Every single PC that comes through my shop door gets the same treatment in the first five minutes: I plug it in, open CrystalDiskInfo, and wait about three seconds to know if I have a storage problem on my hands. In sixteen years of running Cincinnati PC Repair LLC, no tool has saved more client data — or more awkward conversations — than this one.
⚠️ 2026 SSD Price Alert
SSD prices have risen sharply since late 2025 due to a global NAND flash shortage driven by AI data center demand. NVMe SSDs are up roughly 100-115% and SATA drives around 75% compared to mid-2025 lows. A 1TB NVMe that cost $60-$70 last year now runs $130-$260 depending on the model. Prices are expected to remain elevated through at least mid-2026. The picks in this guide are still the best options available — but expect to pay significantly more than historical prices. We are updating prices as they shift.
The problem is that most people who actually run CrystalDiskInfo have no idea how to read what it’s telling them. They see a wall of numbers, a color they don’t like, and then they either panic unnecessarily or, worse, dismiss a genuine warning. This guide is the plain-English walkthrough I give to every client who asks me what those numbers mean.
By the end of this article, you’ll know exactly how to use CrystalDiskInfo, how to interpret every health status, which S.M.A.R.T. attributes actually matter, and what steps to take the moment your drive shows any sign of trouble.
What Is CrystalDiskInfo?
CrystalDiskInfo is a free, open-source utility developed by Noriyuki Miyazaki (Crystal Dew World) that reads the S.M.A.R.T. data built into virtually every modern hard drive and solid-state drive. S.M.A.R.T. stands for Self-Monitoring, Analysis, and Reporting Technology — it’s a system that your drive maintains internally, tracking dozens of performance and reliability metrics over its entire lifespan.
Think of it as your drive’s medical chart. The drive is constantly taking its own vitals. CrystalDiskInfo just reads those vitals and translates them into something you can actually look at on screen. Without a tool like this, all that diagnostic data sits locked inside the drive where you can’t see it.
Here’s what CrystalDiskInfo can read and report:
- Overall health status (Good, Caution, Bad, or Unknown)
- Drive temperature in real time
- Power-On Hours and Power-On Count
- All S.M.A.R.T. attribute values, including raw values and thresholds
- Drive interface (SATA, NVMe, USB)
- Firmware version and buffer size
- Remaining life percentage for SSDs
It supports HDDs, SATA SSDs, NVMe SSDs, and even some USB-connected drives. It also works with Intel RAID configurations. The tool is completely free — no subscription, no premium tier, no watermarked reports.
Where to Download CrystalDiskInfo
Only download CrystalDiskInfo from two places: the official site at crystalmark.info/en/software/crystaldiskinfo/ or directly from the Microsoft Store. The Microsoft Store version is actually the cleanest option right now — no installer ads, verified by Microsoft, and it updates automatically. The official site offers a standard installer edition and a portable (.zip) edition that runs without installing anything, which is my preference when I’m working on a client machine.
Avoid third-party download sites. The tool is small and completely free, so any site bundling it with “extras” is not doing you any favors.
Understanding the Health Status Colors
When you open CrystalDiskInfo, the most prominent information is the health status shown in the top-left area of the interface. This is the single most important thing on the screen. There are four possible statuses, and each one has a specific meaning.
Good (Blue)
Blue means the drive is operating within normal parameters. All monitored S.M.A.R.T. attributes are within acceptable ranges, no critical thresholds have been crossed, and there are no pending sector errors. This is what you want to see. A “Good” status doesn’t guarantee a drive will never fail — drives can fail suddenly without warning — but it means there are no measurable signs of deterioration right now.
In practice, when I see a blue Good status, I still glance at the Power-On Hours and temperature before moving on. A 10-year-old drive showing Good is still a 10-year-old drive.
Caution (Yellow)
Yellow means the tool has detected at least one S.M.A.R.T. attribute where the raw value has exceeded the manufacturer-set threshold for a critical parameter. For HDDs, CrystalDiskInfo flags Caution based on three specific attributes: Reallocated Sectors Count (0x05), Current Pending Sector Count (0xC5), and Uncorrectable Sector Count (0xC6). For SSDs, it also triggers Caution when the remaining life percentage drops to 10% or below.
Caution does not mean the drive is about to die this afternoon. What it means is that you have a measurable problem that is likely to get worse over time. I tell clients the same thing every time: Caution is your warning window. You have time to act, but the clock is running.
Bad (Red)
Red means a critical S.M.A.R.T. attribute’s current value has dropped below its manufacturer-defined threshold. In S.M.A.R.T. terms, this is the official definition of a failing drive. The drive itself has essentially reported that it is failing. For SSDs, Bad triggers when the remaining life drops below the vendor-defined threshold value.
