Wise to use HDD as offline storage?

[Rod Speed]

I was referring to the standby power line that you mention a bit
further on. On that line, it's a regular 5V; when looking at the
total current draw from the PSU, it appears as a "trickle" there.

Still, enough to get into trouble.

Not with the unplugging of hard drives it isnt.

I saw a particular line of power supplies that would tend
to heat up when ATX "off" (the top of the case would be
warmer than when the PC was running), and sometimes
these would go Bang! and blow a cap when switched on.

One completely dud power supply design is completely
irrelevant to what is being discussed, whether hard drives
are powered from the +5VSB, they arent, even indirectly.

This would then spike the +12V and typically kill motherboard,
RAM, HD, and the CD-ROM would often fail a week or so later.

One possible mechanism for this failure pattern might be that the
lower standby current with no fan caused the PSU to "run" hotter
than normal operations with the fan running. Then when the initial
demands of full power are applied to this heated circuitry, pop!

How odd that normal ATX power supplys
dont even get warm in standby mode.

Or it could be some other mechanism entirely, unrelated to the
warmer temperature noted when "just" supplying standby power.

Yep, most do fail on power on.

So what we need to know is where that line goes.,

Its obvious where that goes, to components that need to
be powered in standby mode. Thats the whole point of it.

And you can check that trivially with the
keyboard and mouse with optical mice.

and whether power from that line emerges
through other ciruitry as data line activity.

Corse it doesnt, there is no 'data line activity' in standby mode.

And its completely trivial to check that with a multimeter.

Well, it could be some sort of "ghost power" arrangement I suppose,
Nope.

much as some USBs which have external power connections
"don't need external power" and can run without it.

Different situation entirely. The USB standard allows for USB
devices to be powered from the USB cable if they dont use
too much power. No 'ghost power' involved whatever.

I notice some mobos support wake-up from various IRQs,
Yes.

including the IDE controllers
Nope.

(wake-on-demand is not a not a feature I use, so I
don't spend much time in that part of CMOS setup).

And clearly know nothing about it.

+5VSB doesnt even make it to the ribbon cable.

If the mobo can sense events on the IDE,

It cant, and there are no 'events' in that sense.

then that suggests power on the IDE.

Fanciful, and its completely trivial to use a multimeter
and prove that there is no power on the IDE in standby.

Again, maybe and maybe not. It's quite possible to circuit something
that's safe for isolated power, i.e. a logic chip that derives power
from one of the IDE lines and doesn't rely on other HD components
that would have to formally powered via the molex.
Waffle.

I don't have links, but I reckon if there's doubt on this,

There isnt, and its completely trivial to eliminate all doubt with a multimeter.

it's safer not to assume it's safe to plug and unplug
HDs while the ATX power is connected but "switched off".

Its safer to wear a seat belt and superglue your sox
on and wear a hazmat suit too. Most of us dont bother.

You're welcome to do whatever you want.

You arent welcome to advise others to do what you are stupid enough to do.

We do know that while "off", power enters the PC as a whole,
and the mobo too; we are debating whether it gets to the HDs.

And its completely trivial to prove that it doesnt.

Until certain it's safe, I'd rather unplug the power

Your problem.

a good idea to always disconnect the power cord of /anything/ you open."

Yep. In the old days of "real" power switches, there was
a school of thought that suggested it was safer for the PC's
parts if mains was plugged in but switched off, in that the system
would still be grounded via the earth line of the mains supply.

Its still true. Static damage is the only
possibility when unplugging the hard drive.

I'd say that's changed now, in the ATX era;

You're wrong.

if it ever was good advice,

Corse it was, and is.

it's less so these days.

Wrong again.

Even when plugging and unplugging PCI cards, which is certainly
undesirable with the system in standby mode, its better to turn
the mains off at the switch near the mains cord connector than
to unplug the mains cord, so the system is still grounded.

 

[J. Clarke]

Yep.


YOU are being pedantic in fact.

How so? I'm not the one quibbling over the definition of "trickle".

Corse it isnt when discussing whether its safe to unplug a hard
drive with an ATX system shutdown and the mains still connected.

It is pointless to argue whether the power is "a trickle" or "full current
on the +5VSB". You are welcome to think that is it a matter of great
importance to argue that point. If you do however, eventually everyone
will become bored with that subject and stop paying attention to anything
you post.

How odd that its never happened to me once.

Don't worry, if you continue to swap components around with the machine in
ATX off it will eventually. Probably at the least opportune time.

Not with just changing a hard drive it isnt. THAT is what was being
discussed.

With _anything_ that is not specifically documented to be hot-swappable.

 

[TonyB]

Nope.

How so?

You're the one splitting hairs on the use of the word trickle.

I'm not the one quibbling over the definition of "trickle".

All that means is that you arent the only one being pedantic.

It is pointless to argue whether the power
is "a trickle" or "full current on the +5VSB".

And you did that yourself in spades.

You are welcome to think that is it a matter
of great importance to argue that point.

Never even said anything about 'a matter of great
importance' just said that it isnt a pointless argument.

If you do however, eventually everyone will become bored
with that subject and stop paying attention to anything you post.

Anyone is always welcome to be selective about what they choose to read.

Don't worry, if you continue to swap components
around with the machine in ATX off it will eventually.

I've likely done it a lot more than you have thanks.

I have enough of a clue to realise when the mains needs to be unplugged
and when it doesnt. It doesnt when unplugging or plugging hard drives.

Probably at the least opportune time.

Never ever happens to be, because I understand when
the mains needs to be disconnected and when it doesnt.

With _anything_ that is not specifically documented to be hot-swappable.

Wrong again. I know that the hard drives arent powered in standby mode thanks.

I also know that the PCI cards and ram are.

 

[Ge]

To sum up: Connecting or disconnectin HDD power when voltages are
present is extremely risky ...

