Circuits for the Charge Push-through Electronics: Power Efficient Signal Processing Inside the Artificial Cochlear Implant

Available online at www.sciencedirect.com

ScienceDirect Procedia Engineering 168 (2016) 1710 – 1713

30th Eurosensors Conference, EUROSENSORS 2016

Circuits for the Charge Push-Through Electronics: Power Efficient Signal Processing Inside the Artificial Cochlear Implant Jaromir Zaka *, Jaromir Hubaleka, Jan Praseka, Jan Pekareka, Vojtech Svatosa, Zdenek Hadasb, Daniel Dusekb a

Faculty of Electrical Engineering and Communications, Brno University of Technology, Technicka 10, CZ-61600 Brno, Czech Republic b Faculty of Mechanical Engineering, Brno University of Technology, Technicka 2, CZ-61600 Brno, Czech Republic

Abstract This work deals with the new technique called Charge Push-Through technology which is more energy efficient than the currently used approaches. The new energy efficient design regarding utilization of emerging technology such as Energy harvesting (EH) power sources is very promising for future development of energy independent (zero-power) and nonobtrusive cochlear implant. This work presents the solution of Charge Push-Through circuits using the component level design of artificial cochlea. © by Elsevier Ltd. This is an openLtd. access article under the CC BY-NC-ND license © 2016 2016Published The Authors. Published by Elsevier (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference. Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference Keywords: artificial cochlea; Charge Push Through; zero power system;

1. Introduction Today‘s artificial cochlear implants basically use two different approaches for design of signal processing electronics. The first approach is based on frequency decomposition of microphone analogue signal by band pass filters. Decomposed signal is processed by standard analogue techniques to drive the nerve stimulatory output. The second approach translates analogue signal into a digital value which is divided into single frequencies by Fast Fourier Transform (FFT) in real time and processed by digital signal processing unit (DSP). Against these principles, this work deals with the new technique called Charge Push-Through [1] which is more energy efficient. The power consumption of most electronic devices including artificial cochlear implants can be divided to useful (output) current and negligible (internal) currents [2]. The output current value depends on the bio-electric interface respectively on the stimulation current required by nerve terminals inside the human cochlea [3]. * Corresponding author. Tel.: +420 54114 6288

E-mail address: [email protected]

1877-7058 © 2016 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of the 30th Eurosensors Conference

doi:10.1016/j.proeng.2016.11.496

Jaromir Zak et al. / Procedia Engineering 168 (2016) 1710 – 1713

1711

It has not to be decreased to ensure the appropriate function of artificial cochlear implant. This power consumption depends on the ambient acoustic pressure and it is not stable in time. The maximal peak value of bipolar current pulse is approx. hundreds of micro-amperes during the micro-seconds pulse widths [4]. The bipolar pulses have to be charge balanced but the pulse shape is not strictly specified [4]. The main power losses usually consist of negligible power supply currents of internal electronic circuits. These currents should be minimized without an effect on the device output. The new Charge Push-Through technique is based on the idea that output current is also used for internal circuits powering. In this case, the charge drawn from power supply is pushed-through the internal circuits (it works as power supply current) into the output electrode (it works as useful stimulating current) without negligible leakage current [1]. To minimize the power consumption in practical point of view, the standard versatile design techniques cannot be used. The circuits have to be designed in close context to the final application using special techniques such as the above mentioned Charge-Push Through technique. The basic principles of the Charge-Push Through technique adapted to the target application - artificial cochlear implant, as well as behavioral simulation model of the device were already defined [5]. The results of these simulations were used as start point for future design of the component level model described in this work. 2. Electronic Circuits Design A simplified block diagram of the circuit is shown in Fig. 1. It consists of acoustic sensors with AC voltage output, operational transconductance amplifier (OTA) with rectifying output, integration capacitor, control comparator and output buffer [5]. The main power source is used only for the acoustic sensor powering and as the charge source for the first stage of the electronic circuits – OTA. The OTA output current is proportional to the level of acoustic pressure and it is collected in the integration capacitor. The charge from the capacitor is used for powering of the output bipolar current pulse generator - output buffer which is controlled by comparator. The output buffer is switched on when its threshold voltage is reached. The charge of the generated pulse is defined by internal capacitor value and hysteresis of the comparator.

Fig. 1. Block diagram of the Charge Push-Through electronics.

