Memtest86+ is always right? So it must be a software problem

RayLopez99 wrote:

Update. I think you're right Paul, it was something in the BIOS.
But I'll never know since now I cannot even get to POST in the desktop
in question (I am using my laptop to write this message). Very long
story somewhat short. After screwing around in the BIOS, seeing if
I accidentally overclocked (since another program SiSoft, indicated
perhaps there was a BIOS mistake/overclock), I exited BIOS without
saving, but then I got what is notorious for this mobo, an Asus mobo
(P8H67), the dreaded "USB device over current status detected will shut
down in 15 seconds" message.

Long story a bit shorter: I bet this ASUS mobo is crappy. I tried
resetting the CMOS via jumper (drains the 'battery' charge for the BIOS),
tried a PS/2 keyboard, new mouse, unplugging the front USB panel, checking
that my mobo is not touching the metal case, and still no POST. This
particular company ASUS is popular with overclockers, and perhaps I set
(or it was set) to overclock incorrectly (the system always was a bit
weird, even when I bought it at a Thai shop, so it could be the tech who
built it had it overclocked) and perhaps that and in combination with a
short somewhere, the uP will not boot up (apparently Intel has a switch
that will shut off the chip if there's any short, so you won't even get
into POST). I will take the entire desktop to the local PC shop and try
another motherboard.

Question: I assume that you can switch out a mobo, connect all the cables,
and voila! your system is back to normal, right? I don't see why not...
any dissents?

I'll post here later, probably in a few days, when the repairs are made.

RL

More info from the net:

USB device over current status detected will shut down in 15 seconds

http://www.tomshardware.com/forum/29...m-shut-seconds

What system do you have and how old is it? Sometimes that is actually
caused by the USB device, but sometimes it is caused by a motherboard issue.

http://www.tomshardware.com/forum/31...tatus-detected

I have newly purchased (1 month) a computer with AMD FX 8120 CPU &
VIP-500W smps, Odyssey Cabinet, Sata 500GB. Recently I face usb over
current issue to ASUS M5A88-M MB.

I read the thread & other posts. Many a times when I boot pc after 6-7 hours
it starts well. But for another boot it shows same prob. Also I had removed
all cables and put it on paper and then connected.. I also changed the
jumper (for 2-3 seconds to clear cmos).

It only shows the USB 3.0 connected ports. But suddenly it started again
with the same prob. I also updated the bios through usb ez flash but no use..

http://img.tomshardware.com/forum/uk/icones/smilies/non...

I'm really fed up with this. Instead of this my other 2.4 GZ intel PC
with chinese mb is going OK..

Now also I'll need to pay Rs.175/- to the shop from where I purchased pc
as traveling charges. In future I'll not prefer Asus.. Nobody should have
such trouble.. I don't know after repairing for how many times I'll face
same problem...

http://www.whoho.com/2011/03/usb-dev...-detected.html

Causes:

Screws used to mount the motherboard are too tight
Rogue mounting bolts attached to your case are touching the underside
of your motherboard
Some kind of conductive material is under or on your motherboard
causing short-circuits
Wrong installation of front-panel cables

http://www.tomshardware.com/forum/id...-detected.html

Hey everyone I have never posted here because I have not needed to yet.
I have been working on computers for about 10 years so I have a lot of
knowledge. Today I received a computer in that wouldn't boot.. wouldn't
even post. I tore the computer down to just the motherboard, 24 pin power
and 4 pin (cpu) power and the monitor into the onboard graphics. I got the
computer to post but it showed the error below. I have looked through
tons of forums to find that it's as I expected usually caused by the front
usb ports plugged into the motherboard or a dirty or bent usb port. I
cleaned out the usb ports on the board and as far as I can tell there is
no debris or problems with them. All that's plugged into the board is power
and one piece of ram and I am getting this error. I have tried everything
I can think of can anyone assist? The motherboard is running outside of
the case right now but gives the same error screwed down to the case or out
of the case. This motherboard has a jumper to clear CMOS I tried that. I
even took out the battery to try reseting that way. This error is given with
or without the keyboard plugged in.