No contest here. But the question was whether there are any voltages
present at the harddisk with the ATX power supply in standby mode -- which
you confirm that there aren't any.

Since this would be complete nonsense, I doubt there is any link.
Why spend the extra money to give the controller a power-switch
that switches the controller over between +5VSB and +5V, when the
controller is perfectly happy to work on +V5?
Exactly...

Actually it is good advice because +5VSB does go to some devices like
the PCI cards.

I wouldn't advise exchanging cards while on standby power. But the question
was about exchanging drives while on standby.

Gerhard

 

[Ge]

I was referring to the standby power line that you mention a bit
further on. On that line, it's a regular 5V; when looking at the
total current draw from the PSU, it appears as a "trickle" there.

But (even though someone else who doesn't seem to get the point called me
"pedantic") the current /capacity/ is not the factor that controls the
current /draw/ for standby power that works as designed. The actual current
drawn is determined by the connected load, not by the current capacity of
the standby supply. The moment the current gets actually limited by the
power supply, you are in trouble.

That's exactly the reverse for what's commonly called "trickle power": here
the current drawn is limited by the trickle supply. It's the power supply
that determines the current, not the load -- by design. That's the thing
about "trickle" vs. "standby".

Of course I realize that this has little to do with the core question of
whether or not it is safe to (un)plug harddisks while on standby power. I
just thought I'd contribute to the general level of education ...

Well, it could be some sort of "ghost power" arrangement I suppose,
much as some USBs which have external power connections "don't need
external power" and can run without it.

The term "ghost power" is usually used when the I/O lines of a device have
a signal while the power is down, and internally to the device there's a
current path that now tries to power the device from the I/O line -- which
generally hasn't enough current capacity to fully power the device, and the
internal current path often only powers parts of the device. That's usually
a design flaw (if the situation is a normal operating condition) or a
misuse (if the specs say you shouldn't power the I/O without powering the
device).

(This also applies to your suspicion that IDE connectors may carry signals
even though the harddisk power connectors don't carry power. I'm pretty
sure that harddisk manufacturers would call this a "misuse", because they
don't want to carry the burden of designing their devices to always
withstand this situation. Ask one, if you don't believe me...

In the case of USB, a power line is part of the bus. Devices are allowed to
power themselves through the bus if they don't draw more current than
specified in the USB standard and can be provided by upstream devices.
(That's important, because e.g. an external hub with its own power supply
will be able to provide the full spec current to its downstream devices,
but an external hub that works on USB supply will only be able to provide a
fraction of the full spec current to its downstream devices -- depending on
the current capacity of the upstream device it is connected to.)

I notice some mobos support wake-up from various IRQs, including the
IDE controllers

I've never seen a mobo that supports wake-up on IDE controller activity.
Care to share a make/model?

I've also never seen an ATX power supply that would provide the harddisk
power connectors with power in standby mode (that would have to be both 5 V
and 12 V). That would be necessary for wake-up on IDE activity making any
sense.

And, as said above, I'm pretty sure harddisk manufacturers don't like
signal activity on the IDE connector while the disk is not supplied with
power. So any mobo that does this is probably in violation of (most)
harddisks' specs.

Gerhard

 

[Ge]

You're being pedantic here IMO. +5VSB is a "trickle" compared to the normal
operating power. It seems a pointless argument that you are raising.

I'm not raising an argument about "trickle" vs. "standby". I'm just trying
to explain to people willing to learn what the better (because more
universal, at least among the people who understand about power supplies)
terminology is -- and why. See my other post here for more details (if
you're interested, of course).

The point is that there is power in the case and on the motherboard with the
ATX power in the "off" state.

Correct, in that there is power in the case. Wrong, in that this exactly
was the point.

And depending on what you're doing that standby power can become full
power very easily and very quickly

Exactly -- depending on what you are doing.

I've inserted boards into or removed boards from a machine and had it
power up while I was doing it

See, that's something I wouldn't do, or advise doing -- and I'm not
surprised that it did create problems. Sometimes, understanding the
workings of power supplies /does/ help

(and don't tell me it can't happen based on some fancy-Dan analytical
model--in the real world it can and does)

See, what you call my "fancy-Dan analytical model" is actually the
understanding of the inner workings of power supplies and the like. This
tells me that you should /not/ (un)plug PCI boards under standby power
unless you know the mobo and the PCI board very intimately and know what
you're doing. OTOH, it also tells me that it is safe to (un)plug an IDE
harddrive.

When adding or removing hardware, the best policy is a cold chassis,
that means either unplugged or hard-off.

That's why I wrote the below, and here I seem to be in complete agreement
with you (even though you didn't seem to see it that way):

"As long as you're not knowing what you're doing, it's a good idea to
always disconnect the power cord of /anything/ you open."

Gerhard

 

[J. Clarke]

But (even though someone else who doesn't seem to get the point called me
"pedantic") the current /capacity/ is not the factor that controls the
current /draw/ for standby power that works as designed. The actual
current drawn is determined by the connected load, not by the current
capacity of the standby supply. The moment the current gets actually
limited by the power supply, you are in trouble.

That's exactly the reverse for what's commonly called "trickle power":
here the current drawn is limited by the trickle supply. It's the power
supply that determines the current, not the load -- by design. That's the
thing about "trickle" vs. "standby".

Nobody was talking about "trickle power". He was talking about "a trickle
of power". You're being pedantic because you're trying to turn a standard
usage of the word "trickle", to refer to a small flow of something, into a
persnickety technical point about the nomenclature of power supplies.

Of course I realize that this has little to do with the core question of
whether or not it is safe to (un)plug harddisks while on standby power. I
just thought I'd contribute to the general level of education ...



