Do external USB drive enclosures with battery backup actually work?

I have never seen such a product, but to be fair I haven't ever looked for one either.

I see a few reviews but only one mentions that the power loss protection actually seems to work in their testing, the others are pretty ambiguous if they actually tested or not.

There are plenty of SSDs with power loss protection as well if you don't want to trust a random enclosure. If you just want something with a fancy screen integrated, I actually got 2 of these for $25 before the price went up. https://www.ebay.com/itm/156139876944
So far they work as advertised.

Marketing Gimmick. Lets say sure it powers the drive for the case of an OTG connection to a Cellphone.

But when you connect it to a laptop. Your Bus better have enough power to power the drive via USB connection in the first place. And power loss also means your system should have UPS power protection as just keeping the drive powered on it completely useless if the source in a copy procedure just decides to shutdown during a write.

I'll assume this power protection they are talking about isn't something like giving the SSD enough power to finish a write from buffer in this kind of scenario.

Andrewcw said:

I'll assume this power protection they are talking about isn't something like giving the SSD enough power to finish a write from buffer in this kind of scenario.

Click to expand...

That smells exactly like what they're talking about from the description.

An SSD needs a fraction of a second's worth of power to finish writing data, and you can get drives that have their own protection for that if it's really needed. Five (or ten, both are claimed) seconds of backup power might be more useful for a mechanical drive to finish an operation. But it's really not needed for a USB-powered drive of either type in MOST situations. The only time such a drive will lose power is if you yank the cable out or the host loses power, and you should have backup power on the host if it's that important and be ensuring the cable can't be accidentally pulled. If you don't expect to encounter such situations (like if you need to work with random machines you don't control) and it's really important, then sure this might be useful to some degree, to prevent file system corruption.

They do make claims about what this enables that either don't actually happen or refer to things the drives already do without their extra power protection. The claim that it will protect against "SSD failure" is questionable. An SSD won't be damaged by having power pulled; the only risk is data not being flushed from the cache and written, resulting in data loss. The file system being damaged in that case is VERY low risk if you use NTFS or another journaling filesystem which would be the norm for an external drive like this, because the file table doesn't get updated until the write is finished; the data in the cache just gets lost. The window where file table damage could occur is extremely small. The claims about what it enables for a mechanical drive are also iffy, because for more than 2 decades drives have had built-in capabilities to park their heads safely when power is lost; it's basically spring-loaded with a dead-man switch. Again, the only thing that might benefit is giving it time for data to be flushed from the cache, but of course with a mechanical drive that pulls more power than an SSD, there is more risk that there might not be enough time to finish it all, depending on where the heads are and where the data needs to go and then getting back to modify the file table.

The display/touch screen is a personal preference on whether there's any point to it. I certainly wouldn't pay money for that in normal usage but maybe it would be more useful if you were only using the drive on a smartphone or ChromeBook where there don't seem to be (easy to use) tools to let you view drive health status for a USB drive.

Also, wow, it uses a VIA chipset. I didn't know they were still alive. Seems to be a completely different company as far as products now.

continuum said:

I see a few reviews but only one mentions that the power loss protection actually seems to work in their testing, the others are pretty ambiguous if they actually tested or not.

Click to expand...

I imagine it's fairly hard to test this. You already know that whatever file transfer was in progress is likely to fail if the drive is unplugged suddenly, because the OS isn't just going to check and continue from where it was if power is restored or it's plugged back in, and there's no way to check whether some specific bytes of data that were in cache were actually written to the drive. (Using a special file copy utility COULD resume a transfer like that, but in that case it wouldn't matter if a few bytes didn't get flushed from the cache.) The only thing you could do is yank it out during transfers and then plug it back in and check the filesystem, and do that repeatedly and see if you ever get corruption. But you can do the same thing with a drive with no protection, even a shitty flash drive, and still never happen to get any corruption, so it's hard to make any real claims that it actually is doing anything. In reality, the usefulness on a consumer system of having a drive that can flush the cache on power loss is very limited. Even if file data gets flushed, there's nothing to acknowledge that and update the file table to record the changes. The final file table updates for a transfer are about the only thing it would really be useful to ensure got flushed.

