Bad Sectors Data Recovery
Growing Defect Lists and What They Mean for Your Data
Bad sectors are locations on a hard drive's platter surface where the magnetic coating can no longer hold data reliably. When the count is growing, the drive is physically degrading. Software cannot fix physical platter damage. Running repair tools like chkdsk or SpinRite on a drive with growing bad sectors accelerates the failure and can make your data unrecoverable. Our hard drive data recovery workflow starts with imaging on a DeepSpar Disk Imager before the remaining good sectors degrade further.
What Are Bad Sectors?
A hard drive stores data on spinning platters coated with a thin magnetic layer. The read/write heads float nanometers above this surface, reading and writing magnetic signals. A bad sector is a location where this magnetic layer is too damaged, worn, or contaminated for the heads to read or write data reliably.
Bad sectors come in two forms. Logical bad sectors result from corrupted error-correction data or interrupted writes. These can sometimes be fixed by rewriting the sector. Physical bad sectors result from actual damage to the platter coating: scratches, head contact marks, particle contamination, or age-related magnetic decay. Physical bad sectors cannot be repaired by any software because the underlying recording surface is destroyed.
The critical distinction: if bad sector counts are growing over time, the problem is physical. The drive's heads or platters are degrading, and continued use accelerates the damage.
P-list vs G-list: How Drives Manage Defects
Every hard drive ships with some defective sectors. The drive's firmware uses two lists to track and hide them from the operating system.
Primary Defect List (P-list)
Created at the factory during manufacturing. The P-list records sectors that failed quality testing before the drive shipped. These are permanently mapped out by the firmware's translator module. The operating system and SMART data cannot see them.
A typical drive ships with thousands of P-list entries. This is normal. Modern platter manufacturing cannot produce a defect-free surface at the densities used in current drives (over 1 terabit per square inch on high-capacity models).
Grown Defect List (G-list)
Populated during the drive's operational life. When a sector fails a read or write during normal use, the firmware adds it to the G-list and remaps the logical block address (LBA) to a spare sector from the reserve pool.
G-list growth is what SMART attribute 05 (Reallocated Sector Count) tracks. A growing G-list is the warning signal that the drive's recording surface or heads are degrading in the field.
Why the reserve pool matters: Every drive has a limited number of spare sectors (typically a few thousand). When the G-list exhausts this reserve, the drive can no longer remap bad sectors. At that point, read/write errors start surfacing directly to the operating system as I/O failures, file corruption, and blue screens.
Why Bad Sectors Grow Over Time
Head Degradation
Read/write heads have a finite lifespan. As the heads weaken from thermal cycling, vibration, or age, they lose the ability to maintain a stable magnetic signal. Sectors that were readable with fresh heads become unreadable as the head's signal-to-noise ratio degrades. This is the most common cause of progressively growing bad sector counts.
Platter Surface Contamination
Microscopic particles inside the drive enclosure (from outgassing or worn components) can settle on the platter surface. At the nanometer-scale flying heights of modern heads, even a sub-micron particle causes read errors. The debris contaminates adjacent tracks as the platters spin, spreading the damage outward from the initial impact site.
Magnetic Coating Decay
The cobalt-alloy magnetic layer on platters degrades over years of thermal cycling. High operating temperatures (sustained above 50°C) accelerate this process. Areas of weakened magnetic retention lose their ability to hold a stable bit pattern, producing read errors that appear as bad sectors to the firmware.
Which SMART Attributes Indicate Bad Sectors?
Three SMART attributes track bad sector activity on a hard drive: Reallocated Sector Count (05), Current Pending Sector Count (C5/197), and Offline Uncorrectable Sector Count (C6/198). Each attribute reports a normalized Current value, a Worst value, a Threshold, and a Raw value. The Raw value is the actual count. When Current drops below Threshold, the drive flags a pre-failure warning.
- Attribute 05: Reallocated Sector Count
The Raw value is the exact count of sectors the firmware has already remapped to the reserve pool. Each remapped sector consumed one spare. The normalized Current value starts at 100 (or 200 on some manufacturers) and decreases as more sectors are remapped. When Current drops below the Threshold (typically 36 on Seagate, 140 on Western Digital), the drive sets a pre-failure flag in SMART status.