The term "ghost power" is usually used when the I/O lines of a device have
a signal while the power is down, and internally to the device there's a
current path that now tries to power the device from the I/O line -- which
generally hasn't enough current capacity to fully power the device, and
the internal current path often only powers parts of the device. That's
usually a design flaw (if the situation is a normal operating condition)
or a misuse (if the specs say you shouldn't power the I/O without powering
the device).

(This also applies to your suspicion that IDE connectors may carry signals
even though the harddisk power connectors don't carry power. I'm pretty
sure that harddisk manufacturers would call this a "misuse", because they
don't want to carry the burden of designing their devices to always
withstand this situation. Ask one, if you don't believe me...

They do carry signal--power down one drive on a cable and the other still
works. Sometimes, however, the drive that is powered down will cause
signal corruption on the line.

In the case of USB, a power line is part of the bus. Devices are allowed
to power themselves through the bus if they don't draw more current than
specified in the USB standard and can be provided by upstream devices.
(That's important, because e.g. an external hub with its own power supply
will be able to provide the full spec current to its downstream devices,
but an external hub that works on USB supply will only be able to provide
a fraction of the full spec current to its downstream devices -- depending
on the current capacity of the upstream device it is connected to.)



I've never seen a mobo that supports wake-up on IDE controller activity.
Care to share a make/model?

Wake on IRQ. Just tell it to wake on IRQ15 or whatever turns out to be
assigned to the IDE host adapter (interesting that you're so persnickety
about "trickle" but call a host adapter a "controller").

I've also never seen an ATX power supply that would provide the harddisk
power connectors with power in standby mode (that would have to be both 5
V and 12 V). That would be necessary for wake-up on IDE activity making
any sense.

And, as said above, I'm pretty sure harddisk manufacturers don't like
signal activity on the IDE connector while the disk is not supplied with
power. So any mobo that does this is probably in violation of (most)
harddisks' specs.

How does the motherboard know if there is power being supplied to the disk?

 

[J. Clarke]

Ge,

Are you a native speaker of English? The reason I ask is that I find a
certain amount to object to in this post but most of it could be explained
by a slightly imperfect command of the English language. I note that
you're posting from globo, whose home page seems to be in Portuguese, hence
the question. Instead of going on with my comments I wanted to make sure I
was addressing the right issue.

 

[Arno Wagner]

On 8/15/05 13:11:23, Arno Wagner wrote:

No contest here. But the question was whether there are any voltages
present at the harddisk with the ATX power supply in standby mode -- which
you confirm that there aren't any.

True. I should haveread the posting more carefully.

I wouldn't advise exchanging cards while on standby power. But the question
was about exchanging drives while on standby.

As you said, no problem. Even unlikely to be a problem if there
was a stupidly designed mainboard with signals on the IDE
bus while in standby.

Arno

 

[Ge]

Are you a native speaker of English?

No. My native language is German.

The reason I ask is that I find a certain amount to object to in this
post but most of it could be explained by a slightly imperfect command
of the English language.

I would appreciate if you found the time to point out what you think is
expressed (or understood) not quite right.

OTOH, I'm pretty sure that my /technical/ English is pretty good. I've been
writing technical papers for US companies for some time -- and I've been
reviewing technical papers written by native speakers. While it seems that
I couldn't manage to completely get rid of creating some kind of "odd
feeling" on occasion, the terminology and the grammatical constructs I'm
using seem to be correct.

Gerhard

 

[Ge]

Nobody was talking about "trickle power". He was talking about "a trickle
of power". You're being pedantic because you're trying to turn a standard
usage of the word "trickle", to refer to a small flow of something, into a
persnickety technical point about the nomenclature of power supplies.

I find your opinion that I'm "trying to turn a standard usage of the word
'trickle' [...] into a persnickety technical point about the nomenclature
of power supplies" not quite compatible with the paragraph of mine that was
right following your opinion. I wouldn't assert that "this has little to do
with the core question of whether or not it is safe to (un)plug harddisks
while on standby power" if I really wanted to turn this into a "persnickety
technical point".

The reason I was trying to point out the difference is that from a
technical (electronic) point of view, there seems to be quite a confusion
in this thread about ATX standby power and its "powers", so to speak. So I
tried to bring in some foundation knowledge. It's yours to take or leave.

I don't quite understand why you object, or to what. It seems that you
don't object to the technical content; at least you haven't brought forth
any technical objections to my explanations of the differences between
trickle and standby power. It may not help you, but it may help others --
if it is correct. If it is not correct, then please start saying what's
wrong. If it is correct, maybe stop objecting and focus on what you want to
object to.

They do carry signal--power down one drive on a cable and the other still
works. Sometimes, however, the drive that is powered down will cause
signal corruption on the line.

I'm not sure what you mean. The situation where the harddisk power
connectors of an ATX power supply don't carry power is in standby mode and
when off. In neither of these situations, the IDE connectors of the mobo
carry any signals (unless you can show some evidence that they do) -- and
definitely no other harddisks are working.

When you say that sometimes the powered-down drive will cause signal
corruption on the IDE bus is a confirmation of my point (see below) that a
drive connected to the IDE bus without being connected to power is in an
operation mode not supported or guaranteed to work by the drive
manufacturer.

Wake on IRQ. Just tell it to wake on IRQ15 or whatever turns out to be
assigned to the IDE host adapter (interesting that you're so persnickety
about "trickle" but call a host adapter a "controller").

Just because you can tell your mobo to wake up on the IDE controller's
interrupt (more to the use of the term "IDE controller" below) doesn't mean
it actually will wake up. In standby, there's no power to the disks,
therefore the disks won't initiate any activity on the IDE bus, therefore
the IDE controller won't create any interrupt -- if the IDE controller is
powered up at all in standby, which is doubtful.

