[SOLVED] Expert opinions about a hardware failure?

[Qqmike]

I have an ASUS H97 PLUS motherboard, I built this PC in March 2015. My PC was on, and there was an electrical power outage on Good Friday morning at 7:30 am. Shortly thereafter, the power came back on, the PC remained off, I pressed the Power-ON button of the PC, everything started up seeming normal (normal fans, normal sounds, normal lights e.g. on the HDD), but nothing on the monitor. I do a hard shutdown (by pressing and holding the Power-ON button), I turn off the power supply, unplug it, and turn off power to the monitor. Wait a couple minutes. Turn all power back on, press the Power-ON button of the PC, and the same thing happens—it starts up but no display on the monitor.

I try this a few times and finally everything works well.

I leave the PC ON, use it heavily, and everything is OK. I then shut down the PC. When I try to turn it back on, the cycle (described in the paragraph above) starts again except it takes longer to get the PC to finally work OK, until finally the PC and monitor start up OK and everything works. Once it does start OK, it will run even overnight. But I have noticed high CPU usage, high fan activity, and “noise.”

This process of starting up the PC continues, about the 3^rd cycle it takes rather long time to get the PC to start up right. On one cycle, the LED standby light on the motherboard didn't even turn on (which it should be on if there is power to the board from the power supply).

OK, I know what you're thinking by now … is the monitor damaged? The power supply damaged? Well, I can't be sure, of course, but I tried a known good monitor and a known good power supply, and neither effort solved the “bad startup issue.” I even reset the CMOS-firmware (ouch, that wiped out a lot of my UEFI firmware settings).

Bottom-line:

I think the motherboard is damaged—something to do with the power management, the flow of juice to the video output, something like that. And NOW, I do not trust the CPU or the memory, if I had to bet on either of them, as they may have incurred some “collateral.”

What do you think? What's the “safe” diagnosis? To scrap MB, CPU, and maybe the 8 GB memory? … and start over?

The ASUS tech rep thought it could be the board—given my facts above, given my “testing” which included changing out the power plug strip, too. It is just slightly out of warranty (3 years). BUT … testing this stuff is not easy, can be hard and/or expensive (have you every removed an Intel CPU heatsink/fan glued down to its CPU in socket?), and so on.

So what do you think – scrap the board, the CPU, and the memory and start over?

In the meantime, not wanting to use my wife's Windows 10 excellent ASUS laptop, not wanting to use my Samsung 8” tablet, and I don't own a smart phone, and not wanting to borrow the wife's, I scrapped together an old PC box (BIOS, circa 2009, using some components from 2005-2009), and here I am posting this and studying up on the complicated and long list of new ASUS mainstream boards.

[Bings]

Well according to this page https://www.techwalla.com/articles/h...-a-power-surge the first thing to look at is your power supply after your surge. I would also say that is the first thing to look at because I've always found that if a computer is misbehaving irregularly then it's generally the PSU. I've also... I mean a mate of mine once short circuited a motherboard on a case when building a pc and when that motherboard and graphic card got electricity they shouldn't they died instantly.

Also the best thing about accusing the PSU is you can buy one and if that's not the problem, you can put it back in the box for the inevitable rainy day.

[Snowhog]

Well according to this page https://www.techwalla.com/articles/h...-a-power-surge the first thing to look at is your power supply after your surge.

He already did that:

"...but I tried a known good monitor and a known good power supply, and neither effort solved the “bad startup issue.”"

[GreyGeek]

Have you tried the RAM test from the Grub menu?
If it is good then it looks like you are in for a new mobo.

Last summer I had a lightening strike take out my Eth0 port. Rfkill didn't show the port. I bought a USB Eth0 stick, which worked great. A couple months later, while using rfkill to check another device I noticed that it saw my Eth0 port but it was blocked. I unblocked it and used Ifconfig to turn it on and test it. It worked!

My suspicion is that a capacitor dialectic was punctured but slowly healed itself in a couple months. I've used it several times since then and it works great.

[woodsmoke]

QqMike
As you well know; the old woodsmoker has never claimed to be more than a hardware kinda guy and there are those who would dispute that!

You have done everything needed.

You have
a) eliminated a boot up situation
b) eliminated an "OS situation"
c) You have CORRECTLY diagnosed that it could be i) power supply ( many times correct, it can be "intermittent' in a NON-perceivable way ) ii) oh horrors...the MOBO! PLEEZE GREAT GODS OF TRON...not that!! iii) ROM iv) RAM.

