S.M.A.R.T. stands for Self-Monitoring, Analysis and Reporting Technology. CrystalDiskInfo displays the values a drive exposes, but the meaning of a raw number can vary by manufacturer and device family.
| Column | What it means | How to read it |
|---|---|---|
| ID | A hexadecimal attribute identifier. | Useful for recognizing common attributes such as 05, C5, C6 or C7. |
| Attribute Name | The human-readable label assigned to the attribute. | Names can vary slightly by drive type and vendor. |
| Current | A normalized vendor-defined score. | Often higher is better, but the scale is not universal. |
| Worst | The worst normalized score recorded by the drive. | Useful for seeing if the normalized score has degraded. |
| Threshold | A vendor-defined limit for the normalized score. | Crossing it can indicate failure according to the drive firmware. |
| Raw Value | The underlying counter or encoded value reported by the device. | Often the most useful field, but it can use vendor-specific units or encoding. |
Why raw values are not universal
A single raw number cannot always be interpreted identically between manufacturers. One drive may expose a simple count, another may pack multiple pieces of information into the same field, and an SSD may use attributes that do not exist on an HDD.
Normalized values versus counters
Current, Worst and Threshold are normalized values chosen by firmware. Raw Value is closer to the original telemetry, but it may require model-specific interpretation. When diagnosing warnings, record both the raw value and whether it changes over time.
| ID / Field | Attribute | Meaning | Watch |
|---|---|---|---|
| 05 | Reallocated Sectors Count | Sectors that have been remapped to spare area on many HDDs. | Any non-zero value deserves attention; increases over time are more concerning. |
| C5 | Current Pending Sector Count | Sectors waiting to be re-read or remapped. | A rising count can indicate unstable media. |
| C6 | Uncorrectable Sector Count | Sectors the drive could not correct. | Treat increases as a serious warning signal. |
| C7 | Interface CRC Error Count | Communication errors between a SATA drive and host. | Often points to cable, connector or electrical-contact issues. |
Raw values need context
The raw-value column is often the most interesting column, but it is also the easiest one to misread. Some raw values are simple counters, while others contain encoded information or vendor-specific scales. A value may be cumulative and remain recorded forever even after the original cause is no longer active.
Current, Worst and Threshold are normalized fields. They are not directly comparable with raw values and they are not equally meaningful across vendors. A high normalized value may be good, while a high raw value can be good, neutral or problematic depending on the attribute.
Better questions to ask
- Which attribute triggered the warning?
- Is the raw value new, stable or increasing?
- Does the interpretation match an HDD, SATA SSD or NVMe SSD?
- Are there temperature, connection or media-error signals as well?
- Is there a backup before any long diagnostic test is started?
The safest interpretation combines the attribute name, raw value, trend, drive type and role of the drive. A data disk with important personal files deserves a more conservative response than a replaceable test drive with no important data.
What S.M.A.R.T. cannot do
S.M.A.R.T. is not a complete health contract between the drive and the user. Some failures happen suddenly without a clear counter increasing first. Other counters look dramatic but remain stable after the original cause has been fixed. Use the values as technical evidence for backup, replacement planning and further diagnosis, not as the only decision input.