So, to summarize, for a "wake-up on IDE" feature to work, these are the
necessary conditions:
- There must be the possibility of IDE activity in standby mode.
- This requires that there is a device on the IDE bus that is powered in
standby mode.
- It also requires that the IDE controller is powered in standby mode.
- The mobo/BIOS must support wake up on activity from the IDE controller.

I don't know an ATX mobo/power supply combo that would satisfy these
conditions. I doubt you know one, either. If you do, it should be easy to
post a link or two to the manufacturer's site where these outstanding and
quite rare features are described.


Now to your question about "host adapter" vs. "controller". They are not
the same, and I did use the term "IDE controller" quite on purpose. I was
not talking about all of the IDE host adapter, I was talking specifically
about the IDE controller. Every IDE host adapter contains an IDE
controller. Google for "IDE controller" (use the quotes to limit the
results) and you'll see that there is in fact something like that out there
-- and that's what I was referring to.

To make a somewhat longer story short, the "IDE controller" of a mobo is
usually an IC (or part of an IC) that performs the functions of the IDE
interface controller -- in an IDE host adapter. Maybe something like this
http://www.smsc.com/main/catalog/slc90e66.html (note the term "IDE
Controller" right in the title of the page).

(Who's trying to be "persnickety" about the use of the term "controller"?

How does the motherboard know if there is power being supplied to the disk?

This is all a discussion about ATX standby power, remember? So of course
the mobo knows when the power supply is in standby mode, and it knows that
in standby mode the disks are not supplied with power from the ATX power
supply. It also knows when the power supply is in normal mode, and it
infers that any harddisk connected to its IDE interfaces is supplied with
proper power.

Of course you can connect a harddisk to the IDE interface and /not/ connect
it to the power supply. In this case, there is indeed signal activity on
the harddisk's IDE interface while the disk is not supplied with power. But
as I said, I think that you'll find that most disk manufacturers will
consider this an out-of-spec operation mode and won't guarantee proper
function of the drive or its IDE interface. I'll be happy to read up on
harddisk manufacturer links stating the contrary.

Gerhard

 

[J. Clarke]

No. My native language is German.


I would appreciate if you found the time to point out what you think is
expressed (or understood) not quite right.

OTOH, I'm pretty sure that my /technical/ English is pretty good. I've
been writing technical papers for US companies for some time -- and I've
been reviewing technical papers written by native speakers. While it seems
that I couldn't manage to completely get rid of creating some kind of "odd
feeling" on occasion, the terminology and the grammatical constructs I'm
using seem to be correct.

As far as it goes your English is fine, probably better than mine. But I
think that the difficulty may be that it's formal technical English and not
conversational.

I'm not raising an argument about "trickle" vs. "standby". I'm just trying
to explain to people willing to learn what the better (because more
universal, at least among the people who understand about power supplies)
terminology is -- and why. See my other post here for more details (if
you're interested, of course).

Let's start out with this. "Trickle", according to the Oxford (two of the
standard reference dictionaries for English are the Oxford for British and
Merriam-Webster for American--not just any Webster, but with the "Merriam"
hyphenated--there's also one for Australian but I forget the publisher),
has a number of meanings. The most commonly used are as a noun "a small or
flowing drop; a tear; a small quantity of liquid; a small or fitful stream"
and as a verb there are several variants, most of them suggesting flow in a
"small" or "fitful" "stream". This is not a technical or abstruse word in
any sense, it is in the vocabulary of most Americans and British and its
use is not at all uncommon.

It is most commonly used as a noun or verb--this is the usage that any
schoolchild can recognize. "Trickle charger" is a relatively recent
addition to the language and uses the word as an adjective.

There's nothing in the definition of "trickle" that suggests that it may or
may not be adequate for a given purpose. There are some purposes for which
a trickle is sufficient, others for which it is not. If the power that a
power supply delivers is small then it's a "trickle", regardless of the
capabilities of the power supply. One that is designed to deliver no more
than a trickle and will either shut down or be damaged if it is called on
to deliver more than that can be called a "trickle power supply" if you
wish, but that does not mean that a much larger one that is delivering only
a tiny fraction of its power is delivering anything other than a "trickle
of power".

Thus, arguing that 5VSB is not "a trickle" basically seems to be quibbling
over whether 5 VA is or is not "a trickle" and that sort of argument is in
the same nature as how many angels dance on the head of a pin.

Correct, in that there is power in the case. Wrong, in that this exactly
was the point.

Now, this makes no sense. What is wrong?

Exactly -- depending on what you are doing.


See, that's something I wouldn't do, or advise doing -- and I'm not
surprised that it did create problems. Sometimes, understanding the
workings of power supplies does help

Now, here we hit an issue that seems more related to experience than to
language. How long had you been working with PCs when the first ATX
machines became available? I had been working with them for several
decades and in all that time, turning off the power switch meant that the
power was physically disconnected from the mains, so having power on the
chassis with the power supply turned off was something that occurred only
when the power supply was broken (or when one was working on a Mac). Until
I had taken apart an ATX machine and discovered in the process that the
power switch was connected to the motherboard rather than to the power
supply, there was no reason to suspect that that had changed with ATX.

This has nothing to do with "understanding the workings of power supplies"
and everything to do with the system specs being changed with out anybody
getting the word out. It took several years after ATX became available for
the A+ exam for certifying technicians to be altered to address this
issue--at the time that ATX shipped the correct answer to a question about
leaving a power cord plugged in was that one should so as to ensure that
there is a good chassis ground for static protection purposes.

See, what you call my "fancy-Dan analytical model" is actually the
understanding of the inner workings of power supplies and the like. This
tells me that you should not (un)plug PCI boards under standby power
unless you know the mobo and the PCI board very intimately and know what
you're doing. OTOH, it also tells me that it is safe to (un)plug an IDE
harddrive.

