How Donor Drives Are Matched for Head Swaps

A head swap requires transplanting the entire Head Stack Assembly (HSA) from a compatible donor drive into the patient drive (the drive with failed heads). The donor heads must be mechanically compatible with the patient's platters and electronically compatible with the patient's firmware and preamp circuitry. Matching by model number alone is not sufficient. Hard drive manufacturers produce the same model number across multiple hardware revisions, with different head components, preamp chips, and firmware variations within a single model line. Donor matching is a prerequisite for head-swap cases in our hard drive data recovery workflow.

Six PC-3000 Donor Matching Criteria

Donor selection at the PC-3000 bench resolves to six criteria, applied in order of strictness. A donor that passes all six produces stable reads on first power-up. A donor that fails one criterion may still image with intervention from PC-3000 Portable III; a donor that fails two or more is set aside before the patient cover is opened. The criteria below extend the firmware, head map, and preamp matching framework documented in the following sections.

1. Preamp gain matching

The preamp on the HSA flex cable amplifies head signals at a calibrated gain stage. When donor and patient preamp revisions differ, the read-back signal arrives at the controller channel at the wrong amplitude. Too-low gain produces signal amplitude collapse, where the Viterbi detector cannot resolve transitions and the read returns unrecoverable errors across every sector. Too-high gain saturates the analog front end and clips the signal, which also defeats detection. Preamp gain is not a tunable parameter from the host side; it is fixed by the preamp revision and matched at factory calibration.

2. Channel coefficients (Viterbi and FIR equalizer)

The read channel uses a Finite Impulse Response (FIR) equalizer to flatten the partial-response waveform recovered from the preamp, followed by a Viterbi detector that resolves the equalized signal into NRZ bit decisions. FIR tap coefficients and Viterbi branch metrics are head-specific adaptive values stored in the System Area. The patient's coefficients were trained against the patient's heads; loading them against donor heads produces equalization residuals that the Viterbi detector mis-decodes as bit errors. PC-3000 reads the donor's channel coefficients from the donor SA before the swap and writes them into the patient firmware as part of adaptive transfer, so the equalizer is targeted at the heads that will actually be installed.

3. Microjog offsets

Per-head radial position deltas stored in adaptive tables. Seagate F3 encodes microjog as a Q8 fractional track unit (1/256 track increments); WD Marvell encodes it as DAC values inside Module 47. Mismatched microjog produces Adjacent Track Erasure on write and off-track SNR collapse on read. Detailed mechanics are documented in the SA Adaptives and Microjog Calibration section below.

4. ZAP tables (Zone-Adaptive Parameters)

Modern HDDs divide the platter surface into recording zones; linear density and bit cell size differ between zones because the linear velocity at the head changes with radius. ZAP tables store per-zone, per-head gain, timing, and equalization compensation values that the controller loads when the actuator crosses a zone boundary. On WD Marvell drives ZAP data is part of Module 47; on Seagate F3 drives it is embedded in the SA adaptive tables with a ROM backup. After a head swap, the patient ZAP values were calibrated against the original heads and produce zone-boundary read errors when the donor heads cross from one zone to the next. PC-3000 regenerates or transfers ZAP entries during adaptive transfer.

5. HSA family and firmware revision compatibility

A donor labeled with the same model number routinely fails because the HSA family or firmware sub-revision differs. Seagate Rosewood drives split into preamp generations C202 and 8202 that are not interchangeable; WD Marvell drives split on the 5th DCM character, which encodes the head stack supplier; Toshiba MQ splits on country of manufacture. Firmware revision is matched as a precondition, but firmware-revision match alone is not sufficient when the HSA family within the revision has changed. This is why same-model donors fail more often than the model number alone suggests.

6. Site code and HGA generation for multi-platter helium drives

Helium-sealed multi-platter families (WD Ultrastar DC HC series, Seagate Exos X) iterate Head Gimbal Assembly generations more frequently than air-filled families because each revision ships as a separate SKU rather than as a running change to a single SKU. Site Code (WU, SU, TK, AMK) match becomes a load-bearing criterion rather than a preference, and the donor window tightens from months to weeks. The HGA generation is not printed on the label; it is inferred from Part Number prefix, Site Code, and Date of Manufacture together. Helium head swap pricing uses $3,000–$4,500 per the helium HDD pricing tiers.

Donor Selection Decision Cascade

The six criteria above describe what is matched. The cascade below describes the order in which a technician resolves them at the bench, from label-only inspection through ROM extraction to internal preamp verification. Each phase eliminates donors that fail the earlier checks before any clean-bench work is committed.

1. Phase A. Patient Label Analytics. Transcribe the metadata that determines candidate filtering. For Western Digital units, record the MDL string (16-digit or 17-digit format) and the Drive Configuration Matrix (DCM), which encodes head stack vendor and preamp tuning in specific character positions. For Seagate units, record Model Number, Part Number prefix, Site Code (WU, SU, TK), and Date Code. Seagate Date Codes use a proprietary fiscal-year format and must be converted to a standard calendar date before the 12-week donor window can be evaluated.
2. Phase B. Candidate Donor Lookup. Inside the PC-3000 Utility status window, the Search donor drives function auto-populates from the connected patient profile and queries integrated donor inventory databases. The DonorDrives.com cross-reference is used as an industry-wide lookup that maps external labels to deeper PC-3000-extracted parameters such as microjog values and preamp generations, so candidate drives can be filtered before they are physically acquired.
3. Phase C. External Label Cross-Check. When a candidate donor arrives, the WD model string is re-verified against the patient. On a 16-digit WD MDL such as WD5000AAKX-083CA0, the 3rd, 4th, and 5th characters after the hyphen must match. On a 17-digit WD MDL such as WD5000LPVX-22V0TT0, the 3rd, 4th, 5th, and 6th characters after the hyphen must match. DCM alignment uses the J or 2 anchor character near the end of the DCM string. The anchor and the single character immediately preceding it encode the HSA supplier and preamp tuning, and must match exactly between patient and donor.
4. Phase D. ROM Dump and Shadow Module Verification. Before the donor seal is broken, PC-3000 reads the donor ROM through vendor-specific commands. WD Marvell drives are forced into Kernel Mode and an LDR loader is uploaded to controller RAM, which permits reading Module 102 (factory head map backup), Module 103 (adaptive parameters), Module 104 or 109 (microprogram version), and Module 190 (SMR zone map, when evaluating shingled families). Seagate F3 drives are accessed via the serial diagnostic COM port at 38,400 baud, elevated to 460,800 or 921,600 baud for faster transfer. The SPI flash ROM is first extracted over Y-Modem in Boot Code mode. PC-3000 then generates a Tech Mode unlock patch from that image, which is loaded into RAM to bypass the Diagnostic Port Lock for subsequent terminal access.
5. Phase E. Internal Preamp and Adaptive Verification. If the ROM dump confirms compatibility, the donor is staged for internal inspection. On modern locked Seagate Rosewood architectures the preamp vendor and revision are not exposed by the controller via standard terminal output and must be verified by cross-referencing the HSA flex cable markings against the donor inventory record before any clean-bench commitment. On WD units, the Heads map changing feature reads Module 47 microjog values; a donor whose microjog variance exceeds roughly 300 DAC points from the patient baseline is rejected, because the read channel will not hold servo lock at the patient track pitch.