The review you linked even seems to make it sound like being able to power the drive with a plug, when it's not even connected to a device, is somehow beneficial. I can't imagine how it would be, since the drive isn't DOING anything during that time. All it might do is charge the battery so that the display can be used. The ten-second PLP is probably a separate capacitor and take moments to recharge after it's drained, unless there IS no battery to power the display and it only actually works when plugged in.

It's a SSD maker so who knows, but Transcend does talk about about this, both in terms of full-on power loss protection:

As well as what they call "power shield" which is more limited in scope but still designed to limit potential areas of corruption to flash.

Kingston seems to call their approach to the latter to be "firmware PLP" as opposed to calling the former "hardware PLP."

I thought Anandtech or someone years ago had a much more detailed article on the impact of power loss protection and ensuring data integrity but I can't find it at the moment.

Seems to be a reasonable discussion here at STH but not something I have vetted:

Lord Evermore said:

An SSD needs a fraction of a second's worth of power to finish writing data, and you can get drives that have their own protection for that if it's really needed.

Click to expand...

Older article:

But has actual numbers as far as time-- 20ms to 40ms.

More:

But yeah, as stated, I think the actual risk in reality is not as significant as feared. Consumer drives generally removed PLP what, about 10 years ago? And we have not see a rash of data corruption, undetected-that-manifested-only-later, or otherwise.

continuum said:

we have not see a rash of data corruption, undetected-that-manifested-only-later, or otherwise

Click to expand...

Well, would we even know? Most of us would just assume there was a Microsoft patch applied the night before. :-D

continuum said:

I have never seen such a product, but to be fair I haven't ever looked for one either.

I see a few reviews but only one mentions that the power loss protection actually seems to work in their testing, the others are pretty ambiguous if they actually tested or not.

Dockcase Explorer M.2 SSD Enclosure Review - Straightforward and Practical

The Dockcase Explorer M.2 enclosure offers realtime stats for your NVMe drive as well as PLP, is this device worth its price tag?

wccftech.com

Click to expand...

Unless something has changed, and that article doesn't suggest anything has, wccftech is not exactly a paragon of reliability and expertise. They didn't even see what would happen with the drive in another enclosure if it was subjected to the same "abuse". The 20ms to 40ms you mention later? The USB-C connector has the power & ground pins longer than the rest specifically so that USB-C devices can detect an unplug event before they actually lose power.

continuum said:

But yeah, as stated, I think the actual risk in reality is not as significant as feared. Consumer drives generally removed PLP what, about 10 years ago? And we have not see a rash of data corruption, undetected-that-manifested-only-later, or otherwise.

Click to expand...

A little more than that. Intel's last consumer SSDs with enough capacitors to serve as PLP were the 320 and 710 from back in 2011. The 510 that came out at the same time as the 320 lacked those PLP capacitors, and so did the 520 and 330 from 2012. Intel was pretty much the only one with those PLP caps at the consumer level; other early consumer SSD brands like Crucial, Corsair, and OCZ never had them.

cogwheel said:

The USB-C connector has the power & ground pins longer than the rest specifically so that USB-C devices can detect an unplug event before they actually lose power.

Click to expand...

That's not new with USB-C. Power and data have always been different lengths to allow hot-plug. But that amount of time is SO tiny it's irrelevant if there's no backup power, when it comes to being able to write data from the cache.

Lord Evermore said:

That's not new with USB-C. Power and data have always been different lengths to allow hot-plug.

Click to expand...

True. I debated mentioning it, but the post was already long enough.

Lord Evermore said:

But that amount of time is SO tiny it's irrelevant if there's no backup power, when it comes to being able to write data from the cache.

Click to expand...

0.5mm over 40ms is pulling the plug out at 12.5mm/s, ignoring acceleration. Likely enough if you pull the plug normally, possibly not enough if you trip over a cord. It'd probably take testing to characterize how useful it is and in which situations.