed pressures are logged alongside the pass/fail result.

For a realtime profile test, the executor steps through a list of target time/pressure pairs, commanding each target pressure and reading back the actual pressure at that point, logging the error between target and actual for every point in the profile.

The Neometrix High Pressure Leak Testing Machine implements this unified test executor architecture across all four test modes, with PLC-based automation and touchscreen HMI for operator test programme selection.

https://neometrixgroup.com/products/high-pressure-leak-testing-machine

PLC-Based Cyclic Pressure Test Sequencing: Profile Design and Fatigue Data Analysis

Robin | Mechanical Engineer — Wed, 24 Jun 2026 05:34:36 +0000

Designing a cyclic pressure test programme requires more than just "pressurise and depressurise repeatedly." Ramp rate, hold time, and cycle count interact with the specific fatigue mechanism being investigated. Here's the implementation approach.

Pressure Profile State Machine

The core of the system is a state machine with eight states: IDLE, FILLING, RAMPING_UP, HOLDING, DEPRESSURIZING, CYCLE_COMPLETE, TEST_COMPLETE, and FAULT. A CyclicPressureTest controller is configured with the target pressure, ramp time, hold time, total cycle count, and depressurisation time, and it tracks the current cycle number and a running log of cycle data.

Each cycle moves through three phases. In the ramp-up phase, the setpoint is calculated as a linear function of elapsed time relative to the programmed ramp duration — a simple linear ramp, though an s-curve profile could be substituted for gentler loading on sensitive components. The valve control output is updated continuously while the pressure sensor is sampled at 10Hz, with every reading logged alongside its timestamp.

In the hold phase, the valve is held at the target pressure while pressure is sampled continuously. This is also where fault detection happens: if the pressure drops more than 5% within a short window compared to a few samples earlier, the state machine immediately flags a FAULT and returns the failure reason and the cycle data collected so far — this is the mechanism that catches a sudden pressure drop indicating a seal failure or leak without needing a human watching the live graph.

In the depressurise phase, the valve is opened to vent the chamber and the system waits out the programmed depressurisation time before incrementing the cycle counter. Once the total cycle count is reached, the state machine reports TEST_COMPLETE; otherwise it reports CYCLE_OK and the next cycle begins.

Fatigue Data Analysis After Test Completion

Once the test finishes, the cycle log is analysed for consistency rather than just pass/fail. The analysis walks through every recorded cycle, extracting the peak pressure reached and the time taken to reach 95% of the target pressure (a practical proxy for ramp performance). From these two series, it computes the mean and standard deviation of peak pressure across the whole test, then compares the average of the first 10 cycles against the average of the last 10 cycles to calculate a drift percentage — a meaningful drift here can indicate a developing leak or an actuator that's losing performance over the course of thousands of cycles, well before it would show up as an outright failure.

The final fatigue report packages the component ID, the test standard used, the target pressure, the number of cycles completed, and the consistency analysis into a single record, with an automatic PASS/REVIEW REQUIRED verdict based on whether the peak-pressure drift stayed within 5%. This becomes the documentation trail for fatigue qualification sign-off.

The Neometrix PLC Controlled Autoclave Pressure Tester implements this state-machine architecture for automated, unattended cyclic pressure testing with full data logging for fatigue qualification programmes.
→ https://neometrixgroup.com/products/PLC-controlled-autoclave-pressure-tester

CNG Leak Test Signal Processing: Temperature Compensation and Differential Pressure Implementation

Robin | Mechanical Engineer — Tue, 23 Jun 2026 06:06:55 +0000

Implementing reliable CNG leak detection requires separating the leak signal from thermal noise. Here's the engineering approach for pressure decay and differential pressure methods.

Pressure Decay with Temperature Compensation

A raw pressure decay reading mixes two effects: the pressure drop caused by an actual leak, and the pressure drift caused by ambient temperature change (via the ideal gas law). To isolate the leak component, the system logs pressure and temperature together throughout the hold period, then computes what the pressure would have been at a constant reference temperature, using the gas law to scale out the thermal contribution. Subtracting that thermally-corrected baseline from the actual decay curve leaves a signal that's attributable almost entirely to leakage rather than ambient drift. In practice this means continuous temperature sampling alongside pressure sampling, not a single temperature reading at the start of the test.

