Chapter 8 — Quiz

15 questions: 11 multiple choice, 2 true/false, 2 short answer. Answers and a scoring band are at the bottom. No peeking until you have committed to an answer — this chapter rewards restraint.

Multiple choice

Q1. Of all the sounds a failing hard drive can make, which is the single worst, and what is the correct immediate action? - A) The "click of death"; keep retrying to see if it stabilizes - B) A continuous grinding/scraping; power off immediately - C) A high-pitched beep; put it in the freezer - D) Total silence; open the drive to inspect the heads

Q2. A drive shows total silence at power-on — no spin, no LED, no sound at all. Why is this often described as the best news? - A) It means the platters are guaranteed undamaged and the data will always be recoverable - B) It usually indicates an electronics/PCB failure, the cheapest and most DIY-friendly family, with a typically intact HDA - C) It means the file system is merely corrupt and chkdsk will fix it - D) It proves the heads have safely parked and the drive can be reopened later

Q3. Why must you transfer the original drive's ROM when swapping a modern PCB onto it? - A) The ROM contains the user's files - B) The ROM stores per-drive adaptives (calibration unique to that drive's heads and platters) that the firmware needs to read the platters correctly - C) Without it the SATA connector will not fit - D) It is not necessary on any modern drive

Q4. What should the first ddrescue pass on a failing drive do? - A) Aggressively retry every bad sector until it reads - B) Run chkdsk first to repair the file system - C) Copy all the easy-to-read data first, with no scraping and no retries, to capture the bulk before the drive worsens - D) Read the disk from the end toward the beginning

Q5. A drive spins up smoothly, is quiet, and makes no clicking — yet the BIOS reports a capacity of 0 ("LBA0") or hangs busy. The most likely cause is: - A) A head crash - B) A seized spindle motor - C) A firmware / service-area fault (e.g., translator or log problem) - D) A scored platter

Q6. The Seagate 7200.11 "BSY" failure was famously fixed by: - A) Replacing the head-stack assembly in a clean room - B) Freezing the drive for several hours - C) Connecting to the drive's diagnostic serial port (3.3 V TTL) and issuing commands to clear an overflowed log and regenerate the translator - D) Swapping the PCB without transferring the ROM

Q7. What is the single clearest bright line separating "DIY-reasonable" from "send it to a lab"? - A) Whether the drive is larger than 1 TB - B) Whether the fix requires opening the sealed head-disk assembly (HDA) - C) Whether the drive is a Seagate or a Western Digital - D) Whether the client is in a hurry

Q8. A modern read/write head flies just a few nanometers above the platter. Why does that fact make a clean room non-negotiable for head work? - A) Because nanometer gaps generate dangerous static - B) Because a smoke particle (~250 nm), dust (~1,000–10,000 nm), or a hair (~70,000 nm) is enormous by comparison and will cause a head crash - C) Because the platters spin too slowly to clear debris - D) It does not — any clean desk is fine

Q9. Before quoting any physical recovery, the question that should always be asked first is: - A) "What is the drive's model number?" - B) "How fast do you need it?" - C) "Does a copy of this data exist anywhere else (a backup)?" - D) "What operating system was it formatted with?"

Q10. Which of the following is genuinely DIY-reasonable for a competent technician? - A) A clicking drive holding a client's only data - B) A grinding drive with a suspected head crash - C) Removing or replacing a confirmed shorted TVS diode on the PCB - D) Opening the HDA to swap heads on a single-platter drive at your bench

Q11. In a ddrescue mapfile, the status character + on a block means: - A) The block failed and is a bad sector - B) The block was read successfully (finished / good) - C) The block has not yet been tried - D) The block was trimmed but not scraped

True/False

Q12. Running chkdsk (or fsck, or Disk Utility First Aid) is a safe, sensible first step on a physically failing drive. (True / False)

Q13. Hashing the image you pull from a failing drive certifies the contents of the original physical drive. (True / False)

Short answer

Q14. In one or two sentences, explain what is physically happening when a drive grinds, and why that sound means "stop now" rather than "try once more."

Q15. Name the three factors that determine whether professional recovery is worth it (not merely possible), and state which one should be checked first.

---

Answer key

Q1 — B. Grinding means a head is in contact with the spinning platter, machining away the magnetic layer; every second destroys more data, so you kill power instantly. (The freezer is folklore; clicking is bad but grinding is worse.)

Q2 — B. Silence usually points to a dead PCB/electronics — the cheapest, most DIY-accessible family — with the sealed HDA typically intact. It is not a guarantee of recoverability, but it is the friendliest starting hand.

Q3 — B. Modern drives store per-drive adaptives (head/zone calibration) in the PCB's serial-flash ROM; the firmware needs this drive's values to interpret the platter signal. A donor board with the wrong adaptives may not initialize, may click, or may report the wrong capacity.

Q4 — C. The good-data-first strategy: grab the bulk on a fast, no-retry, no-scrape pass, then return for the dangerous areas later — so you capture the readable majority before the drive degrades further.

Q5 — C. A healthy-sounding, undetected (or 0-capacity / busy) drive is the classic signature of a firmware/service-area fault. The mechanics are fine; the data is locked out, not lost.

Q6 — C. The 7200.11 BSY bug was a service-area log overflow, repaired by talking to the drive's serial diagnostic port (3.3 V TTL, commonly 38400 baud) to clear the log and regenerate the translator — pure firmware surgery, no clean room.

Q7 — B. The sealed-HDA line is the bright line: anything outside the seal (cables, enclosure, PCB, ROM, host-side ddrescue imaging) is fair DIY game; anything requiring the HDA opened is clean-room/lab territory.

Q8 — B. The fly-height is smaller than common airborne particles, so any contaminant becomes a "boulder" under the head and causes a crash — hence ISO Class 5 / Class 100 conditions, created in practice by a laminar-flow clean bench.

Q9 — C. The cheapest recovery is the backup you already have; confirm whether another copy exists before spending a dollar on physical recovery.

Q10 — C. A confirmed shorted TVS diode can be removed or replaced outside the HDA — DIY-reasonable. The other three all involve clicking/grinding or opening the HDA, which are lab jobs.

Q11 — B. + = finished/recovered. (- = bad-sector, ? = non-tried, * = non-trimmed, / = non-scraped.)

Q12 — False. Repair tools write to the disk and force heavy head activity to walk file-system structures; on dying hardware this accelerates failure and can overwrite recoverable data. They are for logical problems on healthy drives only.

Q13 — False. The hash certifies the image you produced — that it has not changed since acquisition. On a degrading drive you cannot hash the original meaningfully (it is changing/unreadable), which is exactly why your first complete image becomes the best evidence and why every unreadable region must be documented.

Q14. A head has touched the spinning platter and is gouging the magnetic coating, turning data into dust (which then contaminates the rest of the drive); because the damage is permanent and cumulative with every rotation, the only correct action is to cut power immediately rather than risk one more pass.

Q15. (1) The value of the data (financial and emotional), (2) the probability of success (from the lab's evaluation), and (3) whether a copy exists anywhere else — and #3 should be checked first, because an existing backup makes recovery unnecessary.

Scoring: 13–15 — you can triage a dead drive by ear and know when to stop; you're ready for Chapter 9. 10–12 — solid; re-read the four-families and DIY-vs-lab sections. 7–9 — review diagnosis-by-sound and the ddrescue/imaging workflow before moving on. Below 7 — re-read the chapter, focusing on the logical-vs-physical fork and the "stop and think" restraint rules; this is the chapter where guessing costs data.