Cherryville - SSD 520

[johnw]

Thanks. Is there a power cap

I apologize if this sounds like a nitpick, but it is really not. A "power capacitor" is a capacitor that is usually used together with a power supply to filter or smooth out the signal.

What we are wondering about here is a power-loss-protection capacitor. I suppose we could call it a PLP cap.

[Anvil]

It's often called super capacitor.
It was mentioned in an article where the V3 was discussed (very early, before release) and iirc it was said that there would be no super capacitor on the SF2281.

[johnw]

It's often called super capacitor.

No. A supercapacitor is a type of capacitor with very high charge density that can be used as a PLP cap. But super capacitor is NOT another name for a PLP cap. Actually, in some ways a super capacitor is a poor choice for a PLP cap, since it very expensive and makes a single point of failure. If there is room on the circuit board to achieve the same capacitance with multiple inexpensive capacitors, that is a better design. That is how Intel designed the 320 series (except that the firmware initially had problems with shutting down the SSD under PLP cap power)

As for whether the SF2281 supports PLP capacitor(s), I would be surprised if it cannot. It certainly has not been enabled on any of the 2281 SSDs that I have heard of, but perhaps Intel insisted on it. Or not. It would be interesting to know.

[Anvil]

I didn't say it was another name for "PLP"

"super capacitor" is often used as a term when describing protection mechanisms against data corruption/data loss when power is lost.

[johnw]

"super capacitor" is often used as a term when describing protection mechanisms against data corruption/data loss when power is lost.

Mistakenly used. A super capacitor is a specific class of device that can be used for the application of power loss protection, but so can many other kinds of capacitors. It is similar to saying "Kleenex" when you actually mean "tissue". That sort of thing is acceptable for discussing mundane things like tissues, but it is prone to cause confusion in technical discussions.

By the way, here is a nice picture of the six capacitors (not super capacitors) that Intel used for PLP in the 320 series.

[Anvil]

Then there is a lot to fix on various review sites as well as wikipedia

I'm sure that "PLP" is a more correct term but not often used.

[Ao1]

@ Johnw You beat me to it

The Intel 320 controller holds very little data in cache, but it still has an array of six 470µF “caps” in parallel. An SF drive holds a lot more data in cache so it would be surprising if it doesn’t have “caps”. AFAIK all SF drives have the capability to deploy “caps”

The shot of the SF cap below is from anandtech review.

[johnw]

Then there is a lot to fix on various review sites

I agree, the quality of most Internet journalism is extemely poor. I have often stated this.

As for wikipedia, please provide a link of where it uses "super capacitor" when it really means "capacitor" for power loss protection.

[johnw]

The Intel 320 controller holds very little data in cache, but it still has an array of six 470µF “caps” in parallel. An SF drive holds a lot more data in cache so it would be surprising if it doesn’t have “caps”.

Hmmm, I thought the only cache Sandforce SSDs use is on-controller cache. But I do not know how the size of the SF2281 on-controller cache compares to the DRAM cache chip used in the Intel 320 series.

[Anvil]

The SF2582 and SF2682 both support "Power Fail Circuit Support"

None of the "Client" drives initially supported this, it might have changed?

edit;
@johnw
Here's a link to Wikipedia on SSDs, you'll find a snippet on "Battery or super capacitor" Link a bit down the page.

[johnw]

The SF2582 and SF2682 both support "Power Fail Circuit Support"

None of the "Client" drives initially supported this, it might have changed?

The question I have long had is how much the actual silicon differs between Sandforce's different chips, and how much is just features being turned on or off (with firmware or other methods), which features are still present on the chip even if disabled.

[johnw]

@johnw
Here's a link to Wikipedia on SSDs, you'll find a snippet on "Battery or super capacitor" Link a bit down the page.

Notice that they got it right in the first sentence, but then slipped and made a mistake in the last sentence. They should have said just "capacitor" both times.

Another component in higher performing SSDs is a capacitor or some form of battery. These are necessary to maintain data integrity such that the data in the cache can be flushed to the drive when power is dropped; some may even hold power long enough to maintain data in the cache until power is resumed.[citation needed] In the case of MLC flash memory, a problem called lower page corruption can occur when MLC flash memory loses power while programming an upper page. The result is data written previously and presumed safe can be corrupted if the memory is not supported by a super capacitor in the event of a sudden power loss.