What Disqualifies a Donor Before the Clean Bench

Donor drives cost from hundreds to thousands of dollars, and a misjudged candidate destroys both the donor heads and the patient platters on first spin-up. The following conditions remove a donor from consideration before any HSA extraction begins. Each is observable from the label, the ROM dump, or a microscopic platter inspection performed with the patient cover off but the donor still sealed.

Mismatched physical head count or geometry

PC-3000 displays the active head assignment as a Column PH (Physical Heads) array. A donor may have more heads than the patient as long as the active heads sit in the same spatial positions across the platter stack. A donor with fewer heads, or with an active head map that does not align geometrically with the patient chassis, is rejected. A re-routed head map cannot be made to work by selectively disabling elements in firmware, because the actuator radial calibration was performed against the original geometry.

SMR versus CMR topological mismatch

WD Palmer and Spyglass families are exclusively Drive-Managed Shingled Magnetic Recording architectures, managed by Module 190 acting as the T2 translator. Module 32 holds the Grown Defect List (G-List) and is not a translator. When evaluating donors across SMR and CMR families more generally, the read elements differ physically: SMR heads carry asymmetrical magnetic shielding to read narrower overlapping tracks, while CMR heads are wider and read fully isolated tracks. Transplanting a CMR donor HSA into an SMR patient (or the reverse) produces adjacent-track interference and uncorrectable bit error rates the firmware cannot recover from.

Visible platter damage and crash-rings

Before HSA extraction, the patient platters are inspected under the 0.02 micron ULPA clean bench. A concentric crash-ring scored through the CoCrPt magnetic coating means the failed head has lifted particulate matter into the HDA. Installing a pristine donor HSA into a contaminated HDA destroys the donor heads on the first spindle rotation. Recovery on a crash-ring patient requires platter burnishing and media cleaning under the bench before the donor HSA is committed.

Date of Manufacture beyond the 12-week window

The industry baseline is a 12-week DOM window between patient and donor. Inside that window, manufacturers continue to iterate slider mass, actuator-arm metallurgy, and aerodynamic lift profiles silently between production runs. A donor outside the window may carry heads tuned for a different platter overcoat thickness or a different fly-height target; the patient adaptives no longer match the donor head flight dynamics, and the read channel will not lock the signal to the servo marks.

DCM anchor character deviation (WD)

The Western Digital DCM is not arbitrary. The J or 2 anchor character near the end of the DCM string, and the single character immediately preceding it, encode the HSA vendor and preamp tuning configuration. A donor whose anchor pair does not match the patient anchor pair is disqualified, even when the model number, firmware revision, and Date of Manufacture all match.

Head Map Verification on the PC-3000

The logical head map tells the controller which physical read/write elements to route I/O through. Verifying the donor head map against the patient head map before committing the swap, and selectively disabling weak heads on the donor after the swap, are both performed through the PC-3000 utility against specific firmware modules. Procedural context for the swap itself is documented in what a head swap involves.

Head map module locations

WD Marvell architectures store the active head map in Module 0A on the platter Service Area, with a factory-defined backup in Module 102 inside the ROM shadow. If Service Area corruption prevents Module 0A from reading, the map can be reconstructed from the ROM backup. Seagate F3 architectures bind the physical head map natively within the drive's primary ROM image, while the Read Adaptive Parameters are held in SysFile 06.

Verification sequence

1. Force the drive into Kernel mode under PC-3000 control so the Service Area is reachable while the controller is bypassing normal boot.
2. Navigate Tools, Utility extensions, Modules directory to inspect Service Area integrity.
3. Read Module 0A as a binary head map. A bitmask such as 2,3,4,5,6,7 indicates which physical head positions are active in the patient configuration.
4. Compare the patient head map against the donor ROM head map extracted in Phase D. The donor must support the patient configuration exactly, or a superset of it.

Selective head disablement after the swap

If a transplanted donor HSA presents one marginal head, PC-3000 can logically retire that element from the active translation pool. The path on WD is the RAM Head map editing protocol; on Seagate F3 it is the Zones and Heads Inactivation through the P-List menu. Disabling a single head on a 4-head 2 TB Rosewood drive removes access to one platter surface, roughly 500 GB of the 2 TB volume. The remaining stable heads are imaged first under DeepSpar Disk Imager with conservative timeouts, and the disabled head is re-engaged in a later pass with aggressive timeout parameters to recover the surface it controls.

Firmware Revision Matching

The drive's firmware revision is the primary matching criterion. The firmware controls how the drive communicates with the heads: signal timing, read channel calibration, servo decoding parameters, and write current profiles. Heads from a donor with a different firmware revision may not work because the patient drive's firmware expects specific electrical characteristics from the heads that the donor heads do not provide.

