ALLPCB
1. Background and Overview
1.1 Development objective
China has a large population and limited medical resources, and hospitals are frequently overcrowded. Intravenous infusion is a common treatment method used by many patients. However, many infusion devices still provide only basic functions and lack intelligent or informational control. This contributes to disorder in infusion wards, heavy workloads for nursing staff, and inconvenience for patients. Current hanging infusion sets often have flow-control mechanisms but cannot indicate remaining infusion time or provide end-of-infusion alerts, which prevents patients from arranging their time and requires them to call nurses when infusion is finished, creating noise and extra effort.
1.2 Functional features
The designed device can adjust patient infusion flow rate in real time, display remaining infusion time to indicate infusion progress clearly, and automatically issue a warning when infusion is about to finish.
1.3 Novelty and practical value
Compared with traditional infusion devices that only provide infusion and simple speed adjustment, this device adds time monitoring and end-of-infusion alert functions, allowing patients to set infusion duration and coordinate other activities. For hospitals, the device provides timely infusion status information to support orderly replenishment and completion of infusions, improving workflow efficiency. The device is designed for easy installation at the top hook of existing infusion stands and does not require modification of traditional devices.
2. Design Principles
2.1 Overall design principle
A suspended infusion bottle is subjected to two main forces: the downward gravitational force G and the upward tensile force F. Since F = G, the bottle remains stationary in the air.
The liquid inside the infusion bottle can be analyzed by standard hydrostatic relations. From the derivation it follows that the tensile force F exerted on the hanging bottle is linearly related to the liquid volume V in the bottle. By measuring F, the liquid volume V can be obtained.
Based on this relation, the system measures the change in tensile force F over a short time interval and uses the current tensile force F to compute the estimated time t remaining until infusion completion.
Note: F denotes the tensile force corresponding to the liquid mass in the bottle, excluding the tare weight of the bottle.
2.2 Hardware design principle
The hardware challenge is accurate measurement of the liquid weight. To address this, a high-precision 3 kg load cell (force sensor) is used. The sensor is a bridge-type pressure sensor.
Resistance-bridge pressure sensors work by attaching strain gauges to an elastic element. When the elastic element deforms under load, the strain gauges change resistance. The resistance changes are converted to voltage changes, and by measuring output voltage and applying calibration, the weight of the measured object can be derived.
Although the sensor has high precision, its output voltage is still very low and cannot be read directly by a microcontroller ADC. The chosen solution uses the HX711 sensor front-end.
HX711 is a 24-bit analog-to-digital converter designed for high-precision electronic scales. Compared with similar chips, HX711 integrates necessary peripheral functions such as a regulator and an on-chip clock, offering high integration, fast response, and strong anti-interference performance. This reduces system cost and improves reliability. The HX711 interface and programming with an MCU are simple: control signals are driven from pins without requiring internal register programming. Input channel selection connects to a low-programmable-gain amplifier. Channel A provides programmable gain of 128 or 64, corresponding to full-scale differential input signals of ±20 mV or ±40 mV respectively. Channel B is fixed at gain 64, used for system parameter checking. The internal regulator can provide power to the external sensor and the ADC, eliminating the need for a separate analog supply on the system board. The on-chip clock oscillator requires no external components. The chip also provides an automatic reset on power-up to simplify initialization.
After 128x internal amplification and 24-bit ADC conversion by the HX711, the resulting digital value meets the precision requirements of this project.
Because of the sensor's high precision, the power supply must be stable. A 5 V linear regulator is used to provide a stable, low-ripple supply.
2.3 Software design principle
The software challenge is accurately obtaining infusion weight and rejecting external interference.
Following the overall principle, the software subtracts the load cell tare and the empty bottle weight from the total measured weight to obtain the liquid weight. By sampling the liquid weight every 1 s and calculating the weight change, the remaining infusion time can be estimated. In practice, infusion rate is very slow and the per-second weight change may be below sensor resolution. To address this, the algorithm waits until the weight has changed by one digital unit within n one-second sampling periods, then multiplies n by the per-unit time to estimate remaining time.
Because the bottle may be subject to external forces and motion, the software applies a simple first-order filter to the computed weight-change rate to reduce high-frequency noise. To reduce the effect of large transient disturbances (for example, a manual shake), the program accepts data only when three consecutive computed remaining-time values are consistent. This approach reduces false readings caused by external disturbances.
3. Project Design
3.1 Hardware design
Power is derived from 220 V mains to a 5–12 V supply, then regulated to 5 V using an LM2940-5.0 linear regulator to power the MCU, sensor, display, and wireless module.
Because the change in output from the load cell is small, amplification is required before ADC. Since the HX711 provides an internal programmable amplifier of up to 128x, no external amplifier is needed.
For wireless communication in a compact infusion ward with many hanging points, Zigbee would be appropriate. At the time of development, an XL02-232AP1 module was used instead.
The LCD used for development was a 128x64 0.96-inch yellow-blue display available in the lab.
Due to time constraints, no PCB was fabricated. A minimal system was constructed on an existing development board using jumper wires. Photographs of the prototype were taken.
3.2 Software design
The software is divided into two parts: the lower-level MCU firmware and the upper-level host application. The lower-level firmware samples the infusion weight at the configured interval, computes remaining time using the algorithm, updates the local display, and transmits data via the wireless module to the host. The host application collects data from multiple lower-level nodes, classifies and sorts the data, and displays it on a user interface.
Lower-level firmware
The MCU firmware performs initialization (timers, serial port, load cell tare, etc.) and then waits for the 1 s sampling period. At each sampling event, the MCU reads the HX711, processes the data, and discards readings that do not meet validity criteria.
Host application
The host application opens a specific communication port, waits for data from the lower-level nodes, validates incoming packets against a simple protocol (A5 + ID + AA + remaining time), and, if valid, sorts and displays the data; invalid packets are discarded.
4. Test Results
After debugging, measurements were performed using medical infusion bottles.
Short-term error is near zero. Over longer periods, the maximum error observed in our tests was 15.7%, corresponding to a maximum time deviation of 11 minutes, which is acceptable for many clinical scenarios.
5. Conclusion
The design centers on a Shenlianhua microcontroller with a load-cell sensor and a wireless module to provide real-time infusion time feedback to staff and patients. The system addresses long-standing issues in infusion management by providing remaining-time estimates and end-of-infusion alerts, which can improve order and workflow in infusion areas. Prototype measurements produced acceptable accuracy. The design is compact, low-cost, and extensible for broader deployment in clinical environments.
