Performance Testing and Shucking the WD Elements External Drive - is it worth it?

[@yangxu2496]

Very critical finding at 10:30 mark. Thanks for sharing.

[@jovankadoo]

Even tho I would never go down the "shuck lane" for my home "production" (went with EXOS with 5y/warranty which replaced REDs with 5y/warranty), I appreciate the "how to open it" part, it will be useful sooner or later! :) cheers!

[@godminnette2]

The one 14tb drive I my home server was shucked from a WD shell about four years ago. It was a crazy steal at the time - the drive was like 40% off at microcenter, iirc. I just bought a second 14tb to back it up to, as I'm concerned about losing some of my more... Obscure Linux ISOs.

[@sometechguy]

Those Linux ISOs need preserving. 🥳

[@FennecTECH]

The boards from these drives are still very useful for drive testing and getting at data on old stacked in the corner drives.

[@sometechguy]

They are. I saw some report they didn’t work on non-WD drives but I tested it on a WD Black and a Seagate Skyhawk and seems fine. I didn’t yet finish testing these two drives comparative performance against SATA yet. So may be interesting.

[@BoraHorzaGobuchul]

Though buying a half decent drive dock is way more convenient if you have to deal with drives more or less often

[@jodajackson4489]

Last I checked, around the holiday season stateside, a new 20TB WD Elements drive was about ~$250 USD. This was comparable to a number of 20TB re-certified drives from serverpartdeals. IIRC the latter had a longer warranty than the former.

[@frankwong9486]

Is there any chance USB device in windows have additional/ different cache mechanisms compare to sata ?

[@sometechguy]

You can configure additional caching for USB devices, which I believe uses system RAM but then requires more careful removal of the storage. This wasn’t added for this test however, and not sure how much impact that would have over a full disk write.

[@TerminalWorld]

It would be great if you released your test suit so that we could compare out drives to whatever we can see in the videos.

[@sometechguy]

I may do this at some point. Maintaining it for others is a bit more of a burden than for myself, and some of the setup of the test config files feels like it isn’t as neat as it could be. I also started running tests on a load of NVme SSDs for some content I am working on and there may be some functionality tweaks there.

[@amon_69]

i wonder how different the performance under linux would be as some of the performance hits seem to be OS or driver related.

[@sometechguy]

I am actually going to do this. My test rig is dual boot and the test tooling is portable as it uses python, so I should be able to test with all the hardware the same between Windows and Linux. My gut feel is that it should be marginal, but nothing tells the story like actually testing it and the results can be surprising. I wasn’t really expecting such a difference here between the enclosure and the SATA attachment.

[@DaystromDataConcepts]

Great video. However, I found that Seagate Exos drives typically are even cheaper than WD Elements units and offer far better potential reliability. Also, you made no mention of the other thing I've seen with some higher end WD Elements units where it appears that one of the SATA pins on the drive is altered in a way to disable the drive when directly connected to SATA i.e a way of discouraging shucking. The fix is to tape over the pin on the drive to permit spin up.

[@sometechguy]

I know about the pin3 issue, but I have personally never experienced it. It’s more dependant on the power supply, but yes I could have mentioned it. More likely to happen if the drive is based on Enterprise drive I guess.l, and if you are running an older power supply, which can happen in a lab box.

[@seansingh4421]

Buddy Seagate and Reliability are two words that don’t usually go together 😂😂😂

[@RandallRoman]

Ran into this issue myself with a shucked WD drive that I wanted for a NAS. It was a white label. Had to tape the pin. I’ve decided to pass on all WD drives for now and have switched to Toshiba NAS models. Still 7200 RPM and 256 cache at the price of WD Red Plus which are 5400 RPM. Much cheaper than Red Pros

[@DaystromDataConcepts]

@@seansingh4421 I would have agreed with you until a few years ago when I was a WD guy. The SMR WD debackle had me bite the bullet and opt for the Ironwolf drives. So far, having 8 of them has resulted in zero failures. Seagate had a deserved poor reputation but I think if you look at the Back Blaze data you will find that they have sorted out their house and improved reliability quite significantly.

[@seansingh4421]

@ But is all of that worth dealing with Seagate’s RMA process ? Like WD is very nice when it comes to all of that. Seagate support has basically straight up tried to rip me off by saying “return device not received” even though the tracking shows delivered.

[@wiedapp]

So, if someone just needs bulk storage, these drives can be an option. If someone needs bulk storage with better R/W performance, they better look for the right drive from the get go.

