A fundamental component of this microelectrode driver is a high-voltage switch and it is necessary to
characterize the implemented switch for different stimulation current levels. An analog CMOS switch
can be an n-type or p-type MOSFET which is operated either in cut-off (OFF state) or ohmic region
(ON state). The switch can be characterized by a fixed linear transconductance, gds , which is strongly
signal-dependent and expressed according to Eq. 8 
where, μn and μp are electron and hole mobility respectively in two types of transistors, Cox is the
gate-oxide capacitance, W n and W p are the dimensions of transistors, VDD is the power supply
voltage, and ...view middle of the document...
In our proposed MED, there are 32-high-voltage switches (the
main switch-matrix) which are configured to be connected directly to electrodes. There are four such
switches in each of Node1 and Node2 for establishing connections between these nodes, and Istim1, 2,
VHH , VLL and GND. The output impedance of each group of switch is different from another group,
depending on their locations within the system. We have considered two different scenarios.
Only two switches in the switch matrix, S1 and S2 are connected to the electrical equivalent model of
two electrodes as presented in Figure 7 (left) for one site microstimulation.
Figure 7: Schematic diagrams for (a) calculating the on resistance of different switches in the first scenario (one
monopolar or bipolar stimulation), (b) high-voltage transmission gate, and (c) equivalent circuit of two electrodes.
Figure 8: Schematic diagram for calculating the on resistance of switches in the second scenario (4 monopolar or
Eight switches in the switch matrix, S11 to S24 , illustrated in Figure 8, have been configured for four
sites microstimulation. The same electrical equivalent model of electrode has been used between each
pair of switches.
In both cases, Node1 switch SN11 , dedicated for connecting Istim1, 2 and Node2 switch SN23 , al-
located for connecting VLL have been considered for the simplicity. Switch on resistances for different
groups of switches have been calculated by applying different levels of stimulation currents and pre-
sented in Section 4.
. MICROELECTRODE MODEL
The simplified electrical equivalent model for the electrode, used for post-layout simulations, is pre-
sented in Figure 7(c) . The electrode model is made up of passive elements, Ce and Re , where Ce
is double-layer capacitance at the electrode-electrolyte interface, and Re is the leakage resistance cre-
ated due to the charge carriers crossing the double-layer . Both of these parameters are frequency
dependent. In  the reported values of Ce and Re are 1.14nF and 140kΩ respectively. In our case,
the value of Re has been reduced to 50kΩ to comply with iridium-oxide coated silicon microelectrode
from Blacrock Microsystems. Rm is the resistance of microelectrode metallic part between the back of
the microelectrode and the microelectrode driver, and Rs is the saline spreading resistance. The values
of Rm and Rs are set to 1.5Ω and 11.7kΩ respectively .
Hasanuzzaman, Md., Rabin Raut, and Mohamad Sawan, "A high-impedance microelectrode driver dedicated for visual intracortical microstimulation", 2012 IEEE 55th International Midwest Symposium on Circuits and Systems (MWSCAS).
 Liu X, Demosthenous A, Donaldson N. An integrated implantable stimulator that is fail-safe with-
out off-chip blocking-capacitors, IEEE Transactions on Biomedical Circuits and Systems, 2008;
 Boyer S, Sawan M, Abdel-Gawad M, Robin S, Elhilali...