Now here you're arguing like an engineer, not like a technician. To a
technician, if you don't _need_ power then you disconnect it no matter what
your understanding might be. There are several reasons for this. One is
that you're usually inside the machine because it's broken, and broken
electronics do not behave like the book says they are supposed to--if they
did they wouldn't be broken. Another is that one does not want to develop
bad habits. If you disconnect the power every time then you won't forget.
If you disconnect it sometime and not others then one day you'll go through
a long spate of repairs where you don't disconnect it and then come to one
where you need to and you will forget. Far safer to only _connect_ when
you need to than only _dis_connect when you need to. A third is that even
if everything is working as it is supposed to, jostling the wrong board or
cable can bring the power up when it was not intended--one should not risk
that without compelling reason.

Engineers tend to assume that their superior technical knowledge will
preserve them from such difficulties. It doesn't. Most experienced techs
have cleaned up after enough engineer-created smoking holes to know better.

Note that I work as a technician by choice, but I am also a degreed engineer
who worked as an engineer for many years before coming to the understanding
that I enjoy fixing things more than I enjoy the politics of engineering,
so I have seen this first-hand from both sides.

That's why I wrote the below, and here I seem to be in complete agreement
with you (even though you didn't seem to see it that way):

"As long as you're not knowing what you're doing, it's a good idea to
always disconnect the power cord of anything you open."

Now this goes with your comment about "understanding how power supplies
work" and I was about to flame you over it until it occurred to me that you
might not understand why I was doing it. That is what led me to check your
headers and find that you were posting from a non-English-language ISP.

"As long as you're not knowing what you're doing" suggests a value judgment
that I suspect was not intended. It gives the impression that you feel
that anyone who unplugs the machine every time does not know what he's
doing and the mark of a l33t type such as yourself is unplugging sometimes
but not others. Your earlier comments about "understanding how power
supplies work" further advance that impression. Perhaps in German to a
German these do not give offense, in English they can.

If your experience is relevant, then the way to address it is to simply
state your experience and why you think it relevant if that is not obvious
from context. Otherwise, it's best not to mention it.

 

[Ge]

On 8/16/05 10:05:48, J. Clarke wrote:

Let me just clarify this a bit more...

Nobody was talking about "trickle power". He was talking about "a trickle
of power".

You are of course correct here.

You're being pedantic ...

Maybe... but that's not a crime, is it? And with technical issues, it's
better to be pedantic and correct than to be lax and wrong.

... because you're trying to turn a standard usage of the word "trickle", to
refer to a small flow of something, into a persnickety technical point
about the nomenclature of power supplies.

Definitely not.

The problem I saw (and tried to get out of the way of this discussion) is
that the term "trickle of power" seems to create in some minds the notion
that while this "trickle of power" is fed into the mobo there is some kind
of a trickle of power everywhere on the mobo.

That's why I tried to explain that there isn't really a "trickle"; there is
a well-defined standby power supply. While it is in fact a trickle compared
to the full power fed into the system when not in standby mode, it doesn't
trickle all over the mobo. It is only present in well-defined locations
(even though they may not always be as well-known as they are
well-defined).

Everything else -- everything that's not connected to the ATX 5 V standby
power line -- is off. No ifs and buts, not trickles, no nothing -- just
off. Specifically, the peripheral power connectors (the ones with the four
wires; usually red, yellow and black) are off, and so are all the harddisks
connected to them. Not even a trickle of power here.

That's why I wanted to get the term "trickle" out of the way. I think it
hinders the factual discussion about the ATX standby mode and its
implications and creates misunderstandings.

Gerhard

 

[J. Clarke]

Nobody was talking about "trickle power". He was talking about "a
trickle
of power". You're being pedantic because you're trying to turn a
standard usage of the word "trickle", to refer to a small flow of
something, into a persnickety technical point about the nomenclature of
power supplies.

I find your opinion that I'm "trying to turn a standard usage of the word
'trickle' [...] into a persnickety technical point about the nomenclature
of power supplies" not quite compatible with the paragraph of mine that
was right following your opinion. I wouldn't assert that "this has little
to do with the core question of whether or not it is safe to (un)plug
harddisks while on standby power" if I really wanted to turn this into a
"persnickety technical point".

The reason I was trying to point out the difference is that from a
technical (electronic) point of view, there seems to be quite a confusion
in this thread about ATX standby power and its "powers", so to speak. So I
tried to bring in some foundation knowledge. It's yours to take or leave.

I don't quite understand why you object, or to what. It seems that you
don't object to the technical content; at least you haven't brought forth
any technical objections to my explanations of the differences between
trickle and standby power. It may not help you, but it may help others --
if it is correct. If it is not correct, then please start saying what's
wrong. If it is correct, maybe stop objecting and focus on what you want
to object to.

I've already addressed this in another post.

I'm not sure what you mean. The situation where the harddisk power
connectors of an ATX power supply don't carry power is in standby mode and
when off. In neither of these situations, the IDE connectors of the mobo
carry any signals (unless you can show some evidence that they do) -- and
definitely no other harddisks are working.

You neglect the possibility that the signal cable is connected but the power
cable is not. There are very few motherboards that can detect this
situation.

When you say that sometimes the powered-down drive will cause signal
corruption on the IDE bus is a confirmation of my point (see below) that a
drive connected to the IDE bus without being connected to power is in an
operation mode not supported or guaranteed to work by the drive
manufacturer.

In dealing with practical matters of technology, one is concerned with what
is, not with what manufacturers "support or guarantee".

Just because you can tell your mobo to wake up on the IDE controller's
interrupt (more to the use of the term "IDE controller" below) doesn't
mean it actually will wake up. In standby, there's no power to the disks,
therefore the disks won't initiate any activity on the IDE bus, therefore
the IDE controller won't create any interrupt -- if the IDE controller is
powered up at all in standby, which is doubtful.

