I think there are two design strategies for low field MRI detection coils. Strategy 1 is to wind a lot of coils to increase the induced voltage. Strategy 2 is to increase the SNR with a high Q with a small number of turns.
Strategy 1 is often used in proton-magnet meters for geomagnetic measurements. If the magnetic field is low and it is not affected by the skin effect or proximity effect, it is better to wind a lot of coils.
In the case of a low magnetic field, it is better to wind the coil so that the thermal noise due to the resistance of the coil and the input conversion noise of the preamplifier are about the same.
I think Strategy 2 will be used in low field MRI with some magnetic field. In the range I investigated, the number of turns of the 6 to 10 mT MRI receiving coil is about 10 ~60 turns.
http://www.imagtech.it/?q=products/RFCoils
(10mT 30turns)
https://www.researchgate.net/publication/282809852_Very_Low_Field_MRI_A_fast_system_compatible_with_Magnetoencephalography
https://www.nature.com/articles/srep15177
(6mT 30turns)
https://onlinelibrary.wiley.com/doi/pdf/10.1002/mrm.28396
(50mT 40turns)
The reason for the low number of turns is that the skin effect and proximity effect increase resistance. To avoid the proximity effect, the space between the wires of the receiving coil should have a clearance of 1 to 1.5 wires.
Another reason is that the larger the number of turns, the greater the effect of parasitic capacitance. If the number of turns is large, the capacitance of the resonance capacitor becomes small, and the resonance frequency tends to change significantly due to the change in the parasitic capacitance when the imaging target is changed. Therefore, there is an upper limit (Larmor frequency = self-resonant frequency) for the number of coil turns.
Bandwidth is limited by the Q of the coil being too high, but it is not a good idea to solve this by connecting resistors in series (SNR deteriorates). Active Q damping methods that do not worsen SNR have been proposed.
https://hal.archives-ouvertes.fr/hal-02069159v2/document
Saddle type coils are often used for operational convenience, but it is better to use solenoid type coils because SNR is a particular problem in low field MRI.
An MRI simulator can be created from the Bloch equations and the magnetization formula of water. By creating a hand shape with 3D CAD, dividing it into voxels, and calculating the voltage induced in the RF coil, it is possible to verify the validity of the designed RF coil.
Assuming that the Q factor is 70 and the preamplifier is 60 dB, the voltage input to the ADC will be about 0.1 to 0.5 V. (In the case of my hand)