On 6/7/2016 3:24 PM, Ken Moffat wrote:
On 7 June 2016 at 20:09, Aleksandar Kuktin <[email protected]
<mailto:[email protected]>> wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
>On Tue, 7 Jun 2016 19:36:43 +0100
>Ken Moffat <[email protected]
<mailto:[email protected]>> wrote:
>
> On 7 June 2016 at 18:18, Aleksandar Kuktin <[email protected]
<mailto:[email protected]>> wrote:
> I'm not familiar with the sound a dead psu makes. I had one a few
> years ago, all I remember is the burnt smell. The sound is
> high-pitched, probably in the high hundreds of Hz, or maybe 1 KHz.
> But before I sent this reply I googled for the sound of a dead PSU -
> one forum post where a new PSU solved the problem does seem to match
> what I'm hearing.
I would have been "happier" with, say, 50 kHz because switching power
supplies (the kind that gets put into PSUs) usually operate internally
at about that frequency.
I don't think I would hear even 15 KHz.
[...]
>> I'm not clear about the actual symptoms you are having. I'll post my
>> understanding, you correct my mistakes.
> For the test machine: as soon as I turn the switch on the PSU to 'on'
> the sound starts. If I then press the power switch on the case the
> motherboard and CPU fans both start, but I seem to not have any video
> output (that might be luser error - the KVM switch is dieing, only one
> position works reliably, maybe I misconnected after failing to boot
> the test machine to try to do a clean shutdown andlater swapping it
> back).
Does it beep? A properly operating motherboard will beep twice. In this
case, the most important beep is the first one. That's the motherboard
firmware saying "I'm alive". Fans are simple devices and will operate
even if the PSU outputs "wrong electricity", say a train of 100Hz
pulses instead of a steady voltage.
No beep. But I don't think it has any speaker connected.
Most motherboards will still have a little black can, a "piezo squeaker
can" as one of my friends calls it. As for PSUs, I had one in a friend's
machine that I had to replace with a very high pitched noise like that.
Nothing had failed in that system, so I can't be too confident on yours,
but I am leaning towards the possibility that its just the PSUs.
> [snip]
>
> Getting space to open them is a pain, but yes I have other machines
> (using one of them at the moment).
You'll probably need to open the machines.
The traditional way of doing this is to have a known good computer and
then swap in parts from a broken computer, waiting to see when it will
fail. In theory, I'd first concentrate on the disks and make sure they
are alive. If the disks failed (got fried), then you will at least
discover that early. But in practice, it's probably good enough to
replace the PSU with a known good one and see if the machine boots.
Make sure the KVM magic is working well. :) If it does not boot, I'd do
the whole transplant-bit-by-bit ritual. Beware hasty conclusions.
Yes, I've done this in the past. That test machine is the one I would
have used to test a new or doubtful disk (light case, easy to access in
its current location) - using one of the others will take a bit of time.
> The server would have been running and I won't be surprised by a
> failed PSU. But the test machine was using minimal power (nominally
> powered off, probably 5 Watts maximum) and it is perhaps a year old,
> maybe two, but unlikey to be more than that.
Hmm.. bizarre. Yet, everything is possible. Note that the fact the PSU
was switched off may not be of any significance. There's no way to know
without knowing more about its internals.
Thanks. I'll see about looking at the SSD from the test machine, then
probably get a PSU or two. For the moment I'm not touching the server,
I seem to have developed an ability to break things recently.
I seem to have developed that ability recently too. Had a few things die
recently that I have been able to repair... after more issues arose.
ĸen
--
Douglas R. Reno
--LFS/BLFS systemd maintainer
--
http://lists.linuxfromscratch.org/listinfo/lfs-chat
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
