The study of cardiac dynamics is a topic of great clinic and experimental concern. However, the ability to apply methods from the emerging field of nonlinear dynamics to cardiac time series has been limited by the demand for large amounts of relatively artifact-free data. Both the ability to collect such data and the capacity to analysis such large data sets represents limits. In the present paper we describe a system that allows for the collection of large amounts of high quality data and the analysis of large data sets. The recording system consists of a miniature, single-module electrocardiogram-recording device. This module consists of an integrated three-electrode device that is attached to the chest of the subject. A low power 8-bit micro-controller detects the R-spike and stores the time between R-spikes in milliseconds on a 512 KB EEPROM. This system can record continuously for over four days. The output of this system is down-loaded to a PC and the RR intervals fed to a suite of digital signal processing programs. Among other things, estimates of Approximate Entropy and Poincaire plots are generated. This system will expand the capability of researchers and clinicians to investigate nonlinear cardiac dynamics in ambulatory subjects.