By the way, wikipedia has a decent article on super capacitors (aka electric double-layer capacitors):

http://en.wikipedia.org/wiki/Electri...ayer_capacitor

[Ao1]

Not a 100% match but as close as I can find for the SF cap.



http://www.cap-xx.com/products/products.php

[Ao1]

The question I have long had is how much the actual silicon differs between Sandforce's different chips, and how much is just features being turned on or off (with firmware or other methods), which features are still present on the chip even if disabled.

I'm almost 100% sure that its just a case of fetaures being turned off or on. Just like most SF dives not having temp sensors. The capability is there but its not used. SF1xxx controllers have the capability to run at SATA 3 speeds, but its disabled in f/w.

[DooRules]

[johnw]

No large capacitors of any kind visible on the circuit board. I am completely disappointed with the 520. The incompressible sequential write speeds are lower than the 510, and it has no PLP capacitors. I don't see the point of the "upgrade" from 510 to 520. I suppose for people who need high QD random write performance. But that must be a very small market segment.

[m.oreilly]

yay, a lot less pedantry and a little more pcb pr0n!

[Ao1]

The pcb appears to be completely different to the reference SF board. I agree though the 520 is way too late to market and they don’t seem to have done anything special with performance. With the amount of work that seems to have gone into this drive you just wonder why Intel bothered with SF.

[Christopher]

I'm... a little disappointed, but until I get mine (at this point I have no way of knowing when it will ship) I will reserve judgement.

I was hoping for a little more, and I had always said "I hope it's N.J.A.Fing.SF."

Perhaps there is something redeeming about it... some magic longevity advantage that can only be known through endurance testing....


I won't be selling my 510 on Ebay anytime soon.

[Computurd]

Actually, in some ways a super capacitor is a poor choice for a PLP cap, since it very expensive and makes a single point of failure.

and most are not 'endurant' to high heat, for server environments and ruggedized units, Tantalum is a must

[Christopher]

I know I'm a lot more excited about the 711 series which should be out shortly as well. A new 711 34nm SLC 32GB will be SATA II, somewhere in the neighborhood of $350. If random performance is improved, I'm getting one.

Actually, I just looked -- C'ville is back in stock but the retailer has removed the 711 listings.

[bluestang]

John, can you find me one component manufacturer that calls it a PLP or Power Loss Protection capacitor? Have fun on that one.

They are either called supercapacitors, just plain old capacitors, and sometimes charge capacitors. They are either Electrolytic or Tantalum.
Now how you use or implement them in an apllication (for power loss or correction) is different than what they are named. Don't confuse the two.

BTW, if you want, I can charge up some of the ones we have in our RF Transmitters here, ship it to you and it would restart a heart if need be.

Just my 2 cents (oh, and the electronics industry as well).

[johnw]

John, can you find me one component manufacturer that calls it a PLP or Power Loss Protection capacitor? Have fun on that one.

They are either called supercapacitors, just plain old capacitors, and sometimes charge capacitors. They are either Electrolytic or Tantalum.
Now how you use or implement them in an apllication (for power loss or correction) is different than what they are named. Don't confuse the two.

BTW, if you want, I can charge up some of the ones we have in our RF Transmitters here, ship it to you and it would restart a heart if need be.

Just my 2 cents (oh, and the electronics industry as well).

Bluestang, can you find one quote where I claimed that any component manufacturer calls a capacitor being used for power loss protection a "power loss protection capacitor". Have fun on that one.

They are NOT called supercapacitors if they are not supercapacitors. They are called supercapacitors if they are supercapacitors. In all cases, they can be called capacitors, since they are capacitors.

Do you really find that so difficult to understand? Don't confuse yourself.

Just my 2 cents (oh, and people who can read, too).

[bluestang]

John...it's Friday!...check PM.

[lm358]

John is right, a super capacitor is a super capacitor and may or may not be used for power-loss protection. The latest Samsung SM825 review on storagereview explains it pretty well. They call it simply "Cache Power Protection" which essentially describes the functional purpose, however I do think power-loss protection is a nicer term for it. BTW, comparing the size and number of the capacitors with that on the Intel 320 is rather telling