Optimization of 4S System

I spearheaded the design and fabrication of a fully automated water‐collection and flushing subsystem that interfaces seamlessly with the 4S test tower. Mechanically, I installed a 120 VAC solenoid valve on the collection manifold, enabling rapid purging of residual fluid between test runs. The plumbing network channels spent coolant into a modular reservoir, where an inline strainer prevents debris ingress. I sourced heavy‐duty fittings and corrosion‐resistant tubing to withstand repeated cycles of wet/dry operation.

On the electronics side, I developed a multi‐stage power distribution breakout board.

• I implemented a logic‐level relay driver that uses 3.3 V control signals to reliably switch the 120 VAC solenoid, enabling precise and repeatable fluid‐level management.
  
  

• A precision op‐amp front end that conditions the 4–20 mA signal from the 0–2 psi pressure transducer, converting it into a 0–3.3 V ADC range on the Arduino Nano,
  
  

• 0.1 % tolerance resistors to guarantee accurate current‐to‐voltage conversion,
  
  

• Varistors on the solenoid input and sensor lines to protect against inductive spikes and EMI when relay switching.
  
  

The firmware, written in Arduino C++, orchestrates sensor diagnostics, data logging, and valve control:

1. Startup Calibration: Compares raw ADC readings against factory transducer curves and stores calibration coefficients in EEPROM.
   
   

2. Real‐Time Measurement: Samples pressure at 100 Hz, applies a 20‐point moving‐average filter, and calculates instantaneous flow rate (L/min) using the inverse of the manufacturer’s flow–pressure relationship.
   
   

3. Adaptive Logging: During active flushing or test phases, logs at 10 Hz; when idle, automatically drops to 1 Hz to conserve SD card space and reduce wear.
   
   

4. Event Triggering: Upon reaching user‐defined pressure thresholds, the code toggles the solenoid valve and annotates the log with “Flush Start”/“Flush End” markers.
   
   

All measurements and status flags are written as timestamped CSV entries to the on‐board microSD card, enabling post‐processing in MATLAB or Python. I validated system accuracy through bench testing against a calibrated pressure gauge, achieving ±0.02 psi agreement across the full 0–2 psi range. This integrated solution demonstrates my end-to-end capability in component selection, PCB design, embedded firmware development, and system validation, resulting in a robust, field-deployable module that enhanced the tower’s data fidelity and operational uptime.