This paper outlines the theory and procedures for calibrating an echo integration acoustic system with a standard sphere. It presents the results of an extensive calibration of a Simrad EK500 scientific echosounder with a 120 kHz split-beam transducer in a refrigerated 10 m deep tank. Calibration parameters are studied in relation to sphere material (WC and Cu), water temperature (0.5 to 5.5°C), transmitted pulse length (0.1, 0.3 and 1.0 ms), target depth (0.8 to 7.5 m), and time (149 days). The total range in TS gain, including the effects of temperature, standard sphere and time, is 2.9 dB. Gain values calibrated with the Cu sphere are, with the exception of the long pulse length, lower than those from the WC sphere. The general trends are consistent between SA gain and TS gain. The total spread in SA gain, including the effects of temperature, standard sphere and time is 2.1 dB. Thus, the accuracy of the standard sphere as a reference TS value, the pulse length, the water temperature range, and equipment instabilities during the duration of a survey can contribute significant errors to the accuracy and precision of an echo integration acoustic survey. To minimise these effects, the TS gain and SA gain parameters should be meticulously measured, measured frequently, and matched to the pulse length used and the water temperature in the survey area.