Temperature behavior under maximum and real load
The two thermographs show the WD Red Pro 26 TB after one hour of stress testing in an open setup. The drive was apparently written to continuously in order to test the thermal behavior under maximum load. The measurements were taken with a thermal imaging camera at an emission factor of 0.95, which is appropriate for painted metal surfaces and printed labels. The front cover of the hard disk, typically a smoothly closed cover without active components, reached an average temperature of 43.9 °C according to the measurement, with a maximum value of 44.8 °C and a minimum of 30.8 °C in the edge area. These values are completely uncritical. The homogeneous heat distribution indicates an even heat input from internal components, presumably via the panel stack housing and the supporting structure.
A different picture emerges on the rear: Clear thermal hotspots can be seen here, especially near the top right corner, where the controller IC and possibly the motor connection are typically located. The maximum value is 50.3 °C, the minimum 29.1 °C, the average surface temperature is 45.0 °C. This is also still within the tolerable range, but shows that the electronics become noticeably warmer than the mechanical part under continuous load. The spot heating in the corner is characteristic of active voltage converters and controller chips.
The next IR image shows the WD Red Pro 26 TB after 45 minutes of real use in the productive environment of the OBS server. Two parallel streams with continuous recording were running on the drive, which corresponds to a typical scenario for a permanently high, but not maximum burst-like write load. The measured average surface temperature is 36.0 °C, with a maximum value of 40.4 °C and a minimum of 29.7 °C at the cooler edge zones. The temperature distribution in this image is significantly more homogeneous than in the previous stress test under synthetic load. The lack of pronounced hotspots, both on the cover plate and in the surrounding area of the mounting cage, is particularly noticeable. This indicates effective heat dissipation due to the installation in the server frame and a stable thermal connection, presumably supported by airflow within the housing. The measured temperatures remain well below the 45-degree mark and are therefore completely uncritical.
In comparison to the previously recorded stress test values with temperatures of up to 50.3 °C on the rear, it is very clear here how much the thermal status depends on the actual usage profile and the housing environment. While synthetic benchmarks specifically target maximum stress on the mechanics and electronics, the real recording load is at a moderate level with even activity. The lower temperatures also indicate that the drive does not run continuously in full write mode, but is temporarily relieved by buffering, caching and data stream-related idle phases. Even though the rear is not visible in this image, the low temperature gradient on the front suggests that the electronics have not reached critical values either. The 40.4 °C measured as the local maximum value indicates convective heat displacement from the interior rather than an active electronic hotspot.
This measurement thus shows that the use of the WD Red Pro 26 TB in the OBS server is thermally completely unproblematic in real streaming operation. The moderate surface temperature confirms a stable thermal balance in the overall system, which indicates successful integration into the server housing and a functioning ventilation concept. We can state that no thermal restrictions are to be expected, even under continuous load in everyday streaming, as long as the ventilation concept is maintained.
The temperature curve in the AIDA64 stress test is remarkably stable and forms the average value. At the start of the test, the temperature rises rapidly to around 39 °C within the first five minutes. The temperature then stabilizes at exactly 40 °C and remains constant there for the entire duration of the test, without any noticeable fluctuations or intermediate rises. This temperature is in the absolutely uncritical range for a mechanical drive and indicates a functioning cooling system with an even thermal load. No thermal saturation effects or reheating phases can be seen, which speaks for a good thermal design.
The lower part of the image shows the CPU load during the test period. This averaged around 3 to 5 percent over the entire period, with occasional outliers up to a maximum of around 6 percent. The very low and stable curve confirms that the load from the hard disk test had practically no influence on the overall system and that the CPU never became a bottleneck. It can be deduced from this that no bottlenecks occurred in the I/O subsystem or interrupt processing, even under continuous load.
Surface vs. internal temperature and evaluation
The measured surface temperatures do not reflect the actual internal temperatures 1:1. In the area of the spindle, the read/write heads and in particular the central motor (voice coil and spindle), temperatures can be significantly higher than the external temperatures under continuous load, typically by 10 to 15 K. The central disk stack can also reach 55-60 °C internally, even if only 45 °C is visible externally. The electronics of these hard disk models are usually mounted on the base plate with thermal pads, so that there is a relatively good thermal connection.
Western Digital specifies an operating temperature range of 0 to 65 °C for the Red Pro series, whereby a maximum of 60 °C is usually specified as the recommended continuous load temperature. Above 55 °C, the risk of thermally induced ageing effects increases, particularly in the electronics and bearings. Temperatures above 60 °C in continuous operation should be avoided as they can significantly worsen the MTBF values (Mean Time Between Failures). Short-term peaks of up to 65 °C are tolerable, but lead to degradation over a long period of time.
Kommentieren