Check NVMe SSD Health on Windows 10 and 11: PowerShell, SMART Data and GSmartControl

As Oluwademilade Afolabi recently chronicled in MakeUseOf after inspecting his own 512GB Samsung NVMe SSD, relying solely on Windows' native one-word verdict hides the fascinating—and occasionally alarming—paperwork your drive keeps behind the scenes. By pairing native Windows PowerShell commands with an open-source diagnostic utility like GSmartControl, you can pull back the curtain on what your storage has actually endured, learn which scary-looking numbers are completely harmless, and verify your drive's integrity before a sudden failure catches you off guard.

The Baseline: What Windows PowerShell Tells You (And What It Leaves Out)​

Before reaching for third-party diagnostic software, Windows users can run a native baseline check right from PowerShell. Microsoft's Storage module includes the Get-PhysicalDisk cmdlet, which queries any physical disk objects visible across the system's active Storage Management Providers.

To see how Windows summarizes your connected drives, open PowerShell and run:

Get-PhysicalDisk | Format-Table FriendlyName, MediaType, HealthStatus, OperationalStatus, Size -AutoSize

According to Microsoft Learn documentation, Get-PhysicalDisk evaluates drives against four possible HealthStatus states: Healthy, Warning, Unhealthy, and Unknown. In Afolabi's test on his 512GB Samsung NVMe SSD, PowerShell returned a HealthStatus of Healthy alongside an OperationalStatus of OK.

For a five-second sanity check, that output is serviceable. However, Get-PhysicalDisk is designed for high-level storage pool and volume administration rather than granular hardware forensics. While Windows can surface a few basic NVMe indicators in the Settings app on supported hardware, or expose additional counters if you pipe Get-PhysicalDisk into Get-StorageReliabilityCounter, the standard health check omits the nuanced historical logs that tell you how hard the drive has worked and how gracefully it is aging.

Bringing in GSmartControl: A Graphical Window into smartctl​

To read the full autobiography of a drive, you need a tool that queries Self-Monitoring, Analysis, and Reporting Technology (SMART) and NVMe health logs directly. While many Windows enthusiasts default to CrystalDiskInfo, GSmartControl offers a particularly thorough, cross-platform alternative that acts as a graphical front end for the venerable smartctl (smartmontools) command-line engine.

According to the official GSmartControl project portal (gsmartcontrol.shaduri.dev), the current release is GSmartControl 2.0.2, which supports Windows 7 and later—including Windows 10 and Windows 11—via 64-bit and 32-bit installers or portable ZIP archives. Because it wraps smartctl in a point-and-click interface, you get deep diagnostic readouts and self-test triggers without memorizing terminal flags.

Important Security and Hardware Caveats​

If you decide to add GSmartControl to your troubleshooting toolkit, keep two vital caveats in mind:

• Avoid the Phishing Domain: The GSmartControl project explicitly warns users that gsmartcontrol.com is a phishing website. Always obtain the software from the official project site (gsmartcontrol.shaduri.dev) or its official GitHub repository.
• Controllers and USB Bridges Can Block Telemetry: When GSmartControl launches, it scans all connected storage devices, allowing you to double-click a drive to inspect its attributes. While the utility read Afolabi's internal Samsung NVMe drive effortlessly, he noted that a second device connected via a JMicron bridge reported SMART as unsupported. According to GSmartControl's official hardware documentation, the software supports ATA (SATA/PATA) and NVMe drives, as well as drives behind some USB bridges and select RAID controllers (such as Intel Matrix Storage CSMI, Areca, LSI 3ware, and LSI MegaRAID on Windows). If an external enclosure or storage controller does not pass smartctl commands through properly, the drive may show blank or unsupported health data even when the underlying SSD is perfectly fine.

Section Summary: Windows PowerShell's Get-PhysicalDisk provides a fast Healthy or Unhealthy status check, while GSmartControl 2.0.2 unlocks the detailed smartctl telemetry and self-test capabilities of supported internal and external drives.

Decoding the NVMe Numbers: Why the Biggest Counter Isn't Always the Bad One​

Opening the NVMe Attributes tab on a well-used SSD is a bit like reading a car's onboard diagnostic computer after five years of city driving. Without context, raw hexadecimal and decimal counters can easily trigger unnecessary panic. Looking closely at the real-world telemetry reported by Afolabi's 512GB Samsung NVMe SSD illustrates how to separate routine wear from genuine danger.

Endurance: Percentage Used and Data Units Written​

In the NVM Express specification, Percentage Used is a manufacturer-defined estimate of how much of the drive's rated write lifespan has been consumed. On Afolabi's drive, 7% Percentage Used paired with 100% Available Spare shows a drive with plenty of reserve flash blocks left. As NVM Express technical documentation explains, Percentage Used is not a strict countdown timer to death; a drive reaching 100% has simply met its warrantied Terabytes Written (TBW) or Drive Writes Per Day (DWPD) threshold, and the counter is allowed to climb past 100% (up to 255) while the drive continues operating normally.