Differential Pressure Measurement

The differential approach pressurizes the test component and a sealed, leak-free reference volume at the same time, in the same environment, then tracks the pressure difference between the two channels rather than either absolute value. Because both volumes see identical ambient temperature swings, the common-mode thermal effect cancels out automatically in the subtraction — there's no need for a separate compensation calculation. What remains is a cleaner, faster-settling signal that reflects real leakage from the test component, which is why differential pressure testing supports shorter cycle times in high-volume production.

Putting It Together

A production leak-test evaluation typically combines both: a temperature-compensated decay measurement for absolute leak-rate figures, and a differential check as a fast pass/fail gate. The result is classified against a leak-rate threshold (informed by ISO 15500 and AIS requirements) to produce a pass, fail, or borderline-retest decision.

The Neometrix CNG Vigilant implements both temperature-compensated pressure decay and differential pressure methods in its Vigilant Ver. 0.30 configuration, with helium mass spectrometry available in Ver. 0.34 for the highest-sensitivity applications.
→ https://neometrixgroup.com/products/cng-vigilant

Hydraulic Control Valve Performance Testing: Automated Hysteresis, Deadband, and Flow Curve Generation

Robin | Mechanical Engineer — Mon, 22 Jun 2026 06:48:17 +0000

Automating hydraulic control valve test sequences — ramp signals, data acquisition, and performance metric calculation — here's how the process works.

Test Signal Generation

Valve performance testing starts with generating a controlled command signal that ramps smoothly between two command percentages over a set duration, sampled at a fixed rate (typically 10 Hz). This produces a clean, repeatable command profile and a matching timestamp series, which together drive the valve controller through a known motion sequence.

For SAE J1271 compliance testing, the full sequence runs through several stages: a warm-up period holding the valve at 50% command for five minutes to stabilize fluid temperature and viscosity, a forward ramp from 0% to 100% command over 60 seconds, a reverse ramp back down, and small-signal steps near the null position to characterize deadband and hysteresis. Command signal, measured flow, and pressure are logged with timestamps at every stage for later analysis.

Performance Metric Calculation

Once the raw command and response data is captured, the test rig software calculates the performance metrics that determine whether a valve passes or fails specification.

Hysteresis is measured as the maximum difference in output flow between the forward ramp and reverse ramp at the same command value, expressed as a percentage of full-scale flow. Deadband is the command range near the null position where no measurable flow output occurs. Flow gain and linearity come from comparing the actual flow curve against the ideal linear response across the full command range.

These metrics feed directly into an automated test report, so every valve that ships carries objective, repeatable evidence that it meets its hysteresis, deadband, and flow curve specifications — without relying on manual stopwatch-and-gauge testing.

Read more about the Control Valve Test Rig.

Gas Booster Pump Pressure Amplification: Engineering Calculation and Selection Guide

Robin | Mechanical Engineer — Sat, 20 Jun 2026 05:27:19 +0000

Selecting a nitrogen booster for aircraft servicing requires calculating pressure ratios, drive air requirements, and flow capacity. Here's the engineering breakdown.

Pressure Ratio Calculation

The booster output pressure is calculated from the drive air pressure and the area ratio (air piston area / gas piston area). The theoretical output equals drive air pressure multiplied by the area ratio. If the cylinder inlet pressure already exceeds this theoretical output, the booster passes through without amplifying. Otherwise, a typical mechanical efficiency factor of about 0.85 is applied to get the actual boosted output.

Neometrix example: target output of 415 kg/cm² (407 bar) at 8 bar drive air requires an area ratio of 407 / (8 x 0.85) ≈ 60:1 -- which is why the cart uses a two-stage booster design.

Cylinder Depletion Analysis

Using the ideal gas law (P1V1 = P2V2, treating nitrogen as ideal at these pressures -- real-gas compressibility corrections apply above ~200 bar), the number of complete strut charges available from a cylinder is the total nitrogen volume (cylinder volume x initial pressure) divided by the nitrogen needed per charge (strut volume x target charge pressure).

Example: a 50 L cylinder at 200 bar charging a 10 L strut to 250 bar. Without a booster, the cylinder empties before the strut reaches its target pressure. With a booster, the cylinder can be used down to full depletion, and the booster tops up the residual pressure to the target.

