SMART Drive Health
Technician holding an internal hard drive above a workbench with several other drives stacked nearby
Reference index

The SMART Attribute Database, Explained

Every attribute below is listed by ID, current common meaning, and how Seagate, WD, Toshiba, Samsung, Kingston, and Crucial expose it. Where vendors disagree, both interpretations are shown.

Start with how to read an entry Open the Drive Health Calculator

How to read these entries

Find the number, read the meaning, then check the threshold guidance. The order is deliberate: a name alone will not tell you whether a reading is calm or serious.

Attribute names are not standardized, so the same number can carry different labels between vendors and even between drive families. The three values beside each entry tell you what the drive is actually reporting.

Raw value

The underlying count the drive keeps. Zero is good for an error counter and meaningless for temperature, so read the type before the number.

Normalized value

The drive's own score, usually starting at 100 and falling as the condition worsens. Vendors choose where it lands, so compare it only within one drive.

Threshold

The point where the drive reports failure for that attribute. When the normalized value reaches it, the vendor considers the attribute failed.

Normalized value and threshold travel together; the raw value is the history.

The rules that don't bend

A stable small count and a count that doubles in three weeks describe two different drives.

Treat the direction of travel as the reading. One old reallocation that never grows is a footnote; a pending sector that survives a reboot is a decision deadline.

Four attributes most people search for

If you only look at three things on a failing drive, look at these. Each one explains a different failure mode, and each vendor reports them slightly differently.

05

ID 05 · also written 0x05

Reallocated Sectors Count

The count of sectors already remapped to spare space after read or write problems. A non-zero raw value usually means the drive encountered bad sectors at some point in its life. An increasing value is treated as a stronger warning than a one-time event, because it shows the spare pool is still being drawn down.

It is commonly cited as attribute 05 and appears across Seagate, WD, Toshiba, and Crucial HDD documentation under the same name. Where a vendor adds its own wording, the underlying meaning is unchanged.

Watch for

Any growth from a stable baseline, and growth that keeps pace rather than stalling.

197

ID 197 · SATA convention

Current Pending Sector Count

This counts sectors the drive marked unstable and is waiting to re-test or reallocate. If a later write succeeds, the sector may clear without reallocation. If it fails again, it is remapped and moves into the reallocated pool.

Pending sectors are not just data points; they are a live warning that the drive cannot reliably read some areas. A pending count that returns to zero after a rewrite is a different story from one that stays put or climbs.

Watch for

A non-zero value, and especially one that is still rising when you check again.

09

ID 09 · raw in hours

Power-On Hours

Total hours the drive has been powered. The raw value is in hours on virtually every consumer drive. What it means depends on the drive class: a NAS drive in a Toshiba N300 runs continuously by design, while a WD Blue in a desktop may accumulate far fewer hours over the same calendar life.

Age alone rarely triggers replacement. It is context for the other attributes: a high hour count next to a clean error record is a well-used drive, while a low hour count next to growing errors is a problem.

Watch for

Context-dependent. Read it together with the error counters, not on its own.

194

ID 194 · also 190

Temperature attributes: 194 and 190

Temperature is reported as attribute 194 on many drives and 190 on others, and some vendors use both. The raw value is degrees Celsius. Sustained heat affects HDD platters and SSD NAND in different ways, but heat is a stress that compounds whatever else is wearing the drive.

The drive health calculator treats a hot drive as a watch item rather than an instant failure, because a single spike during a large copy is not the same as a drive that idles at 60°C for months.

Watch for

Sustained readings above 50°C while the drive is under normal workload.

Threshold values table

Thresholds vary by vendor and model. The table below is the general reference used across this site; always cross-check against your drive's own declared threshold before acting on a reading.

Compare Manufacturer Mappings
Attribute ID Raw meaning Typical threshold signal
Reallocated Sectors 05 Remapped sectors Any growth from a stable baseline
Current Pending 197 Unstable sectors Non-zero, especially if rising
Power-On Hours 09 Total running hours Context-dependent
Temperature 194 / 190 Degrees Celsius Sustained above 50°C

A normalized value that reaches the vendor's declared threshold is the drive's own signal that the attribute has failed — not a general rule you can carry between models.

Technician at a workroom bench connecting a 2.5-inch solid-state drive to a bench adapter

SSD-specific: wear level and life left

SSDs replace surface-error logic with endurance logic. There is no platter to scratch; there is a limited number of writes per cell, and the controller tracks how much of that budget has been used.