The REAL problem with hardware is that it really can be "intermittent" and "I don't have a stick of ram laying around".

GG posted a great solution but...welll... it was working during the INSTANTANEOUS test...and then the cat walked on the keyboard and not so much.

a) open the case and "take a whiff"...really... MANY times BUT NOT ALL... if something has gone wrong with the MOBO one can smell... " a whiff of...metal" other than the 'reguilar metal smell"...and just btw WOODSMOKE what is the "regular smell of metal...!!!!! "...it is the solder... it really does have a unique smell...

and all of the above is USELESS unless you actually really DO SMELL..."metal" / :"hot metal" / "acid"

because what one really smells is the "rosin / resin " that is the "carrier for the metal of solder"...if you smell it then... maybe take it to a person to check out the MOBO...

if not...

then... given that is really IS a problem with the MOBO and you can't "smell it" then it "MIGHT" be the ROM or the RAM...

Both of them are "finiky" but the ROM can go south first because it is "hit first" by lightening surge and also the RAM has the ...kind of doofus but real...the literal interface of the connexion of the clips of where you put the RAM into the slots...that physical interface really can be a "last barrier" to lightening...

so as to ROM...

yes do the boot up test for ROM but it provides and INSTANTANEOUS... milliseconds...test and is VALID but only for those milliseconds... heat may get in the way just a few milliseconds later when the ROM ..."delaminates" --- spreads apart --- for a few milliseconds...

THAT is why you pay $60 USD for a "bench fee"... the guy or gal literally wanders around doing "stuff" and starting and stopping the computer to see what happens...and it can take a LONG TIME...

So...

what to do, what to do...

if you do not already have in hand sticks of RAM to trade out all of the slots...buy some...

a) first replace ALL of the sticks
If it works
you know the problem
b) if there is still a problem then it is...

ok ...NOW we get into permutations and combinations...

you have to WRITE DOWN ON PAPER...

trade out stick 01, stick 02 , stick 03...

BUT it is like the old "Christmas tree" lights "in series' from the 1930's ...which stick is causing the problem...?

so you change out "this stick" and wait to see what happens...nada or the problem, then you change out stick -02 ...samo samo...

but...what about sticks 01 AND...02...

and if you have THREE sticks!! or FOUR!!

It is get your better half to cuddle up with you on the bear rug while you go through all of the combinations of sticks...

And guess what...

after spending all that time with your better half on the bar rug and... changing the sticks produce no observable results...

it is "pin 02" on the power supply for the fan...

WELCOME to the FASCINATING...and frustrating...world of "being a hardware kinda guy"...

BUT...YOU HAVE DONE THE RIGHT THING... you have approached it logically maybe now... purchase another power supply or borrow one and purchase some RAM or borrow some...

and...get a bottle of bubbly and some nice candles from Bath and Body Works and a FAKE FUR RUG...lol

wood... hardware kinda guys...a dying breed... so ...sad...the woodsmoker is of little worth...smoke

The woodsmoker is building is ...latest...LAST... computer... a wall computer... on the wall...lights...fans...

and for the FIRST TIME... water cooled...

hardware kinda guy...sad...smoke

[Rod J]

Yes, as GreyGeek suggested run the memory test (preferably overnight). If it shows no errors then I think you could be pretty certain the memory is OK.

Some ideas I have:

Have you tried just putting the system into standby overnight and resuming the next day rather than cold booting it, would that work?

I would have thought for sure it's the PSU to blame but as you swapped it for another it seems not.

Definitely odd that the LED light on the motherboard didn't come on when you had the PSU turned on. That definitely points to a problem with the motherboard I think (as long as the PSU is good).

[Qqmike]

Rod J: Definitely odd that the LED light on the motherboard didn't come on when you had the PSU turned on. That definitely points to a problem with the motherboard I think (as long as the PSU is good).

Yes, that's concerning. And the fact that the PC will run if it is now running, but it won't start up if I shut it off--actually it starts but no video output ... suggests a video output thing on the mobo, caused by damaged circuits. And no standby LED mobo light ... mobo damage. And I do know/think that the power supply and the monitor and the power strip are all OK.

Woodsmoke: yes, you hit the nail on the head: trying all the permutations/combinations can be maddening. And it CAN be just one pin or one capacitor or one small hidden circuit.

I'm waiting to see what--if anything--the claims rep from the power company may offer (to help fix this). This is the 2nd mobo blown by these local power company failures. I don't want to get too "invasive" until talking to her I hope today. Problem with the mobo, as you guys know, everything is connected to the darned thing! No small task to remove it ...