Firmware revisions are printed on the drive label. For Seagate drives, this is a four-character code (e.g., CC26, SDM1, 0001). For Western Digital, the firmware revision is part of the extended model number. The firmware revision indicates the generation of controller code and, by extension, the generation of head technology the firmware is calibrated for.

Within a firmware revision, there can be sub-revisions that affect compatibility. Two Grenada drives with firmware "CC26" manufactured six months apart may have different micro-code patches. In most cases, same firmware revision is sufficient. In some Seagate Rosewood drives, even drives with the same firmware revision but from different manufacturing sites (identifiable by the Site Code on the label) have different head compatibility.

Head Map and Head Count

A drive's head map specifies which heads are installed and active. A two-platter drive can have 2, 3, or 4 heads depending on the capacity variant. A Seagate Rosewood ST2000LM007 (2 TB) uses 4 heads across 2 platters. The ST500LM030 (500 GB) uses 2 heads on 1 platter. Both are "Rosewood" drives, but their head assemblies are physically different.

The head map is stored in the drive's firmware and defines which physical head positions are active and how the firmware addresses them. The donor must have the same head map as the patient: same number of heads in the same physical positions. A 3-head donor HSA cannot be used in a 4-head patient drive because the firmware expects to address a head that does not exist, and the physical mounting may differ.

Preamp Chip Compatibility

The preamp is a small IC mounted on the HSA flex cable, inside the sealed drive cavity. It amplifies the microvolt signals from the read heads and drives the write current to the write heads. The preamp model must match between donor and patient because the controller's read channel is calibrated for the specific signal characteristics of that preamp.

Preamp models can be identified by the marking on the chip (visible when the drive is opened) or inferred from the firmware revision and drive family. Common preamp manufacturers include Texas Instruments (TLS-series mixed-signal preamps), STMicroelectronics, Broadcom (Avago), and Renesas. Renesas has become the dominant preamp vendor in modern WD and Toshiba drives. Silicon Systems read-channel devices appear in legacy hardware. Within the same drive model line, a preamp change usually coincides with a firmware revision change.

Streaming vs Non-Streaming Preamp Variants

Surveillance and AV drive families (WD Purple, Seagate SkyHawk) use preamp variants tuned for sustained sequential writes from continuously recording camera systems. Compared to a consumer desktop preamp, the streaming variant typically has a deeper write buffer, a flatter write-current (Iw) profile across sustained workloads, and thermal management calibrated for continuous duty. Consumer preamps are tuned for bursty, read-heavy office workloads and run hotter under sustained sequential write pressure.

Cross-matching a streaming preamp with a non-streaming HSA (or the reverse) produces write-element burnout, a Diag Err state during SA initialization, or an immediate head park after a short read burst. The write-current parameter is adaptive and coercivity-tuned per preamp revision; the controller applies an Iw value stored in the SA adaptives that assumes a specific preamp write-path gain. When the installed preamp drives the write element at a mismatched current, the head coil can fail thermally within a few write cycles.

The STMicroelectronics SMOOTH L7250 family is worth calling out separately. L7250 is a motor controller rather than a preamp, but it sits on the same PCB and its revision often changes alongside preamp revisions within a Seagate generation. Labs track L7250 revisions on the PCB as a proxy signal when preamp markings are obscured.

Adaptive Parameters and Post-Swap Calibration

Every hard drive generates a unique set of adaptive parameters during factory self-scan. These parameters record the specific calibration data for the drive's individual heads: optimal read channel settings, write current values, fly height compensation, and servo tracking offsets. Adaptive parameters are stored in the drive's System Area on the platters and backed up in the ROM chip on the PCB.

After a head swap, the patient drive's adaptive parameters no longer match the installed heads. The parameters were calibrated for the original heads, not the donor heads. In some cases, the drive can still read with minor degradation. In other cases, the mismatch prevents the drive from reading its own System Area, causing it to fail initialization.

PC-3000 can modify adaptive parameters after a head swap. The technician can load the donor drive's adaptive parameters into the patient drive's firmware, aligning the calibration with the installed heads. This is not always necessary (some drive families tolerate mismatched adaptives well enough for imaging), but it improves read stability and reduces errors on drives where the mismatch causes issues.

Manufacturing Date and Factory Site

The closer the donor is to the patient in manufacturing date and production site, the higher the compatibility probability. Drives manufactured in the same batch at the same factory use the same component lots: same head wafer, same preamp batch, same platter lot. Component-level consistency within a production batch is high.

Western Digital encodes manufacturing information in the DCM (Drive Configuration Matrix) printed on the label. The 5th and 6th characters of the DCM indicate the head stack supplier and preamp configuration. Two drives with the same model number and firmware revision but different DCM codes may have incompatible heads.

Seagate uses a combination of the Part Number (PN), Site Code, and Date of Manufacture printed on the label. The Site Code identifies the factory (e.g., WU for Wuxi, SU for Suzhou, TK for Thailand). Two Seagate drives with matching model numbers and firmware revisions but different Site Codes or Part Numbers may have incompatible head assemblies.

How Labs Maintain Donor Inventory

Professional recovery labs maintain an inventory of donor drives organized by manufacturer, model family, firmware revision, head map, and manufacturing date range. Common drive families that fail frequently (Seagate Rosewood, WD Blue/Green 2.5", Samsung Spinpoint M8) are stocked in higher quantities.

When a patient drive arrives, the technician identifies the required donor specifications from the drive label and firmware. If the lab has a matching donor in stock, the head swap can proceed immediately. If not, the lab sources one from supplier networks, which may take 1-5 days depending on the drive's rarity.

Donor drives are purchased specifically as parts inventory. They are functional drives that have been verified to read and write normally. Using a donor from a failed drive (e.g., a drive that had bad sectors but working heads) is possible but risky: the heads may be degraded even if they currently function.

Firmware Family Architecture by Manufacturer

Hard drive manufacturers iterate on base firmware architectures across multiple product generations. Recognizing which firmware family a drive belongs to is the first step in donor matching, because cross-family donors are incompatible even when the model number prefix looks similar. Each family uses a distinct read channel configuration, SA module layout, & head addressing scheme.

