Re: is my power supply dieing?

Top Page
Attachments:
Message as email
+ (text/plain)
+ (text/html)
+ (text/plain)
Delete this message
Reply to this message
Author: Michael Butash via PLUG-discuss
Date:  
To: Main PLUG discussion list
CC: Michael Butash
Subject: Re: is my power supply dieing?
There's really no way to tell from the software side if your psu is
failing. If the psu randomly stops, your mobo and os dies with it. Your
mobo may detect bad power as part of kernel dmesg's to the cpu, but likely
not.

Problem with PSU's are generally bad electrical components like a blown out
capacitor(s), meaning you're not handing power burst demands from your
motherboard and/or gpu. Why it's so hard to tell if/when they really do
go, they tend to only under certain loads or thresholds. You more need a
volt/ohm/capacitance meter to see this.

Any old PSU you can replace it with from a working device in the past 15
years or so should tell you if it reboots in the same duration as current.
Beg/borrow/steal one for a week. Most stores have a few week return
policy, worst case buy one and return it later. Craigslist, letgo, etc too.

If not that, I'd suspect either memory, gpu (if separate), or mobo. Your
linux bootloader with grub should include memtest as a menu option by
default, let that run and test patterns across your ram. If an nvidia gpu,
install nvtop and psensor to watch your gpu use and temps. Your mobo, look
for big cylinder components (google capacitor) on the board that have
exploded or are leaking goo out of them.

-mb


On Wed, Jul 15, 2020 at 11:44 AM Michael via PLUG-discuss <
> wrote:

