- Sudden power loss causes damage?
- Posted by John Doe on February 2nd, 2005
Are modern hard disk drives susceptible to hardware damage or data
corruption from sudden power loss?
By "sudden power loss", I mean like if you pull the power supply plug
out off the wall socket, or if the hard disk drive power cable is
disconnected.
Partly curious. Thank you.
- Posted by Alexander Grigoriev on February 2nd, 2005
Modern journalling file system are sufficiently robust against power loss,
but any unwritten DATA may be lost, causing corruption in the user's files
structure.
Don't take chances and buy an UPS.
"John Doe" <jdoe@usenet.is.the.real.thing.com> wrote in message
news:Xns95F12C69A29D1wisdomfolly@151.164.30.42...
- Posted by Eric Gisin on February 2nd, 2005
If the drive is writing, the current sector will be bad. Documented in many
drive manuals, and also the servo chipset docs.
The drive will always retract the heads to landing zone or offload ramp.
"John Doe" <jdoe@usenet.is.the.real.thing.com> wrote in message
news:Xns95F12C69A29D1wisdomfolly@151.164.30.42...
- Posted by Rod Speed on February 2nd, 2005
John Doe <jdoe@usenet.is.the.real.thing.com> wrote in
message news:Xns95F12C69A29D1wisdomfolly@151.164.30.42...
Nope, they should all handle that with no data loss or damage.
You know what that did to the cat dont you ?
- Posted by Arno Wagner on February 2nd, 2005
Previously John Doe <jdoe@usenet.is.the.real.thing.com> wrote:
There was a time where certain drives suffered bad sectors when
this happens. The problem should be solved now.
Writing a sector takes something like 10-20 us. The drive buffer
capacitor should have enough power to keep the electronics working
that long. The disk will detect the power-fail a bit before
it has problems with the voltage levels.
If you short out the power leads, it may be a different story.
Arno
--
For email address: lastname AT tik DOT ee DOT ethz DOT ch
GnuPG: ID:1E25338F FP:0C30 5782 9D93 F785 E79C 0296 797F 6B50 1E25 338F
"The more corrupt the state, the more numerous the laws" - Tacitus
- Posted by Eric Gisin on February 2nd, 2005
"Arno Wagner" <me@privacy.net> wrote in message
news:36cr8tF4qap71U7@individual.net...
There is a chipset doc online that says low voltage simply disables the write
head.
- Posted by CJT on February 2nd, 2005
Alexander Grigoriev wrote:
I think that's expressed poorly, because I can interpret it to be either
true or a mis-statement. Vague terms include "unwritten" and "user's
file structure."
A Google of "two phase commitment" might be relevant.
--
The e-mail address in our reply-to line is reversed in an attempt to
minimize spam. Our true address is of the form che...@prodigy.net.
- Posted by John Doe on February 3rd, 2005
"Eric Gisin" <ericgisin@hotmail.com> wrote:
Instead of a troll, why don't you post a link to that chipset dock.
- Posted by Eric Gisin on February 3rd, 2005
"John Doe" <jdoe@usenet.is.the.real.thing.com> wrote in message
news:Xns95F1D90494E53wisdomfolly@151.164.30.42...
The title is Texas Instruments Servo/MSC Product Line User's Guide (PDF).
This is over 5 years old, so things may have changed.
You are better of reading hard drive manuals. IBM's describe the "at most one
bad sector" behavior.
Arnie's "drive buffer capacitor" is a fabrication. You cannot store enough
energy on those dinky caps.
- Posted by Alexander Grigoriev on February 3rd, 2005
After 10 us, 100 uF, 1A, the voltage drops by 100 mV. 100 uF is not a
terribly big capacitor.
"Eric Gisin" <ericgisin@hotmail.com> wrote in message
news:ctsasd0sd4@enews1.newsguy.com...
- Posted by Arno Wagner on February 3rd, 2005
Previously Alexander Grigoriev <alegr@earthlink.net> wrote:
Also you just need to power the write amplifier and circuit
by it, so 1A is probably on the high side.