Seagate F3 Architecture (Grenada, Rosewood, Makara)

Seagate's F3 architecture covers drives where the firmware revision is a short alphanumeric code without a period (e.g., CC49, SBK2, SDM1). Within F3, three sub-families have distinct matching requirements:

Grenada (7200.14 desktop)

Desktop 3.5" drives including the Barracuda ST1000DM003 & ST2000DM001. Grenada matching depends on firmware revision, head map, & preamp chip. The preamp revision is visible on the HSA flex cable when the drive is opened. Certain preamp substitutions are possible within the same generation (e.g., a B2 preamp can sometimes substitute for a CA in the same firmware revision), but cross-generation swaps fail.

Rosewood (2.5" mobile)

The Rosewood family includes the ST1000LM035, ST2000LM007, ST1000LM048, & ST500LM030. These are among the most common drives in recovery labs. Rosewood matching is complicated because the preamp revision is not printed on the drive label, and the serial terminal is locked on most units. The preamp must be identified by opening the drive or estimated from the Date of Manufacture. Two common Rosewood preamp revisions (C202 & 8202) are not interchangeable; a donor with C202 heads will not produce stable reads in a patient that originally used 8202, even with identical firmware.

Makara (enterprise & high-capacity)

Shares structural DNA with Grenada but uses different head map logic & servo calibration optimized for sustained sequential workloads. Makara donors are not interchangeable with Grenada despite similar firmware revision formats. The SA module layout differs, and the adaptive parameter tables use different calibration ranges.

Western Digital Marvell-Based Architecture

Western Digital transitioned from older Caviar IDE/SATA controllers to Marvell-based MCU architectures. Modern WD drives are categorized into Marvell Version 1 & Version 2, distinguished by the family code in the model number.

WD donor matching requires aligning three parameters: the full model number, the physical head map, & specific characters within the DCM (Drive Configuration Matrix). The 5th character of the DCM encodes the head stack supplier. Two drives with identical model numbers & firmware but different 5th DCM characters have physically different head assemblies from different component vendors.

WD firmware stores the head map in ROM Module 0A. Module 30 contains the SA translator (initial preamp calibration values & compiler data), and Module 47 holds the SA Adaptives (servo parameters & head-specific calibration). For modern Shingled Magnetic Recording (SMR) drives such as the WD10SPZX & WD20SPZX, Module 190 contains SMR-specific operational data. A donor for an SMR drive must have compatible Module 190 data, adding an additional matching constraint that CMR drives do not require.

Toshiba MQ & MK Series

Toshiba drives have the most forgiving donor matching requirements among the three major manufacturers. For the older MK series (MK6475GSX), matching the full model number & PCB family number is typically sufficient. For the modern MQ series (MQ01ABD, MQ04ABF), matching the full model number & the HDD Code printed on the label covers most compatibility cases.

One edge case: drives manufactured in different countries (China vs. Philippines) occasionally differ in read channel calibration. If a model-matched donor produces unstable reads, the next donor should match both the model number & country of manufacture.

SA Adaptives & Microjog Calibration

Service Area (SA) adaptive parameters are head-specific calibration values generated during factory testing. After a head swap, these parameters no longer match the installed donor heads. PC-3000 can transfer the donor's adaptives into the patient drive's firmware to restore read stability, but the process varies by manufacturer & firmware family.

On Western Digital Marvell-based drives, SA Adaptives are stored in Module 47. This module contains voltage settings, read channel gain profiles, & servo calibration values specific to each head position. If the donor heads deviate from the patient's Module 47 values, the drive may fail to read its own Service Area on power-up, resulting in a clicking loop or immediate head park. The technician uses PC-3000 to replace the patient's Module 47 with the donor's version, allowing the firmware to calibrate for the installed heads.

Microjog values are a subset of the adaptive data. During factory calibration, the drive measures microscopic alignment offsets for each head on the actuator arm & records compensation values. These values correct for physical variation in head placement that is unavoidable during HSA manufacturing. When a donor HSA has microjog values close to the patient's original values, the swap is more likely to produce stable reads without additional intervention. Large deviations in microjog values indicate the donor heads are physically misaligned relative to what the patient firmware expects, which causes read errors & sector instability during imaging.

Seagate F3 drives handle adaptives differently. The adaptive tables are embedded in the SA modules on the platter surface, and the ROM chip on the PCB contains a backup. PC-3000's Seagate F3 utility can read, edit, & write these modules. For Rosewood drives, the technician may need to use a hot-swap technique (powering on with the donor HSA, then physically swapping to the patient platters while the controller remains initialized) to bypass the locked terminal.

Microjog Storage Units and Deviation Tolerance

Each manufacturer encodes microjog differently. Seagate F3 drives store per-head microjog as a Q8 fractional track unit inside the SA adaptive tables, giving the firmware a sub-track offset in 1/256 track increments. Western Digital stores microjog compensation inside Module 47 as DAC values that drive the fine-position servo loop. Both schemes encode the same physical quantity: the offset between the head's mechanical position on the actuator arm and the track center the servo system expects.

The primary failure mode from a microjog mismatch is Adjacent Track Erasure (ATE). If a write fires while the installed donor head is off-track relative to the patient firmware's expected microjog, the write pole overlaps the neighboring track and erases user data on that adjacent cylinder. ATE is silent at the host level until the adjacent track is read back and returns unrecoverable errors.

The secondary failure mode is sector re-read storms and off-track SNR collapse. The read channel's Continuous Time Analog Filter (CTAF) is tuned for a specific head alignment; when the installed donor head sits outside the CTAF's tolerance band, the analog signal-to-noise ratio drops below the Viterbi detector's error-correction threshold and the controller issues repeated re-read retries. These retries show up as excessive head dwell time and slow imaging rates during read passes.

PC-3000 Portable III can average donor and patient microjog values when an exact match donor is not available. The averaged value is written into the patient firmware's SA adaptive table, splitting the servo offset between the two sets of calibration data. This stabilizes reads enough for imaging on weak-match donors at the cost of some read rate, but does not correct ATE risk if writes are issued to the drive.