The output buffer is designed as a part of the main comparator. The power consumption is minimized by reduction of the components using this combination (see Fig. 2). The comparator model was already designed and discussed [6]. The output buffer consists of the complementary pair of low V th MOSFETs and output capacitor (CBUF). The output capacitor decouples direct current but its main purpose is to generate negative part of output bipolar pulse. The capacitor is charged by output current when the positive pulse is generated by MOSFET transistors. After that, the input electrode of capacitor is grounded by buffer transistor and accumulated charge is discharged through the implanted electrode. The discharge current forms the negative part of the pulse. By this technique, the total charge balance of the output pulse is guaranteed and the charge drained from the power supply is half of the total charge of the output pulse. The capacity of the internal capacitor can be calculated from known parameters of the output bipolar pulse and electrical parameters of the comparator. The total charge of the positive pulse QPOS was determined by standard output pulse current and width to 7.5 nC [4]. The hysteresis voltage U HYST was set by comparator simulations to 0.2 V [6]. The ideal internal capacity was established by (1) to 37.5 nF.

‫ܥ‬ூே் ൌ

ொುೀೄ ௎ಹೊೄ೅

ൌ

ூುೀೄ ή௧ುೀೄ ௎ಹೊೄ೅

ൌ

଻Ǥହήଵ଴షవ ଴Ǥଶ

ൌ ͵͹Ǥͷ݊‫ܨ‬

(1)

1712

Jaromir Zak et al. / Procedia Engineering 168 (2016) 1710 – 1713

Fig. 2. Schematic diagram of the comparator and output buffer inside the processing electronics.

The OTA construction and parameters depend on the C INT capacity, acoustic sensors characteristics and dependency of the output pulses period on the acoustic pressure. The CINT capacity was calculated above (1) and the pulse period dependency is defined by bio-electric interface. The sensor used in the new concept of the cochlear implant was developed instead of this work. It was fabricated as mechanically trimmed frequency selective nitride membrane using MEMS technologies [7]. This concept eliminates losses caused by electrical frequency decomposition because the frequency of the signal captured by the sensor is mechanically preselected. The new sensor was connected to the Wheatstone’s bridge with reference structures and then it was characterized by the automated workplace which consists of anechoic box, acoustic generator and measuring devices connected to the personal computer and appropriate software programmed especially for this purpose. In this configuration, the acoustic pressure is automatically trimmed to the value required for measurement in the real time. The electric signal generated by acoustic sensor under characterization process is pre-amplified by low-noise amplifier as closest to the sensor as possible. The next amplification is realized by Lock-In amplifier on the defined frequency which is matched to the frequency of driving acoustic signal. The reference frequency is directly given by generator synchronization output. By the measurements, the coefficient K of sensor resistance change compared to the acoustic pressure change were established to the value of 556 ppm·Pa-1. The reference output voltage of the sensor for the human talking acoustic level was calculated to the 22 μV P-P by (2). ܷ ൌ ܷௌ௨௣௣௟௬ ή ቆͳ െ

ଵ ಼ή೛ೌ೎೚ೠೞ೟೔೎ మ

ଵା

ቇ

(2)

The parameters of the OTA can be calculated using values described above. AC current generated by OTA have to charge internal capacitor (CINT) to the comparator hysteresis voltage UHYST in the time which is equal to the period between output pulses as it is defined by (3). The peak current during the human talking has to be 11.8 μA as it is calculated using (4). In comparison with voltage generated by MEMS sensor during human talking, transconductancy of the OTA has to be 1 S as it is shown in (5) for one-way rectification. ‫ܫ‬ൌ

஼಺ಿ೅ ή௎ಹೊೄ೅

‫ܫ‬௉ா஺௄ ൌ ݃௠ ൌ

ൌ

௧ ூήగ ଶ

ൌ

ଶήூಾಲ೉ ௎ೄಶಿೄ

ଷ଻Ǥହήଵ଴షవ ή଴Ǥଶ ଵήଵ଴షయ

଻Ǥହήଵ଴షల ήଷǤଵସ

ൌ

ଶ ଵଵǤ଼ήଵ଴షల ଶଶήଵ଴షల

ൌ ͹Ǥͷߤ‫ܣ‬

ൌ ͳͳǤͺߤ‫ܣ‬

ൌ ͳܵ

(3) (4) (5)

The OTA is designed as a standard differential mode instrumentation amplifier (INA) with voltage to current converter as it is shown in Fig. 3. There are ultra-low power operational amplifiers used for this purpose. The power consumption of each amplifier is below 500 nW [8]. The output current of the OTA is driven into internal load (RINT) resistor and this current is mirrored by rectifying current mirror to the internal capacitor (C INT). The total power consumption without output current of the whole OTA is below 2.5 μW.

