Getting 7200 rpm drives when expecting 5400 rpm drives may result in increased power consumtion and heat issues.
Pluse 7200 rpm drives are noticably more noisy than 5400 rpm drives.
Many EMAZ and EZAZ seems to be affected.
It seems that the iOS app Phyphox can be used to determine if your drive is running at 7200 rpm or 5400 rpm using acoustic analysis.
From the comments:
> "There is an app called phyphox that allows you to access all of your phone's sensors. If you download it and use the Acceleration spectrum feature, you can measure the speed of your drive by simply placing your phone on top of the drive. The application is surprisingly accurate and I use it to check the health of my drives."
7200 RPM drives indeed have faster access (in terms of seek latency) and higher sustained throughput but, heat and power is not simple variables as one might think.
Especially RED drives are marketed for NAS systems (from home to SMB size). When a drives are installed into a closure, things like power and thermal budget comes into play.
NAS boxes are designed with strict limits and if they cannot cool themselves (and the drives) off or cannot supply the required power to drives; data corruption, performance/life span degradation may occur.
In extreme cases (with large storage arrays) calculated overall power budget can skew and may cause blackouts in some parts of the data center. There's a presumed ~2W difference in power consumption between 5.4K and 7.2K drives. We can safely install 60 of them in a 4U enclosure today. If we add 4 enclosures to a rack, that creates a ~480W difference, which is a one HPC server or 2-3 low-load, 1U or 2U servers. This is indeed significant.
While some of these drives are in MyBook enclosures, REDs may end up in very different places and may cause problems. In that case, head or spindle noise is not even the smallest concern.
If we're talking about noise, continuous (spindle) noise is much more degrading than head/seek noise. For me, even benign noises like USB fans have much more noticable effects after the day than mechanical keyboard noise for example.
If we consider WD80EFAX [0], which is one of the suspect drives, that figure is for idle power consumption where disk is spinning but heads are not doing anything.
So, that difference is for sustaining higher spindle speeds.
The peak power requirement is 1.79A (~21.5W) which is probably consumed when starting from full-stop state (stand-by, sleep or power-off where motor is turned off).
Ever drive a manual transmission? At very low RPM the engine will seem to shake more. It isn't so much that there is less vibration at higher speeds but that it occurs at a higher frequency. That means the amount the part physically moves from one side of the oscillation to another is less, just as a higher-revving engine doesn't bounce around as much as it would at lower RPM. The total amount of noise/vibration may be similar but the change in frequency can still have other effects.
Hmm, that app actually is pretty neat. It's published (for free, and under GPL) by the University of Aachen to support physics experiments. Recommended, at least from what I've seen so far.
The data from this tool is fascinating! When you run the Acceleration Spectrum experiment there's a history view that makes a 120hz vibration extremely trivial to see.
It is when it blows through your power budget and heating budget. Also, some array performance degraded significantly when using different speed drives, to much below the slowest drive (that’s not common anymore, but used to be an issue and might still be in some setups)
If I bought the drive for active work, yeah I'd probably be happy about the speed for the price (IF they don't artificially limit the speed!).
But if I buy these for NAS/backups, I want them to be quiet, cool and longer lasting - in my experience 5400 RPM drives do heat less and last longer than 7200 RPM.
Not sure if they still do, but WDs had this IntelliPark feature, which slammed the heads into parking a few times a minute (because otherwise your drive would last a decade, I guess), and their 7200rpm drives were clicking so loud I often thought "that's it, it failed". But no, just "normal" operation...
In this case it is. High capacity 7200 rpm drives can be unbelievably noisy, to the level that they are not recommended for home theatre NAS as you can hear the seeking sound during a movie or from the next room at night.
Speaking of seeking, if the data was laid out sequentially on disk, and the computer was not accessing any other files on the disk in question (spinning disks for movies and other data, SSD or m.2 for OS itself), then it would be very silent. But my impression is that a lot of the time we end up with files fragmented more than what is desirable. That's another wish of mine, that I would have a system that would keep this in mind and which would be better tuned to my use. Maybe one day.