A distinct concept from microjog is the servo-burst field encoding. Modern HDDs use embedded servo patterns of alternating A/B/C/D burst fields written onto reserved wedges of each track at the factory servo-track writer. Head production lots differ in pole-tip width and magnetoresistive stripe geometry, which changes the analog amplitude the head returns when crossing those burst fields. The firmware decodes Position Error Signal (PES) from the ratio of A/B and C/D burst amplitudes; when a donor head from a different production lot returns a burst amplitude curve the PES decoder was not calibrated for, the servo loop interprets a valid on-track position as off-track and issues corrective actuator current. The result is micro-oscillation around track center that degrades the read SNR even when microjog values appear correct. Labs treat same-production-lot donors as preferable over same-firmware-revision donors when both are available, because burst field response is a physical head property that does not appear in any adaptive table.

Adaptive ROM Binding: Why the Patient's ROM Transplants to the Donor

A common question from customers is whether a head swap is accompanied by a ROM swap in the same direction. It is not. The ROM contents follow the patient's heads and platters; never the donor's. This is because the ROM holds adaptive data that was burned during factory calibration against the original HDA, and the firmware expects that data to match the physical media it is reading. Transplanting a donor ROM wholesale onto a patient HDA produces controller errors that look identical to head failure but are actually firmware misconfiguration.

On Seagate F3 drives, the ROM chip holds three critical modules: Module 0A contains the boot code, Module 28 contains the translator stub used before the full translator is loaded from the SA, and Module 02 contains the adaptive seed that tells the firmware how to interpret the first sectors of the SA. The patient's heads were calibrated against the patient's platters, so the patient ROM must move onto the donor PCB (or equivalently, the donor ROM is replaced with the patient's ROM contents via a programmer). A wholesale donor ROM swap commonly produces the LED 000000CC code on the F3 diagnostic port, or a Diag Err state where the controller cannot locate its own SA.

On Western Digital Marvell-based drives, the ROM module set is richer. Module 01 holds the module directory, Module 02 holds configuration and drive ID data, Module 0A holds the head map, Module 30 holds the SA translator, and Module 47 holds the SA adaptives. All five pin adaptive data to the physical controller and its factory-calibrated HDA. Using a donor ROM wholesale causes the controller to apply donor servo parameters to the patient HDA, which produces a hardware retry loop (repeated read-channel calibration attempts that never converge) or a slow-responding error where the drive takes several seconds to answer IDENTIFY DEVICE and then fails to initialize the user area.

The practical workflow at our Austin data recovery lab is to read the patient ROM with a programmer clip before the HSA swap is attempted, store the dump, and write the patient ROM contents onto the donor PCB. If the donor PCB is incompatible at the motor-driver level, the donor's electrical components are populated onto the patient's PCB instead, preserving the patient's ROM in its original socket.

Date-Code Window and Site-Code Fingerprinting

Date of Manufacture is one of the most load-bearing external matching criteria. Industry practice is to match donors within approximately 3 months (12 weeks) of the patient's Date of Manufacture. Some labs stretch this window to 6 months when rare drive families leave no tighter option. Match rate degrades as the window widens because manufacturers iterate head wafer design, change preamp vendors, or alter platter carbon overcoat thickness between production batches. A donor from outside the window often looks electrically similar on paper and still produces unstable reads in the patient drive.

The full fingerprint is a combination of three fields. Seagate Part Number prefix (e.g., 9YN162-021, 1ER164-300) identifies the head stack design revision. The Site Code (WU for Wuxi, SU for Suzhou, TK for Thailand, AMK for the Singapore Ang Mo Kio site) identifies the factory and, by extension, the component vendors the factory was sourcing during that production window. The Date of Manufacture anchors the batch within the manufacturer's component-change timeline. Together, Part Number prefix + Site Code + DOM act as an HSA lot fingerprint.

Helium drive donor windows are tighter. Helium-sealed families (WD Ultrastar DC HC series, Seagate Exos X) iterate internal components more frequently within a single model line because the sealed-cavity design makes post-production component revisions cheaper to ship as a separate SKU rather than as a running change. A helium donor window measured in weeks rather than months is standard practice, and the Site Code match is more frequently required than on air-filled drives.

Pre-Swap Clean Bench Verification

Before the patient drive is opened, the vertical laminar flow bench is validated and the donor HSA is staged. The 0.02 micron ULPA-filtered bench delivers an ISO 14644-1 Class 4 or better localized environment at the open drive. ULPA filtration captures particles down to 20 nanometers; HEPA filtration stops at 0.3 microns and is not sufficient for head-swap work because modern head fly-height is at the nanometer scale. The lab uses a particle counter to verify the bench before opening the drive; the TSI P-Trak 8525 ultrafine particle counter is the instrument of record in most recovery labs.

Visual platter inspection runs under high-magnification illumination and checks for three primary defects. Rotational scoring appears as concentric rings around the platter, indicating a head has touched the surface during prior operation. Crash rings are darker zones where the head has scraped the CoCrPt magnetic coating off the substrate; crash rings mean the donor HSA will be sacrificed if installed. Embedded debris is any foreign particle adhered to the platter surface; a single particle requires cleaning with a platter-safe procedure before the donor HSA is brought into contact with the disk.

Slider count verification compares the physical head count on the patient HSA to the logical head map reported by firmware. A 6-head drive must show 6 intact sliders on the HSA arm comb. A missing or damaged slider indicates the head map and the physical assembly are out of sync, and the donor selection must be revisited against the actual physical configuration rather than the label-reported head count.

A FLIR thermal camera establishes a pre-power baseline on the motor controller (e.g., the STMicroelectronics SMOOTH L7250 or a TI TLS-series driver) and the preamp region of the PCB. If either runs hot on initial power-up, the lab can rule out seized spindle stiction masquerading as head failure before the HSA swap is committed. A stuck spindle motor draws excess current through the motor driver and can mimic a clicking-drive symptom that would otherwise be attributed to failed heads.