A non-zero Raw value that remains stable for weeks is a warning to back up but not an emergency. A Raw value that increases between checks means the drive is actively degrading.
- Attribute C5/197: Current Pending Sector Count
The most critical attribute for triage. The Raw value counts sectors that are failing reads right now but have not been reallocated yet. The drive is waiting for a successful write to that LBA to trigger reallocation. When the operating system hits a pending sector, the drive enters aggressive retry loops (20+ attempts per sector), causing multi-second hangs and I/O timeouts that freeze applications.
Any non-zero value in attribute 197 means the drive is struggling to read specific locations on the platter surface. Back up accessible data and stop using the drive.
- Attribute C6/198: Offline Uncorrectable Sector Count
Counts sectors that failed during the drive's offline self-test (an internal scan the firmware runs during idle time). A high Raw value means the magnetic domains at those locations have degraded beyond the ECC polynomial's correction capacity. The drive's Reed-Solomon or LDPC error correction can tolerate a certain number of flipped bits per sector; once the error density exceeds that limit, the sector is uncorrectable.
Attribute 198 often tracks closely with attribute 197. If both are climbing, the media degradation is accelerating.
How Bad Sectors Progress to Head Crash
Bad sectors on a hard drive follow a predictable degradation chain. Isolated media defects grow into G-list saturation, which stresses the read/write heads until they degrade, and degraded heads eventually lose aerodynamic lift and crash into the platter surface. Each stage narrows the recovery options and increases the cost.
- Isolated media degradation. Small areas of weakened magnetic coating produce occasional read errors. The firmware compensates with ECC correction and retry loops. SMART attribute 05 may show a low, stable count. The drive functions normally from the user's perspective, but the damaged areas are permanent.
- G-list saturation. As more sectors fail, the firmware remaps them to the reserve pool. Once the reserve pool is exhausted (typically a few thousand spare sectors), the translator module can no longer hide bad sectors from the operating system. Read errors surface as I/O failures, file corruption, and application freezes. SMART attribute 197 (pending sectors) climbs.
- Head degradation. Repeated reads over slow and damaged zones stress the magnetoresistive read element in the head slider. The head's signal-to-noise ratio drops, making it harder to resolve weak magnetic transitions on the platter. Sectors that were marginal become unreadable. The bad sector count accelerates because the head itself is now contributing to read failures, not just the media.
- Terminal head crash. A degraded head loses stable aerodynamic lift. The slider contacts the platter surface at rotational speed (5,400 to 7,200 RPM on consumer drives), gouging the magnetic coating and generating metallic debris. This debris contaminates adjacent platters and heads. The drive produces a clicking or scraping sound and stops responding. Recovery at this stage requires a head swap with platter cleaning on a 0.02µm ULPA-filtered clean bench.
Sending a drive in at stage 1 or 2 costs less than half of what a stage 4 head swap with surface damage costs. The earlier the drive is powered down and shipped, the more data is recoverable at lower cost.
SMR Drives and Bad Sector Complications
Shingled Magnetic Recording (SMR) drives overlap their write tracks like roof shingles to increase storage density. This design creates a complication when bad sectors appear: rewriting a single sector on an SMR drive requires rewriting the entire shingled zone (typically 256 MiB of adjacent tracks) because the overlapping tracks must be reconstructed in sequence.
When a bad sector appears in the middle of an SMR band, the drive's firmware must read the entire band, replace the damaged sector with data from the spare pool, and rewrite all overlapping tracks. This write amplification puts substantial stress on the heads and can trigger additional bad sectors in adjacent tracks that were previously healthy.
On some consumer SMR drives (particularly WD and Seagate models sold without SMR labeling), the translator module that manages the shingled-to-logical mapping can become corrupted during bad sector reallocation. This makes sectors appear "bad" at the logical layer even when the magnetic surface is intact. Recovery from translator corruption requires PC-3000 firmware access and cannot be performed with consumer software.
Data Recovery Standards & Verification
Our Austin lab operates on a transparency-first model. We use industry-standard recovery tools, including PC-3000 and DeepSpar, combined with strict environmental controls to make sure your hard drive is handled safely and properly. This approach allows us to serve clients nationwide with consistent technical standards.