$100 power backups might work, some say they don't always work, though, and as I understand it, they are "On" 7x24 drawing juice, charging batteries? But, three things come to mind, for the next PC: (1) Get a decent plug bar (not the $10 variety) with surge protection; (2) Buy a good power supply with built-in protection; and (3) Fact is, the ASUS boards DO have surge protection, too; and some sort of 5-way protection built in. Those efforts should prevent most power surges.

First things first, though. I'll keep you posted. Many thanks to GG, Woodsmoke, and Rod J for some very good ideas.

[jlittle]

...I think the motherboard is damaged...

Well, if it's not the power supply, that points to the motherboard. Your procedure to get the computer to go I am familiar with, you have to get things warmed up to go. If my experience is a guide, it will get worse over months before dying.

So what do you think – scrap the board, the CPU, and the memory and start over?

I suggest checking out a cheap motherboard that will support your board and memory. I don't know if such exists, but there might be one much cheaper than yours was in 2015.

Have you tried the RAM test from the Grub menu?

If the machine uses an EFI boot, there will be no memory test on the grub menu, because the freely distributed memtest86+ does not support EFI. You can download a free version from Passmark that will run from EFI, or use a legacy boot from something.

[GreyGeek]

Woodsmoke, are you going to tell us that after snuggling on the rug with your significant other you are actually thinking about that hardware test you started earlier? Really? LOL. [emoji23]

[Qqmike]

Whew ... pooped tonight. No claim to be made to the local power company because "it was an equipment failure and not their fault." Seems that someone (like a student texting, driving from the nearby high school) hit a power pole with their car, knocked it over, and knocked out power to 10,000 customers. At least that's some closure; now I can start spending my own money with no hopes of recovery from the electric company, and no arguments with them either or waiting for their phone calls. I started with a better power strip from Home Depot; although I was using a power strip w/surge protection when this happened.

Seems memtest is not easy to run with UEFI installed (16b vs 64b, or something). There is memtester, can be run in the Kubuntu environment; but tonight my reading on it (man memtester) suggests you might have to be an electrical engineer to run it.

After several attempts, I do have the broken system booted again, it is running now, I got some important documents and files copied off it to a flash drive, even though I will probably just install that HDD into a new system I will build soon.

Got the old system running (the one I quickyl pieced together and will use until I can build something better), and something has me foxed about hplip. The version in the repos is 3.14.13-Oubuntu3.4. But it won't work with my new-ish cheap-o HP printer, HP Deskjet 2132, so I got from HP a new version of hplip, 3.18.13, and the printer will now print. BUT, I can't see how to control the printer, no gui, no hplip-gui for it. BUT, on the broken system, I ran the same printer using hplip-3.16.19 and it DOES have a gui (the standard HP Device Mgr or whatever it is called). This is useful for scanning, for controlling the print queue, etc. Both PCs are running K 14.04. Why can't I get some gui in this old but working PC with the very new hplip-3.18.13? I've got to look into this maybe tomorrow if the hangover isn't too rough. Related: how to run hplip from the command line with full controls? To-do's for later.

About memtest on the broken-but-now-booted PC, thing is, I don't readily see how to run memtest, but fact is I can run the hell out of that broken PC (after I finally get it to boot up, that is), running up the CPU etc., and it does fine--none of the usual symptoms of a bad memory (frozen screens etc.).

I started to spec & source the components for a new system today, may post on that sometime in the future.

That's the update tonight. Thanks for all your comments and tips.

[woodsmoke]

GG:
lol

But back to the thread...

I used to work as a welder first as a job welder then in a production facility.
The job of note was the job shop.
The man had inherited it from his dad and the shop had been in operation for many decades.
In an "addition" he had added what seemed to be an obsessive compulsive set of..."normal 220 breakers" in SERIES before the electricity was distributed to the various stations and there was a circuit breaker at each station.

Let me REstate that in an attempt to have a visual...

incoming electricity breaker 1 breaker 2 breaker 3 breaker 4 breakder 5 breaker 6 interior wiring breaker at work station welder.

A LOT of people laughed at the man ... I mean... a LOT of people laughed at him about that but...

In an era and location where about "$8 / hour" USD was the norm he charged $16 USD per hour...yes that is SIXTEEN dollars per hour. He didn't do stuff for the walk in that wanted a cheap job UNLESS it was a person who really NEEDED the help, like a woman with a broke down car that needed a broken trailer hitch fixed or something and then he just charged what he wanted.