HSA Extraction and Transplant Procedure

The physical transfer of a Head Stack Assembly from donor to patient is a bench procedure, not a single motion. Every step is performed inside the 0.02 micron ULPA vertical laminar flow bench with the drive cavity open. The tooling is model-family-specific, and the risk at each step is contact between slider and platter surface. Once a slider lands on a spinning or stopped platter with any lateral force, the CoCrPt magnetic coating is scored and the platter track is lost for that revolution; repeated contact creates the crash ring pattern described in the clean-bench verification section above.

1. Head comb staging. A head comb is a set of thin polymer or shim-steel fingers sized to the exact inter-platter gap of the target drive family. The comb slides between the platters and separates the read/write heads before the HSA is lifted off the pivot. Comb dimensions are specific to drive family: a 3.5" enterprise comb will not fit a 2.5" Rosewood, and a Rosewood comb will bind in an Ultrastar helium cavity. The correct comb is selected by drive family before the patient cover is removed and kept under the laminar flow until installation.
2. Actuator pivot release. The HSA is held to the baseplate by a pivot screw or magnetic latch (depending on family). The head stack is parked on the ramp load/unload structure or on the inner crash stop, and the comb is inserted between the platters to capture the sliders. Only after the comb is seated can the pivot be released. Lifting the HSA without the comb lets the slider suspension arms spring inward, which drives the sliders into the platter edge and damages both the heads and the platter outer diameter.
3. Platter rotation lock. On drives with free-spinning spindles (some 3.5" desktop families), a spindle lock pin is inserted through the motor hub to prevent platter rotation during HSA removal. Any rotation while the comb is between the platters drags the sliders against the comb surface and abrades the air-bearing face. On sealed helium drives, the spindle is mechanically captive and no lock is required, but the cover reseal procedure after head swap requires helium refill from a calibrated line per our helium-HDD recovery workflow.
4. Donor HSA preparation. The donor drive is opened in parallel under the same bench. The donor HSA is combed and lifted identically. The donor ROM is not transferred; only the physical head stack is moved. The patient ROM contents have already been programmed onto the donor PCB (or the donor PCB has been swapped onto the patient body) before this step, per the ROM binding procedure in the section above.
5. Transplant and comb withdrawal. The donor HSA is seated onto the patient pivot with the comb still holding the sliders apart. The pivot screw or latch is reattached. The comb is withdrawn in a single smooth motion parallel to the platter surface; a tilted withdrawal brings a slider into momentary contact with the platter and can score the landing zone. The cover is replaced before power is applied to prevent airflow contamination of the newly transplanted heads.
6. First power cycle under PC-3000 control. The drive is connected to the PC-3000 Portable III or Express terminal, not to a normal SATA port. First power is issued with the terminal recording the boot sequence so that any SA read failure, adaptive mismatch, or preamp error is captured in the log before the firmware commits a write. If the drive fails to spin up, the procedure aborts and the donor selection is revisited before a second attempt is staged.

Extraction procedure varies across families. Seagate F3 Rosewood drives require a four-finger comb sized for the 2.5" dual-platter cavity and a pivot screw torque that is lower than the desktop Grenada family, because the Rosewood pivot bearing uses a smaller retention surface. WD 2.5" Marvell drives (WD10SPZX, WD20SPZX) use a magnetic latch instead of a screw on some sub-revisions, which the technician identifies visually before lifting. Toshiba MQ drives use a standard screw pivot across the family. Helium drives (Ultrastar DC HC series, Exos X) require the additional step of breaking the cover hermetic seal under the bench and refilling the cavity with helium from a calibrated line before the cover is re-bonded. Helium head swaps use $3,000–$4,500; surface damage cases use $4,000–$5,000, plus helium and donor costs.

PC-3000 Head Testing Workflow

After installing a donor HSA, the technician uses PC-3000 to evaluate whether the donor heads can read the patient drive's data. This is a structured diagnostic sequence, not a single pass/fail test. The workflow determines head compatibility, identifies weak heads, & guides the imaging strategy.

1. SA Module Read Test. Power on the patient drive with the donor HSA installed. PC-3000 attempts to read the Service Area modules (the firmware stored on reserved platter tracks). If the drive reads its SA successfully, the donor heads are electrically compatible with the patient's controller & preamp circuit. If the SA read fails, the heads are incompatible or the adaptives need transfer.
2. Head Stability Evaluation. PC-3000 reads sample sectors from each head position & reports the error rate per head. Donor heads are never perfectly calibrated for the patient's platters; some read degradation is expected. The technician evaluates whether each head reads well enough for sustained imaging or if specific heads need to be disabled in the head map to prevent platter damage from a weak head dragging.
3. Conditional Adaptive Transfer. If the SA read succeeded but the error rate is high, the technician transfers the donor's adaptive parameters into the patient firmware. On WD drives, this means replacing Module 47. On Seagate F3 drives, this means editing the SA adaptive tables. The drive is then power-cycled & the head stability test is repeated.
4. Selective Head Map Configuration. If one or more heads are unstable after adaptive transfer, PC-3000 can disable specific heads in the firmware head map. The imaging proceeds using only the stable heads, recovering data from the platter surfaces those heads can reach. Data on surfaces served by the disabled heads may require a second donor attempt with a better-matched HSA.

Post-Swap Diagnostic Log Signals

The PC-3000 terminal log and the drive's SMART table are the primary sources used to classify post-swap failure modes. On Seagate F3 drives, two terminal errors appear frequently after a head swap and each points to a specific remediation path. SIM Error 1002 indicates defect list corruption, where the P-list or G-list stored in the SA has been written by the controller before the adaptive transfer stabilized. SIM Error 2044 indicates damaged translator tables, where the LBA-to-PBA mapping is no longer readable by the installed heads. Both require translator regeneration in PC-3000 before imaging can proceed; imaging against a corrupted translator returns scrambled data even when the heads are physically stable.

