WD10SPZX-21Z10T0 - Is this a firmware problem?

[Pimpom]

This 1TB WD HDD came in an Acer laptop that I bought for my daughter in January 2021. Everything was painfully slow from the beginning but lockdowns prevented me from returning the laptop within the return window. I tracked down the cause to the HDD and have since replaced it with an SSD, after which it behaved normally.

The HDD does not show any error in SMART, no reallocated sectors, no CRC error, no spin retry, no read/write error. Power on hours count is reasonable for the few months my daughter was forced to use it in its painfully slow condition. No bad sectors with surface scan. No start-up problem except that it takes minutes for the login screen to come up and minutes more for the desktop to appear. Opening MS Word, browser and even small fast apps like Irfanview are all excruciatingly slow.

A fresh installation of Windows makes makes it somewhat better at first but quickly deteriorates with use.

HDTune does show weird benchmark curves that keep changing with use. CrystalDiskMark shows unrealistically high speeds with small data sizes (like >300MB/s with 64MB) and more reasonable results with larger data sizes. Re-partitioning, full format and zero fill do not cure the problem. They just return the HDTune curve to a basically constant ~250 MB/s instead of the usual stepped downward curve. This curve changes to a widely fluctuating one with use.

Is this likely to be a problem with the firmware? If so, is there anything an end user can do?

[WebClaw]

Are you performing these tests in the ACER laptop?
Did you check to see if BIOS shows HDD is AHCI mode?

[Pimpom]

I performed the tests numerous times - in the laptop, as an external USB drive on my desktop and as an internal drive in the desktop. In AHCI and IDE modes. Same thing for the re-partitioning, format, zero fill, Windows 10 installation. All with essentially the same results.

Oh, and the HDTune benchmarks show very low access times (<1ms), especially after a fresh prepping. I'm inclined to think that there's something wrong with the interaction between the cache and the platter. Hence my inexpert opinion that it has something to do with the firmware.

[pepe]

Hello,

This model is SMR and i suspect the 2nd level translator being the culprit. High transfer speed with small amount of data is due to caching both on OS side and within the drive itself, so these are not representative.
My experience with these is that it performs fine for writing a few 10-100GB of data, then, when the translator population reaches a limit, it starts to defragment it, which leads to a dramatic fall in the transfer speeds. It all depends on the number and average size of written blocks.
I would not recommend such drive for installing OS on it, as the OS (windows at least) writes frequently on its host drive, triggering the above described symptoms soon.
It might be just fine to store user data or use it for backup purposes in an external enclosure, but it will always demand some idle time to sort its things out internally.

pepe

[Pimpom]

(Since I'm still in my first few posts, they are badly out of sync with others while they await approval. So please bear with me).

Thanks for the reply. I've used SMR drives before and am familiar with their characteristics in actual use. I've even used them for the OS drive. The weird behaviour of this drive goes way beyond that.

The drive is still in warranty but I don't know how thoroughly RMA-ed drives are tested at the service center. I'm wondering if they'll just return it as healthy after a quick test.

Here are a few sample screenshots out of many that I saved. There were still many more that I didn't save.

Attachments