Open-drive work is performed in a ULPA-filtered laminar-flow bench, validated to 0.02 µm particle count, verified using TSI P-Trak instrumentation.
Transparent History
Serving clients nationwide via mail-in service since 2008. Our lead engineer holds PC-3000 and HEX Akademia certifications for hard drive firmware repair and mechanical recovery.
Media Coverage
Our repair work has been covered by The Wall Street Journal and Business Insider, with CBC News reporting on our pricing transparency. Louis Rossmann has testified in Right to Repair hearings in multiple states and founded the Repair Preservation Group.
Aligned Incentives
Our "No Data, No Charge" policy means we assume the risk of the recovery attempt, not the client.
Technical Oversight
Louis Rossmann
Louis Rossmann's well trained staff review our lab protocols to ensure technical accuracy and honest service. Since 2008, his focus has been on clear technical communication and accurate diagnostics rather than sales-driven explanations.
We believe in proving standards rather than just stating them. We use TSI P-Trak instrumentation to verify that clean-air benchmarks are met before any drive is opened.
See our clean bench validation data and particle test videoWhy "Bad Sector Repair" Software Makes It Worse
Tools like SpinRite, HDD Regenerator, Victoria HDD, and Windows chkdsk all share the same fundamental approach: they scan the drive surface sector by sector and attempt to read, rewrite, or remap bad sectors. On a drive with physical media damage, this is the worst possible approach.
A full-surface scan forces the heads to traverse every track on every platter, including the damaged areas. Each pass over a damaged zone puts mechanical stress on heads that may already be weakened. The heads heat up from continuous seek operations. Debris from damaged areas gets redistributed across the platter by the airflow inside the drive enclosure.
Professional recovery takes the opposite approach. The PC-3000 images good sectors first by skipping damaged areas entirely on the initial pass. It then works through damaged regions on subsequent passes with controlled retry counts (typically 1-3 retries per sector vs the 20+ retries consumer tools use). Between passes, the heads are parked to cool. This recovers more data while putting less total stress on the drive.
Tools That Damage Drives with Bad Sectors
- ×chkdsk /r: Scans every sector, deletes unreadable file entries, and overwrites metadata needed for recovery.
- ×SpinRite: Forces up to 2,000 read/write attempts per sector. On a degrading drive, this hammers the weakest heads over the most damaged areas.
- ×HDD Regenerator: Claims to "regenerate" bad sectors by rewriting them with specific magnetic patterns. Physical platter damage cannot be repaired by rewriting data.
- ×Low-level format: Rewrites every sector on the drive. Destroys all existing data and does not fix physical damage.
How We Recover Data from Drives with Bad Sectors
The approach depends on whether the bad sectors are isolated or indicate broader head/platter degradation.
Stable Bad Sectors (Drive Still Reads)
- Connect the drive to PC-3000 Portable III in read-only mode (ATA command level, below the OS)
- Build a head map to identify which head(s) are producing the most errors
- Image healthy areas first at full speed, skipping sectors that timeout on the initial pass
- Run targeted passes over damaged sectors with 1-3 retry attempts per sector and automatic head parking between passes
- Reconstruct the file system from the cloned image using R-Studio or UFS Explorer
Typical cost: $250 to $900
Progressed to Head Failure (Clicking/Not Detected)
- Open the drive on our 0.02µm ULPA-filtered clean bench
- Replace the failed head stack assembly with matched donor heads (same model, same firmware revision, same head map)
- Transfer adaptive parameters (head fly height calibration, preamp gain settings) from the original ROM to the donor heads via PC-3000
- Image the drive with the new heads, using the same multi-pass approach with retry control
Typical cost: $1,200 to $1,500
How Multi-Pass Imaging Recovers Data from Bad Sectors
Professional recovery from a drive with bad sectors uses hardware-level imaging with millisecond timeout control, not software running through the operating system. The PC-3000 and DeepSpar Disk Imager (DDI) communicate directly with the drive over the PHY link at the ATA command level, bypassing the OS storage driver and its 30+ second default timeout. This allows the imager to skip a failing sector in milliseconds instead of waiting half a minute per bad read.
