Portable ECG Monitor: Front-end Amplifier
2.4 Isolation between Digital and Analog Circuits
A portable monitor mixes digital and analog circuitry, so routing must address mutual interference. Digital circuits operate at higher frequencies while analog sections are sensitive. Keep high-frequency signal traces as far as possible from sensitive analog components. Although the PCB has only one connection to external ground, inside the PCB the digital ground and analog ground should be routed separately and only tied together at the external interface point. A ferrite bead, also called an EMI/RFI suppressor, provides good high-frequency suppression and can shield conductive interference on wires and cables. In this design a 6.8 uH ferrite bead is used to isolate the analog and digital power supplies.
2.5 Component Placement and PCB Layout
A PCB can contain small-signal analog circuits, high-speed digital circuits, and high-frequency GPIO/RS signals. To avoid coupling and interference between these sections, separate placement for different circuit types is a fundamental PCB design principle. Maintain sufficient spacing between sections and route their traces separately.
Power routing, including VDD and ground (VSS), is a critical part of EMI control. VDD and VSS areas should be maximized to reduce electromagnetic emission and to ensure low impedance between high-frequency signals and ground.
2.6 Shielding for Analog Circuits
Weak ECG signals are buried in interference induced by ambient electromagnetic fields. Those induced voltages couple to the body and lead wires through capacitance or inductance. High-frequency digital or RF signals can substantially affect the small analog signals feeding the ADC.
In addition to separating analog and digital sections on the PCB, add a metal shield over the analog circuitry to further reduce interference.
3 Software Strategies for Interference Suppression
When random interference mixes with the input signal, analog filters can remove unwanted components to improve signal quality. Analog filtering is difficult at low and very low frequencies; digital filters do not have those limitations. Digital filters offer high precision, reliability, and stability, and are widely used to reduce random noise. Digital filtering can be implemented in the frequency domain, using FFT to perform discrete Fourier transforms and filter by desired frequency characteristics, or in the time domain, using difference equations on sampled data. The time-domain approach is simple and practical. Digital filters are classified by impulse response into finite impulse response (FIR) and infinite impulse response (IIR). IIR filters have effectively infinite memory but fewer computational terms, and for the same order they can provide higher selectivity. To balance filtering quality with computational speed, an IIR filter is used here.
3.1 Low-pass Filter Design
To remove high-frequency components above the ECG band, a low-pass filter is used. The chosen sampling frequency is fs = 200 Hz, so T = 5 ms. Based on the ECG frequency range, select a cutoff frequency fc = 99 Hz. Using a filter design table, select passband ripple rp = 0.1 dB and stopband attenuation rs = 60 dB, with filter order n = 4. According to Matlab design results, the algorithm formulas are shown below.
3.2 Notch Filter Design
Experiments show that an analog notch filter alone is not sufficient to eliminate 50 Hz mains interference. A digital filter is therefore used to suppress mains noise. With sampling frequency fs = 200 Hz (T = 5 ms), an elliptic filter is chosen. Design parameters: passband edges wp1 = 45, wp2 = 55; stopband edges ws1 = 49.8, ws2 = 50.2; passband ripple rp = 0.1 dB; stopband attenuation rs = 60 dB. Matlab design results yield the algorithm formulas shown below.
Hemodialysis Pump and Fluid Control (System Description)
Two identical chambers are separated by a high-elasticity composite diaphragm. Under hydraulic pressure the diaphragm flexes left or right. Each chamber has inlet and outlet ports connected to a three-way solenoid valve. The three-way valves control switching of dialysis fluid in and out of the chamber. During operation the lower and middle ports of valve V1 are connected and the pump drives liquid into chamber A, expanding volume to the left. At the same time valve V2 connects its upper and lower ports so the fluid in chamber B is expelled under diaphragm pressure and passes through a delivery sensor. The delivery sensor monitors flow in the chamber; when a chamber is emptied it signals to switch valves V1 and V2 so that flow alternates between chambers. With fixed chamber volumes, counting valve switches detected by the delivery sensor enables accurate calculation of dispensed dialysis fluid volume.
The dialysis fluid from the delivery sensor enters an average flow control unit composed of a flow control valve and a flow feedback valve. The flow control valve is a three-way controllable valve; the outlet flow equals inlet flow minus feedback. By adjusting the feedback amount, average dialysis fluid flow is controlled. The feedback valve is driven by a DC motor, with precision about ±5 ml/min. Average flow is continuously adjustable from 400 to 600 ml/min. When A-fluid sensors detect conductivity and temperature within set ranges, bypass three-way valve V14 connects the left and right ends and valve V15 opens to allow fluid into the dialyzer. Spent fluid after ion exchange exits the dialyzer, passes filters, V15, and a dialysis pressure sensor into a blood leak detector. The blood leak detector senses red blood cells to prevent dialysate leakage into blood if the dialyzer membrane ruptures.
The detection sensitivity range is 200 to 1000 ppm, i.e., 0.2 to 1 ml of blood per liter of dialysate can be detected and will trigger an alarm. The dialysis pressure sensor monitors dialysate pressure to prevent excessive pressure that could rupture the membrane and compromise safety.
Spent fluid from the blood leak detector enters a deaeration chamber via a negative-pressure pump. Air accumulates at the top of the sealed chamber; a float switch opens valve V8 to purge air, ensuring accurate flow through the discharge chamber. Spent fluid then passes through the discharge chamber, a discharge sensor, a pressure offset device (POD), and a heat exchanger before leaving the machine. The negative-pressure pump is a DC variable-speed pump; pump speed is controlled according to the set patient fluid removal rate (0 to 5.99 L/h) with control voltage from 8 V to 21 V. Pre-pump pressure (dialysate pressure) ranges from +70 mmHg to -400 mmHg. The discharge chamber and discharge sensor operate like the delivery chamber and sensor, enabling precise measurement of expelled fluid. The patient's actual fluid removal equals discharged volume minus delivered volume; by precisely controlling pump speed, the removal rate and total removal can be controlled with accuracy around ±30 ml/h. The POD is connected via valve V7 to the patient's venous pressure port to reduce TMP fluctuations caused by venous pressure changes, improving ultrafiltration accuracy. The heat exchanger uses the temperature of spent fluid to preheat incoming reverse-osmosis water.
In "disinfection" or "acid wash" states the main water path remains operational, but bypass valve V14 connects the right and upper ports and valve V15 is closed to prevent fluid entering the dialyzer for safety. A- and B-fluid suction tubes are placed into their respective wash chambers; valves V9, V10, V11 open and disinfectant or acid wash is pumped through the system for cleaning.
In "rinse" state, the main water path is the same as in disinfection but A and B pumps reverse, flushing the main line and wash chambers and discharging the wash fluid at the outlet. This process also rinses the suction tubes and chambers.
The JMSS DS-20 hemodialysis water circuit differs from other brands by omitting an ultrafiltration pump. A negative-pressure pump performs both deaeration and ultrafiltration functions. Together with inlet/outlet chambers and the POD, this simplifies the water circuit and reduces machine cost while retaining the same functional performance.
4 Conclusion
As a portable monitor, the hardware should be simple and compact. To meet these requirements, this article proposed a unipolar supply approach for the front-end amplifier and discussed ECG interference mitigation strategies. The emphasis here was on hardware solutions for interference suppression. Software-based interference suppression is also important for portable instruments and will be covered in detail in subsequent work on ECG preprocessing.