Right beside that percentage sits Data Units Written. The NVMe standard defines one data unit as 512,000 bytes. Thus, Afolabi's logged count of 98,948,652 Data Units Written multiplies out to approximately 50.66 decimal terabytes (50.7 TB) of lifetime writes across 11,080 power-on hours and 5,089 power cycles. Comparing that total against your specific SSD model's published TBW rating tells you far more about longevity than a generic health badge ever could.

The Scary Number vs. The Real Danger Sign​

What happens when your eyes drift down the page and spot 25,546 Error Information Log Entries? For most users, seeing twenty-five thousand errors on a storage drive is enough to induce cold sweats.

Yet this is where expert interpretation matters. According to NVM Express documentation on SSD error reporting, the Error Log Page records command and protocol discrepancies—such as when an operating system driver, filesystem tool, or diagnostic utility sends a command with an unsupported parameter. In Afolabi's readout, the visible entries consisted mostly of Invalid Log Page or Invalid Field in Command notices referencing Logical Block Address (LBA) zero.

Crucially, the drive's Media and Data Integrity Errors counter sat at 0. That means the NAND flash media itself hadn't suffered unrecoverable read/write corruption or bad sectors. In NVMe telemetry, a non-zero Media and Data Integrity Error count or a tripped Critical Warning bit is a genuine five-alarm fire requiring immediate data evacuation; a bloated Error Information Log filled with invalid command fields is merely harmless administrative chatter between Windows and the drive controller.

Unsafe Shutdowns and Thermal History​

The drive also reported 97 Unsafe Shutdowns. An unsafe shutdown is recorded whenever main power vanishes before the host operating system sends a clean shutdown notification to the NVMe controller—think power outages, tripped circuit breakers, Blue Screens of Death, or holding down the PC case power button. Because modern SSDs rely on volatile DRAM caches or pseudo-SLC write buffers to juggle mapping tables in flight, unexpected power cuts carry a risk of metadata corruption. While 97 historical events over five thousand power cycles isn't cause for panic on its own, a rapidly climbing unsafe-shutdown counter accompanied by missing files, sudden read-only lockups, or a drive vanishing from the BIOS demands urgent investigation.

Finally, checking the thermal telemetry confirmed the SSD was running at a comfortable 42°C, with zero minutes spent above either its warning or critical composite-temperature thresholds.

Section Summary: Raw NVMe logs require careful reading—high Error Information Log counts caused by Invalid Field in Command entries are typically benign if Media and Data Integrity Errors remain at zero, while Data Units Written and Available Spare provide the true measure of flash wear.

Making the SSD Prove Its Claim: Built-In Self-Tests vs. Windows Repair Tools​

Reading historical counters is passive; forcing the drive to run an active diagnostic routine gives you real-time verification. The NVMe specification (specifically the Device Self-Test command standardized in NVMe 1.4) allows host software to trigger offline controller and media integrity checks.

In GSmartControl 2.x, you can initiate both Short and Extended Self-Tests directly from the interface—provided your underlying environment uses smartctl 7.4 or newer for NVMe self-test support. A Short Self-Test typically finishes within two minutes or less, checking SMART critical warning bits, validating controller metadata, and testing key operating parameters. When Afolabi triggered a Short Self-Test on his Samsung SSD, it passed without errors at the 11,080-hour mark.

Knowing Which Tool Fixes Which Problem​

It is vital for Windows administrators and everyday users alike not to conflate hardware self-tests with operating system repair utilities, nor with speed benchmarks:

1. SMART / NVMe Self-Tests (Hardware Layer): Checks the physical drive controller, internal metadata, spare blocks, and flash media health.
2. CHKDSK (Filesystem Layer): Scans and repairs NTFS or ReFS logical volume structures, directory indexes, and file allocation tables. Running chkdsk will not fix dying NAND flash, and a SMART test will not fix a scrambled Windows partition table. (Similarly, Windows tools like SFC and DISM target corrupted Windows system files and component store images, not hardware wear.)
3. Performance Benchmarks (Throughput Layer): SMART counters only track reliability and endurance. If your SSD feels sluggish during large file transfers or game loads, health logs won't show you throughput drops—you will need a dedicated benchmarking tool to measure sequential and random read/write speeds.

Practical Guidance: Backups Always Trump Telemetry​

What should you do if you open GSmartControl today and discover that your Available Spare is dropping, Media and Data Integrity Errors has ticked above zero, or a Short Self-Test fails?

Stop running tests immediately and copy your irreplaceable files to a separate drive or cloud backup.

When an SSD enters genuine hardware decline, subjecting it to lengthy extended self-tests or aggressive chkdsk /r surface scans only places heavier stress on failing flash cells or an unstable controller. And even when your drive passes every test with flying colors—boasting 0 media errors and 100% spare capacity—telemetry is an audit of the past, not a promise of tomorrow. AsNVM Express engineering documentation points out, sudden firmware bugs or catastrophic component failures on the PCB can brick a drive without ever incrementing a wear counter.

Windows can keep its reassuring one-word "Healthy" badge for casual glances. For anyone who values their data, taking five minutes every few months to inspect the real NVMe paperwork—and maintaining an independent, verified backup outside of Windows' basic recovery defaults—is the only storage strategy that truly holds up under pressure.