The SMART attributes monitored during the first imaging pass after a swap report on donor-head stability. Attribute 5 (Reallocated Sector Count) rising during imaging indicates the donor heads are exhausting the patient drive's spare sector pool, which usually means the microjog averaging did not bring the heads close enough to track center and the controller is trying to remap marginal sectors. Attribute 197 (Current Pending Sector Count) rising indicates the read channel is struggling to verify data integrity under the installed heads and is deferring the reallocation decision on a growing list of marginal sectors. Attribute 193 (Load/Unload Cycle Count) climbing rapidly during the read pass indicates the donor heads are retreating to the ramp because the servo is losing sync; excessive parking during imaging is a precursor to a complete SA-read failure and usually means a second donor attempt is warranted.

Modern 20TB+ Enterprise Families and Donor Matching

Enterprise HDDs released from 2024 onward introduce recording-layer technologies that tighten donor matching beyond the rules that apply to PMR and helium-PMR families. The three platforms below dominate current high-capacity recovery casework, and each one changes which fields on the patient label become hard match criteria. Donor windows shrink, cross-family substitution is ruled out, and additional verification steps run on the PC-3000 Portable III before any head-comb work begins. The procedures referenced here run inside our in-house mechanical recovery lab under the same 0.02 micron ULPA clean bench used for consumer drive head swaps.

WD Ultrastar DC HC560 and HC580 (TSA HelioSeal)

The Ultrastar DC HC560 (20 TB) and HC580 (24 TB CMR) families add a Triple-Stage Actuator (TSA) on top of the primary VCM and the secondary milli-actuator. The TSA is a tertiary micro-actuation layer integrated into each slider and is calibrated per HSA during factory test. A donor must carry the matching TSA calibration generation, because the TSA bias currents and the servo loop integration with the primary VCM differ between calibration batches. The HelioSeal recipe iterates per SKU on these drives, so donors that differ in helium-fill batch can also differ in internal mechanical clearances.

The practical effect on donor selection is a donor window measured in weeks rather than the standard 3-month window used for older PMR helium drives. The Site Code and the full DCM string on the label become hard match criteria. A site-mismatched donor inside the HC560 family is treated as a different drive for matching purposes, even when the firmware revision is identical.

Seagate Exos X22 and X24 (TDMR Dual-Reader)

The Exos X22 (22 TB) and X24 (24 TB) families use Two-Dimensional Magnetic Recording (TDMR) heads. Each read head carries two physically offset reader elements, and the read channel combines the two streams to cancel Adjacent Track Interference (ATI) at high linear density. Matching the dual-reader generation is a hard requirement, because the dual-stream cancellation kernels stored in the SA adaptive tables are specific to the reader pair geometry. A donor with an older single-reader head or with a different TDMR generation produces a partial spin-up where IDENTIFY DEVICE answers, but the LBA range covering the affected zone returns unrecoverable errors because the channel cannot resolve TDMR streams against single-reader signals.

Donor verification on the Exos X22 and X24 includes a check of the head generation field in the SA, read through PC-3000 with the Seagate F3 utility, before the patient cover is opened. The check confirms the dual-reader generation matches the patient, not only the model number and firmware revision.

Toshiba MG10 and MG11 (FC-MAMR)

The Toshiba MG10 (20 TB) and MG11 (24 TB) families use Flux Control Microwave-Assisted Magnetic Recording (FC-MAMR). Each write head carries a spin-torque oscillator that emits a microwave field at the gap, lowering the effective switching field of the recording layer so smaller bit cells can be written reliably. Donor matching uses the HDD Code printed on the drive label as a hard criterion, and the first six characters of the serial number encode the head count and HSA layout. The two fields together define the head generation, and a mismatch in either field disqualifies the donor at the label inspection stage.

The microwave bias parameters for the spin-torque oscillator are stored in the SA adaptive tables and transferred during adaptive transfer on PC-3000, the same way standard PMR adaptive parameters are transferred. A donor that matches the HDD Code and the head-count serial prefix typically requires standard adaptive transfer; a donor that does not match either field is set aside before HSA extraction.

Modern HDD Family Quick Reference

OptiNAND PCB and iNAND Migration

Western Digital OptiNAND drives move metadata that older designs kept inside the platter Service Area onto an iNAND UFS chip soldered to the PCB. The metadata includes the Repeatable Run-Out (RRO) tables, the Adjacent Track Interference (ATI) refresh accounting, and the ArmorCache emergency-power-off cache. A standard SPI ROM transplant procedure is not sufficient on these drives, because the EAMR head positioning servo loop depends on data that no longer lives on the platters. Donor PCB swaps and ROM transfers on OptiNAND drives require treating the iNAND as a separately migrated component.

The correct PCB transplant on an OptiNAND drive moves the iNAND UFS chip onto the donor PCB so the donor controller reads patient-specific RRO and ATI metadata, or moves the SoC and the iNAND together as a paired unit. The iNAND footprint on the PCB is a discrete UFS package next to the controller, and it is reworked under the same Atten 862 hot air station used for SPI ROM transplants, with FLIR thermal monitoring to keep the package within the manufacturer reflow profile. After the migration, the controller reads RRO and ATI from the patient iNAND and the original head positioning model is restored.

An iNAND failure that is independent of the heads is a distinct failure mode from a classic head crash. The patient HSA can be mechanically healthy, but with the iNAND unreadable the ATI and RRO indices are lost and the heads cannot stabilize tracking against recorded data. The diagnostic signature is a drive that spins up, completes IDENTIFY DEVICE, and then issues continuous head-recalibration retries with no measurable user-area read progress. The recovery path is to image the iNAND on the PC-3000 in UFS mode, repair the metadata where possible, and reinstall the iNAND on the donor PCB before any HSA work.

Mozaic HAMR Donor Matching Constraints

The Seagate Mozaic HAMR HSA carries an edge-emitting nanophotonic laser diode, an integrated optical waveguide, and a plasmonic near-field transducer that heats a sub-50-nanometer spot on the recording layer during writes. The platters are an iron-platinum (FePt) superlattice deposited on a glass-ceramic substrate rather than aluminum, and the published areal density target is 1.742 TB per square inch on the first-generation Mozaic 3+ family. Donor matching on Mozaic HAMR is the strictest case in current production HDDs.