> if the data was laid out sequentially on disk, and the computer was not accessing any other files on the disk in question (spinning disks for movies and other data, SSD or m.2 for OS itself), then it would be very silent
That's not true. Even a spinning sound of a 5400 rpm disk makes quite loud.
Whether you can hear it due to casing or distance is another matter. Most of the time fan noise will make you ignore it. When fan is not spinning I can hear laptop drives clearly.
It is very hard to trust hard drive manufacturers these days. I have a few of these drives mentioned here https://www.seagate.com/www-content/datasheets/pdfs/ironwolf... with the specified "typical" noise level of 25 dB for seek and it is definitely much much more than that.
The terrible sin was putting SMR (shingled magnetic recording) disks in drives sold as conventional magnetic recording. This may not be nearly at that level, but it's concerning that the habit of mis-labeling drives continues.
It is. 5400rpm drives have specific uses. It also hurts sales of competitors selling real 5400rpm drives. It may even hurt the whole 5400rpm drive market, if WD competitors can't match the price of the fake 5400rpm drives and decide to cease production.
I also heard that it might cause problems mixing 5400 rpm and 7200 rpm drives in a single NAS box due to vibration interference. Not sure if true.
But the general trend that Western Digital seems to be OK with lying on specifications is concerning.
Because “faster” is not the only variable involved. If customers wanted 7200 they’d buy that, but the difference in power and heat between a 5400 and 7200 can reach 100% which is an issue when you have sized and enclosure for one type and get the other.
Noise is also an issue, if you were looking for an HTPC drive for instance you don’t need the increase in performances but the noise pollution is a problem.
WD provides a data sheet with the power and noise specs for their individual drives. If you are buying to use in a situation where these matter shouldn't you be consulting the published specs instead of just choosing 5400 vs 7200? After all these can be greatly variable even between drives of the same RPM.
I mean... doesn't that same sheet of published specs state that it's running at 5400 RPM?
If they lie about that on their spec sheet, how can you believe the power/noise levels written on there?
Also, I've always had a fair amount of scepticism for digital product spec sheets. I've been looking for a projector with low input lag for a while, and the measured value has constantly been way off the claimed value (> factor of 2). Admittedly, this seems to most often be the case with some of the cheaper chinese brands, but definitely not exclusively.
They don't actually claim it is running at 5400 RPM. They claim performance is in the 5400RPM performance class. In other words it will perform similar to a drive running at 5400RPM. Nowhere do they mention that it runs at 5400RPM.
For the people who can't hear the vacuum cleaner noise from their PC because they use headphones yes, getting 7200 rpm at the price of 5400 might be nice.
For the people who want to run cool, quiet and less power hungry devices, they'd like to get exactly what they ordered. Nobody buys a 5400 rpm hard drive for speed.
Stick a 7200 rpm drive in a console and see how the cooling system copes, for example. Or a low power PC with a PicoPSU. Or a NAS.
Some portable hard drives can be powered off USB. If you get a 7200 advertised at 5400 drive, it won't run because it's not getting enough power.
These aren't home drives, where you have ample cooling in a PC case and power.
These are for NAS's/Servers where power and heat is a controlling factor, if admins wanted/could they would buy the 7200rpm drives and there would be a market for it, but they chose the 5400rpm because of constraints.
This is WD being lazy/cheap. Instead of selling them as 7200rpm drives, they're shovelling them into the 5400rpm lines as a presumably cost cutting measure.
They did not learn their lesson with the SMR/CMR scandal. Label your drives clearly.
The hard drive industry is such a weirdly complex marketplace. For example, there's people dedicated to "shucking" the external hard drives for the purpose of extracting the drives inside for use in a NAS or wherever, which would normally sell for higher than those with the external enclosure included, then chronicling the types and attributes of the drives they've found. The https://www.reddit.com/r/DataHoarder/ subreddit is a strange place.
Not the mybooks which is what people are shucking. I have about 20 in my nas and they’re going strong. For ~$120/8TB hard to go wrong when they’re essentially red drives inside (they’ve switched to white label drives with the same-ish specs).
