A digital signal processor then processes the digital data for monitoring or diagnosis applications. Biomedical signal acquisition systems typically consist of a low-noise amplifier (LNA), a bandpass filter, an analog sample-and-hold, and an analog-to-digital converter (ADC), as shown in Fig. 1(a). While the architecture shown in Fig. 1(a) is typically used, in some cases chopping technique is used to reduce the impact of the flicker noise, as shown in Fig. 1(b).
With the advancement of CMOS technology, the supply voltage is being reduced, which decreases the voltage headroom for analog block of an IC. Although the technology scaling leads to lower power consumption and higher performance in digital circuits; many parameters [such as signal-to-noise ratio (SNR), dynamic range, gain, and so on] of the analog parts of an IC are negatively impacted. Therefore, it is desirable to find new architectures, in which more digital blocks are used. Recently, a few methods, which are based on digital techniques, are introduced. The block diagram of the system designed is shown in Fig. 2. In this circuit, many of the functions that are typically implemented by analog blocks are performed by digital circuits. Using this digitally enhanced approach can help increase the flexibility of the system in removing unwanted interferences. Moreover, digital calibration techniques can be used more easily.
- Power consumption is high
Fig. 3(a) shows the block diagram of the proposed fully digital architecture. In this structure, the processing of the bio-signal is performed in the time and digital domain. Hence, the advantages of digital CMOS technology are utilized. The analog bio-signal coming from the electrode is directly connected to the front-end circuit and is converted to time with a voltage-to-time converter (VTC). From this point on in the circuit, the signal information is in the phase of the VTC output signal. The output of the VTC is applied to the time-mode processing block, in which the anti-aliasing and offset cancellation are done in time domain. Then, a time-to-digital converter (TDC) transfers the time-mode signal into digital domain where other processes (digital filtering, data compression/reduction and so on) are performed.
The proposed digital architecture is shown in Fig. 3(b). It consists of an active electrode, two digital-to-current converters (DCCs), a moving average VTC (MA-VTC), a control logic block, a counter, and a demultiplexer. In this architecture, ac coupling capacitors are removed, and the impact of the electrode offset on the circuit is cancelled via a feedback loop. The technique used for the offset cancellation will be described in Section III. As explained earlier (Fig. 1), in conventional bio-signal acquisition systems, an LNA is used after the electrode. In the proposed architecture, this block is removed. In the following text, each of the blocks of the proposed architecture is explained.
Voltage to Time Converter (VTC)
In the proposed digital implementation, the analog input voltage is converted to a measurable time via a VTC at the first stage. The signal information is now in the delay of the clock signal (CLK).
Moving Average Filtering
Since, VTCs work with a clock and are broadband compared with the signal bandwidth, not using an anti-aliasing filter before VTCs would lead to out-of-band noise aliasing. To prevent aliasing and to avoid having an analog filter in the design, we have developed the structure shown in Fig. 5 for converting the voltage-to-time as well as anti-aliasing filtering.
Architecture for offset cancellation technique
The block diagram of the proposed offset cancellation technique is shown in Fig. 6. It contains two 5-bit DCCs, control logic circuit, 5-bit counter, and 5–32 demultiplexer (Demux). Since the offset voltage changes very slowly, the frequency of the clock signal used for the counter (Clkc) is ∼10 times less than the clock of the rest of the circuit.
- Low power consumption
- Tanner tools