If you operate a Siemens CT scanner — Sensation, Definition AS/AS+/Edge, or Flash, the IRSmx reconstruction tower family — you will eventually see this pop-up: “Fatal Error in Image Reconstruction System. Please shut down, switch off, and restart the system.”It's alarming, it stops scanning, and it's genuinely the single most common IRS-related event reported by field engineers. Here's what's actually happening, and the order in which to check it.

It's usually not the tower — it's a resync problem

The event log typically shows codes such as CT_ISV 29, CT_SSQ 328, CT_ISV 3 or CT_ISV 31 alongside this message. That pattern is documented in Siemens service manuals as a communication loss between the console (ICS) and the reconstruction tower (IRS) — not a hardware death. It happens when the console, tower and UPS lose synchronization, most often from an out-of-sequence power-down: one side stays energized while the other drops, or a UPS event leaves the two out of step.

What clears it, most of the time: a full, controlled power-down of the console, the tower and the UPS to completely de-energized, a pause of roughly 30 seconds, then a clean power-up so the systems resynchronize together. Follow that with the Hardware Test → Network check in Local Service to confirm the link is healthy before you scan.

This is worth internalizing before your next service call, because the instinct when a “Fatal Error” pop-up appears is to assume the tower needs replacing. In our experience and in Siemens' own service documentation, that's the wrong first move far more often than it's the right one.

The dashboard has a trap: a dead tower can look silent, not alarmed

The tower has a small LED panel most engineers glance at during triage — Power On, Power Fail, Fan Fail/Overheat, LAN, SATA/DVD. It's tempting to read “no lights flashing” as “nothing's wrong,” but there is a specific failure mode where that reading is wrong.

A fully dead power-supply set cannot light its own fault indicator.The Power Fail LED is powered by the same supplies it reports on — if all of them are gone, the annunciator has no power to blink with. A dark panel with mains present at the input is not proof the supplies are healthy; it can be exactly what a total power-supply failure looks like. If a tower is unresponsive and the panel looks unremarkable, don't rule out the power-supply modules on LED evidence alone — check for actual output at the branch, not just the indicator.

The honest ranked list of what actually fails

  1. Console-to-tower communication/resync fault— an event, not a hardware failure, and the most common trigger of the “Fatal Error” pop-up.
  2. Disks and the RAID array — the leading true wear item. Boot failure, hardware-test disk failure, or slow and failed reconstruction are the usual symptoms.
  3. Power-supply modules — a blinking Power Fail LED (faster blink means more than one module affected) is the normal presentation; a fully dead set can present as no indication at all, as above.
  4. The fiber-optic link between tower and gantry— field consensus among service engineers is that this connection “can look fine but is usually the fault” when receiver boards show a signal error. Reseat and inspect rather than trusting a visual check.
  5. Fans and cooling— a solid Fan-Fail/Overheat LED means the mainboard is overheating; a flashing one means a fan has failed. Two different problems sharing one indicator, so read the pattern, not just whether it's lit.
  6. Software version mismatch between console and tower — happens after a reinstall or a tower swap, and produces symptoms that mimic hardware faults. Matching versions exactly resolves it.

How often does an IRS actually fail?

There is no published failure-rate or MTBF figure for the Siemens IRS — not from Siemens, not from ECRI, not from any parts vendor we've checked.Any vendor quoting you a specific failure percentage for this system is asserting a number, not citing one, and it's worth asking for their source. What can be said honestly, based on how deeply this part is stocked across every generation of the tower by every major independent supplier: in a well-powered, fixed-site installation the reconstruction system tends to be one of the more reliable subsystems on the scanner, and a full tower failure is uncommon.

Mobile and trailer-mounted CT is a different environment, and it's reasonable to expect more events there— transport vibration, thermal cycling, condensation and less stable shore or generator power are all present at higher intensity than the fixed-site conditions most reliability assumptions are built on. We treat “an IRS-related event every one to three years rather than every five to ten” as a defensible planning assumption for mobile operations — stated plainly as our judgment, not as a manufacturer's published figure, because none exists.

What actually protects a reconstruction system

  1. Disciplined shutdown sequencing, every time.This single habit prevents the most common documented fault. Make sure your UPS configuration can't leave the tower energized while the console drops power, or the reverse.
  2. Power quality, especially on a mobile platform. On any trailer installation, verify the shore/generator transfer-switch neutral bonding. An incorrectly bonded neutral exposes every identically-rated supply in the branch to the same overvoltage event at once — a recognized mechanism for multiple power-supply modules failing together, which is a far worse outcome than a single-module failure. Logging power events gives your next service call evidence instead of a guess.
  3. A proper online double-conversion UPS with isolationahead of the reconstruction system's power branch, with batteries replaced on a calendar schedule — sooner in a temperature-cycled trailer.
  4. Filter and fan service. Dust loading is a real, documented issue on this tower family. Treat filter service as routine, not optional.
  5. Fiber inspection after any work that flexes the runto the gantry, given how often that link is the quiet root cause of a “board” fault.
  6. Keep console and tower software versions matched, and re-verify after any reinstall or tower swap.
  7. Monitor disks for early warning signs and replace proactively rather than waiting for a hardware-test failure — standard enterprise practice, not a manufacturer-mandated interval.

Seeing this error right now? Call 236-988-6666before you order a tower — we can walk you through the resync sequence on the phone, and if it is hardware, we'll tell you plainly what's actually failed. Or request service.