Power supply systems which are currently installed in the two major experiments at the LHC, ATLAS and CMS (which represent the state of the art in the field), are based upon an architecture with three levels of voltage reduction. The first layer implements the conversion from alternating current to direct current (AC / DC conversion), with values ??of the output voltage in the range from 300 VDC to 48 VDC, and it is usually installed out the hostile area. The second layer reduces the voltage to a value around 5 VDC, and it is installed in mildly hostile area, a few meters or tens of meters from the detectors to be powered. The third layer is usually on the detectors, i.e. on the boards of the front-end electronics, and it is made ??by linear low-dropout voltage regulators (LDO), without any remote control and with low efficiency, providing the correct voltage to Front-end boards, typically ranging from 1.2 V to 3.3 V.
The connections between the different levels are implemented with electrical cables of suitable cross section, and the distances covered by the latter, related to the size of the above experiments, are such that approximately one third of the total output power is dispersed in the environment instead of being used front-end electronics, with serious consequences on the detector thermal stability and on the sizing of environmental conditioning systems. To give a picture, the total power delivered by the first level of AC / DC converters of the muon detectors of ATLAS is of about 150kW.
Currently some elementary integrated components (chips) for the DC / DC conversion are under analysis, whose performance in terms of radiation and magnetic field tolerance seems promising. These components, suitably inserted in a power supply system, may allow the reduction of the dispersion of power of a factor of 10.
The power supply for a PET-MRI system is in principle simpler compared to power supply systems for large experiments of high energy physics. The number of detectors is less (order of 10000 channels) and the power consumption, of the order of kW, is for instance considerably lower than 150 kW supplied from the first supply stage of ATLAS.
The current strategy is to use power supplies that provide voltages of 12 V to the back-end cards that are located a few meters away from the PET-MRI system (in the low magnetic field area). On these cards a series of LDO are mounted, properly shielded, to generate the required voltages to the Front-End circuits which are located within the magnet.
The power for the photodetectors is generated by dedicated power supplies and flows directly into the high-field using shielded cables and connectors which are not magnetic. The voltages of the active components placed in the PET system, which is located inside the magnet, vary from 1.2V to 3.3V, while the power for the photodetectors (in the specific case Silicon photomultipliers, photodiodes operating in Geiger regime of high gain, extreme compactness and insensitivity to magnetic fields) varies from 25 V to 35 V. Furthermore, since the performance of SiPM is temperature dependent, these detectors require current monitoring and active stabilization of the gain via remote control.
In the combined PET-MRI technology the strict requirements are related to the compactness, power dissipation and noise induced in the MRI system. The PET detectors and electronics of the front-end are located within the magnet and must occupy a volume which allows a comfortable positioning of the patient. For example, in the system dedicated to PET-MR neurological imaging in development for the TRIMAGE project, the available space for the detectors ring of the PET system is extremely reduced. The PET system is made ??of an insert placed in the core of a superconducting magnet of 1.5 T and confined in a crown of only 13 cm inner diameter and 30 cm outer diameter. The options for shielding and heat dissipation are very limited in this geometry.