The SIGNA platform: generations, subsystems and the room as this family builds them

Identify which SIGNA generation and configuration is in front of you, name its magnet, gradient, RF, computer and cooling blocks as GE packages them, and know which parts of the room belong to the scanner.

Where this course starts

In MRI Service — Level 1 you learned the scanner as four subsystems, the safety zones, cryogenics, quench, the room and image QC in general terms. In the Sumitomo cryocooler course you took the compressor and cold head apart. This course does not repeat either. It takes the GE SIGNA 1.5T and 3T family as you will meet it in hospitals that run without a manufacturer contract: HD, HDe, HDx and HDxt, the older Excite systems, the compact Brivo/Creator/Explorer line and the newer Pioneer and Voyager. We go only as far as the sources support, and where a value or a procedure belongs to GE, we say so and send you to the documentation for your system.

The first habit to build is simple: 'Signa' is a brand, not a model. Engineers lose days because a part, a heat load figure or a fault history was looked up for the wrong system. Before you diagnose anything, write down the exact system name and field strength from the data plate and the console, the software release, the magnet type, the cabinet layout and the coils present. A colleague can only help you, and a parts supplier can only quote you, when that line is right.

A second habit: know what has been changed. Many older SIGNA systems have been upgraded over their life, sometimes very deeply, with new gradients, table and host on an original magnet. The software then carries a hardware configuration that must match what is really installed. A mismatch between the two produces faults that look like hardware but are not.

Generations at a glance

For service purposes engineers split the family into three groups. This is a field map, not GE's own classification and not a specification: exact contents depend on the system, its options and its release, and you confirm them on the machine and in its documentation.

The HD group (Excite, HD, HDx, HDxt, and the compact HDe) is the cabinet-based generation. Gradients, RF and the acquisition electronics sit in separate cabinets in the equipment room, the console has a host computer and a separate reconstruction engine, and cooling comes from a chiller loop plus a gradient heat exchanger. The HDe is the economical 1.5T member with a smaller gradient system (a third-party comparison lists about 50 T/m/s slew rate for the HDe against 120 for the HD and HDxt); colleagues repeatedly note that heavy diffusion protocols push it close to its limits, which matters when you judge a 'gradient fault' on an HDe.

The compact group (Brivo MR355, Optima MR360, Creator, Explorer) packs more into fewer cabinets; GE registered Creator and Explorer under the Optima MR360/Brivo MR355 name. In field reports you meet names such as the ICN (the reconstruction computer), a 'driver module' in the cabinet electronics, and a table with its own power supply in the rack. Pioneer (3T) and Voyager (1.5T) belong to the newest group covered here: two main cabinets in the equipment room, an Integrated System Cabinet and an Integrated Cooling Cabinet, both fitted against the shield wall through a copper mesh tunnel, and cables largely pre-attached to the magnet at the factory.

• Data plate: system name, field strength, serial number, date of manufacture.
• Console: software release and the hardware configuration the software believes is installed.
• Magnet: cryostat type, magnet monitor, cold head and compressor model, quench pipe route.
• Equipment room: number of cabinets, how gradients and RF are packaged, where the chiller or cooling cabinet sits.
• Coils and table: which coils exist, which ports they use, fixed or removable table top.
• History: upgrades, board swaps, software reloads, last helium fill, last cold head exchange.

Magnet and cryogenics on this family

The magnet is a superconducting, actively shielded, helium-cooled cryostat; the physics is in mri-l1. What is specific is how GE lets you watch it. A magnet monitor (MagMon in field reports and on the Pioneer/Voyager cable labels, MON in that system's component list) reads the helium level probe, cryostat pressure and cold head and compressor status and keeps a history. It runs even when the scanner electronics are off, and on most sites it must keep running. Its readings are the first thing you check on any visit, and lesson 2 shows how to read and record them.

The cold head sits on top of the cryostat and is driven by a helium compressor through flexible lines; the compressor needs cooling water from the chiller loop. On Pioneer and Voyager the cryocooler compressor is fitted inside the cooling cabinet but takes its power separately from the rest of the system, so a system power-off does not by itself stop it. Know how your site is wired: a compressor left off for hours costs helium. The compressor and cold head themselves are covered in the Sumitomo course; here we only add how their state appears on the GE magnet monitor.

Also in the magnet room is the Magnet Rundown Unit (MRU), the guarded control that deliberately quenches the magnet in an emergency. Every GE superconducting system has one, and it must work. In 2015 GE issued a field correction for its superconducting MR systems after finding units where the rundown unit had been disconnected, and asked sites to test it; that history is the reason this course treats any tampering with the MRU as a red line. The quench pipe leads the helium out of the building; its route, outlet and weather cap are part of the system you inspect.

