In MRI systems the radio frequency (RF) transmitter produces the RF current (oscillator) and delivers it to the transmitting coil (antenna). The RF signal produced by the transmitter is used to excite the protons in the imaging volume.

The schematic figures of a pulse sequence timing diagram illustrate the steps of basic hardware activity that are incorporated into a pulse sequence. Time during sequence execution is indicated along the horizontal axes. Each line belongs to a different hardware component. One line is needed for the radio frequency transmitter and also one for each gradient (G_s = slice selection gradient x, G_f = phase encoding gradient y, G_f = frequency encoding gradient z, also called readout gradient).
In picture 1, a timing diagram for a 2D pulse sequence is shown.
Slice selection and signal detection are repeated in duration, relative timing and amplitude, each time the sequence is repeated. A single phase encoding component is present each time the sequence is executed.
Additional lines are added for ADC (Analog to Digital Converter) and sampling. A gradient pulse is shown as a deviation above or below the horizontal line. Simultaneous component activities such as the RF pulse and slice selection gradient are indicated as a non-zero deviation from both lines at the same horizontal position. Simple deviations from zero show constant amplitude gradient pulse. Gradient amplitudes that change during the measurement, e.g. phase encoding are represented as hatched regions.

The second picture shows a timing diagram for a 3D pulse sequence.
Volume excitation and signal detection are repeated in duration, relative timing and amplitude, each time the sequence is repeated. Two phase encoding components are present, one in the phase encoding direction and the other in slice selection direction (irrespectively incremented in amplitude) in each time the sequence is executed. A description of the comparison of hardware activity between different pulse sequences.

(SAR) The Specific Absorption Rate is defined as the RF power absorbed per unit of mass of an object, and is measured in watts per kilogram (W/kg).
The SAR describes the potential for heating of the patient's tissue due to the application of the RF energy necessary to produce the MR signal. Inhomogeneity of the RF field leads to a local exposure where most of the absorbed energy is applied to one body region rather than the entire person, leading to the concept of a local SAR. Hot spots may occur in the exposed tissue, to avoid or at least minimize effects of such theoretical complications, the frequency and the power of the radio frequency irradiation should be kept at the lowest possible level. Averaging over the whole body leads to the global SAR.
It increases with field strength, radio frequency power and duty cycle, transmitter-coil type and body size. The doubling of the field strength from 1.5 Tesla (1.5T) to 3 Tesla (3T) leads to a quadrupling of SAR. In high and ultrahigh fields, some of the multiple echo, multiple-slice pulse sequences may create a higher SAR than recommended by the agencies. SAR can be reduced by lower flip angle and longer repetition times, which could potentially affect image contrast.
Normally no threatening increase in temperature could be shown. Even in high magnetic fields, the local temperature increases not more than 1°C. 2.1°C is the highest measured increase in skin temperature. Eddy currents may heat up implants and thus may cause local heating.

FDA SAR limits:

IEC (International Electrotechnical Commission) SAR limits of some European countries:
All limits are averaged over 6 minutes. (For more details: IEC 60601-2-33 (2002))

In most countries standard MRI systems are limited to a maximum SAR of 4 W/kg, so most scanning in level II is impossible.
For Level I, in addition to routine monitoring, particular caution must be exercised for patients who are sensitive to temperature increases or to RF energy.
For Japan different SAR limits are valid.

A coil is a large inductor with a considerable dimension and a defined wavelength, commonly used in configurations for MR imaging. The frequency of the radio frequency coil is defined by the Larmor relationship.
The MRI image quality depends on the signal to noise ratio (SNR) of the acquired signal from the patient. Several MR imaging coils are necessary to handle the diversity of applications. Large coils have a large measurement field, but low signal intensity and vice versa (see also coil diameter). The closer the coil to the object, the stronger the signal - the smaller the volume, the higher the SNR. SNR is very important in obtaining clear images of the human body. The shape of the coil depends on the image sampling. The best available homogeneity can be reached by choice of the appropriate coil type and correct coil positioning. Orientation is critical to the sensitivity of the RF coil and therefore the coil should be perpendicular to the static magnetic field.

RF coils can be differentiated by there function into three general categories:
The RF signal is in the range of 10 to 100 MHz. During a typical set of clinical image measurements, the entire frequency spectrum of interest is of the order 10 kHz, which is an extremely narrow band, considering that the center frequency is about 100 MHz. This allows the use of single-frequency matching techniques for coils because their inherent bandwidth always exceeds the image bandwidth. The multi turn solenoid, bird cage coil, single turn solenoid, and saddle coil are typically operated as the transmitter and receiver of RF energy. The surface and phased array coils are typically operated as a receive only coil.

See also the related poll result: '3rd party coils are better than the original manufacturer coils'

Duty cycle is the time during which the gradient system can be run at maximum power. The duty cycle is based on the total time and includes the cool down phase. The duty cycle on the RF pulse during MRI is restricted based on the specific absorption rate (SAR) limit.
SAR limits restrict radio frequency heating effects. The specific absorption rate increases with field strength, radio frequency power and duty cycle, type of the transmitter coil and body size. The especially in high and ultrahigh magnetic fields, important SAR issue can be readily addressed by reducing the RF duty cycle due to longer repetition times (TR) and the use of parallel imaging techniques. A TR longer than the minimum needed provides time for the tissue to cool down, but for the cost of a longer scan time. A parallel imaging technique reduces the RF exposure and the scan time.

See also High Field MRI.