Personally I find it funny that people who don't know about buffer
capacitors presume to talk about how electronics work. It is one of
the most fundamental things you need to have to get digital
electronics to work right. (In analog electronics they more serve
as filters than as power storage, depending on the circuit
functionality.) I guess these people also never heard of
wire impedances...
Arno
--
For email address: lastname AT tik DOT ee DOT ethz DOT ch
GnuPG: ID:1E25338F FP:0C30 5782 9D93 F785 E79C 0296 797F 6B50 1E25 338F
"The more corrupt the state, the more numerous the laws" - Tacitus
- Posted by Eric Gisin on February 3rd, 2005
Then you and Arnie can show us the picture of a HD controller with that
capacitor, right?
I have never seen an electolytic cap on a HD. They cost money, so you would
never see it on IDE.
"Alexander Grigoriev" <alegr@earthlink.net> wrote in message
news:GyiMd.5208$cl1.1273@newsread3.news.pas.earthl ink.net...
- Posted by Folkert Rienstra on February 3rd, 2005
Well, well, well, someone calling himself "John Doe", calling people troll,
is trying to redo a subject that last time this was discussed resulted
in a thread with a massive 180 posts.
"John Doe" <jdoe@usenet.is.the.real.thing.com> wrote in message news:Xns95F12C69A29D1wisdomfolly@151.164.30.42
- Posted by Eric Gisin on February 3rd, 2005
"Arno Wagner" <me@privacy.net> wrote in message
news:36ejktF4qcgs3U1@individual.net...
I have a drive with two filter caps marked 6u8. 6.8uF is not going do anthing.
Besides, you also need a blocking diode. Never seen one of those either.
- Posted by Folkert Rienstra on February 3rd, 2005
"Arno Wagner" <me@privacy.net> wrote in message news:36ejktF4qcgs3U1@individual.net
Yeah, no need to keep the heads tracking, obviously.
Who cares where that write goes, right Arnie?
Yes Arnie, that is so funny, especially coming from you.
Gosh, makes one wonder where Power Amps get their power from and
how they can work with AC signals using asymetric power sourcing.
And then Arnie went from a rush into a delirium ....
- Posted by John Doe on February 4th, 2005
Off-topic troll.
"Folkert Rienstra" <see_reply-to@myweb.nl> wrote:
- Posted by Eric Gisin on February 4th, 2005
No, you are the troll. Pretty evident from you fake name and address.
"John Doe" <jdoe@usenet.is.the.real.thing.com> wrote in message
news:Xns95F2BCAAA3B5wisdomfolly@151.164.30.44...
- Posted by Folkert Rienstra on February 4th, 2005
"John Doe" <jdoe@usenet.is.the.real.thing.com> wrote in message news:Xns95F2BCAAA3B5wisdomfolly@151.164.30.44
Says he who posts message headers containing email addresses left
right and center all over the internet for the spammers to see..
ews.prodigy.com!prodigy.net!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail
- Posted by RPR on February 4th, 2005
This is correct. There are comparators on the +5 and +12V supply lines.
As soon as one detects an out of limits supply, the write gate is
turned off, for fear that the head may go offtrack and eat into a
neighboring track since the actuator voice coil needs quite a bit of
current (several amps peak - no way you can buffer that with a
capacitor a disk drive manufacturer can afford.) Hence the currently
written sector is incomplete and will result in a read error. It isn't
bad per se, you can overwrite it and it will be fine.
The emergency retract (to the landing zone or load ramp) is powered by
the rotating disk stack, through windage or the motor's back EMF,
sometimes assisted by a spring or a magnet.
Ralf-Peter
- Posted by Eric Gisin on February 4th, 2005
I did google for some modern disk controller chips. The best I found is:
http://www.st.com/stonline/prodpres/...r/datastor.htm
I didn't find detailed info for the system-on-chip disk controllers. Their koi
servo controller does have a data sheet:
http://www.st.com/stonline/prodpres/...mb/powcomb.htm
"RPR" <rohbeck@yahoo.com> wrote in message
news:1107544568.508350.273780@o13g2000cwo.googlegr oups.com...