The "IDE controller" is to a substantial extent a PCI-to-ISA bridge. Why
would it _not_ be powered to the same extent that the rest of the PCI bus
is powered?

So, to summarize, for a "wake-up on IDE" feature to work, these are the
necessary conditions:
- There must be the possibility of IDE activity in standby mode.

Only to the extent that an interrupt can be generated.

- This requires that there is a device on the IDE bus that is powered in
standby mode.

Only to the extent that an interrupt can be generated.

- It also requires that the IDE controller is powered in standby mode.

Host adapter.

- The mobo/BIOS must support wake up on activity from the IDE controller.

No. Only on activity on a particular IRQ.

I don't know an ATX mobo/power supply combo that would satisfy these
conditions. I doubt you know one, either. If you do, it should be easy to
post a link or two to the manufacturer's site where these outstanding and
quite rare features are described.

Only if it was "outstanding or quite rare". Wake on interrupt is a common
feature.

Now to your question about "host adapter" vs. "controller". They are not
the same, and I did use the term "IDE controller" quite on purpose. I was
not talking about all of the IDE host adapter, I was talking specifically
about the IDE controller. Every IDE host adapter contains an IDE
controller. Google for "IDE controller" (use the quotes to limit the
results) and you'll see that there is in fact something like that out
there -- and that's what I was referring to.

The whole point of IDE was that it could be very cheaply connected to an ISA
bus. You could take a first generation IDE drive and an ISA motherboard
and unsolder a connector from the motherboard, solder in a cable, and plug
the other end of that cable into a drive and it would work fine. IDE host
adapters for the ISA bus were simple boards that carried the ISA card edge
connector into a pin header for a cable.

When PCI came about they changed into ISA-to-PCI bridges. But the
controller remains on the disk.

To make a somewhat longer story short, the "IDE controller" of a mobo is
usually an IC (or part of an IC) that performs the functions of the IDE
interface controller -- in an IDE host adapter. Maybe something like this
http://www.smsc.com/main/catalog/slc90e66.html (note the term "IDE
Controller" right in the title of the page).

Sloppy terminology. All it does it pass commands from the PCI bus to the
IDE bus. All the controlling is done on the disk.

(Who's trying to be "persnickety" about the use of the term "controller"?

You're the one who started demanding precise use of language.

This is all a discussion about ATX standby power, remember? So of course
the mobo knows when the power supply is in standby mode, and it knows that
in standby mode the disks are not supplied with power from the ATX power
supply. It also knows when the power supply is in normal mode, and it
infers that any harddisk connected to its IDE interfaces is supplied with
proper power.

Of course you can connect a harddisk to the IDE interface and /not/
connect it to the power supply. In this case, there is indeed signal
activity on the harddisk's IDE interface while the disk is not supplied
with power. But as I said, I think that you'll find that most disk
manufacturers will consider this an out-of-spec operation mode and won't
guarantee proper function of the drive or its IDE interface. I'll be happy
to read up on harddisk manufacturer links stating the contrary.

So when that is done, how does the motherboard know that is has been done so
as to not send signals?

 

[Rod Speed]

J. Clarke wrote

No. My native language is German.

I would appreciate if you found the time to point out what
you think is expressed (or understood) not quite right.

OTOH, I'm pretty sure that my /technical/ English is pretty good.
I've been writing technical papers for US companies for some time
-- and I've been reviewing technical papers written by native speakers.
While it seems that I couldn't manage to completely get rid of creating
some kind of "odd feeling" on occasion, the terminology and the
grammatical constructs I'm using seem to be correct.

Yeah, he is just being nasty, your english is quite a bit better
than many native english speakers can manage, including him.

 

[Ge]

As far as it goes your English is fine, probably better than mine. But I
think that the difficulty may be that it's formal technical English and not
conversational.

That may well be. But then, we're trying to solve here a formal technical
problem, no?

Let's start out with this. ...

I'd like to start with continuing this part of the conversation (/why/ I
started talking about "trickle") from the message I sent previously where I
more explicitly -- and hopefully in proper "conversational language" terms
-- explained why I did it. I repeatedly tried to explain to you why I did
it, yet you never ever reacted to these explanations. Either you ignore
them on purpose (whatever purpose that may be) or you just missed them --
for that case I wrote up a summary of them in a separate message.

Now, this makes no sense. What is wrong?

Wrong is that this was the point, as I wrote before. This was /not/ the
point -- the point was whether or not it is safe to disconnect a harddisk
from an ATX mobo's IDE interface while the ATX power supply is in standby
mode. Whether or not there is power somewhere "in the case or on the mobo"
may or not be relevant -- which is exactly what this discussion is about.
It never was contested that it /is/ -- it in fact is, we always agreed on
this. It was contested that this is /relevant/ -- and this is the point.
Sorry that you didn't get that earlier... but I don't think that the fact
of me being a non-native speaker had much to do with that.

Now, here we hit an issue that seems more related to experience than to
language. How long had you been working with PCs when the first ATX
machines became available?

I've been working with "PCs" before they were called that way.

Until I had taken apart an ATX machine and discovered in the process that
the power switch was connected to the motherboard rather than to the
power supply, there was no reason to suspect that that had changed with
ATX.

Not sure why that was, but maybe because you didn't read the specs. Or
because you fingertips are not sensitive enough to feel the difference
between the old-style power switch (usually a switch that toggles
mechanically between on and off) and the ATX-style power switch (usually
only a push button, which doesn't toggle).

This has nothing to do with "understanding the workings of power supplies"
and everything to do with the system specs being changed with out anybody
getting the word out.

Not sure they actually didn't publish the supply specs when coming out with
them. I would think that they were available. I know that I found them when
I needed them -- but I also know that this was quite some time after they
came out. In any case, /now/ we have them.

