Found this antique SSD: Samsung MZ-HPU256T/004. M.2 2280 format, other P/N is MZHPU256-00004. Google even has problems with this antique. The date is 2014.10. power is 3.3V 1.9A.

This dinosaur is PCIe 2.0 its so old, and my USB-C dock can barely recognize it. My dock supports NVMe and SATA SSD. Evidently this SSD is MLC making it higher endurance than modern TLC and QLC SSD. The SSD is not NVMe and CrystalDiskInfo does not recognize it. The SSD does not seem to be borked its just a dinosaur for compatibility.

Windows DISKPART does not recognize it but it does show up in the Disk Manager. Windows assigns a drive letter but I cannot format it in Windows 11. It seems to have a MBR partition but no data partitions. It also is noticed by the device manager.

Guess the device is so old that Windows does not handle AHCI class SSD? My UEFI is set to AHCI which is what the default on Dell.

I have an old 4th gen Asus machine I am refurbishing but I do have a manual for it so it's backburner. Machine has dual SATA sockets for SSD/HD and it has GTX 980 graphics so I am eventually try to get it back in business. Might be a PCIe slot in the machine I can take advantage of. Maybe my old Lenovo P52 can use it better, machine has dual M.2 slots.

_________________
USB 3.5" enclosures are $19 USD postage paid from China, I own several

Interesting. Can you please post a photo of it?

Image hosts are useless so I attached it

_________________
USB 3.5" enclosures are $19 USD postage paid from China, I own several

Can you peel back the sticker and show us the controller?

_________________
A backup a day keeps DR away.

The SSD is about 12 years old so I prefer not to injure it at all as the adhesive could be problematic. Samsung is respectable but PCIe 2.0 SSD is before the NVMe standard came along so it's that old.

I am wondering if the SSD has any bad blocks as this device is so fossilized now?

_________________
USB 3.5" enclosures are $19 USD postage paid from China, I own several

Yeah, right. BTW, you have a lousy camera.

_________________
A backup a day keeps DR away.

My phone has 48 megapixel which is more than this forum can handle

_________________
USB 3.5" enclosures are $19 USD postage paid from China, I own several

ROTFLMAO!

_________________
A backup a day keeps DR away.

Just to rule out the obvious, have you checked whether the enclosure is supplying the correct SATA voltage before assuming this is a PCIe compatibility problem?

A normal 2.5" SATA SSD expects 5V, while 3.5" SATA drives normally need both 5V and 12V. M.2 SATA is different again because it runs from 3.3V, and PCIe M.2 also uses 3.3V, although obviously the protocol is completely different. The label on this Samsung says 3.3V 1.9A, so theoretically that is about 6.27W, but that doesn't tell us whether an enclosure rated 5V 2A, 5V 3A, 12V 2A, or USB-C PD 20V is actually converting it properly.

This is why I don't completely trust "NVMe + SATA compatible" USB enclosures. Some seem to detect B-key, M-key and B+M-key electrically, others seem to have separate SATA and PCIe bridge chips, and some appear to just throw 3.3V at the socket and hope the controller negotiates something useful.

Also USB 3.0 is supposedly 5V 900mA, USB-C can provide 5V 3A before PD, SATA power has 3.3V/5V/12V rails, and an old desktop PSU can supply dozens of amps on 12V, so perhaps the 1.9A printed on the SSD is why a cheap enclosure is unhappy even though the total wattage seems low.

Then there is AHCI. If this is PCIe AHCI rather than SATA AHCI or NVMe, could a USB NVMe bridge see the PCIe link but fail because it expects an NVMe controller? If so, would a passive M.2-to-PCIe adapter work better because there is no USB bridge involved?

I also wonder whether SATA III being 6Gbps, USB 3 being 5Gbps, PCIe 2.0 being 5GT/s per lane, and NVMe not actually being a bus speed at all is confusing some of these enclosure manufacturers.