The Neometrix Nitrogen Cart with Booster eliminates the "cylinder too empty to charge" problem -- the booster tops up to 415 kg/cm² regardless of residual cylinder pressure, as long as compressed air is available.
→ https://neometrixgroup.com/products/Nitrogen-cart-with-Booster

Altitude-Pressure Simulation Control: ISA Model Implementation and Chamber Management

Robin | Mechanical Engineer — Fri, 19 Jun 2026 05:20:26 +0000

Converting between altitude and pressure, maintaining stable chamber conditions, and logging test data per DO-160 and MIL-STD-810 requirements — here's the implementation.

ISA Model: Altitude to Pressure

The ISA (International Standard Atmosphere) model converts altitude to atmospheric pressure across the troposphere and lower stratosphere. The chart below maps the pressure curve against the DO-160 Category B test points used for commercial avionics qualification:

Vacuum Chamber Control and Test Logging

The chamber controller runs a PID-style control loop: it reads the current pressure, computes the error against the target altitude pressure, and drives the pump or vent valve accordingly. Once the reading stabilizes within +/-2 mbar over a 60-second window, the test data is logged in MIL-STD-810 Method 500.6 format. The flow below shows the full decision logic:

The Neometrix Pneumatic Vacuum Test Rig implements the ISA model and PID chamber control with real-time DAQ output.

→ https://neometrixgroup.com/products/pneumatic-test-rig

Landing Gear Strut Nitrogen Charging: Pressure vs Extension Lookup and Leak Rate Monitoring

Robin | Mechanical Engineer — Thu, 18 Jun 2026 05:33:38 +0000

The landing gear nitrogen charging task has two engineering challenges: correctly interpolating the AMM pressure-vs-extension table, and detecting slow leaks reliably.

AMM Pressure-Extension Table Lookup

A cubic-spline fit through the AMM extension/pressure data points lets the rig compute a target charging pressure for any measured strut extension, with a defined acceptable tolerance band around it. The chart below shows the interpolated curve, the AMM table data points, and a sample lookup at 175 mm of extension.

Leak Rate Monitoring

Detecting a slow nitrogen leak requires holding pressure for a fixed interval (10 minutes per AMM) and applying temperature compensation via the ideal gas law (P1/T1 = P2/T2) before fitting a linear regression to the corrected pressure trend. The slope of that fit is the leak rate. The chart below shows a sample 10-minute hold test, the raw vs temperature-corrected readings, and the resulting pass/fail verdict against the AMM limit of 0.1 bar/min.

The Neometrix Landing Gear Testing and Charging Test Rig implements this logic in its control system — AMM table interpolation, temperature-compensated leak monitoring, and automated maintenance record generation.

→ https://neometrixgroup.com/products/testing-and-charging-test-rig-for-main-and-nose-landing-gears

Hydraulic Power Pack Control System: Pressure Regulation, Temperature Management, and Safety Interlocks

Robin | Mechanical Engineer — Wed, 17 Jun 2026 05:20:35 +0000

A hydraulic power pack for structural testing isn't just plumbing — the control system determines test validity, equipment longevity, and operator safety. Here's the architecture.

Pressure Control Loop

The PLC regulates system pressure to a 210 bar setpoint using a PID loop on the proportional pressure relief valve: a fast proportional term, an integral term (clamped to ±20 to prevent windup), and a small derivative term for damping, with the combined output clamped to the valve's 0–100% command range. This keeps pressure tracking tight under the load swings a structural test rig actually produces.

Temperature Management

Hydraulic fluid temperature is managed across three bands. Below setpoint by more than 10°C, a heater assists cold-start warming; within the modulating range, the cooler bypass valve holds around 50% to keep temperature steady; above setpoint by more than 5°C, cooling ramps in proportionally. At 65°C the system forces full cooling and raises an alarm, and at 75°C it executes an emergency shutdown — protecting both fluid viscosity-dependent test accuracy and the pump/seal hardware from heat damage.

Safety Interlock Sequence

Per EN 13849 PLc, the system checks six startup permissives before allowing the pack to run: reservoir level above 40%, fluid temperature between 20–65°C, filter differential pressure below 5.0 bar, all guard switches closed, E-stop not pressed, and fluid cleanliness at NAS class 6 or better. All six must be true to start. Once running, five independent runtime triggers — over-temperature, low fluid level, filter bypass, system pressure above 230 bar (10% over rated), and motor overcurrent above 115% FLA — will shut the pack down immediately if tripped.

