1
\$\begingroup\$

From Wikipedia:

The Power Good signal prevents a computer from attempting to operate on improper voltages and damaging itself by alerting it to improper power supply. The ATX specification defines the Power-Good signal as a +5 volt (V) signal generated in the power supply when it has passed its internal self-tests and the outputs have stabilized. This normally takes between 0.1 and 0.5 seconds after the power supply is switched on. The signal is then sent to the motherboard, where it is received by the processor timer chip that controls the reset line to the processor.

We know than an HDD has as complex as an entire motherboard and has two or three CPUs with special purposes. I can't understand why the POWER_OK is required by motherboard and not provided to HDDs.

If the HDDs has the logic to check the power signal by themselves, or are using a self timer before setting the CPUs why the same has not been done for the central motherboards.

It was told here that "Peripherals can be fitted with their own voltage supervisory circuits" as the mobos has voltage tests shown in DMIbios that is also shown in some bioses settings menu. It is reasonable that both has their protection so I cant't figure how why POWER OK exists if every hardware in the case is self caring.

Moreover the electronics that must check other electronics not to go crazy for bad power how can be trustable if is itself bad powered? Maybe that is the scope of power_ok since power supply can know something is wrong before the capacitors gets completely empty?

It seems reasonable that microcontrolles has a startup delay to spend before POST waiting for reset command. I suppose, please confirm, that after start up the design is (in case of power problems) to reset the motherboard by dropping POWER OK. At this point is the motherboard responsable to send reset over sata bus?

In case of bad power, if the CPU is not left "going crazy" for a long time since is reset there is less possibility of data corruption. After the reset the CPU is not started if powerok is not up again, right? Problem is if molex remains powered up since SMART is doing read and write job (during idle, since smart III it also rewrites older sectors). I am also wondering if hard drives has some capacitors for some instant of residual power.

Is seems to me an incomplete design: or power ok is not necessary at all or modern hard drives operating by themselves are incompatibile with that design. ( I tested by myself drives from HFST, Segate, WD and older IBM doing SMART activity with no data cable inserted so no initial reset received).

\$\endgroup\$
3
  • \$\begingroup\$ Hi all, is seems to me that ATX power ok design is somehow illogical or not complete. While mobos knows when power supply is ready to run and when voltage levels drops below certain levels hard drives has to deal with that potential problems by themselves. So question can turn in: why power_ok exists if also motherboard could deal the same way as hard drives. \$\endgroup\$ Jan 13, 2015 at 10:34
  • \$\begingroup\$ Maybe molex connectors are powered on only when power ok is up? HDDs are the only component which can have permanent data damage in case of power problems. \$\endgroup\$ Jan 13, 2015 at 12:12
  • \$\begingroup\$ And they have had such problems. A few years ago HDDs started to have RAM caches. And Windows 7 (or was it Vista?) did shutdown so quickly that the HDDs still had unwritten data in their internal caches when the power was switched off. The Microsoft guys actually had to slow down the shutdown of windows to prevent data loss. \$\endgroup\$
    – PkP
    Jan 14, 2015 at 4:39

2 Answers 2

2
\$\begingroup\$

The hard drive power connector ("Molex") predates the PC. It was in use already in 1976 and manufactured by Amp (not Molex). The connector was an established standard even before any of the PC, PC/XT, AT or ATX power supplies became into existence.

Hard drives and such are connected to the motherboard via a control bus (IDE, SATA,...) which provides them the necessary resets and commands. It wasn't really necessary to force any changes to the established power connector, because the hard drive is essentially slow to start up and doesn't need to do anything before it receives a reset from the motherboard, although they can spin up the disk if they so choose. And in any case, hard drives were already manufactured without the need for any "power good" signaling from the power supply.

Peripherals can be fitted with their own voltage supervisory circuits, if needed. Also the power supply could be designed to not to provide power to the Molex connector before the +5 volt and +12 volt supplies become stable.

\$\endgroup\$
5
  • \$\begingroup\$ Hi, also the motherboards are fitted with voltage supervisory circuits as seen on DMI. Moreover even without reset over SATA the SMART attributes and periodic test are performed so the HDD logic is working. \$\endgroup\$ Jan 12, 2015 at 19:17
  • \$\begingroup\$ Right, also I think the origin of the power good signal is that the power supply might have better knowledge of the power conditions than what the motherboard can detect. So the motherboard first gets the power good signal from the power supply and then the motherboard knows it can start its operation and start all the different peripherals etc. \$\endgroup\$
    – PkP
    Jan 12, 2015 at 19:48
  • 1
    \$\begingroup\$ Unlikely SMART attributes are updated even if no SATA cord is plugged. No reset command is necessary for HDD to operate. It seems to use two different logic for motherboard and peripericals which seems not ideal. \$\endgroup\$ Jan 12, 2015 at 20:59
  • 1
    \$\begingroup\$ @AlessandroCantelli-Forti, I did some research and rewrote the answer. \$\endgroup\$
    – PkP
    Jan 13, 2015 at 14:11
  • \$\begingroup\$ It would be very nice to know if the known as molex are powered after power ok is up. Also the at and atx connectors were already on marked and made by molex but they decided to wire the powerok. Any test can be done with standard equipment? \$\endgroup\$ Jan 15, 2015 at 16:06
1
\$\begingroup\$

A hard drive takes fractionally longer to spin up than for the power supply to stabilize. During that time the controller is doing very little.

Also the hard drive doesn't need to be "ready" as quickly as the motherboard, since it's not probed for some seconds into the POST sequence.

The controller can quite happily idle for a second or so during power-up before it actually starts processing. The majority of microcontrollers and similar control systems have such a "Power On Timer".

\$\endgroup\$
1
  • \$\begingroup\$ Ciao :), during spin up time a lot of work is done and SMART attributes updated, ssd doesn't have spin up time moreover. About the microcontroller "power up timer" why it is not implemented the same way over the motherboard? That approach is not efficient against power flickering or values going lower than threshold, I suppose. \$\endgroup\$ Jan 12, 2015 at 19:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.