A Bad status is a five-alarm fire. The drive may still be functioning, but it is not reliable, and that status can change at any moment. I cover what to do in detail in the sections below.
Unknown (Gray)
Gray means CrystalDiskInfo could not retrieve S.M.A.R.T. data from the drive. This happens with some USB enclosures, certain RAID controllers, and occasionally with older or non-standard drives. It doesn’t mean the drive is bad — it means you simply can’t read its health. If you see Unknown on a drive you rely on, consider testing it with the manufacturer’s own diagnostic tool instead.
The Critical S.M.A.R.T. Attributes to Watch
CrystalDiskInfo displays a full table of S.M.A.R.T. attributes, each with an ID, name, Current value, Worst value, Threshold, and Raw Value. Most of those attributes are informational. Five of them are the ones I focus on in every diagnosis.
05 — Reallocated Sectors Count (The Most Important One)
This is the single most predictive attribute for HDD failure. When the read/write head of your hard drive encounters a bad sector — one it cannot reliably write to — it “reallocates” the data from that bad sector to a spare sector held in reserve. The Reallocated Sectors Count tracks how many times this has happened over the life of the drive.
The Raw Value is the number you care about. If it’s anything above zero, pay attention. If it’s rising over time, start preparing for replacement. According to Backblaze’s large-scale drive failure research, this attribute — along with four others — is among the most reliable predictors of imminent drive failure. A single reallocated sector isn’t necessarily a death sentence, but any non-zero value means the drive’s surface has developed at least one area it can no longer trust.
The practical threshold I use in my shop: any Raw Value above 0 gets flagged for monitoring, and any value above 10 puts the drive on the replacement list regardless of other attributes.
C5 — Current Pending Sector Count
This attribute tracks the number of sectors that the drive has flagged as “suspect” but hasn’t yet been able to confirm as bad. These are sectors where a read error occurred and the drive is waiting for a chance to test them again. If the next test succeeds, the count can drop back down. If it fails, those sectors get reallocated and the Reallocated Sectors Count goes up.
A non-zero Pending Sector Count means your drive has sectors it’s uncertain about right now. Again, any value above zero deserves attention. The danger here is that data sitting in a pending sector is at risk — if the drive can’t read that sector, that data is gone unless you have a backup.
C6 — Uncorrectable Sector Count
This is the most urgent of the three critical sector attributes. Uncorrectable sectors are sectors where the drive attempted error correction and failed completely. The data in those sectors is unrecoverable. Any non-zero value here means data has already been lost at the hardware level, even if you haven’t noticed it yet in Windows.
In my experience, by the time I see a non-zero Uncorrectable Sector Count, the drive is already in serious trouble. It’s not a “watch and wait” situation — it’s a “back up everything today” situation.
09 — Power-On Hours
This is the drive’s odometer — the total number of hours it has been powered on since manufacture. It doesn’t directly predict failure, but it gives critical context for everything else. A drive showing 500 hours with a reallocated sector is a very different situation from a drive showing 35,000 hours with a reallocated sector.
For HDDs, most manufacturers rate drives for 3–5 years of typical use. At roughly 8 hours per day, that’s around 9,000–15,000 hours. A drive past 30,000 hours on an HDD is living on borrowed time regardless of what the other attributes say. For SSDs, the endurance model is different (see the next section), so Power-On Hours matters less than the wear metrics.
C2 — Temperature
CrystalDiskInfo displays your drive’s current temperature prominently at the top of the interface. For HDDs, the safe operating range is generally 0–60°C, with 30–45°C being the healthy working range. For SSDs, safe operating temperature is typically 0–70°C. Anything above 50°C for an HDD during normal operation is worth investigating — check case airflow and cable management.
Sustained high temperatures accelerate every other form of drive wear. I’ve seen drives in cramped, poorly ventilated cases accumulate reallocated sectors at double the rate of identical drives in well-cooled systems. Temperature management is drive longevity management.
Reading HDD Results vs. SSD Results
One of the most common points of confusion with CrystalDiskInfo is that HDDs and SSDs fail in fundamentally different ways, so the attributes that matter most are different for each.
For HDDs: Focus on Sector Health
Hard drives have spinning magnetic platters and physical read/write heads. They fail primarily due to mechanical wear: head crashes, platter degradation, and the slow accumulation of bad sectors as the magnetic coating degrades over time. For HDDs, the three critical sector attributes (05, C5, C6) are your main warning system, backed up by temperature and Power-On Hours.