It took several years after ATX became available for the A+ exam for
certifying technicians to be altered to address this issue

I'll just assume you are correct on this, but IMO it's quite irrelevant to
this discussion. Any type of certification tends to lag behind technology
development; this doesn't surprise me.

Now here you're arguing like an engineer, not like a technician.

Not really sure what you mean by that. I am an engineer, I have been a
technician, and I think a good engineer can think not only like you may say
an engineer thinks, but also like you may say a technician thinks, and also
in some other ways. The art is in putting it all together -- which is why I
don't tend to think in such categories.

There are several reasons for this. One is that you're usually inside
the machine because it's broken

Now this is a different beast. You're making a lot of assumptions that we
didn't make in the first place. I for my part am usually (dis)connecting
disks because I'm exchanging disks, not because anything is broken.

Another is that one does not want to develop bad habits.

I tend to try to do things out of reason, not out of habit. Any habit in
technology is a bad habit (and you may again say that I'm thinking like an
engineer -- but there's really no bad in that, is it?

Far safer to only _connect_ when you need to than only _dis_connect when
you need to.

Probably. But this is from the quite specific point of view of a repair
technician who does nothing but fixing PCs all day long. That's not my
position, and I don't know why I should do like I possibly would do if I
were one.

A third is that even if everything is working as it is supposed to,
jostling the wrong board or cable can bring the power up when it was not
intended--one should not risk that without compelling reason.

This is the only /real/ point -- and I was well aware of it. And that's why
I wrote that you should know what you're doing if you don't disconnect the
power cable. It's as simple as that.

Engineers tend to assume that their superior technical knowledge will
preserve them from such difficulties.

I hope you don't mean "all engineers" when you say "engineers". I don't
think you're qualified to make such an assertion. In any case, it's not
relevant to the question we're discussing.

Now this goes with your comment about "understanding how power supplies
work" and I was about to flame you over it until it occurred to me that you
might not understand why I was doing it.

I don't think I have a problem with your reasons. From your POV they
probably make sense. But you seem to fail to see that the reasons that make
sense in your specific situation may not make the same sense for everyone
else.

"As long as you're not knowing what you're doing" suggests a value judgment
that I suspect was not intended.

See, here you don't seem to understand how /this/ engineer thinks. I have a
deep respect for "things" and their "life". I don't pretend to know how
they behave -- but I have in some cases a pretty deep knowledge and broad
enough experience to make relatively safe judgment calls. Actually, that's
what I get paid for to do. I can't play it safe all the time; engineering
is mostly about doing what you don't know, but having a reasonable insight
about your odds. When I feel I don't know enough to not destroy anything
(including myself), I unplug the cable -- not only with PCs. Both sides of
the condition are very familiar to me and it's part of my daily life to
determine on which side I am in a particular situation.

It gives the impression that you feel ...

It gives /you/ that impression, possibly because you managed yourself into
a defensive position (without need).

the mark of a l33t type such as yourself

Now here you hit this not-native speaker... I don't have a clue what a
"133t type" is. Nor why you would think that I think that I am one.

If your experience is relevant, then the way to address it is to simply
state your experience and why you think it relevant if that is not
obvious from context.

That's what I did. You managed to read it how you want to see it, not how I
wrote it. That's not a language issue, let me tell you.

Gerhard

 

[Ge]

You neglect the possibility that the signal cable is connected but the power
cable is not.

Of course. To reiterate: the question was whether it is safe to disconnect
a harddisk from the IDE interface of an ATX mobo while the ATX power supply
is in standby mode. You said no, I said yes.

Whether or not the power cable is connected to the disk while doing so is
completely irrelevant (because it doesn't supply any power anyway while the
power supply is in standby).

In dealing with practical matters of technology, one is concerned with what
is, not with what manufacturers "support or guarantee".

Staying inside of what manufacturers guarantee is usually safer than going
outside of it. The odds that things go well are /much/ higher when devices
are used inside their specs than if you use them in situations they are not
designed for -- let alone situations that the manufacturer tells you will
damage the device. You need to be /very/ familiar with a device's inner
workings if you want to use it /safely/ outside of its specs. I don't know
why you object here -- this should be obvious.

I've never seen a mobo that supports wake-up on IDE controller activity.
Care to share a make/model?

[...]
So, to summarize, for a "wake-up on IDE" feature to work, these are the
necessary conditions: [...]
I don't know an ATX mobo/power supply combo that would satisfy these
conditions. I doubt you know one, either. If you do, it should be easy to
post a link or two to the manufacturer's site where these outstanding and
quite rare features are described.

Only if it was "outstanding or quite rare". Wake on interrupt is a common
feature.

You haven't been able to bring forth a single example for a system that
implements "wake-up (from ATX standby mode) on IDE activity" and confirm
that you actually saw it waking up from IDE activity. A few nit-picking
comments is all that is there.

Sloppy terminology. All it does it pass commands from the PCI bus to the
IDE bus. All the controlling is done on the disk.

You are of course right in that all the controlling /of the disk/ is done
by the controller on the disk. But the controlling (however minimal that
may be) of the IDE /interface/ is done by the IDE controller of the IDE
host adapter. That's the common terminology for this part of the host
adapter (whether it being not much more than an ISA-to-PCI bridge or not,
but I'm sure you're enough engineer to know that even that is not exactly
trivial) -- just check out this
http://www.google.com/search?q="ide+controller" ; the term is /very/
common. It may not be to your liking, but it's commonly used.

You're the one who started demanding precise use of language.

No, not "demanding". I never demanded anything. You started criticizing my
precise use, for whatever reason, so I explained the reasons for doing so.

Of course you can connect a harddisk to the IDE interface and /not/
connect it to the power supply. [...]

So when that is done, how does the motherboard know that is has been
done so as to not send signals?

It doesn't, that's why this is an "out of spec" situation. You shouldn't do
that (connecting a hard disk to the IDE interface without connecting it to
the power supply). You already confirmed that you had problems when doing
so, so why not simply agree with me here -- no matter how hard it may be?