His reasoning was that the shop made SO MUCH money per hour that he could not afford to "wait for an electrician" to come repair a possible ...lightening strike...

And...guess what... a year into working for him we had a massive electrical storm over the town where...

this after those breakers had just hung on the wall for years, ( I kind of remember that he had said that there had been one breaker blown a few years earlier but I may be wrong on that).

get this... a strike was SO DIRECT on a house that it BLEW UP a "tube type" t.v. that blew out a hole in the wall of the house...it was a friend of ours and I went to help clean up the mess.

It looked as if the whole town was down... the mainline power was back up within a few hours...but...guess what...

The lightening had literally MELTED the first two breakers and then blew...can't remember, two or three of the rest...

He was busy wiring around the blown out breakers when I arrived and we opened for business maybe... a half hour late...

And he got the last laugh because people were LINED UP to get work done because the several other welders in town were down for the day.

So...to this very day... when the whole thing about the "power strips" / "surge stripsl" with breakers in them appeared I have always had TWO of them in series ahead of a computer and never have had a blown out computer from a power surge.

Yeah, kinda obsessive compulsive but I am working after the main line power comes back on and the other teachers are complaining about messed up computers...

or...they may be LYING to not have to do stuff for a while. lol

P.S. I actually did the uninteruptable power supply thing once...second hand... and just well...traded it off.

woodobsessiveaboutacompterthatiasssembleandmodsmok e

[jlittle]

Ah. the language that divides us...

Whew ... pooped tonight.

Now, some hours after I first read this, I imagine that you were tired.

At first I thought something more scatological, which had me confused.

[dibl]

About 4 or 5 years ago my main desktop rig started exhibiting some strange pixellation "glitches" on the display. It didn't seem to affect the functioning of anything, but obviously something was wrong in the graphics department, or so I assumed. I opened the box and reseated the graphics card, a fairly expensive Nvidia GPU for that day. That didn't help. I reseated all the memory modules and ran the memtest, and that didn't fix it. I threw up my hands, declared the Nvidia card broken, and dropped a couple hundred bucks for a new display adapter. Guess what? Boo hoo hoo. And at that point my LGA 775 Core 2 CPU was sufficiently aged that there were no high-performance motherboards available for it. So I bought an Asus Z7 motherboard, a new LGA 1156 i7 CPU, and new memory, and moved the new Nvidia GPU onto that.

The glitching problem was fixed, of course -- obviously it was a motherboard fault of some sort that induced the glitch through a working GPU. Testing confirmed the old GPU worked perfectly -- I sold it. So my money is on a fault on your motherboard, and I'm sorry to say you're the stuckee for the power problem that you didn't cause.

[Qqmike]

Whew ... pooped tonight.

Now, some hours after I first read this, I imagine that you were tired.
At first I thought something more scatological, which had me confused.

Ha! That broke my serious state this morning and added some laughter to start the day today!
As for the scatological issues, I solved those several years ago: simply go 80+% vegetarian, which converts the issue to one of readiness (being ready each morning between 5 am and 7 am).

Thanks dibl for your philosophy, too. And it IS largely a philosophical issue. Computer breaks down ... panic sets in ... oh how will I get along without it? ... more panic ... like, maybe, get a RealLife^TM? I agree with you that motherboard problems can cause weird and subtle problems.

Good idea Woodsmoke, I toyed with that, too, placing a couple surge protectors in series, easy enough to do. AND, placing a couple PCs in parallel!

The young people get by with a d&*!m&d smart phone -- no PC, no laptop, and NO Printer (you know, the kind that makes symbols on paper, like hard copy). I still haven't gotten use to fully relying on my 8" tablet, as I find it too small, not comfortable, not used to copy-and-pasting, or saving docs to be exported, and so on. Maybe I'll take small steps and try a laptop (in parallel to my regular ATX desktop ...) ...

Oh, what will today bring?!?

[Qqmike]

Nope. She just ain't gonna go.

Again, this morning:
Removed memory cards, cleaned contacts, re-set.
No go.
Reset CMOS.
No go.
Reset CMOS again, this time removing the battery while doing so.
No go.
Reset all cables, esp power and CPU 4-pin.
No go.
Futzed with I/O shield at back of case, checked for anything & obvious faults, checked with intense lighting all capacitors for visual signs of blowing, etc.
No go.

I believe I am ready to let this one go now.

Thanks for your input. Thread has many good ideas in it.