Pay particular attention to whether the bad sector attributes are stable or increasing. A drive with five reallocated sectors that has been stable for two years is in a different situation from a drive where that number jumped from 3 to 47 in the last month.
For SSDs: Focus on Wear and Remaining Life
Solid-state drives don’t have moving parts, so they don’t fail the same way. Instead, SSDs fail because NAND flash memory cells can only be written to a finite number of times before they wear out. CrystalDiskInfo tracks this through SSD-specific attributes:
- Wear Leveling Count — Tracks how evenly write cycles are being distributed across the NAND cells. Higher is generally better (meaning more balanced distribution and more life remaining).
- Total Host Writes / Total Host Reads — The cumulative amount of data written to and read from the drive over its lifetime, measured in gigabytes or terabytes. Compare this to the manufacturer’s TBW (Terabytes Written) rating to estimate how much life is left.
- Remaining Life % — Many SSDs report this directly. CrystalDiskInfo displays it prominently and uses it as the primary trigger for Caution (≤10%) and Bad status for SSD health assessment.
For example, if your SSD is rated for 300 TBW and CrystalDiskInfo shows it has processed 280 TBW of Total Host Writes, you’re at about 93% consumed endurance. The Remaining Life percentage will reflect that. An SSD at 7% remaining life should be on your replacement list, even if it still shows “Good” technically.
The sector-based attributes (05, C5, C6) are less relevant for SSDs in most cases, though you’ll still see them listed. If they show non-zero values on an SSD, that’s actually more serious than on an HDD, because SSDs aren’t supposed to have bad sectors in the traditional sense — it often indicates controller or NAND issues.
If you’re thinking about upgrading to a new SSD and want current recommendations, check out our guide to the best SSDs to buy this year (best SSDs for every budget — coming soon). For laptop-specific upgrades, my guide on the Dell laptop RAM upgrade covers the process of opening up your machine and working with internal components safely.
What to Do When You See “Caution”
A yellow Caution status is a warning, not a verdict. In my shop, I’ve seen Caution drives run reliably for another 18 months, and I’ve seen them fail within a week. The problem is you can’t know which scenario you’re in. Here’s exactly what I do, in order, every time a client’s drive comes up Caution.
Step 1: Back Up Immediately
Before you do anything else, back up your important data. This is not optional. Use an external drive, cloud storage, or both. If you’re copying large amounts of data, use small batches rather than one giant transfer — large sequential writes can stress a struggling drive and accelerate failure. Prioritize irreplaceable files: documents, photos, anything that can’t be re-downloaded.
Step 2: Identify Which Attribute Triggered the Caution
Look at the S.M.A.R.T. table and find the attribute highlighted in yellow. If it’s 05 (Reallocated Sectors Count), C5 (Current Pending Sector Count), or C6 (Uncorrectable Sector Count) with a non-zero raw value, you have a physical drive surface problem. If it’s temperature-related or Power-On Hours, the situation is less immediately critical but still worth addressing.
Step 3: Run chkdsk
Open Command Prompt as Administrator and run: chkdsk C: /f /r /x (substitute your drive letter if needed). This scans the file system for errors and attempts to recover data from bad sectors. It won’t fix the underlying hardware problem, but it can prevent file system corruption from compounding the issue. It also forces the drive to test its pending sectors, which may resolve some C5 counts.
Allow chkdsk to complete fully — on a large drive, this can take several hours.
Step 4: Plan for Replacement
A Caution drive should be considered a temporary drive. Even if your backups are solid, you don’t want to be relying on it indefinitely. Start shopping for a replacement. You likely have days to weeks to months before the drive actually fails — but you don’t know which. Having a replacement ready means you control the timing of the swap rather than the drive controlling it for you.
For most users, I recommend replacing a Caution HDD within 30–60 days. For an SSD showing Caution because of low remaining life, you have a bit more predictability, but the same timeline applies.
What to Do When You See “Bad”
A red Bad status means the drive has formally failed according to its own diagnostic system. The current value of at least one critical S.M.A.R.T. attribute has dropped below its manufacturer-defined threshold. In plain terms: the drive itself is saying it cannot be trusted.
Stop Using the Drive for Anything New
Do not install software, run updates, or save new documents to this drive. Every write operation on a failing drive is a gamble. The less you ask of it, the better your chances of recovering what’s already there.
Recover Your Data First — Before Replacement
If the drive is still responsive, this is your window. Copy data off it immediately, prioritizing the most important files first. If the drive is already failing to mount or appears in Windows but becomes unresponsive during file access, do not run chkdsk — it can make a failing drive worse. Instead, use a data recovery tool like Recuva (free) or professional software like R-Studio.
