Li-Ion Charger for Implantable Devices - Selection of optimal charge algorithm and implementation

Abstract

Deep Brain Stimulation (DBS) is a method for treating diseases like Parkinson’s Disease (PD) and various other movement disorders. An Implantable Pulse Generator (IPG) is placed in the chest and wires are run through the neck to provide electrical pulses to a certain target in the brain via an electrode. Results with this treatment are good but today’s stimulators are bulky and last only three to five years. Furthermore the electrode and pulse generator are often placed in two different surgeries. The device should be miniaturized so that it can be placed in the skull while inserting the electrode. This reduces the number of placement surgeries and eliminates lead failures. The size of the device is dominated by the battery volume, however no compromises on battery life can be made since battery replacement requires, again, surgery. Switching to a rechargeable Lithium-Ion battery that can be recharged through the skin is a way to reduce battery size without compromising the lifetime of the device. Lithium-Ion batteries are, however, delicate devices and should be charged and discharged with care. The classic method to ensuring safe charging of these types of batteries is the Constant Current Constant Voltage (CCCV) method. Here a fixed current is pushed into the battery until the battery reaches its maximum voltage. After this the voltage is kept constant and the current slowly decreases to a small fraction of the initial current. Although this method ensures safe and complete charging of Lithium Ion (Li-Ion) batteries it also has some downsides. The abrupt switch from constant current to a constant voltage might introduce stability issues. Furthermore, characteristics of a battery vary depending on their state of charge and this is not taken into account by this algorithm. In this research a possible new algorithm that smoothly regulates charge current is investigated. It turns out that this algorithm is slightly less efficient and slower than the classical CCCV method. The best method to charge the battery of an implantable device is via the CCCV method. An implementation of this method is proposed where both Constant Current (CC) and Constant Voltage (CV) phases are controlled by the same control loop and the transition between the two phases is a smooth hyperbolic tangent function. An inductively powered system was designed that reaches an average efficiency, in circuit simulation, of 87.36% from antenna output to battery. For this implementation, the rectifier is the only part in series with the inductive link and the battery. The efficiency is therefore mostly limited by the diodes used for the AC to DC rectifier. An prototype of the designed system was built and the correct operation of the system was verified by wirelessly charging a 160mAh Li-Ion battery. The results of the performed measurements showed that the inductive link had a poor power transfer efficiency. Mismatch between the resonance capacitor value and coil inductances caused the resonance frequency to deviate from the desired value. The efficiency of the charger was, as expected, mostly determined by losses in the rectifier. Further research should be done in optimization of the inductive link to provide good efficiency even when the device is poorly aligned with the charger. Furthermore the parameters of the system should be tuned to a medical grade battery and the possibility of combining the power transfer with data communication can be investigated.