[@sometechguy]

Yes, I think that’s a good assessment. There are not any stated performance numbers on these disks that I have found, so I don’t think there should be any specific expectations. But I don’t think most would expect the drive to perform worse when SATA attached. If anything you may expect a slight improvement. So this is the real takeaway. But at least with some performance data you can now know what to expect. 😄

[@ricsip]

To be honest I was very surprised, that WD didnt use hard drives with the USB interface already built on the disks controller PCB! When I disassembled a 2,5" external WD box many years ago, I was shocked that the drive inside didnt have a SATA interface, it had a built-in USB interface on its PCB already. So good luck using those as SATA drives. Maybe the 3,5" sized units are not built with integrated USB due to no size-constrainsts in their caddy.

[@sometechguy]

I would expect that multiple disk models are very similar physically and run different firmware, or some minor component changes like vibration sensors to optimise for different use cases, noise, performance, power etc. So it’s cheaper to take a unit that’s already being mass produced and put it in the caddy, and just add the external board, rather than have a different assembly process. But yes, it could just be a different PCB, so maybe there is more to it than this. 🤷‍♂️

[@AlistairBrugsch]

@@ricsip I fell foul of that annoying practice a few years ago. My neighbour had WD passport drive (2.5" spinning rust in a USB enclosure) and the USB device stopped being recognised by any OS. Never mind, I'll get the data off for you by whipping the drive out and connecting it to my SATA... Oh shi.... 😱 I suddenly became very afraid of the near identical passport I was using to back up photos with.

[@frederichardy8844]

Here in France the price difference is no more interesting: for 18TB I can find SAS enterprise drive for about the same price. It was few years ago and I have several (a least 18) shucked drive (in a synology and a HPE DL380 Gen9) with no problem.

[@sometechguy]

Yes, it seems to vary. There are some bargains but I think larger capacity drives, especially enterprise drives got more affordable. Just need to pick out the capacities that give the best price per TB. These external drives do seem to go on offer, and when they do the price can drop more than bare drives however. So some good prices from time to time.

[@clansome]

I have always shucked Elements mainly for my Plex server. Several 8, 10, 12 and now 14 TB pairs. Haven't had a failure yet. Note that the Plex server is really just an ordinary PC, really needs to be transferred into a NAS which I guess I will do soon. I have noticed that Elements drives are no longer the bargain that they used to be.

[@StephenMatrese]

In the United States, federal law forbids the denial of warranty for shucking under the Magnuson-Moss Warranty Act

[@sometechguy]

This is a great call out. Though I hear stories that attempts are made to deny warranty unless you push it, so even more important to understand your rights. This kind of right to repair related legislation is badly needed on this side of the pond also, though I think protection also exists in the EU.

[@dimples282]

Bit surprising that it performed a better and more consistently over a USB link rather than a direct SATA hookup. Surely I thought it would've been the other way around.

[@sometechguy]

Yes, a bit surprising. But having run all the tests repeatedly, they are very consistent. But I suspect it’s just that the firmware is tuned and optimised to this purpose.

[@nikolarun]

interesting findings , I wonder if flashing a different firmware can restore the performance

[@nikolarun]

i also wonder if it performs the same under xfs/ btrfs or unraid etc.

[@ianemery2925]

Heat kills electronics, so running a disk at a cooler temperature should be highlighted a little more; other than that, a great video.

[@esecallum]

MY PC STILL WORKS AFTER 15 YEARS BECAUSE I TOOK THE SIDE PANEL OFF !!!!!

[@ianemery2925]

@@esecallum Hopefully, because you clean out the dust regularly, and bought good quality components. My last 3TB HDD is showing some signs of failure, I bought it waaaaaay back, 12-15 years ago, possibly longer; back when 3TB cost £55; all the electronics test as fine, but it seems to be a little slow in spinning up from cold; and is running 10C hotter than my newer 4TB HDD; mind you, it also has slightly better read and write speeds than the newer drive, despite both spinning at the same RPM; so it might just be it's position inside the PC. CrystalDiskMark and other tests all report it as "Good" though. Higher temperatures cause electronic components to age faster - speaking as an ex-electronics tech.

[@esecallum]

@@ianemery2925 MOTHERBOARD IS h77-3dh intel

[@ianemery2925]

@@esecallum Motherboards tend to last, unless badly overclocked, my daughters PC is built on a similar age Intel board; and until a year ago, my parents were rocking an Intel 775 board - that is 20+ years old.

[@matt4054yt]

Why are external drives cheaper than internal ones in the first place? Obviously the enclosure and USB board should make them more expensive, not less. I tried to look it up and most answers are to the tune of "they use lower quality drives", but there is no hard evidence for that. This video doesn't seem to confirm that hypothesis either. So, what's your take?

