Hello list! This subject is most likely hardware related, but I realize it’s a bit off topic.
In the past I’ve got excellent advise from this forum and my openSuse 11.4 is affected by the shutdown/reboot problem.
If you’re still with me here, both shutdown -h and shutdown -r suddenly started causing a reboot. This Compaq Pentium 4 box is setup with a triple boot of W98, XP, and 11.4. None of these OSs can shutdown the box. They all restart it.
The box does have a BIOS setting for AC power fail, but that’s only for restart, and it works.
What I’m hoping for here is a response from someone that understands the working of the shutdown/reboot process and the role the power supply plays in that.
For example, if a AC power relay failed in the PSU, could that change a shutdown to a reboot? I don’t have another PSU to swap at the moment, but I do have an older working PSU that may have interchangeable parts to experiment with.
A long time ago computers would shut the program down and display on the monitor, the message for the user to manually turn off the power. This leads me to believe that the reboot/shutdown mechanism is independent of the PSU.
I would appreciate a response to my post, but since this post is a ways off topic I’ll understand if I get silence back! Heboland
> I would appreciate a response to my post, but since this post is a ways
> off topic I’ll understand if I get silence back! Heboland
I don’t think it is entirely off-topic, as this is hardware and you use
Linux
I don’t know exactly how it is done, but I can guess. I think that the
motherboard toggles down or up a sensor wire from the PSU, and this
responds by powering off. On a reboot the system either doesn’t toggle, or
the PSU doesn’t respond.
I think that there is another wire to power up the PSU, which in fact
doesn’t fully power off, but sleeps and leaves power on some wires. Thus
you can power up on network activity or a keypress.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
Generally speaking, when multiple OS’ have such a problem it must be hardware related or BIOS setup related. As for the PSU, there is a pin called Power Good, which signals the motherboard it is OK to run and there is a pin the motherboard controls to signal the power supply to turn on or off, based on the front panel switch or resume from suspend and so forth. Nothing is present intended to reboot the computer. So, here are things to check:
Case & Heat sinks can be full of dust and need to be cleaned, including the power supply.
Case and Power Supply fans can go bad and need to be replaced.
Connections can become dirty, needing to re-seated for all adapter cards, cables and memory.
The CMOS battery can go bad causing all manner of issues.
You CMOS setup can become corrupted or changed. Its worth looking for any Wake-On settings such as LAN, Mouse, Modem to make sure all are off.
Power Supply’s do go bad, but just causing a reboot on shutdown seems kind of strange.
Always keep a clean system no matter what but of the things listed above, Item 5 seems the most likely problem to me.
Thanks JD, Robin! I did disable all the wake-up options I found in the bios. Then I took the psu apart, and blew it out along with the mobo. Finally I replaced the CMOS battery all with no change. Inspecting the psu didn’t reveal any relays to shut off the AC power.
The psu has autoswitching so there isn’t a 120/240 mechanical switch. AC power in goes thru a fuse and directly to the bridge rectifier, so the raw dc power is always on. This direct connection seems to support your point 6, JD.
These links apply to an ATX psu. My mobo and psu have the 24-pin connector, but both the connector pinout and color code are different from the example in the bearblain link. Apparently my mobo is not an ATX.
I’m curious what my mobo/psu would be classed as being a 250W pentium 4 from compaq. The PSU has a compaq number of 243890-001.
I did peal the insulation off several psu to mobo wires, I expected to be PS_ON, but there was no change in them with power cycling.
Also I did trace the front panel power switch to the mobo, but it does not connect to the psu. Apparently there is a circuit on the mobo that senses the power switch, then drives PS_ON low. PWR_OK is probably in the 24-pin bundle, driven from the psu, but I hate to peal the insulation off all 24 wires to find either of these signals.
Is there an easier way to find these two psu control wires? Heboland
On 01/30/2012 07:16 AM, heboland wrote:
> a 250W pentium 4
> from compaq. The PSU has a compaq number of 243890-001.