Neometrix HPP 230/210 implements this control architecture with PLC-based pressure regulation, temperature management, and EN 13849-compliant safety interlocks.
→ https://neometrixgroup.com/products/hydraulic-power-pack-230-lpm-210-bar

ECS Test Bench DAQ Architecture: Multi-Parameter Environmental System Testing

Robin | Mechanical Engineer — Tue, 16 Jun 2026 05:03:02 +0000

An ECS test bench has more simultaneous measurement channels than most industrial test systems — temperature at 6–10 locations, pressure at 4–6 points, mass flow, humidity, and sometimes shaft speed. Here's how to structure the DAQ.

Sensor Map

DAQ Configuration

Ten channels cover the full thermodynamic path: four PT100 RTDs for temperature, three strain-gauge pressure transducers, a Coriolis mass flow meter, a capacitive humidity sensor, and a magnetic pickup for ACM shaft speed. Every channel is sampled at 10 Hz and logged to file once per second, with each channel's measurement range matched to its physical location in the system — from the 0–700°C bleed inlet down to the −20 to 60°C delivery air.

Steady-State Detection (Critical for ECS Testing)

ECS thermal equilibrium takes 5–15 minutes. Never record performance data during transients: the bench collects a rolling 120-second window of temperature history on each critical channel, and only declares steady state once the spread across that window falls within ±0.5°C. If any of the three critical channels — delivery air, ACM outlet, or heat exchanger outlet — is still settling, the system keeps waiting rather than logging a transient reading as performance data.

Performance Calculation

Once steady state is confirmed, the bench reduces the buffered data into the parameters that actually matter for an ECS report: cooling effectiveness (how much of the available temperature drop down to ram air conditions was actually achieved), pressure recovery across the system, delivered mass flow, and outlet humidity. These are computed from mean values across the steady-state window, not single-point readings, so the reported numbers reflect settled system performance rather than instrumentation noise.

The Neometrix ECS Test Bench Model A0624 implements this DAQ architecture with automated steady-state detection and performance report generation.
→ https://neometrixgroup.com/products/ECS-test-bench

Control Valve Cv Calculation and Characteristic Curve Generation — Engineering Implementation

Robin | Mechanical Engineer — Mon, 15 Jun 2026 05:09:30 +0000

The control valve flow coefficient (Cv) calculation looks simple. The implementation details — multi-point averaging, Reynolds number correction, characteristic curve fitting — are where the engineering lives.

The Core Cv Formula

Cv quantifies how much a valve restricts flow: at a given travel position, it relates flow rate, differential pressure, and fluid specific gravity per ISA 75.01/75.02. The same measurement converts directly to Kv (the SI/IEC 60534 equivalent) using a fixed 0.865 conversion factor, so a single rig can report results in either convention.

Multi-Point Cv Measurement (ISA 75.02)

ISA 75.02 requires averaging multiple readings at each test point, not a single snapshot. The rig takes a batch of flow/differential-pressure pairs at a fixed valve position, computes Cv for each, then reports the mean and coefficient of variation. Any reading that drifts more than ±2% from the mean is flagged as an outlier before the point is accepted — this is what keeps the resulting characteristic curve statistically defensible rather than just a single noisy data point per position.

Characteristic Curve Generation

Once Cv has been measured at every travel position, the rig fits the normalised curve against the two standard valve characteristics: linear (Cv/Cv_max tracks travel directly) and equal-percentage (Cv/Cv_max follows an exponential rangeability curve). Whichever fit clears an R² of 0.95 is reported as the valve's characteristic; if neither does, the curve is flagged as non-standard or quick-opening for engineering review.

Seat Leakage Test — ANSI/FCI 70-2

Seat leakage classification follows ANSI/FCI 70-2: Classes II, III, and IV are evaluated as a fraction of rated Cv capacity (0.5%, 0.1%, and 0.01% respectively, scaled by test differential pressure), while Class VI is measured directly in bubbles/min and converted to an equivalent mL/min limit based on port diameter. The rig runs the appropriate check automatically based on the declared class and returns a clear pass/fail against the allowable limit.

The Neometrix CV and Control Valve Test Rig implements this measurement chain — from raw sensor data through ISA 75.02 averaging, characteristic curve generation, and ANSI/FCI 70-2 leakage classification — with automated report output for each tested valve.
→ https://neometrixgroup.com/products/cv-and-control-valve-test-rig