> do any of the tools on http://www.ultimatebootcd.com/ (scroll down to
> see the tools) do any dIgnostics on the power supply?
>
> On Wed, Jul 15, 2020 at 12:13 PM Michael <> wrote:
> >
> > the computer gal told me to run it until it dies and then get a new
> computer.
> >
> > On Wed, Jul 15, 2020 at 12:05 PM Michael <> wrote:
> > >
> > > darn.... they couldn't find anything wrong with it... sort of. the
> > > results were inconclusive. Is there an open source diagnostics tool I
> > > could d/l ?
> > >
> > > On Tue, Jul 14, 2020 at 1:01 PM Michael <> wrote:
> > > >
> > > > Thanks for the tip! YEAH.... I took the easy way out; I took it to a
> shop.
> > > >
> > > > On Tue, Jul 14, 2020 at 10:41 AM Donald Mac McCarthy <
> > wrote:
> > > > >
> > > > > I can't confirm really if it would do - I was just pointing out
> that
> > > > > inconsistent power manifests itself in a number of interesting
> ways. So
> > > > > if you suspect the power supply is to blame, it is generally the
> > > > > cheapest in the chain to fix. You could try replacing the CPU
> cooler
> > > > > first. You can get those for $20-30 for a pretty nice one. If that
> > > > > doesn't work, yeah - the PSU sounds like it might be your next
> best bet.
> > > > >
> > > > > Mac
> > > > >
> > > > > Michael wrote on 7/13/20 9:14 PM:
> > > > > > I just opened my box and the power supply fan is spinning as
> well as
> > > > > > the CPU fan so I don't know what inxi is talking about
> > > > > >
> > > > > > On Tue, Jul 14, 2020 at 12:09 AM Michael <>
> wrote:
> > > > > >> Thanks for the tip! I'd love to come get it but the distance
> makes it
> > > > > >> prohibitive. I live in Florida But you think replacing the power
> > > > > >> supply unit will do it?
> > > > > >>
> > > > > >> On Tue, Jul 14, 2020 at 12:00 AM Donald Mac McCarthy <
> > wrote:
> > > > > >>> Some boards have 2 fans for CPU - especially boards designed
> for
> > > > > >>> overclocking. The face that you don't have any voltage/RPMs
> across the
> > > > > >>> fan3 may not be a problem.
> > > > > >>>
> > > > > >>> The only way you are going to be sure to tell it is a power
> supply is
> > > > > >>> generally to replace it. You could replace CPU, memory, or
> board one at
> > > > > >>> a time, but those are generally more expensive than a PSU.
> > > > > >>>
> > > > > >>> I have had many times that the PSU was the issue when running
> a compute
> > > > > >>> cluster years ago. In one case the PSW was not getting enough
> airflow
> > > > > >>> due to the positional design of the chassis vs the fan
> location of a
> > > > > >>> replacement CPU which had undergone a spec change, and
> therefore wouls
> > > > > >>> reach thermal protection shutdown. In another, a thermal
> expansion in a
> > > > > >>> PSU component when under load would cause a short, and the
> system would
> > > > > >>> shutoff. In another case - dirty output due to a power limiting
> > > > > >>> component that was failing cause all kinds of problems,
> eventually
> > > > > >>> causing us to have to replace RAM, CPU, MB, and a RAID card
> because of
> > > > > >>> "brown out" type situations. We were only to know what
> happened after
> > > > > >>> the manufacturer did some testing on the RMAed PSU. When you
> have 850
> > > > > >>> servers all built assembled from components (academic
> environment where
> > > > > >>> the Professor got more systems for his grant money by buying
> components
> > > > > >>> and using undergrad/graduate research assistants to assemble
> them), some
> > > > > >>> interesting things happen, may of them were power supply
> related.
> > > > > >>>
> > > > > >>> Good luck, but I think you may save more in time and effort to
> just
> > > > > >>> replace it. I have a 400W ATX powersupply you can have if you
> want to
> > > > > >>> come get it. I am not sure how much wattage you need, but if
> 400 will
> > > > > >>> cut it - that one can be yours for the low low price of gas
> money.
> > > > > >>>
> > > > > >>> Mac
> > > > > >>>
> > > > > >>> Michael via PLUG-discuss wrote on 7/13/20 8:41 PM:
> > > > > >>>> inxi tells me this about my fans:
> > > > > >>>>
> > > > > >>>> Fan Speeds (RPM): cpu: 0 fan-1: 3139 fan-3: 0
> > > > > >>>>
> > > > > >>>> So my cpu fan isn't working. I thought my computer would
> CRASH/FREEZE
> > > > > >>>> more often if the cpu fan wasn't working. What is: fan-1:
> 3139 fan-3:
> > > > > >>>> 0
> > > > > >>>>
> > > > > >>>> I'm not sure but I think my power supply fan is running slow.
> Is that a thing?
> > > > > >>>>
> > > > > >>>> On Mon, Jul 13, 2020 at 11:26 PM Michael <>
> wrote:
> > > > > >>>>> hey.... I forgot to tell ya all that last night after I put
> the
> > > > > >>>>> system under stress I got it to freeze.
> > > > > >>>>>
> > > > > >>>>> On Mon, Jul 13, 2020 at 9:29 PM Michael <>
> wrote:
> > > > > >>>>>> when I'm just running memory test the temp is 46
> > > > > >>>>>>
> > > > > >>>>>> On Mon, Jul 13, 2020 at 8:06 PM Michael <>
> wrote:
> > > > > >>>>>>> that was it under stress.
> > > > > >>>>>>>
> > > > > >>>>>>> On Mon, Jul 13, 2020 at 4:55 PM Brian Cluff via
> PLUG-discuss
> > > > > >>>>>>> <> wrote:
> > > > > >>>>>>>> If that's your idle temperature, that is terrible. I'll
> bet if you run
> > > > > >>>>>>>> stress while monitoring your temperature you'll see it
> shoot up even
> > > > > >>>>>>>> higher than that.
> > > > > >>>>>>>>
> > > > > >>>>>>>> (I cook chicken sous vide at 60C)
> > > > > >>>>>>>>
> > > > > >>>>>>>> Brian Cluff
> > > > > >>>>>>>>
> > > > > >>>>>>>> On 7/13/20 10:05 AM, Michael via PLUG-discuss wrote:
> > > > > >>>>>>>>> Do 80 is bad? Mine was at 89 when I first started it
> > > > > >>>>>>>> ---------------------------------------------------
> > > > > >>>>>>>> PLUG-discuss mailing list -
>
> > > > > >>>>>>>> To subscribe, unsubscribe, or to change your mail
> settings:
> > > > > >>>>>>>> https://lists.phxlinux.org/mailman/listinfo/plug-discuss
> > > > > >>>>>>>
> > > > > >>>>>>> --
> > > > > >>>>>>> :-)~MIKE~(-:
> > > > > >>>>>>
> > > > > >>>>>> --
> > > > > >>>>>> :-)~MIKE~(-:
> > > > > >>>>>
> > > > > >>>>> --
> > > > > >>>>> :-)~MIKE~(-:
> > > > > >>>>
> > > > > >>> --
> > > > > >>> Donald "Mac" McCarthy
> > > > > >>> Director, Field Operations
> > > > > >>> Open Source Context
> > > > > >>> +1.602.584.4445
> > > > > >>>
> > > > > >>> https://oscontext.com
> > > > > >>>
> > > > > >>
> > > > > >> --
> > > > > >> :-)~MIKE~(-:
> > > > > >
> > > > > >
> > > > >
> > > > > --
> > > > > Donald "Mac" McCarthy
> > > > > Director, Field Operations
> > > > > Open Source Context
> > > > > +1.602.584.4445
> > > > >
> > > > > https://oscontext.com
> > > >
> > > >
> > > >
> > > > --
> > > > :-)~MIKE~(-:
> > >
> > >
> > >
> > > --
> > > :-)~MIKE~(-:
> >
> >
> >
> > --
> > :-)~MIKE~(-:
>
>
>
> --
> :-)~MIKE~(-:
> ---------------------------------------------------
> PLUG-discuss mailing list -
> To subscribe, unsubscribe, or to change your mail settings:
> https://lists.phxlinux.org/mailman/listinfo/plug-discuss

---------------------------------------------------
PLUG-discuss mailing list -
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss