Purpose CZT cameras dedicated to myocardial SPECT are substantially different from conventional Anger ones: they cannot provide planar imaging, and intrinsic performances cannot be evaluated. More over they are equipped with their own specific 3Diterative algorithm dedicated to cardiac imaging. In this scenario, the physicist has to test the performances of the camera in order to verify the correct behaviour, the agreement with the declared technical specifications acquired during the tender and to get the baseline references useful for future comparisons. Starting from the referral literature and the acceptance test protocol of the machine, we present the tests executed. Methods We evaluated the DSPECT (Spectrum dynamic) camera, which has 9 rotating detector column blocks of pixelated CZT detectors with a square-hole (2,26 mm each side) tungsten collimator. So as to perform the intended measurements, we used the following material: Co57 rod source, cylindrical phantom 20 cm diameter, three line sources, Jaszczack phantom, and then a phantom simulating the left ventricle. Results The daily QC test performed with the Co57 rod source gives the following informations: integral uniformity, differential uniformity along X and Y, sensitivity, energy window setting (peaking and energy resolution). Moreover it gives a detector registration index which tells if the detectors are rotating accordingly and shows the relative deviations (a sort of COR alignment test). Then we performed these NEMA tomographic tests: SPECT reconstructed spatial resolution with and without scatter, system volume sensitivity. Using the Jaszczack phantom we obtained: tomographic uniformity, contrast and noise. And finally with the cardiac phantom we checked other interesting parameters more related to the dedicated reconstruction algorithm: count sensitivity, LV wall thickness, contrast between LV wall and inner chamber. Conclusions The tests performed give enough informations to comply the tasks assigned. A dedicated general QC test protocol could be proposed, once having checked its feasibility for other competitor systems. Having a common and easy way of measuring and reporting the main performance parameters for this kind of devices could also help to compare them during the tender.