Do temp sensors go bad?
I have a MacPro 2010 (5,1) dual-processor 3.48GHz machine. It has been running great for over ten years. The OS I use on this is Mojave 10.14.6.
My MacPro is in a studio with the room temp at 68F. The internal ambient idle temp is 88F. When a strong load (30+ tracks in Nuendo) is on the machine, the average processor temp is 125F. This has been the case for several years. For maintenance purposes, I periodically open the case and blow out any dust particles across all three layers/chambers, including the system/processor tray and across the grid-fins of the heat sink of the machine using nitrogen gas. I use the App, "System Monitor" by Marcel Bresink, to monitor this machine's dynamic statistics.
On Friday July 23, 2021, I installed the Safari and Security Update 2021-005. After several soft-reboots the Desktop appeared; I logged in and all appeared to be running normal. At the end of the day I performed the nightly full system backups. Next I shutdown the system for the weekend.
On the following Monday, the machine failed to load the OS to the desktop login after power-on; at approx. 65% during the loader, the machine rebooted itself. This cycle continued for 3 times, prior to my shutting it down and proceeding to diagnose and evaluate the system. The error message reported on screen was than of an unknown system error. I initiated the Apple Hardware Test; an error code relating to a Temp Sensor fault is what triggered the system loader fault. This error also trashed the OS loader on the HDD.
I was able to initialize and install the Mojave OS on a smaller HDD for testing purposes on this machine. All worked correctly. The above aforementioned error only occurs when attempting to execute and apply the Mojave Security Update 2021-005, a cold-system-start.
As a precaution, I will clean, re-apply fresh arctic paste and re-seat both of the processor heat sinks as well as the Northbridge heat sink.
This not so interesting thing about the diagnostic Apple Hardware Test (v3A207), is that it does not designate which sensor is reporting the error, nor does it indicate what the temp was that triggered the purported fault.
I believe this fault interrupt is a false positive created by a sensor failure.
Has anyone else experienced this issue?
