SSD Data Recovery Norwich

When the controller chip gives up, the drive disappears from the bus while the files are still in the NAND. There's nothing left to speak for the memory, and that is where a firmware bench comes in.

It's actually a Phison controller that has got itself stuck in its own boot loader, not a drive called anything of the sort. The files behind the wrong name were never touched, and service-mode work walks them out.

Translator failure. The drive stays polite and reports a size that doesn't belong to anything. We rebuild it in service mode, then image it in full before it's asked to do anything else.

Boot code, power rail or controller, and they all look the same from the outside: nothing on screen. A bench separates them, and it's the first thing the free diagnostic settles.

End-of-life protection stops writes so that what exists is preserved. Not very convenient at a desk, but on an imaging rig it's almost a gift, as the drive can't make things worse while we are copying it.

Flash drives don't usually give a warning before failing. They work fine on Friday and not on Monday, which is why so many calls to us begin with someone saying it was working perfectly the last time they used it.

An interruption mid-write scrambles the translation tables that turn logical addresses into physical ones. A real surge takes out the power stages as well. Both arrive here regularly and both are bench work.

Worn NAND reads back differently from one pass to the next, and SMART only quietly whispers about it. There's only one record worth trusting, the first complete image, so it is taken carefully and only once.

Deleted blocks keep being erased in the background as long as the drive has power. Pull the plug and that stops. If you think about it for too long with it still plugged in, your thinking costs you files.

Thin boards don't stand up to screwdrivers, upgrades and stiff envelopes. Track repair is used for a clean break near the edge. Anything worse goes to chip-off, which takes longer but works just as well.

Samsung's MZ-76E500 and its siblings are commonly found in UK desktops and laptops, so we see the whole range of what happens when one of them dies. The most common way is sudden: the drive disappears from the SATA bus mid-session, then either returns with a tiny capacity or doesn't return at all. No amount of re-cabling will revive it. The MJX controller has lost its translator, but the V-NAND behind it is alive and well. Service-mode work on the controller followed by a full image is the way back. If you've kept the power off since you noticed it acting up, you will always get the best results.

Samsung released a firmware fix after early batches of the MZ-77E1T0 were documented reallocating sectors at an alarming rate. Many of these units still arrive unpatched. The memory reads; it's the map you have to rebuild.

The Crucial MX500 has a well-known habit of showing a rising pending-sector count that worries owners more than it should, until the day it really is joined by read errors. Telling the harmless version from the real one is a bench job, not a guess.

WD's SATA Blue has been known to go into hiding until it gets some firmware attention. A familiar patient, a familiar process, and the NAND is most likely not the problem.

The A400 SA400S37 and its family will typically drop to a few MB, or be completely gone, once the translator gets corrupted. That's service-mode work, and the flash underneath is usually innocent.

They freeze during a write and then stop talking. Some can be revived in service mode; those that can't are read out at chip level instead.

SSUBX and SSPOLARIS blades from 2013 to 2017 Macs fail like any other SSD and are imaged here on native adapters. No donor Mac is used at all, which surprises people.

Since there's no DRAM cache, wear-levelling is clumsy, so a BX500 put under heavy use can get old fast. When that happens, the chips come off the board and the data comes out.

Fakes sold on online marketplaces loop their writes over a small slice of real flash. Anything written within the real capacity can still be recovered, but anything written after the loop was never stored at all.

During a normal swap the plastic tongue snaps and pulls a few pads off with it. It's a board-level repair, and once repaired, it acts like nothing ever happened.

Years of heavy writing use up the drive's rated life, and eventually it locks up or reads unpredictably. There's budget for one attempt at imaging it, so we plan that attempt as such.

The manufacturer's tool stopped mid-update and left a drive that can't even tell you its own name. Either it's reflashed on the bench, or bypassed completely, depending on how much the controller still has to say.

QLC flash runs quickly until its fast cache is full, at which point performance drops dramatically. Users pull the plug when the drive seems to have hung, but it's actually pulling the plug that causes the real harm, not the slower speed.

Each time you read a block, you also nudge the charge in its neighbours. On a drive full of archives that are read often but rarely or never rewritten, these errors build quietly until the error correction is overwhelmed. Re-reading slowly will usually recover much more data than a quick pass over the same area.

Cells keep leaking even after a drive has been powered down, so a drive last used in 2019 may read back as noise. But with repeated reads and patient error correction, you can recover quite a bit of what a single quick pass would have written off.

Over time, warming and cooling put stress on the BGA joints under the controller and memory chips. Rework restores the contact for long enough to take a clean image, which is all that's necessary.

Enterprise and datacentre SATA drives have capacitors built in to give just enough power to write out buffered data when the power goes. These capacitors eventually fail silently. Once they do, the next power cut takes the mapping tables with it, and nobody knew the safety net was gone.

Almost all modern SSDs encrypt data as they write it, whether or not you ever asked for encryption, and the key is stored inside the controller. If that key store gets corrupted, the drive is perfectly readable but completely meaningless. We tell you so plainly rather than selling hope.

More than one SSD maker has released firmware that simply stops the drive at a set number of power-on hours. The hour comes and the drive just sits there. It's repaired in service mode and imaged straight away.

When a self-encrypting SSD's security state glitches, it locks shut. The PSID on its label resets it, but that wipes everything, which is not what anyone wants. So we don't use it, and no one else should either.