PWDIS feature in new SATA drives to watch for.

[nightshade00013]

http://www.tomshardware.com/news/hdd-sata-power-disable-feature,36146.html

Basically most sata drives do not use 3.3 Volts even though it is available per the SATA specs. The power is sent through multiple pins on most legacy connectors but with the PWDIS feature if the drive receives over 2.1V to 3.6V the drive will be powered down. This is something to watch for with new drives and could even cause a problem if you have a backplane that drives the pin to 3.3 volts.

 

[joeschmuck]

I think this topic came up here sometime last month. I don't recall the thread but it was a topic. Someone was likely having a hard drive issue.

I'm not sure if there is a good place to put this data here, maybe in the recommended hardware guide? It could be a warning or note.

 

[nightshade00013]

I actually submitted something for the forum to add in a hardware discussion I think it would be good to have a simple hardware discussion for things like this sort of how we have a 3rd party software discussion.

Didn't notice any other discussions on it but I generally touch in on subjects that I have some knowledge on or places where they have not gotten help after a certain point.

But I am sure the PWDIS issue is going to come up more and more in the future. I haven't read a lot about it yet but it sounds like it will be a way for a backplane to start drives in sequence which can be a good thing when you have 24 drives attempting to start all at once but legacy connectors are going to add a ton of issues into it.

 

[Ericloewe]

But I am sure the PWDIS issue is going to come up more and more in the future. I haven't read a lot about it yet but it sounds like it will be a way for a backplane to start drives in sequence which can be a good thing when you have 24 drives attempting to start all at once but legacy connectors are going to add a ton of issues into it.

It's stupid, there was already a pin for it on the power connector. They just wanted to add a hard-reset connector, for no particularly good reason.

 

[jgreco]

I haven't read a lot about it yet but it sounds like it will be a way for a backplane to start drives in sequence which can be a good thing when you have 24 drives attempting to start all at once but legacy connectors are going to add a ton of issues into it.

Well, that's silly. Staggered spin-up's always been supported.

https://sata-io.org/system/files/member-downloads/Implementing Staggered Spin Up v1.pdf

 

[nightshade00013]

Not completely sure. It also allows the drive to be completely stopped by adding power to the pin but that is already supported as well through the sata connection.

IMHO it is just another addition that has absolutely no purpose other than to mess up what is already out there. Kinda like the SATA power spec requiring a 3.3v rail but nothing I have ever came across actually uses it.

 

[jgreco]

The 3.3v rail is at least understandable. At the time, we had "recently" (mid-90's) gone through the transition from AT to 20-pin ATX, and there was a fairly reasonable line of thinking that the onboard electronics were headed away from 5v to 3.3v. This was followed up with a fairly rapid evolution through variations like ATXGES, EPS12V, and several respecifications of 24-pin ATX that led to things like the infamous "Haswell-compatible" labels on ATX PSU's.

So I'm pretty sure that the industry saw things like 2.5" HDD and SSD coming, and wanted to create a new power standard that would eventually allow for the retirement of legacy 5.25" voltages, and simpler designs for laptops, etc. There was an openness in thinking to ditching legacy stuff at the time.

Unfortunately, what they seem to have failed to recognize is that people don't really mind replacing a PSU at the same time as they replace a mainboard, but peripherals like hard drives, optical drives, chassis fans, etc., have a longer service life, and also they picked a sucktacular connector for SATA power that cannot easily be tapped or "Y"'d, which really did nothing except encourage the continued use of the 3.3v-less AMP Mate-n-lock ("Molex") power standard, where the pins have a phenomenal 11 amps at 12V rating. In the years since introduction, huge strides have been made at deriving secondary voltages, usually from 12V, and so it really isn't outside the realm of possibility that there will never be a significant adoption of the 3.3v. Certainly that's the evidence of the last decade. The result is that no manufacturer really wants to get their feet wet in the whole 3.3v game because it is guaranteed to cause real problems for a nontrivial number of users. That's easier to recognize as a problem in hindsight.

 

[Chris Moore]

I haven't read a lot about it yet

I have.

it sounds like it will be a way for a backplane to start drives in sequence

That is not what it is for. It is intended for massive data-centers where they have a drive in a server that isn't working correctly and instead of walking all the way out on the floor to pull the drive and put it back in, which sometimes corrects a problem in a hardware RAID controller, the can just turn the drive off and back on remotely by selectively applying power to the 3.3v connection. It turns the drive off and back on again, just like it was unplugged and plugged back in. It is all about remote management, but it would require a backplane and firmware that support such. I have not seen hardware support for it yet. It would usually be purchased at the same time as a server full of drives that also support the feature, from a vendor like Dell or HP, where they have spent the time to build the firmware to manage the hardware and charge you an extra licensing fee to use it.

I think this topic came up here sometime last month.

I have responded to user problems with this issue about six times since Nov 2018. Here is a link to further reading:

3.3 volt HDD problem, WD PDF:
https://documents.westerndigital.co...h-brief-western-digital-power-disable-pin.pdf

And a video:
https://www.youtube.com/watch?v=9W3-uOl4ruc

 

[danb35]

I have responded to user problems with this issue about six times since Nov 2018. Here is a link to further reading:

...so starting almost a year after @joeschmuck posted the text you quoted.

 

[jgreco]

...so starting almost a year after @joeschmuck posted the text you quoted.

Please, let's be pleasant.

Basically the earlier posts in the thread summarized the situation well enough, but the link to more materials is potentially useful. I'm not sure what they were thinking when they implemented this - if your firmware is so sucktacular that you need to add electronic circuitry to the drive controller board to power cycle the drive, you should probably fire your firmware folks and seriously consider a different line of business - but creating gratuitous incompatibilities with existing standards is something the standards people love to do. The feature is largely useless-ish without a bunch of scaffolding around it to support the idea of power cycling a drive that's currently in operation, a feature that doesn't seem to exist in most operating systems, and because it interferes with legacy standards, that makes it a misfeature in my opinion. But hard drives are leaning heavily in the direction of needing lots of supporting scaffolding around them - consider host-managed SMR as another example. All I want for xmas are reasonable endurance SATA 2TB SSD's for $100 or less.

Since this thread has probably played out as far as facts are concerned, I'm going to close it with a repeat of a useful link by @Chris Moore

3.3 volt HDD problem, WD PDF:
https://documents.westerndigital.co...h-brief-western-digital-power-disable-pin.pdf