The controller card inside of the enclosure is just a simple usb and power converter.
> USB hdd have dedicated controllers with direct usb bridge
In my (limited) experience, it's mostly the WD portables that have a direct USB connector. It's been going on for many years — I have a WD Passport with USB 2.0 that I've unsuccessfully tried to shuck.
The portable Seagate Backup drives are shuckable, but some of them are SMR.
That's strange. I would assume 7200rpm drive would be costly to produce compared to 5400rpm. What possible reason could they have to falsely advertising which could be a loss.
Producing 7200 RPM drives likely is more costly than 5400 in isolation, but running two full production lines itself is costly, when you could just run one and "fix it" via software.
Plus it allows you to vary the number of 5400/7200 RPM drives you "produce" to exactly what the market wants. Just keep them unlabeled in a warehouse as orders arrive: label, firmware, and ship.
How much mechanical difference is there between (designed as) 5400rpm and 7200rpm drives nowadays? I'm somewhat surprised that they seem to be faking the performance characteristics rather than changing the rotation speed?
No, the drive motor rpm is not controlled by OS drivers but by the drive controller. You'd have to write custom firmware and flash it onto the drive controller (inside the HDD) to change the rpm
Most likely not. The motors are most likely fixed speed. Even if not, the speed is unlikely to be setable through the sata interface, you would need to adjust the firmware --- and it's not usually a good idea to adjust hard drive firmware without deep knowledge.
I bought a WD SSD recently and when I plugged it in, the firmware reported that is was actually a SanDisk, which is I guess a lower-end WD brand at this point. I saw an Internet discussion of other people having this problem with a July 2020 batch of SSDs, so I texted into Amazon support which promised to send a WD drive that didn't report itself as a SanDisk drive to the system.
I'm sure you can guess where this goes. The replacement unit arrived a few days later-- this drive also was marked WD on the outside and identified itself to BIOS etc. as a SanDisk drive. I called Amazon support again and they appeared to consider it a manufacturing defect and I returned it for a refund.
I'm not sure how much I care about whether the drive is SanDisk or WD, but it ought to be consistent. If WD can't get that simple step right I'm not sure why I should trust them to secure and maintain the integrity of a TB of important data.
64 comments
[ 4.3 ms ] story [ 262 ms ] threadMany EMAZ and EZAZ seems to be affected.
It seems that the iOS app Phyphox can be used to determine if your drive is running at 7200 rpm or 5400 rpm using acoustic analysis.
From the comments:
> "There is an app called phyphox that allows you to access all of your phone's sensors. If you download it and use the Acceleration spectrum feature, you can measure the speed of your drive by simply placing your phone on top of the drive. The application is surprisingly accurate and I use it to check the health of my drives."
edit: I'm probably wrong
Especially RED drives are marketed for NAS systems (from home to SMB size). When a drives are installed into a closure, things like power and thermal budget comes into play.
NAS boxes are designed with strict limits and if they cannot cool themselves (and the drives) off or cannot supply the required power to drives; data corruption, performance/life span degradation may occur.
In extreme cases (with large storage arrays) calculated overall power budget can skew and may cause blackouts in some parts of the data center. There's a presumed ~2W difference in power consumption between 5.4K and 7.2K drives. We can safely install 60 of them in a 4U enclosure today. If we add 4 enclosures to a rack, that creates a ~480W difference, which is a one HPC server or 2-3 low-load, 1U or 2U servers. This is indeed significant.
While some of these drives are in MyBook enclosures, REDs may end up in very different places and may cause problems. In that case, head or spindle noise is not even the smallest concern.
If we're talking about noise, continuous (spindle) noise is much more degrading than head/seek noise. For me, even benign noises like USB fans have much more noticable effects after the day than mechanical keyboard noise for example.
Is that purely extra friction due to higher speed? Or is it spin up from sleep needing more power?
So, that difference is for sustaining higher spindle speeds.
The peak power requirement is 1.79A (~21.5W) which is probably consumed when starting from full-stop state (stand-by, sleep or power-off where motor is turned off).
