Simulation and Measurement Results

This work is fabricated in TSMC 0.18µm 1.8V/3.3V process with 1.13mmx1.435mm shown in Fig. 3.15, which includes charge pump (A), non-overlapping clock generator (B), effective PFM block (C), output driver and sensing circuit (D), ring oscillator (E), and frequency divider(F).

The measurement setup is shown in Fig. 3.16, where Agilent E3631A is utilized to provide the fixed 1.8 V for 1.8-V normal device. Agilent 81110A generates non-overlapping duty cycle 10% 800Hz clock signal for VST and VSTB. TDS 3054B is used to observe output current of the stimulus driver. The first measurement step is measure the ring oscillator and frequency divider. The measured frequencies of the ring oscillator and frequency divider are 28.08MHz and 2.83MHz respectfully shown in Fig. 3.17. These frequency is the choice for the proposed PFM block. As the maximum frequency is 28.08MHz, the maximum output voltage is 9.9V when abc is 100 and output loading resistance is 220kΩ. With different capacitance, the electrode voltage would be different. Fig. 3.18 and Fig 3.19 show the electrode voltages when C is 4nF and C is 200nF. With small capacitance, the electrode voltage increases with time obviously. Fig. 3.20 shows the electrode voltage when R is 200kΩ and 290kΩ.

With the resistance difference of 90kΩ, the dc voltage difference is approximately 1.5V. That is, the capacitance affects the change of the electrode voltage with time.

The resistance influences the dc voltage of the electrode voltage. Fig. 3.21 shows the comparison of the stimulus currents flowing through R between different C. Fig. 3.22 shows the comparison of the stimulus current waves between different R.

Fig. 3.15. Die photograph of the new proposed biphasic stimulus driver.

Fig. 3.16. The measurement setup of the power supply, the function generator and the oscilloscope. The outputs of stimulus driver connect to test board with the electrode equivalent circuit.

(a)

(b)

Fig. 3.17. Measured frequency: (a) ring oscillator : 28.08MHz (b) the frequency divider combining Johnson counter and binary counter : 2.83MHz.

Fig. 3.18 The comparison of the measured operating electrode voltages VL1 and VL2 between big C (C=200nF) and small C(C=4nF).

Fig. 3.19 The comparison of the measured electrode voltages VL1 between big C (C=200nF) and small C(C=4nF).

Fig. 3.20 The comparison of the measured electrode voltages VL1 between different loading resistance R=200kΩ and R=290 kΩ.

Fig. 3.21 The measurement result of stimulus current between electrode resistance R=200kΩ and R=290kΩ when electrode capacitance is 4nF and the digital signal abc is 110.

Fig. 3.22 The measurement result of stimulus current between electrode resistance R=200kΩ and R=290kΩ when electrode capacitance is 200nF and the digital signal abc is 110.

As the stimulus current tunability, the proposed design would provide the stimulus current 20µA, 30 µA, and 40µA with different digital signals shown in table 3.1. The measurement results are closed to the specification with different loading impedance. For the measurement results shown in Fig. 3.23 and Fig. 3.24, as the electrode resistance smaller, the stimulus current would be larger when the 3-bit digital signal abc is 111(40µA) and 100(20µA). When the value of electrode resistance is larger than 200kΩ, the amplitude of stimulus current decreases. Since the speed of the current sensing circuit is low, the speed of controlling charge pump(stopping) is low. That is, the value of VCC generator is easily higher than expected. Therefore, the measured stimulus current is larger than simulation results no matter which post-layout or pre-layout shown in table 3.2.

Fig. 3.23 The measurement results of stimulus currents with electrode resistance from 10 kΩ to 473 kΩ when electrode capacitance is 4nF. As abc is 111, the target current is 20uA. As abc is 110, the target current is 30uA. As abc is 100, the target current is 40uA.

Fig. 3.24 The measurement results of stimulus currents with electrode resistance from 10 kΩ to 473 kΩ when electrode capacitance is 4nF. As abc is 111, the target current is 20uA. As abc is 110, the target current is 30uA. As abc is 100, the target current is 40uA.

Fig. 3.25 The measurement results of current mismatch between Icathodic and Ianodic with electrode resistance from 10 kΩ to 473 kΩ when electrode capacitance is 4nF. As abc is 111, the target current is 20uA. As abc is 110, the target current is 30uA.

As abc is 100, the target current is 40uA.

Fig. 3.26 The measurement results of current mismatch between Icathodic and Ianodic with electrode resistance from 10 kΩ to 473 kΩ when electrode capacitance is 200nF As abc is 111, the target current is 20uA. As abc is 110, the target current is 30uA. As abc is 100, the target current is 40uA.

Fig. 3.27 The measurement results of VCC generator with electrode resistance from 10 kΩ to 473 kΩ when electrode capacitance is 200nF. As abc is 111, the target current is 20uA. As abc is 110, the target current is 30uA. As abc is 100, the target current is 40uA.

Fig. 3.25 and Fig. 3.26 are the measurement results of the stimulus current mismatches. As electrode capacitance is 4nF, the current mismatch is 0.568µA in average. As electrode capacitance is 200nF, the current mismatch is 0.51µA in average. The output voltage VCC of VCC generator is shown in Fig. 3.27. As resistance R larger, VCC becomes higher.

Power is generated by VCC generator mostly. When the loading resistance(electrode resistance) is small, the required VCC is small. As VCC is smaller than 5V, the power consumption is smaller than 0.7mW. As VCC is larger than 5V, the power consumption becomes larger. The maximum consumption power is 2.21mW. The minimum power is 0.31mW. The standby power is 166µW.

(a)

(b)

Fig. 3.28 The measurement results of power consumption (a) when the electrode capacitance is 4nF (b) when the electrode capacitance is 200nF. As abc is 111, the target current is 20uA. As abc is 110, the target current is 30uA. As abc is 100, the target current is 40uA.

Table 3.2 Comparison specification, pre-simulation, post-simulation and measurement results.

Table 3.3 Table for the impedance adaptive range.

abc I

stim Adaptive loading range

111 20µA (C,R) = (4nF, 10kΩ) ~ (4nF, 475kΩ) (C,R) = (200nF, 10kΩ) ~ (200nF, 475kΩ) 110 30µA (C,R) = (4nF, 10kΩ) ~ (4nF, 250kΩ)

(C,R) = (200nF, 10kΩ) ~ (200nF, 290kΩ)

100 40µA (C,R) = (4nF, 10kΩ) ~ (4nF, 210kΩ) (C,R) = (200nF, 10kΩ) ~ (200nF, 220kΩ)