Gerhard

 

[Arno Wagner]

On 8/17/05 11:22:03, J. Clarke wrote:

Why don't you two stop arguing. +5VSB is up to 2A in many cases.
While that _is_ "a trickle" compared to what a PC usually needs, it is
quite enough to burn out allmost anything in a computer if applied
wrongly. So the question of whether this is a problem is certainly
a valid question, regardless or wording issues.

Side note: "trickle" is certainly not a defined EE term with fixed
meaning, so get over about your two interpretations being different.

Arno

 

[J. Clarke]

That may well be. But then, we're trying to solve here a formal technical
problem, no?


I'd like to start with continuing this part of the conversation (/why/ I
started talking about "trickle") from the message I sent previously where
I more explicitly -- and hopefully in proper "conversational language"
terms -- explained why I did it. I repeatedly tried to explain to you why
I did it, yet you never ever reacted to these explanations. Either you
ignore them on purpose (whatever purpose that may be) or you just missed
them -- for that case I wrote up a summary of them in a separate message.


Wrong is that this was the point, as I wrote before. This was /not/ the
point -- the point was whether or not it is safe to disconnect a harddisk
from an ATX mobo's IDE interface while the ATX power supply is in standby
mode. Whether or not there is power somewhere "in the case or on the mobo"
may or not be relevant -- which is exactly what this discussion is about.
It never was contested that it /is/ -- it in fact is, we always agreed on
this. It was contested that this is /relevant/ -- and this is the point.
Sorry that you didn't get that earlier... but I don't think that the fact
of me being a non-native speaker had much to do with that.



I've been working with "PCs" before they were called that way.


Not sure why that was, but maybe because you didn't read the specs.

What specs?

Or
because you fingertips are not sensitive enough to feel the difference
between the old-style power switch (usually a switch that toggles
mechanically between on and off) and the ATX-style power switch (usually
only a push button, which doesn't toggle).

Now let's see, you claim to have been fixing personal computers prior to the
time that they were called that and you've never encountered a machine that
uses the power button on the front to control a relay?

Not sure they actually didn't publish the supply specs when coming out
with them. I would think that they were available.

Until one knows that there is a difference one has no reason to look for
specs.

I know that I found
them when I needed them -- but I also know that this was quite some time
after they came out. In any case, /now/ we have them.


I'll just assume you are correct on this, but IMO it's quite irrelevant to
this discussion. Any type of certification tends to lag behind technology
development; this doesn't surprise me.


Not really sure what you mean by that. I am an engineer, I have been a
technician, and I think a good engineer can think not only like you may
say an engineer thinks, but also like you may say a technician thinks, and
also in some other ways. The art is in putting it all together -- which is
why I don't tend to think in such categories.


Now this is a different beast. You're making a lot of assumptions that we
didn't make in the first place. I for my part am usually (dis)connecting
disks because I'm exchanging disks, not because anything is broken.


I tend to try to do things out of reason, not out of habit. Any habit in
technology is a bad habit (and you may again say that I'm thinking like an
engineer -- but there's really no bad in that, is it?

If "thinking like an engineer" results in things getting broken and people
getting hurt then yes, there _is_ harm in it.

Probably. But this is from the quite specific point of view of a repair
technician who does nothing but fixing PCs all day long. That's not my
position, and I don't know why I should do like I possibly would do if I
were one.

So how are you disconnecting these disks if you do not have the case open
plugging and unplugging cables?

This is the only /real/ point -- and I was well aware of it. And that's
why I wrote that you should know what you're doing if you don't disconnect
the power cable. It's as simple as that.

No matter how thoroughly you know what you're doing fingers slip. Maybe you
think that yours don't, but if you believe that then it is difficult to
reconcile that view with your claimed years of experience.

I hope you don't mean "all engineers" when you say "engineers". I don't
think you're qualified to make such an assertion. In any case, it's not
relevant to the question we're discussing.

It is relevant to your assertions.

I don't think I have a problem with your reasons. From your POV they
probably make sense. But you seem to fail to see that the reasons that
make sense in your specific situation may not make the same sense for
everyone else.

So show me that your reasons make sense.

See, here you don't seem to understand how /this/ engineer thinks. I have
a deep respect for "things" and their "life". I don't pretend to know how
they behave -- but I have in some cases a pretty deep knowledge and broad
enough experience to make relatively safe judgment calls. Actually, that's
what I get paid for to do. I can't play it safe all the time; engineering
is mostly about doing what you don't know, but having a reasonable insight
about your odds.

Not in any engineering firm that I am aware of. If you don't _know_ then
you find out.

When I feel I don't know enough to not destroy anything
(including myself), I unplug the cable -- not only with PCs. Both sides of
the condition are very familiar to me and it's part of my daily life to
determine on which side I am in a particular situation.


It gives /you/ that impression, possibly because you managed yourself into
a defensive position (without need).


Now here you hit this not-native speaker... I don't have a clue what a
"133t type" is. Nor why you would think that I think that I am one.


That's what I did. You managed to read it how you want to see it, not how
I wrote it. That's not a language issue, let me tell you.

Actuall it is. You go on talking about how much you know and how little
everyone else does and see how long anybody gives a damn about your
opinion.

I don't anymore.

 

[J. Clarke]

Why don't you two stop arguing. +5VSB is up to 2A in many cases.
While that _is_ "a trickle" compared to what a PC usually needs, it is
quite enough to burn out allmost anything in a computer if applied
wrongly. So the question of whether this is a problem is certainly
a valid question, regardless or wording issues.

Side note: "trickle" is certainly not a defined EE term with fixed
meaning, so get over about your two interpretations being different.

That's my point--it's not a technical term, arguing that it was used
incorrectly is silly.