If the drive has already failed completely and won’t respond at all, you’re looking at professional data recovery. This is expensive (typically $300–$1,500 depending on severity) but it’s often the only option for unreadable drives. This is exactly why I tell every client: back up before you need to, not after.
Replace the Drive Immediately
There is no repairing a Bad-status drive. Formatting it does not fix the underlying hardware problem. The only correct action after data recovery is to replace the drive.
For a solid replacement SSD, the Samsung 870 EVO SATA SSD remains one of the most reliable choices for desktop and laptop replacements, with a proven track record in shop environments. If you’re replacing an NVMe drive, the WD Blue SN5100 NVMe SSD offers excellent performance and a 5-year warranty at a competitive price point.
Setting Up CrystalDiskInfo for Ongoing Monitoring
Running CrystalDiskInfo once is useful. Setting it up to run continuously in the background is far more valuable. Here’s how I configure it for clients who want proactive alerts rather than reactive panic.
Enable Startup Launch
Go to Function → Startup and enable it. This tells Windows to launch CrystalDiskInfo automatically when you log in. The program will sit quietly in the background without consuming meaningful resources.
Enable Resident Mode (System Tray)
Go to Function → Resident to keep CrystalDiskInfo running in the system tray rather than closing when you minimize it. You’ll see a small drive temperature icon in the taskbar notification area that updates in real time. If you have multiple drives, it cycles through their temperatures.
To control what happens when you click the X button, go to Function → Advanced Feature → Resident Style and set it to “Minimize” so closing the window doesn’t kill the background monitor.
Configure Alert Thresholds
Go to Function → Advanced Feature → Health Status Setting. For HDDs, set the threshold for attributes 05, C5, and C6 to 1 — this means any non-zero raw value in those attributes will immediately trigger a Caution flag. For SSDs, I recommend setting the Remaining Life Caution threshold to 15% rather than the default 10%, giving you more runway for a planned replacement.
Set Up Temperature Alerts
Go to Function → Advanced Feature → Alarm Temperature. Set HDD temperature alerts at 50°C and SSD alerts at 60°C. These are conservative but appropriate thresholds for catching cooling problems before they cause damage. You can also enable alert sounds under Function → Alert if you want an audible warning.
Configure Auto-Refresh
Go to Function → Auto Refresh and set it to refresh every 10 minutes during normal use. This keeps the temperature and S.M.A.R.T. data reasonably current without hammering the drive with constant queries. If you have spinning HDDs, keep in mind that very frequent polling can prevent them from spinning down to idle — 10 minutes is a reasonable balance.
Quick Reference: What to Check and When to Worry
Here’s the condensed version of everything above — the reference card I keep in my head for every drive I look at:
| What You See | What It Means | What to Do |
|---|---|---|
| Blue “Good” | Drive is healthy, all attributes within range | Note the Power-On Hours and temperature; re-check in 3–6 months |
| Yellow “Caution” | At least one critical attribute has exceeded threshold | Back up immediately; run chkdsk; plan replacement within 30–60 days |
| Red “Bad” | Drive has formally failed its own diagnostic | Stop writing; recover data now; replace immediately |
| Gray “Unknown” | S.M.A.R.T. data unreadable (USB, RAID, older drive) | Use manufacturer’s diagnostic tool instead |
| Attribute 05 (Raw) > 0 | Bad sectors have been found and remapped | Monitor closely; replace if count is rising |
| Attribute C5 (Raw) > 0 | Suspect sectors pending re-test | Run chkdsk; back up; monitor trend |
| Attribute C6 (Raw) > 0 | Data lost to hardware error | Immediate backup; replacement urgent |
| SSD Remaining Life ≤ 10% | SSD nearing end of rated endurance | Back up; buy replacement; plan migration |
| Temperature > 50°C (HDD) | Overheating risk | Check airflow; add case fan; inspect cable routing |
| Power-On Hours > 30,000 (HDD) | Drive has lived a long life | Treat as aging hardware; keep backups current regardless of status |
CrystalDiskInfo won’t predict every drive failure — nothing will. Drives can and do fail suddenly with no prior S.M.A.R.T. warning, especially from electrical events or physical damage. But for the gradual failures that represent the vast majority of what comes through my shop, this tool gives you a real fighting chance to get ahead of the problem before it becomes a crisis.
My standing advice: run it on every machine you care about, set it up to monitor in the background, and check the results at least once a month. It takes under a minute and it’s free. The alternative — discovering your drive was failing after the fact — is a much more expensive lesson.