Jaromir Zak et al. / Procedia Engineering 168 (2016) 1710 – 1713

Fig. 3. Diagram of the OTA simulation circuit.

All circuits were designed and tested at 1 V power supply voltage. The function of the circuits was proven by complete model simulation. The dependency of the frequency of generated output pulses on acoustic pressure was linear as it was demanded. The total power consumption of the device was depressed to 3.5 μW for each frequency channel. The power consumption is determined mainly by OTA, except 1 μW which is consumed by transducing piezoresistors in Wheatstone’s bridge. This power consumption can be decreased by using piezoelectric transducers. 3. Conclusion There were several blocks of the Charge Push-Through electronics designed as a behavioral model for fully implantable artificial cochlea. These blocks were simulated earlier and their optimal parameters were defined. In this work, the blocks were designed as component level simulation models with respect to previously simulated optimal values of their parameters. The simulation models were proven by functional simulations and their power consumption was tested by modified simulations. The total power consumption was lower than power consumption of standard cochlear implants, approx. 3.5 μA per each channel at 1 V power supply voltage. There are still problems which have to be solved but this technique brings some benefits which can be used in a fully implantable zero-power artificial cochlea together with EH technologies in the near future. Acknowledgements Research described in this paper was financed by Czech Ministry of Education in frame of National Sustainability Program under grant LO1401. For research, infrastructure of the SIX Center was used. References [1] J. Zak, Z. Hadas, J. Pekarek, D. Dusek, V. Svatos, J. Prasek et. al., “Model-based Design of Artificial Zero Power Cochlear Implant,” Mechatronics: The Science of Intelligent Machines. In Press, ISSN 0957-4158. [2] S. M. Kang and Y. Leblebici, “CMOS digital integrated circuits-analysis and design,” Tata McGraw hill, 2003, ISBN 063-9785503910. [3] R. Sacheli, L. Delacroix, P. Vandenackerveken, L. Nguyen and B. Malgrange, “Gene transfer in inner ear cells: a challenging race,” Gene Therapy, 2012, vol. 20(3), pp. 237-247, DOI: 10.1038/gt.2012.51, ISSN 0969-7128. [4] L. Norgia, G. Tognola and C. Svetlo, “Measurement of Electrode Current Pulses from Cochlear Implants,” IMTC 2004 - Instrumentation and Measurement Technology Conference, Como, IT, 2004, pp. 1697 1700, ISBN 0-7803-8248-X. [5] J. Zak, Z. Hadas, D. Dusek, J. Pekarek, V. Svatos, J. Prasek et. al., “Design of the Charge Push-Through Electronics for Fully Implantable Artificial Cochlea,” Progress in Biomedical Optics and Imaging: Bio MEMS and Medical Microdevices II. Bellingham, USA: SPIE Microtechnologies, 2015, vol. 16(50), pp. 1-9, ISSN 1605-7422, ISBN 978-16-284-1641-1. [6] J. Zak, Z. Hadas, D. Dusek, J. Pekarek, V. Svatos, J. Prasek et. al., “The Charge Push-Through Electronics Design for Fully Implantable Artificial Cochlea Powered by Energy Harvesting Technologies,” Microsystem Technologies, pp. 1-15, in press. [7] J. Zak, Z. Hadas, D. Dusek, J. Pekarek, V. Svatos, J. Prasek et. al., “Design and Fabrication of Fully Implantable MEMS Cochlea,” Procedia Engineering: 25th DAAAM International Symposium, Vienna: Elsevier Ltd, 2015, 100, pp. 1224-1231, ISSN 1877-7058, ISBN 978-3-901509-99-5, DOI: 10.1016/j.proeng.2015.01.487. [8] S. Tripurari and V. Bhadauria, “Ultra low-power rail-to-rail linear sub threshold bulk-driven transconductor,” International Conference on Power, Control and Embedded Systems (ICPCES), IEEE, 2014, pp. 1-6. ISBN 978-1-4799-5910-5, DOI: 10.1109/ICPCES.2014.7062826.

1713