Hypercapnic respiratory failure is usually caused by an overload of the respiratory muscles (respiratory pump). After treatment of the underlying disease, mechanical ventilation will achieve optimal treatment success and higher degrees of respiratory muscle unloading will improve the outcome in terms of lower PaCO (2) levels and improved exercise performance. Routinely assisted modes are being used for ventilation, where the patient has to trigger the ventilator with his effort. Controlled ventilation is usually applied in sedated patients lacking spontaneous breathing efforts that are necessary to trigger the ventilator. Controlled ventilation, however, is feasible in awake patients but requires operator expertise. In this process, the respiratory pattern of the ventilator has to be adapted to the patient's own respiratory pattern. Changing conditions require a re-adaptation of parameters. In order to automatise this complex and time-consuming operation, a time-adaptive mode (TA-mode) has been developed. This programmed mode incorporates a self-learning algorithm, primarily detecting the patient's respiratory pattern. The software then calculates a matching flow profile using a motion equation that gives consideration to resistance and compliance. The operator has to pre-select allowed ranges of parameters (especially in- and expiratory pressures, IPAP and EPAP). After detection of a stable respiratory pattern (usually after 10 - 20 breaths), the ventilator will slowly increase the calculated flow profile and achieve controlled ventilation without irritating respiratory centres of the brain. Respiratory drive will cease usually within three to five minutes. Restart of the respiratory drive, for example, after coughing or during REM sleep with an altered respiratory pattern will be detected as ventilator fighting and the programme will return to the analysis algorithm again. After the respiratory pattern has become stable, the ventilator will take over ventilation again. The new mode has been validated in an accreditation study. For this purpose we selected 21 patients with stable hypercapnic respiratory failure, most of whom (20) had previously been ventilated with a controlled T-mode and only one patient had previously been ventilated with an assisted mode and adapted them to the new ventilator under polygraphic surveillance. Each time seven patients were adapted to a T-, ST- and TA-mode, respectively. Two patients, however, could not be adapted to ST-mode ventilation and were switched to TA-mode. PCO (2) values before and after ventilation were not significantly different between modes. Patient satisfaction was rated very good in 34 %, good in 45 % and non-gratifying in 21 % of cases ventilated with TA-mode. Consideration has to be given to the fact that patients previously had been receiving optimal ventilator treatment. The TA-mode is a self-learning system, capable of copying the patients own breathing pattern while awake, in order to achieve complete unloading of the respiratory muscles through controlled ventilation during a circumscribed period.