[@sometechguy]

I think it’s likely just about stock levels and demand. There are going to be economies of scale with production, and external drives open a new market for those that want storage or backup but don’t want to install drives. And the price is probably just getting what the market allows given the currently competitive position with the likes of Seagate. So my take would less about quality, and more about the balance of sales volume vs margin. Also, there will be sales targets in the manufacturers but also at retailers. So at the end of a quarter, they likely drop prices on what’s in stock to close the quarter and make their numbers so they are on track with their forecasts. Public companies especially like to look predictable and that they understand their business, so always pressure to hit sales targets, even if the margin is lower.

[@silverecco]

The first generation of AMD phenoms had a tri-core option that was reused quad-core silicon with a defect in one of the cores. But demand was so high that they ended up just having to use acceptable quad-core silicon with one core turned off to fulfill orders. Could be something similar here. Use lower reliability drives when available, but otherwise just throw whatever is cheapest in there. You'd have to sell a lot of drives to equal the cost of starting a whole dedicated production line. A lot of the good-better-best pricing options are arbitrary and artificial rather than a reflection of actual costs. Consumers just like having options, even if illusory.

[@Don-ii4vm]

It would be nice if the noise drives make were also mentioned.

[@sometechguy]

It’s a quiet drive and has a low rotational speed.

[@Don-ii4vm]

@@sometechguy Awesome, Thanks.

[@sometechguy]

@@Don-ii4vm my test environment is a bit noisy with fans etc so not ideal for actual sound measurement, but I can hear enterprise drives when they are under test. I couldn’t hear anything from this,either in its enclosure or when it was in the SATA attached test enclosure.

[@damienretro4416]

I shucked some 10TB WD drives but they were 7200RPM and run insanely hot in my NAS. Up 20C hotter than Red Plus drives.

[@sometechguy]

Were these also white labelled? Or were they blues?

[@damienretro4416]

@@sometechguy Yeah, white label

[@steven4601]

If your in need of a CMR drive : first dead giveaway is if the drive has only 64/128mb cache .

[@adrianandrews2254]

Why not leave the USB board on and connect to an internal USB3 header ? Ali Express sell suitable cables cheap.

[@sometechguy]

For many people this negates all the reasons for shucking in the first place I would guess. Fitting in a NAS, or using on board RAID. And of course, it won’t go in a drive bay like that either. But if you are going to USB connect it, just use it as it comes out of the box. It’s a solid drive performance wise for a desktop class drive when USB attached.

[@le-zakes]

Judging purely in terms of price/TB, 8TB drives are unlikely to provide better value than larger capacity external drives. 8TB prices are currently around $150 USD or $18.75/TB. This is mostly in line with pricing for larger drives, however recent sales have placed 18 and 20TB drives below the $13/TB price point. Waiting for these sales and buying when prices are ridiculously low is the reason people advocate drive shucking

[@esecallum]

USE old htct drives 20 tb for 40

[@pokiller01]

home NAS users should avoid any drive that is 8TB or less.

[@borlibaer]

In this case it is a swanky HDS 8 TB HDD 👍

[@TnFruit]

Maybe a 280MB/s helium drive could go up to 350MB/s in the USB-Case? 🥴

[@benja1378]

Just buy recertified honesty

[@alain-m2t]

I dropped one on the floor, lost 8TB of data NEVER TRUST mobile device - go NAS

[@cosettapessa6417]

Wow so slow if out of the caddy

[@cosettapessa6417]

And so much cooler

[@OdinReactor]

Dude this is terrible. You fast forward through the important parts or don't have it in the camera shot. 👎 by 8:30.

[@sometechguy]

Thanks for the feedback, but I think all the important parts are fully in shot. The key thing is how to lever the case and separate the clips. There is a brief part where its out of shot when unscrewing a screw on the board, but i don't think this is going to cause any confusion. But of your course, I respect your feedback and opinion and sorry if this made it hard to follow.

[@ricsip]

​@@sometechguyI would also recommend if your video has added value in the topic of disassembly, try to make that process as visible as possible. Multiple angles, zoomed in, slowed down / even pause and highlight the parts that need special attention. But otherwise, very useful info about how not to break those bloody plastic clips.

[@AlistairBrugsch]

​@@sometechguy don't worry. Most people don't need a zoomed in close up slo-mo of undoing a screw. Maybe this is their first ever KZbin tear-down video and need to see every step in excruciating detail, or some people exist just to be a pedant on the internet. As you said, the important bit of unclipping the case without breaking clips was perfectly demonstrated.

[@esecallum]

use 4 cameras. use old moblies. use a extreme close up camera