IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 51, NO. 4, AUGUST 2004 857 Intelligent and Universal Fast Charger for Ni-Cd and Ni-MH Batteries in Portable Applications Juan Díaz, Member, IEEE, Juan A. Martín-Ramos, Member, IEEE, Alberto M. Pernía, Member, IEEE, Fernando Nuño, Member, IEEE, and Francisco Fernández Linera, Member, IEEE Abstract—In this paper, a new fast charger is presented for Ni-Cd and Ni-MH batteries, which are the most frequently used in portable applications. In this charger, the control and supervision of the process has been entrusted to a microcontroller, which pro- vides a powerful and intelligent tool to undertake complex tasks, and reduces the requested circuitry to the microcontroller itself and a few additional components. The resulting charger is able to work out the initial battery state (detecting deteriorated devices), decide the suitable way to charge it (ensuring a long cyclic life), and determine when the charge process must be finished. This way, the state of the battery is always controlled, preventing any damage to it and providing a fully protected operation mode. This paper summarizes the design and construction of the presented charger, as well as shows the experimental results obtained in the prototype tests. Index Terms—Battery charger, design, measurements, software. I. CHARGER OPERATION T HE behavior of Ni-Cd and Ni-MH batteries under charging and discharging process is well known and, also, the temperature and voltage evolution for some fast-charge methods has been studied and described [1], [2]. Bearing these studies in mind, a fast charger capable of fulfilling the operation in about 1 h, minimizing negative effects on the battery life, can be made with commercial purposes. For that, it is necessary to define the following items [3]. A. Fast-Charge Method (FCP) The selected method [2] is based on supplying to the battery a higher value of current than usual. Therefore, the process will finish more quickly. However, to prevent heating damage to the battery, a high charge current cannot be delivered permanently, and it must be interrupted periodically by discharge peaks. Ob- viously, the method itself provides concrete values for the in- tensities and times involved. They are shown in Table I. This current shape determines a voltage evolution, which allows con- tinuously evaluating the charge level and, in this way, detecting when the battery is completely charged, which should occur in about 1 h. B. Intelligent Charge Method In the preceding comments the accent has been put on speed. Nevertheless, an intelligent operation mode has been implicitly Manuscript received January 2001; revised November 18, 2003. Abstract published on the Internet May 20, 2004. This work was supported by FICYT. The authors are with the Área de Tecnología Electrónica, Universidad de Oviedo, Gijón 33204, Spain (e-mail: juan@ate.uniovi.es). Digital Object Identifier 10.1109/TIE.2004.831740 assumed. With it, the cyclic life of the battery is increased by the following coordinated actions: • not to charge half-discharged batteries; • recover overdischarged devices with low charge current; • introduce a novel and safe end of the FCP; • watch temperature damage. To do so, some parameters of the battery must be monitored to acquire information from the process, affording then the chance to correct the energy flux accordingly. These parameters are as follows: • voltage, which gives a measure of the charge level; • current, which must be regulated according to a reference; • temperature, whose increment beyond a limit must inter- rupt the process to prevent thermal damages. The knowledge of these data allows the implementation of a three-step algorithm. 1. To work out the initial state of the battery by means of the measurement of its voltage. That will be done under open circuit conditions at the beginning, and, afterwards, under discharge with a prefixed current (during a prefixed time). 2. To select the sequence of the process. If the battery initial state is known, it is possible to optimize the charge process to increase its cyclic life. The possibilities are: 2.1.- The battery is full-charged (the amount of energy stored on it is above 80%). In this case is not really necessary to per- form the charge. Only a trickle current is supplied to avoid the discharge and to maintain this state. 2.2.- The battery is half-charged (between 10%-80%). Then, to avoid memory effect, it is necessary to discharge it thoroughly before beginning the fast charge process (FCP). The method se- lects the value of the discharging current to minimize the re- quired time in the operation, assuring meantime the absence of any harm. 2.3.- The charge in the battery is low (less than 10%). In this case, if the FCP would be started immediately, the bat- tery could be spoiled. Therefore, it is necessary to perform a previous process of soft-charge (with low current). Under these conditions, the charge level should increase, and when it reaches certain value, FCP can be started. If this would not occur in a prefixed time, it should be assumed that the battery is damaged, and further operation would be useless. Nevertheless, sometimes it is possible to recover the device with several tries. 3. To evaluate when the battery is full charged and the FCP must be finished. The experimentation shows that the process must be cut off when a relative maximum is reached in the battery 0278-0046/04$20.00 © 2004 IEEE