ASUS P7H55-M/CG5275/DP_MB

auto-detecting usb mass storage devices
00 usb mass storage devices found
no keyboard
usb device over current status detected
system will reboot.

http://www.tomshardware.com/answers/...-detected.html

http://www.tomshardware.com/answers/...-detected.html

Yeah i tried a ps/2 keyboard and yes i get the error without anything
connected to it. Im guessing it is shorting out somewhere? By the looks of
it the backplate prongs are not touching anything.

he shouldn't need more standoffs..... You should be able to turn that board on,
outside of a case(no standoffs). A true "short" would most likely prevent
any boot-up/POST. Maybe you try this next.... I know it is a real pain the
@$$ but if you still get the error, you have an issue with your main board
or PSU. I believe most PSU over current protections prevent bootup/POST also.
The main-board's power management utility may be malfunctioning... or

working as intended. were you able to reset the BIOS?

skit75 said:
http://mediapool.getthespec.com/media.pdf?m=TZ3yZfARfgG...

There is a pdf of the manual. Go to section 1.9 and do the steps to clear
the RTC RAM clock & memory and settings. Any luck or have you tried that already?

I tried that already, still nothing. I went ahead and bought a gigabyte mobo.

1337rofl
June 30, 2013 7:11:05 AM

P8H67-M LE - VIP forums are one source of model-specific info.
Newegg or Amazon reviews is another source.

http://vip.asus.com/forum/topic.aspx...age=1&count=67

There are USB Overcurrent threads. As Ulrich notes, there is likely
to be an intermittent short somewhere in the install. For the OP in
this thread, bench testing the motherboard (removal from computer case)
was clean. Problems showed up when it was put back in the computer case.
If it was a BIOS bug, taking the motherboard out of the case would have made
no difference.

http://vip.asus.com/forum/view.aspx?...Language=en-us

This followup thread, says the problem was a screw head. So the motherboard
probably wasn't centered over the holes (standoff holes).

http://vip.asus.com/forum/view.aspx?...Language=en-us

When you install a motherboard, the motherboard has to be moved a fraction of
an inch, up or down, so that cards will insert easily into
the PCI Express slots. But if you're a sloppy installer,
and push the motherboard all the way down ("South"), the screw in the
standoff hole, can contact an adjacent component or copper
track.

It's up to the design engineer, to ensure sufficient clearance
from the nine mounting holes, to any other conductor. It's
up to the shop tech, to not only align the motherboard to
the computer case, but also reject any mobo/case combo that
causes shorting. If you use a super-cheap case (soup can metal)
with poor tolerances, you can't align things properly. Use
another case. Even though cases are cheap, sometimes
a big factory has enough common sense to tolerance them well
(cut the materials accurately).

The nine mounting holes on a full sized motherboard, they're
special in that each "ring" of metal is grounded. In other
words, they're *designed* to "short to the case". The motherboard
designer wanted the metal standoff to touch the ring, and
wanted the screw head to be at the same potential as the
motherboard metal ring. It's all to be grounded. This is
part of emissions control or something.

If a designer does not want that to happen, a ring can be
made to "float" electrically. But motherboard designs
don't do that.

I've done circuit boards, with holes that have no PTH on
them, and that's intended for mechanical mounting points
with no electrical properties. The fiberglass board is
an insulator. I've also had circuits, where the metal
standoff posts we were using, carried the current that
made the board work :-) The standoff posts were electrified.
So engineers can do all sorts of stuff - but on motherboards,
the practice is pretty consistent - grounded rings for the
6 or 9 mounting holes.

In cases where a user inserts an insulator under the ring, to
separate the standoff, I can't tell you why that fixed something.
On a broken motherboard, a tiny flexure of the substrate,
can "make or break" a snapped solder joint, and it could be
just about anything at that point. I've had circuit boards
with as many as 300 broken conductors inside, and poorly
made circuit boards can have a ton of internal bad connections.
You could bend that board all you wanted, and those
300 broken connections weren't all coming back. We stopped
buying PCBs from that supplier. They were blacklisted.