The gradient chain

The gradient chain runs from the sequence controller to a gradient amplifier with one channel for each axis, X, Y and Z, through filters in the penetration panel and heavy cables to the gradient coil in the magnet bore. On the HD generation each axis has its own control board and power stage. Field reports from HD, HDxt and Creator systems describe a gradient power supply with an energy storage bank, in some versions built with supercapacitors, which supplies the peak current of demanding sequences; GE patents describe such storage for peak power.

Two facts drive most gradient faults. First, the load is very uneven: diffusion and fast echo-planar sequences demand the most, and on diffusion the most heavily loaded axis is often the one that fails first. A system that scans routine protocols happily and stops only on diffusion is telling you about the gradient chain's reserve, not about the coil. Second, everything in the chain is liquid-cooled. The gradient coil and the amplifier are cooled by their own water or coolant circuit through a gradient cooling unit or heat exchanger, and that circuit's cleanliness, flow and temperature decide whether the amplifier will run a long perfusion series without an over-temperature pause.

The axis structure is also your best diagnostic tool. Because the three channels are built alike, swapping a control board or a cable between two axes, per the documentation, tells you whether a fault follows the part or stays with the axis. Lesson 4 uses this repeatedly. Remember that the gradient amplifier carries high voltage and stored energy: you never open it or touch its outputs until it is isolated and its storage has discharged by the documented method.

The RF chain and receive path

On the transmit side the exciter produces the RF pulse at the Larmor frequency of the system, about 63.9 MHz at 1.5T and twice that at 3T, and the RF power amplifier raises it to the kilowatt level. High-power coaxial cables carry it into the room to the body coil and, where the system uses them, transmit-receive coils. The installation material for Pioneer and Voyager warns that these transmit cables are stiff and are ruined if pinched or bent tighter than their minimum radius; treat every transmit cable that way. The body coil contains PIN diode switches that detune it during receive, and a fault in those diodes can appear as a transmit or reset failure far from the coil.

On the receive side most coils are receive-only. They plug into ports on the table, and the signal travels through the table cabling to the receivers and the acquisition electronics, then to the reconstruction computer. Contacts in coil plugs and ports, the table harness that flexes every time the table moves, and the coil identification lines are where much of the receive-path trouble lives. A system that 'loses' coils during an exam, or refuses all coils on one port, usually has a contact or cable problem, not a receiver failure.

The RF chain also needs the room. Any opening in the shield, a door that does not close on its contacts, a damaged filter on the penetration panel or an unshielded cable entering the room lets outside noise in, and the scanner shows it as spikes or a raised noise floor. When you meet RF trouble, lesson 4 will have you rule out the room early.

Cabinets, computers and power

The computing side is a small network. There is the operator host, a reconstruction engine (GE calls it the VRE, Volume Reconstruction Engine; on HD-generation and compact systems its computer is the ICN, Image Compute Node), and in the cabinets several embedded controllers that boot from the host over the internal network. Some boards keep their settings in battery-backed memory. When the internal network switch, a controller or its battery fails, the symptom is a system that will not finish booting, with a controller that does not answer on the network.

Power arrives through a main disconnect panel, then a power distribution unit in the system cabinet feeds the gradient power supply, the RF amplifier, the DC supplies and the computers. On the Pioneer/Voyager system cabinet, one bay holds the RF and gradient amplifiers, the gradient power supply and a power monitor, and the other holds power distribution, the acquisition electronics, DC supplies and the reconstruction computer. The installation sequence requires ground resistance checks between the main disconnect panel and the cabinets, between the cabinets themselves, to the operator workspace and to the table. These ground paths matter for safety and for image quality, and you should know where they are.

Older HD systems often run their consoles on a UPS, and many sites put the whole system behind site-level power conditioning. Know what is protected and what is not before you plan any shutdown.

The room and the cooling plant belong to the scanner

The room is part of the system. It includes the RF shield with its door, door contacts and finger strips, the penetration panel with its filters and waveguides, the copper mesh tunnel between the shield and the cabinets on Pioneer/Voyager, the quench pipe, the oxygen monitor in the magnet room, the patient alert and, on many installations, an emergency exhaust fan for cryogen gas. The mobile-unit manual we used describes such a fan in the space behind the magnet, inspected yearly with its airflow measured by a qualified technician.

Cooling is the other half. A chiller, often with a glycol and water mixture, supplies the compressor and the gradient heat exchanger; on Pioneer/Voyager the cooling cabinet contains the compressor, a gradient cooling unit, a cabinet cooling unit, the patient and body coil blowers and its own control box. Room air conditioning and humidity control protect the electronics. The scanner's reliability is never better than that of the chiller and the HVAC, and a summer that brings shutdowns usually brings them through the cooling plant first.

On-site checklist

• System name, field strength, serial and software release recorded
• Hardware configuration in software compared with installed hardware
• Upgrade and repair history collected from the site file
• Magnet monitor readings noted: helium level, pressure, compressor state
• MRU located, guard intact, last test date known
• Cabinet layout sketched: gradient, RF, computers, power distribution, cooling
• Chiller, gradient cooling and room HVAC identified with their own documentation

Check yourself