>
> I did trace the front panel power switch to the mobo, but it does
> not connect to the psu. Apparently there is a circuit on the mobo that
> senses the power switch, then drives PS_ON low.
correct…the power button (or wake-on-lan or other actions) cause the
start sequence to begin, but MotherBoard circuitry (along with the BIOS,
of course) does the work–if that part of the MB/BIOS is fried (or just
ill) you got problems…
> Is there an easier way to find these two psu control wires? Heboland
i think you have confirmed that yours is a hardware problem (independent
of what OS you use) so, and i would strongly suggest you trip over to a
compaq corporate or users forum where you might find a relative mountain
of expert help on how to track down the broken bit in your MB, PSU or
wiring, and thereby solve your unintended start-up issue.
please, when you have a solution (or expert advice to buy a new board,
or junk the machine) please come back here and fill us in (so, those who
google in looking for help find gems in this thread)…
> These links apply to an ATX psu. My mobo and psu have the 24-pin
> connector, but both the connector pinout and color code are different
> from the example in the bearblain link. Apparently my mobo is not an
> ATX.
O rather, it is a Compaq. Compaq is the second name of HP, and HP has a
tradition of doing things differently. You will have to add compaq or hp to
your search parameters.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
Hi Robin, DenverD! This morning I have some egg on my face!
My problem was bios PME wakeup that JD told me to disable. I did disable PME wakeup, but I did that before replacing the CMOS battery. The battery replacement changed some bios settings and I didn’t catch that one. It’s in the Advanced > Device Options.
I tried to mark this thread as solved, but if I did it wrong, feel free to complete the job. Again thanks to you, DenverD and JD for excellent guidance with this problem.
For the record, I found another of the pages on this website Home Page. It shows that the 24-pin connector is pinned the opposite of the 20-pin connector. Also with a good light and a magnifying glass the mobo connector has embossed pin numbers.
Be that as it may, my connector has no wire in pin 8 which is supposed to be the PWR_OK signal. My pin 16 is a white wire with a constant 3V on it.
In the end, I’m glad I don’t have find my mobo pinout. Apparently the CMOS battery caused the original problem, but the bios has to be restored AFTER changing the battery. Heboland.
On 2012-01-30 17:06, heboland wrote:
> In the end, I’m glad I don’t have find my mobo pinout. Apparently the
> CMOS battery caused the original problem, but the bios has to be
> restored AFTER changing the battery. Heboland.
Interesting
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
That is great news to hear heboland. Indeed if you change out the CMOS battery, you must revisit all BIOS settings to make sure they are as you require. I also often load BIOS defaults which is more likely to turn off troubling BIOS settings which can be good if you miss something. IN any event you did good by placing Solved into your New Message Title. I also added the Tag:Solved! you can do at the bottom right of each message thread page. Thanks for letting us know of your success.
So, this is only a test, but I did an edit of the primary title to include Solved. We will not be doing that normally as its not really required, but I wonder if you see that change after say 10 minutes have transpired?
> So, this is only a test, but I did an edit of the primary title to
> include Solved. We will not be doing that normally as its not really
> required, but I wonder if you see that change after say 10 minutes have
> transpired?
Nope, no change. I see it on the web side, but not on the nntp. I use a
local cache, so I would not see it, anyway. So I start another viever
without cache, and as I thought, the titles are unchanged. This is because
once sent to the gateway, a message is not sent to it, the nntp server
keeps its own copy.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
> On 2012-01-31 02:46, jdmcdaniel3 wrote:
>
>> So, this is only a test, but I did an edit of the primary title to
>> include Solved. We will not be doing that normally as its not really
>> required, but I wonder if you see that change after say 10 minutes have
>> transpired?
>
> Nope, no change. I see it on the web side, but not on the nntp. I use a
> local cache, so I would not see it, anyway. So I start another viever
> without cache, and as I thought, the titles are unchanged. This is because
> once sent to the gateway, a message is not sent to it, the nntp server
> keeps its own copy.
The “(Solved)” add-on is in the nntp copy Ijust read…
On 2012-01-31 10:17, DenverD wrote:
> On 01/31/2012 03:03 AM, Carlos E. R. wrote:
>
>> Nope, no change.
>
> i see the moderator edited subject on the nntp side (running online, not
> cached)…
Yes, you are right, I must have been blind.
> and, i can remove/replace it here in my reply…if that is seen on the web
> side fine…if not, it is their loss…
But there is an earlier post with the solved at the start of the sentence
that doesn’t propagate here. I don’t know the number.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