- Forward pass (fast scan). The imager reads sequentially at full speed. Any sector that does not respond within the hardware timeout (configurable down to milliseconds on the DDI) is logged and skipped. This first pass captures all healthy data without stressing the heads over damaged zones.
- Reverse pass. The imager re-reads the skipped LBA ranges in reverse order. Reading logical block addresses in reverse order prevents the drive's read-ahead cache from stalling on consecutive bad sectors. Approaching a damaged zone from a healthy area often recovers boundary sectors that the firmware would otherwise drop during a forward-read hang.
- Head mapping. The DDI generates a per-head error map correlating each physical head to its LBA range and error density. Healthy heads image first. Degraded heads are engaged last, with reduced retry counts and mandatory parking intervals between read bursts to prevent thermal buildup in the preamp.
- Read channel tuning for marginal sectors. For sectors that failed both forward and reverse passes, PC-3000 accesses the drive's microcode to adjust read channel parameters. On modern drives using PRML (Partial Response Maximum Likelihood) signal processing, the analog signal from a degraded magnetic domain is noisy. Lowering the read threshold and modifying equalization filter coefficients in RAM can extract a valid signal from marginal flux transitions that the drive's default settings reject.
Between each pass, the heads park on the ramp to cool. Consumer recovery software does not have access to these hardware-level controls. It reads through the OS storage driver, which waits 30+ seconds per failing sector and has no ability to skip, reverse, or adjust read channel parameters.
PRML Read Channel and Viterbi Detection on Marginal Heads
Modern hard drives do not read bits directly off the platter. The analog waveform produced by the magnetoresistive read element is processed through a Partial Response Maximum Likelihood (PRML) read channel before the drive ever reports a byte to the host. When a head degrades, the analog signal degrades first; the Viterbi detector is what decides whether the drive returns valid data or a UNC error.
The read path on a current-generation drive runs: preamplifier on the head stack assembly, continuous-time analog filter, analog-to-digital converter, finite impulse response (FIR) equalizer, Viterbi detector, then LDPC error correction. The FIR equalizer shapes the sampled waveform to match a known target polynomial. Early drives used PR4 (Partial Response Class 4). Higher areal densities moved to EPR4 and E2PR4, and modern high-capacity drives use Noise-Predictive Maximum Likelihood (NPML) channels that embed a noise-whitening filter directly into the Viterbi branch metric.
A healthy head produces clean flux transitions that the Viterbi trellis resolves with low bit error rate. A degraded head produces low-amplitude, noisy transitions. The drive's factory-calibrated Viterbi thresholds start misclassifying valid bits as noise, and the firmware enters ECC recovery loops. During those loops the drive's adaptive FIR logic continues updating tap coefficients. If the head is scraping over a damaged zone, the firmware will silently retune the equalizer to the noise of the scratch, which ruins readability for the rest of that surface.
PC-3000 Portable III exposes the vendor-specific commands that halt this background adaptation. The technician disables auto-adaptation, loads static FIR coefficients that worked before the degradation started, and loosens the Viterbi decision boundaries so the detector accepts lower-amplitude samples as valid. This raises the raw bit error rate the Viterbi stage emits, but the LDPC or Reed-Solomon ECC downstream absorbs the extra errors and returns a clean sector the firmware can hand off to the host.
Read Adaptive Parameters and Zone-Specific Tuning in the System Area
Each head on a hard drive has its own preamplifier gain, fly-height offset, and equalization profile. Each recording zone has its own linear density and timing calibration. These per-head and per-zone Read Adaptive Parameters (RAP) live in the drive's System Area, not on user-accessible LBAs, and they are written during factory self-calibration.
On Western Digital drives using the Marvell 88i9XXX controller family, the bulk of the adaptive data lives in Module 47. On Seagate F3-architecture drives (including the Rosewood and SkyHawk families), the primary translator lives in SysFile 28, the Non-Resident G-List lives in SysFile 35, and head adaptive parameters are distributed across the SAP, RAP, and CAP modules. PC-3000 accesses these over the serial UART diagnostic port after injecting a ROM patch that unlocks the diagnostic prompt on drives with a locked terminal.