[0]: https://media.flixcar.com/f360cdn/Western_Digital-3805661149...
[0]: https://products.wdc.com/library/SpecSheet/ENG/2879-800002.p...
Yes, I believe this is what Phyphox does for you.
Most people have a microphone in their smartphone. But of course you could use a dedicated microphone and do a frequency analysis yourself instead.
Well, there goes my productivity for the day.
(edit: fixed typo)
But if I buy these for NAS/backups, I want them to be quiet, cool and longer lasting - in my experience 5400 RPM drives do heat less and last longer than 7200 RPM.
Not sure if they still do, but WDs had this IntelliPark feature, which slammed the heads into parking a few times a minute (because otherwise your drive would last a decade, I guess), and their 7200rpm drives were clicking so loud I often thought "that's it, it failed". But no, just "normal" operation...
So I'd rather get the specs I'm paying for.
That's not true. Even a spinning sound of a 5400 rpm disk makes quite loud.
Whether you can hear it due to casing or distance is another matter. Most of the time fan noise will make you ignore it. When fan is not spinning I can hear laptop drives clearly.
Synology forums have several threads like this: https://community.synology.com/enu/forum/1/post/130843
And I understand that these are different companies, but why I even went for Seagate was because of the CMR issues.
IIRC from older threads - seek times are faster but sequential speed is capped.
e.g. https://www.tomshardware.com/news/wd-lists-all-drives-slower...
Noise is also an issue, if you were looking for an HTPC drive for instance you don’t need the increase in performances but the noise pollution is a problem.
If they lie about that on their spec sheet, how can you believe the power/noise levels written on there?
Also, I've always had a fair amount of scepticism for digital product spec sheets. I've been looking for a projector with low input lag for a while, and the measured value has constantly been way off the claimed value (> factor of 2). Admittedly, this seems to most often be the case with some of the cheaper chinese brands, but definitely not exclusively.
> Looking at the spec sheets provided by WD, it seems they don't really list the rotational speed
I still think it's somewhat disingenuous, but my comment above was wrong.
For the people who want to run cool, quiet and less power hungry devices, they'd like to get exactly what they ordered. Nobody buys a 5400 rpm hard drive for speed.
Stick a 7200 rpm drive in a console and see how the cooling system copes, for example. Or a low power PC with a PicoPSU. Or a NAS.
Some portable hard drives can be powered off USB. If you get a 7200 advertised at 5400 drive, it won't run because it's not getting enough power.
These are for NAS's/Servers where power and heat is a controlling factor, if admins wanted/could they would buy the 7200rpm drives and there would be a market for it, but they chose the 5400rpm because of constraints.
This is WD being lazy/cheap. Instead of selling them as 7200rpm drives, they're shovelling them into the 5400rpm lines as a presumably cost cutting measure.
They did not learn their lesson with the SMR/CMR scandal. Label your drives clearly.
The controller card inside of the enclosure is just a simple usb and power converter.
In my (limited) experience, it's mostly the WD portables that have a direct USB connector. It's been going on for many years — I have a WD Passport with USB 2.0 that I've unsuccessfully tried to shuck.
The portable Seagate Backup drives are shuckable, but some of them are SMR.
Producing 7200 RPM drives likely is more costly than 5400 in isolation, but running two full production lines itself is costly, when you could just run one and "fix it" via software.
Plus it allows you to vary the number of 5400/7200 RPM drives you "produce" to exactly what the market wants. Just keep them unlabeled in a warehouse as orders arrive: label, firmware, and ship.
I'm sure you can guess where this goes. The replacement unit arrived a few days later-- this drive also was marked WD on the outside and identified itself to BIOS etc. as a SanDisk drive. I called Amazon support again and they appeared to consider it a manufacturing defect and I returned it for a refund.
I'm not sure how much I care about whether the drive is SanDisk or WD, but it ought to be consistent. If WD can't get that simple step right I'm not sure why I should trust them to secure and maintain the integrity of a TB of important data.