Portable devices for the diagnosis of obstructive sleep apnoea (OSA) are considered to be an acceptable alternative to polysomnography (PSG), but their validation is essential. The aim of our study was to validate a device specifically designed for OSA diagnosis. Twenty nine suspected OSA patients were studied with simultaneous nocturnal PSG and an unattended recording device (MicroDigitrapper-S) (M-S). The device measured body position, snoring sound, oronasal flow, thoracic and abdominal effort, heart rate and percentage arterial oxygen saturation (Sa,O2%). We compared the apnoea plus hypopnoea index (AHI) and Sa,O2% results of PSG with that of the system's automatic analysis (M-SA). We also performed a semiautomatic analysis (M-SS) with visual editing of the raw data. Results at different AHI cut-off levels were analysed to obtain an indication of accuracy in diagnosis and severity. Both M-SA and M-SS showed a sensitivity and specificity of 100% at the cut-off level of AHI > 10. When increasing the cut-off levels, M-SA sensitivity decreased (55% for AHI > 40), while specificity remained high (95%). This was improved to a clinically acceptable level of agreement by M-SS analysis (sensitivity 91% and specificity 94%). In conclusion, the MicroDigitrapper-S device showed a good sensitivity and specificity for the diagnosis of OSA. However, the device could not predict the severity of OSA precisely enough. In severe cases (apnoea plus hypopnoea index > 40), semi-automatic scoring was necessary to obtain a more accurate detection of the severity of the disease.