Two common operations depend on this access. First, reading the patient drive's Module 47 and writing the averaged adaptive values back into the System Area RAM of a drive with matched donor heads, so the donor stack inherits the electrical calibration it needs. Second, editing the logical head map in RAM to disable a single failing head, which lets the drive image the remaining surfaces without hanging on the bad head. Consumer software cannot reach any of these structures.
Shingled Magnetic Recording drives add another layer. The media cache management table lives in SysFile 348 on Seagate F3 SMR drives; incoming writes land in a conventional cache zone and migrate to overlapping bands during idle time. A forced translator rebuild with the standard m0,6,2 command on an SMR drive destroys the media cache mapping and orphans data that was waiting to migrate. Our translator recovery procedure on SMR drives patches the SMP flags to disable background cache migration before touching the translator.
DeepSpar Per-Head Imaging Profiles and Millisecond Hardware Timeouts
The DeepSpar Disk Imager sits between the drive and the host at the SATA PHY layer. It does not see the drive as a block device; it sees it as a controller responding to raw ATA commands. This is what enables imaging a drive with a failing head without finishing the head off.
- Per-head sector bitmaps. DeepSpar maps the full LBA space to physical head geometry and maintains a separate imaging bitmap for each head. The first pass reads only LBAs served by healthy heads at full speed. The failing head is engaged last, under a separate read profile with reduced retry counts and mandatory rest intervals between bursts to let the preamp cool.
- Millisecond timeout enforcement. When a sector fails, the DeepSpar hardware drops the ATA read after a configurable timeout (commonly 150 ms for marginal drives, down to tens of milliseconds for drives that are actively degrading). The operating system default is 30 seconds or longer per read. That 30-second gap is the difference between a salvaged drive and a gouged platter, because the head stays over the damaged zone for the entire wait.
- PHY resets and power cycling. When the drive firmware crashes into a busy state that will not clear, the DeepSpar issues a soft reset, a hard COMRESET at the SATA PHY level, or cuts DC power to the drive to force a full reinitialization. Software imagers have no equivalent; they wait for the kernel to give up.
- Smart Hot Swap for System Area damage. When the patient drive's System Area tracks are unreadable, DeepSpar reads the accessible SA modules from the patient, writes them onto a compatible donor, boots the donor to a stable state, and hot-swaps the SATA data cable to the patient while both drives are spinning. The donor's initialized controller state is what keeps the patient's read channel stable long enough to image the user data.
Turnaround on a typical imaging job runs 3 to 10 business days depending on how many passes the degraded head needs and how much System Area work is required. +$100 rush fee to move to the front of the queue when a case cannot wait in the normal queue.
Why ddrescue Accelerates Head Damage on a Failing Drive
GNU ddrescue is the sanest free option for imaging a drive with stable heads and scattered media errors. On a drive with a failing head, ddrescue causes the damage it is supposed to avoid, because of two architectural limits that cannot be patched around in software.
First, ddrescue reads through the Linux block device layer. When a sector fails, the drive enters internal error recovery and stops responding to the host. ddrescue is waiting on the kernel, and the kernel default timeout for a SCSI or libata read is 30 seconds. For that full 30 seconds the failing head is sweeping over the damaged zone generating heat and debris. Repeat that across a few thousand bad LBAs and the head finishes degrading before the image completes.
Second, ddrescue has no concept of physical head geometry. Errors clustered on head 2 of a four-head drive look the same to ddrescue as errors scattered across the whole surface. It keeps hammering every failing LBA in sequence. HDDSuperClone improves on this by inferring head boundaries from the error pattern and skipping zones served by the failing head, but it still cannot reach the System Area to repair firmware, and cannot recover a drive that has dropped to zero-capacity after a translator failure.
The pattern we see most often in cases arriving after a ddrescue attempt: the initial error count on SMART 197 was small, the customer ran ddrescue for 48 hours, the drive now reports a three-digit reallocated sector count and a noticeably weaker head on imaging. Recovery still works in most of these cases, but the drive has lost healthy margin that was recoverable before ddrescue ran.