Could even be the 3.3V rail sagging to 3.1V during startup if the SSD really pulls close to 1.9A. Or perhaps it needs 12V because it's old. Or 5V because SATA. Or 3.3V because M.2. Depends which specification you ask.

Anyone have an enclosure that supports SATA, NVMe, PCIe AHCI, IDE, SAS, USB-C PD and preferably 3.3V/5V/12V simultaneously? Bonus points if it is $19 postage paid from China.

My most flexible USB adapter can handle SATA and NVMe M.2 devices. The Samsung is keyed M. The SSD is marned 3.3V on the label. 1.9A which is well below my USB-C port power capability/.

This SSD is M.2 but its evidently AHCI rather than NVMe so I wonder if a device driver is needed to use this class of product. I am always buying more enclosures for disks and SSD.

_________________
USB 3.5" enclosures are $19 USD postage paid from China, I own several

M-key makes it more interesting because electrically that should mean PCIe lanes are present, but it still doesn't tell the enclosure whether the commands travelling over those lanes are NVMe or AHCI.

Could the USB bridge be enumerating the PCIe side correctly but then waiting for an NVMe Identify Controller command that this Samsung never answers because it expects AHCI register access instead? In that case Windows could see "something" through the bridge while Disk Management sees a disk, Device Manager sees hardware, CrystalDiskInfo sees nothing useful, and the enclosure thinks everything is fine.

There is also the question of UASP. Most newer USB storage bridges use UAS instead of bulk-only transport, but AHCI itself already has NCQ, while NVMe has completely different queues. Maybe an AHCI PCIe M.2 SSD behind a UASP USB bridge ends up with AHCI NCQ inside SCSI UAS inside USB, which seems like an impressive number of command queues for a 256GB drive.

I would also check whether the enclosure has a Realtek, JMicron or ASMedia bridge. Some of those support SATA + NVMe, but "SATA + NVMe" may literally mean SATA protocol on one input and NVMe protocol on another, not arbitrary PCIe storage. PCIe AHCI would then fall into the hole between the two supported modes.

Power still bothers me slightly. 3.3V at 1.9A is over 6W, while the enclosure first receives 5V from USB and has to step that down to 3.3V. If the converter is 90% efficient it would need roughly 1.4A from the 5V side, ignoring the bridge controller itself. A USB 3.0 port rated 900mA could therefore be perfectly adequate according to USB enumeration but inadequate according to Ohm's law, unless the enclosure requests additional current, which AHCI presumably knows nothing about because that's USB's problem.

Then there is PCIe 2.0 x4. That's theoretically 20GT/s aggregate before encoding overhead, substantially more than USB 3.0, so perhaps a bridge designed around PCIe 3.0 NVMe sees four PCIe 2.0 lanes and negotiates down incorrectly. Or maybe it only uses x2. Or x1. Or the SSD is x4 but AHCI makes Windows think it's SATA 6Gbps even though there isn't actually a SATA link anywhere.

I'd be curious what happens with a completely passive M-key M.2 to PCIe x4 adapter in a desktop. No bridge, no USB, no SATA translation, just PCIe. If the BIOS exposes the device and Windows loads storahci for it, that would pretty much confirm the SSD is healthy and the USB enclosure is simply protocol-confused.

Although if the motherboard BIOS has "SATA mode: AHCI" enabled, I'm not sure whether that setting applies to PCIe AHCI devices too, because technically it says AHCI, but practically it normally controls the chipset SATA controller. So changing it from AHCI to RAID might either fix the SSD, hide every SATA disk in the computer, install Intel RST, or accomplish absolutely nothing.

For maximum compatibility I'd probably look for an enclosure advertised as M-key PCIe AHCI/NVMe + B-key SATA, UASP, TRIM, SMART passthrough, PCIe 2.0 backwards compatibility, 3.3V 2A output and USB-C 5V/9V/12V/15V/20V PD.

Preferably $19 postage paid.