Crucial documents remaining life indicators and endurance-style metrics; attribute 202 and 231, or the vendor equivalents on your drive, track wear. Samsung exposes much of this through its own utility rather than a single attribute ID, so the same condition can look different depending on what tool you opened.

Reading wear honestly

A high wear level with stable error counts is normal aging, not an emergency. What you are looking for is a wear figure that has moved faster than the drive's age suggests, or a life-left indicator that dropped between two checks close together.

Vendors use different IDs and scales for the same wear story. Read your vendor's mapping before comparing two drives side by side.

Manufacturer labels: where they diverge

Confirm the vendor before acting on any reading. A number that means one thing on a Seagate can mean something else on a Samsung, and the label in the tool window is not the same as the meaning inside the drive.

  • Seagate and WD

    They share most attribute IDs but label some of them differently, and the field order in vendor tools does not always match. Check the ID column, not the row position.

  • Toshiba

    The MG Series occasionally maps fields you would expect to find somewhere else, and the NAS-class drives report running hours that look extreme next to a desktop drive. Those hours are the design, not a fault.

  • Samsung and Kingston

    Both mix SATA and NVMe conventions, so a drive from one product line may carry fields you will not see on another from the same brand. Always confirm the interface before judging an entry.

  • Crucial

    Crucial keeps SSD interpretation separate from HDD bad-sector logic. That separation matters: applying platter-style reasoning to a flash drive gives you the wrong answer.

Where a label is genuinely ambiguous across the industry, this page says so rather than picking a meaning. A guessing reference would be worse than a short one.

Uncorrectable and reported errors: 187, 198, 199

Reported Uncorrectable Errors (187), Offline Uncorrectable (198), and UDMA CRC Error Count (199) round out the HDD error set. The first two are about data the drive could not read back; the third is about the path the data travels.

CRC errors usually point to a cable or controller rather than the platters, which is why they are listed separately. Repeated CRC growth is a reason to reseat or replace the SATA cable before blaming the drive itself.

Attribute 241 and vendor-specific fields

Some attributes mean different things on different drives. Attribute 241 reads as total LBAs written on some Samsung drives and as something unrelated elsewhere. Where meaning is genuinely vendor-specific, it is marked as such here rather than guessed.

If you are comparing two drives, sort by ID first and label second. That habit catches most of the mismatches before they cost you a wrong conclusion.

Unplugged SATA data cable beside a desktop hard drive with a screwdriver nearby

Before you replace the drive

Power the machine down, reseat the data cable at both ends, swap it for a known-good one, and check the CRC count again. If the count holds still after that, the drive was not the problem.

Using this database with the calculator

Copy the raw values you care about into the drive health calculator for a plain assessment. The database explains what each number is; the calculator tells you how the numbers combine into one picture.

Start with attributes 05, 197, and temperature if you only check three. Those three cover remapping, unstable reads, and heat — the three most common reasons a drive gets pulled from service before it fails outright.

Open the Drive Health Calculator Read the Diagnostic Guides

The guides cover the two situations people ask about most: telling whether a drive is dying, and checking a used SSD before you buy it.

Questions the database gets most

Short answers with the reasoning kept in. If you need the full method, the guides go further on each one.

What raw value in attribute 05 should trigger replacement?

There is no universal number. A stable small count is common; a count that doubles in weeks, or crosses into the hundreds, warrants backup and planning a replacement.

Are pending sectors more serious than reallocated sectors?

Often yes, because they are unresolved. Reallocated sectors have already been handled; pending sectors are areas the drive still cannot read reliably and may fail on.

Do all SSDs report wear level the same way?

No. Vendors use different IDs and scales. Crucial and Samsung both expose life-left style metrics, but the labels differ. Read your vendor's mapping.

Quiet diagnostics workroom at the end of the day with drives on a bench and a laptop glow

Where to continue

Read the ID, then let the numbers settle the question.

This page is the lookup. The calculator turns raw values into a single assessment, and the guides walk through the two situations that bring most people here: deciding whether a drive is dying, and checking a used SSD before money changes hands.

Open the Drive Health Calculator Read the Diagnostic Guides Ask us about a reading

SMART Drive Health · 452 Silicon Drive, Suite 102, San Jose, CA 95110, USA · +1-408-555-0192 · Monday–Friday, 9:00 AM – 6:00 PM Pacific Time