Bad Sector Recovery Pricing
Cost depends on drive condition when it arrives at our lab. Drives sent in early (before bad sectors progress to head failure) are the least expensive to recover.
Simple Copy
Low complexityYour drive works, you just need the data moved off it
$100
3-5 business days
Functional drive; data transfer to new media
Rush available: +$100
File System Recovery
Low complexityYour drive isn't recognized by your computer, but it's not making unusual sounds
From $250
2-4 weeks
File system corruption. Accessible with professional recovery software but not by the OS
Starting price; final depends on complexity
Firmware Repair
Medium complexityYour drive is completely inaccessible. It may be detected but shows the wrong size or won't respond
$600–$900
3-6 weeks
Firmware corruption: ROM, modules, or translator tables corrupted; requires PC-3000 terminal access
CMR drive: $600. SMR drive: $900.
Head Swap
High complexityMost CommonYour drive is clicking, beeping, or won't spin. The internal read/write heads have failed
$1,200–$1,500
4-8 weeks
Head stack assembly failure. Transplanting heads from a matching donor drive on a clean bench
50% deposit required. CMR: $1,200-$1,500 + donor. SMR: $1,500 + donor.
50% deposit required
Surface / Platter Damage
High complexityYour drive was dropped, has visible damage, or a head crash scraped the platters
$2,000
4-8 weeks
Platter scoring or contamination. Requires platter cleaning and head swap
50% deposit required. Donor parts are consumed in the repair. Most difficult recovery type.
50% deposit required
Hardware Repair vs. Software Locks
Our "no data, no fee" policy applies to hardware recovery. We do not bill for unsuccessful physical repairs. If we replace a hard drive read/write head assembly or repair a liquid-damaged logic board to a bootable state, the hardware repair is complete and standard rates apply. If data remains inaccessible due to user-configured software locks, a forgotten passcode, or a remote wipe command, the physical repair is still billable. We cannot bypass user encryption or activation locks.
No data, no fee. Free evaluation and firm quote before any paid work. Full guarantee details. Head swap and surface damage require a 50% deposit because donor parts are consumed in the attempt.
Rush fee: +$100 rush fee to move to the front of the queue.
Donor drives: Donor drives are matching drives used for parts. Typical donor cost: $50–$150 for common drives, $200–$400 for rare or high-capacity models. We source the cheapest compatible donor available.
Target drive: The destination drive we copy recovered data onto. You can supply your own or we provide one at cost plus a small markup. For larger capacities (8TB, 10TB, 16TB and above), target drives cost $400+ extra. All prices are plus applicable tax.
When to Act: Bad Sector Decision Framework
Use SMART attribute monitoring to decide when to stop using the drive.
Monitor
- Reallocated sectors (SMART 05) under 10
- Pending sectors (SMART 197) at zero
- Counts stable between weekly checks
Action: Back up immediately. Monitor weekly. Plan drive replacement within 1-3 months.
Back Up Now
- Reallocated sectors 10-100 and climbing
- Pending sectors 1-50
- Slow reads or intermittent file access errors
Action: Copy critical data to a different drive. Do not run recovery software. Replace the drive this week.
Stop Using the Drive
- Pending sectors over 50 and growing daily
- Reallocated sectors over 100
- Drive clicking, freezing, or producing read errors on file copy
Action: Power down the drive. Contact a professional data recovery lab for imaging.
Frequently Asked Questions
What is the difference between a bad sector and a reallocated sector?
Can SpinRite or HDD Regenerator fix bad sectors?
Will running chkdsk fix bad sectors on my hard drive?
How many bad sectors is too many?
What causes bad sectors on a hard drive?
Is it safe to keep using a drive with a few bad sectors?
How much does it cost to recover data from a drive with bad sectors?
What is the difference between P-list and G-list bad sectors?
Related Recovery Services
Full HDD recovery service overview
SMART attribute failure prediction guide
Head failure diagnosis and recovery
File system damage from bad sectors
Complete drive failure diagnosis
Transparent cost breakdown
Bad Sectors? Get a Free Diagnosis.
Send us your drive. We image it with PC-3000, working around damaged sectors without making them worse. No data, no charge.