MAGNETOM Symphony — Engineer Field Guide
Tribal knowledge on a 25+ year-old 1.5T that refuses to die. Back to: MAGNETOM Symphony 1.5T.
The three things that kill a Symphony
- Cold head neglect. The cryocooler runs 24/7. Skipped PMs → rising helium boil-off → cryogen fills get expensive fast → eventually a quench. Service the cold head on schedule and monitor helium level weekly. A quench on a Symphony is a $30k–$60k day.
- Gradient cooling faults. The chiller is a consumable. When flow drops, the gradient coil overheats and throttles, scans fail, and if you push through you'll cook the GPA.
- RF amplifier arcing. Late-stage RPA failures announce themselves as intermittent artifact or a sharp pop at high duty cycle. Stop scanning when you hear it.
Cryogen management
- Helium boil-off rate should be <0.03 L/hr on a healthy Symphony. Anything over 0.05 L/hr means the cold head is tired.
- Cold head service interval: OEM says every 3 years. In warm rooms or 24/7 scanning, plan on 2.
- Before any cold head swap: warm up the compressor, have the helium recharge truck on-site, and confirm the shim file is backed up.
- Ramp state: Symphony magnets ramp to field around 472 A. Never, ever vent the bore unless cryogen loss is imminent.
Gradient cooling
- Chiller setpoint: ~18°C supply, 24°C max return.
- Flow alarm = stop scanning immediately. Don't clear the alarm and keep scanning "just to finish the patient."
- Every 6 months: flush the loop and check the inline filter. The mag-room loop collects particulate.
- If the gradient temp trips mid-scan, the GPA log will tell you which axis — replace the loop sensor before the amplifier.
RF chain
- Symphony ships with an 8-channel receiver. Verify channel health in service mode — a channel can die silently and cost you SNR on multi-channel coils.
- RPA (RF Power Amplifier) fault codes: 30-series usually power supply, 40-series usually final stage. A 40-series fault = no more scans until swap.
- Coil connectors: inspect the gold pins at every coil change. Bent = reseat; corroded = stop and clean.
TIM upgrade paths
Symphonies came in multiple TIM tiers — 8x18, 18x46, 32x76 (channels × max coil elements). Buying a refurb:
- Confirm the tier in writing. A Maestro Class
8x18is a very different scanner from a32x76. - Check which coils are included. A Symphony without a body matrix and spine matrix is a gutted system — those are $20k+ each refurbished.
- syngo MR software version sets what applications are licensed. VB17 is a common stopping point; VE-class is newer.
Accepting a refurbished Symphony
- Cryogen level ≥70%, boil-off rate documented
- Cold head hours and last PM date
- Magnet shim file on hand (backup!)
- GPA and RPA fault logs clean, or explained
- Coil inventory matches quote (body matrix, spine, head, neck, flex, extremity)
- syngo MR software version recorded, license dongles present
- Chiller age and last flush date
- Magnet quench history (any? how many?)
Common errors
- "Gradient overtemp" → chiller flow, 95% of the time.
- "RF unit fault" → start at the connectors and cables before pulling the RPA.
- "Shim coil fault" → often the shim power supply, not the coil itself. Cheap fix first.
- "Cold head warning" → not yet critical, but plan the service.
Things nobody tells you
- Helium is the real cost of ownership, not service contracts. A 1000L fill is $20k–$40k depending on market. Sealed-bore systems changed the math for new MRIs — Symphony isn't one of them.
- The magnet itself will outlast the electronics. Most Symphony retirements are because the host computer / syngo MR version is unsupportable, not because the magnet failed.
- RF-shielded rooms leak over time. If SNR drops uniformly, check the door gasket and the penetration panels before you blame the coils.
Contributors
Field-guide entries will be agent-drafted from forum threads + engineer submissions. V1 seeded from community parts networks, community parts sources, and community knowledge.