Donor heads from any PMR or helium-PMR family are strictly incompatible with a Mozaic HAMR chassis. There is no laser path, no optical waveguide, and no plasmonic transducer on PMR heads, and the recording layer on PMR platters does not require heat-assisted writing. A PMR HSA installed on a HAMR chassis cannot write the recording layer at all and cannot read the existing FePt media stably because the read element is not calibrated for the higher coercivity and finer track pitch of the FePt layer.

Within the HAMR family, donor matching requires exact same-batch sourcing at minimum. The plasmonic near-field transducer demands nanometer-scale optical-path alignment between the laser diode, the waveguide, and the slider, and that alignment is set at factory test for each HSA. Manual head-comb insertion under a standard clean bench can introduce torsion on the HSA flex cable that misaligns the optical path past the plasmonic transducer's working tolerance. HAMR head swaps in the lab use a head-comb procedure adapted for the HAMR slider geometry, and the imaging pass is preceded by a full PC-3000 laser-bias adaptive transfer.

Crash damage on the glass-ceramic substrate changes the local thermal response of the surrounding sectors, which alters how the donor laser heats the FePt media during re-imaging. The damaged region of the platter no longer accepts the same plasmonic heating profile that the donor was calibrated to, and the read signal degrades near the crash boundary. This is a physical constraint on the recovery, documented in the case file with platter inspection photos, rather than expressed as a recovery percentage.

Down-Bin Head Depopulation and Larger-Capacity Donors

Manufacturers ship multi-platter HDAs with one or more heads disabled in firmware so a single mechanical platform can fill multiple lower-capacity SKUs. A 4-head HDA may ship as a 3-head or 2-head SKU with the surplus heads marked inactive in the head map. The HDA, the preamp, and the HSA generation are physically identical across the binned SKUs. This opens a procedural option in donor matching: a larger-capacity donor in the same family can be used for a down-binned patient when the active-head positions on the patient align with the physical-head positions on the donor.

On WD Marvell platforms, the active-head bitmask is stored in Module 0A with a ROM backup in Module 102. The bitmask is edited on the PC-3000 so the patient's active positions match the physical heads being installed from the donor HSA, and the surplus donor heads are masked inactive before the first power-up. The procedure is a controlled module edit, not a wholesale module copy, because Module 0A also contains adaptive references that must remain calibrated to the patient.

On Seagate F3 platforms, the Zones and Heads Inactivation P-List path is used to mark unused heads inactive on the donor side before the swap, or to retire a marginal head after the swap when a single donor head proves unstable on the patient's platter surface. The Heads Inactivation entries persist in the SA and survive subsequent power cycles, so the controller stops attempting to read tracks on the masked heads during imaging.

The standard disqualifiers still apply. A larger-capacity donor from a different HSA generation cannot be used regardless of capacity, because the preamp, head attachment geometry, and adaptive parameter set differ. A larger-capacity donor from a different preamp revision cannot be used either, even when the HSA generation matches, for the reasons documented in the preamp compatibility section. Down-bin substitution is a procedural option that expands the donor pool, not a way to bypass the six PC-3000 matching criteria.

Drive Family Matching Pitfalls

Each drive family has specific compatibility traps that general matching rules do not cover. These pitfalls come from undocumented component changes, manufacturing site differences, & firmware sub-revisions that are not reflected on the drive label. The following notes apply to the drive families most commonly seen in recovery labs.

Seagate Rosewood Pitfalls

The Rosewood 2.5" family (ST1000LM035, ST2000LM007, ST1000LM048) accounts for a large share of consumer head-swap cases. The primary matching hazard is the preamp revision. Two preamp generations (C202 & 8202) are in circulation, & neither is printed on the label. Drives manufactured before approximately mid-2017 tend to use C202; drives manufactured from 2018 onward tend to use 8202. The Date of Manufacture printed on the label provides the best external estimate.

A secondary clue is the serial number. Drives with matching 2nd & 3rd serial number characters are more likely to share internal components, though this correlation is not absolute. The only definitive verification is opening the drive & reading the preamp marking on the HSA flex cable, or reading the preamp identifier through PC-3000 after a successful SA initialization.

WD Blue & Green (Marvell Platform)

Modern WD 2.5" drives (WD10SPZX, WD20SPZX, WD10SPCX) use the Marvell controller platform. The DCM is the critical external matching reference. The 5th character identifies the head stack supplier; donors with a different 5th character will have a physically different HSA even when the model number & firmware match.

For SMR variants, the translator layer that maps logical blocks to overlapping physical tracks adds a firmware-level matching requirement. Module 190 in the ROM contains the SMR operational parameters. A donor from a CMR variant of the same model line will not have valid Module 190 data, and the firmware will fail to initialize the translator after the head swap.

Toshiba MQ Series

Toshiba MQ drives (MQ01ABD050, MQ01ABD100, MQ04ABF100) are the most forgiving for donor matching. Matching the full model number is sufficient in most cases. The HDD Code printed on the label serves as a secondary criterion if the first donor produces unstable reads.

The one consistent pitfall is country of manufacture. Toshiba drives from Chinese factories & Philippine factories occasionally differ in read channel calibration. If a model-matched donor fails, the technician tries a donor from the same country of manufacture before escalating to other diagnostic steps.

Samsung / Seagate Momentus Hybrid Drives

After Seagate acquired Samsung's HDD division, several drive models carry Seagate branding but use Samsung-designed internals. The ST1000LM024 is a Samsung Spinpoint M8 with a Seagate label. Donor matching for these drives must follow Samsung logic (full model number & PCB part number), not Seagate F3 logic. Using a Seagate F3 donor for a Seagate-branded Samsung drive will fail because the SA module layout, head addressing, & preamp circuitry are Samsung architecture. PC-3000's Samsung utility (not the Seagate F3 utility) is required for SA access & adaptive parameter management on these drives.

Head swap on Samsung/Seagate hybrid drives follows standard head swap pricing at $1,200–$1,500. 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. +$100 rush fee to move to the front of the queue

For complex donor-sourcing cases, customers can track progress through our in-house mechanical recovery lab, where the donor, adaptive transfer, and imaging steps are all performed under a single PC-3000 workstation without subcontracting.

Frequently Asked Questions