Paul

On Thu, 12 Dec 2013 10:47:42 -0800 (PST), RayLopez99
wrote:

Hey
everyone I have never posted here because I have not needed to yet. I
have been working on computers for about 10 years so I have a lot of
knowledge. Today I received a computer in that wouldn't boot..
wouldn't even post.

-
I wouldn't care even to get into a portion of the hardware
complexities you've pursued without avail with that ASUS MB.

ASUS, however and of course isn't inherently all bad. I've run with a
million ASUS mbs. Sure, they're a bigger presence, nowadays can
probably also command a more pricier product offering;- There's been
QC rumors, however founded or not. All else being equal, any MB, as I
suspect, is not impervious or faultless state.

Hence, I'm always careful of that contingency: the what-if scenario
that requires the MB to be sent back, any dedicated accessories
components also purchased, for a refund or exchange. In fact, one of
my first angles I apprise when considering a system purchase.

Explaining my abilities to diagnose, test and burn-in a system product
purchase to the vendor, least of all concerns me beyond a reasonable
attempt to contact and explain to their representatives product
shortcomings I've encountered beyond a normal scope of expectancy.

And, yes, the normal result may well be an onus they tend not to
acknowledge, an understatement reflected to the purchaser ineptitude
and ignorance;- a delicate position, politically, at times to be in,
as with some vendors there is no less a valid need to be, a priori, in
business, in fact doing business with satisfied customers, both
parties then be willing and able to continue with some compromises
when necessary. Slam dunking them with a well-chosen credit card's
staff, quickly and efficiently able to represent the customer's better
interests, such as I have, may not wholly be in my best interests when
working out a suitable contingency a reputable computer dealer, I
without doubt expect to counter me and proffer by way of a solution.

Something I once heard at a computer club...'if you're not a part of
the solution, then you're just another problem.'

On Friday, December 13, 2013 9:50:31 PM UTC+8, Flasherly wrote:


I wouldn't care even to get into a portion of the hardware

complexities you've pursued without avail with that ASUS MB.



ASUS, however and of course isn't inherently all bad.

Flasherly, Paul:

You won't believe this, but I am 'tickled pink' that I took Paul's advice to reseat the memory--and now things are back to normal! So after all of this it was indeed the memory sticks somehow not being seated correctly--amazing!

What is interesting is that normally I am very tender with handling memory sticks. I make sure my feet are not too dry, that I don't have static electricity (sometimes I even wet my feet a tiny bit), I touch the refrigerator (a kind of ground, though in this country in southeast Asia they don't use a true ground), then I very gingerly seat the memory sticks by gently snapping them into place with the supporting clips. This time however, I figured since the mobo was ruined anyway, that I would pull out the sticks completely, then, holding my hand in back of the mobo so not to bend it too much, roughly 'snapped' the memory sticks in place by pushing them firmly into the slot, and the white handles on each side of the memory stick sockets audibly 'snapped' into place without me even touching them--a departure from my normal gentle protocol. I figured since the mobo was ruined, what's the harm? I was actually inspired by Paul saying 'by the time you read this you'll already have a new board' and figuring why not prove Paul wrong by trying to test the board one last time? I did not expect even a 1% chance of succeeding, since to me it looked like the memory sticks were firmly seated.. But low and behold--I got a POST. Then, since ACHI mode for the SATA drives was not set (since I had drained the CMOS battery by shorting the jumpers earlier) Windows would not boot, but I remembered it was because IDE was set, so I changed it to ACHI and was able to log into Windows upon reboot.. Problem solved! I saved about a half day to a day of work (not to mention having to mess with thermal grease, set off screws for the mobo, cables, etc etc etc and about $65 for a new ASUS or Mega mb).

Now I intend to test using Prime 95...hold on while I do this... and double check that my memory and uP are not double clocked (I wish ASUS had a setting for doing this easily, but it appears to be buried somewhere)... as I type this... all is well with Prime95! All four cores are working fine... "Lucas-Lehmer iterations"...Pass1, Pass2.....Test 1- 14 + all OK, excellent!

You guys, especially Paul, saved me money. And not in 10 years would I have guessed it was the memory, since the sticks looked so firmly seated.

BTW a curious artifact: I had reset the CMOS by draining the charge, shorting the jumper, and waiting at least 10 seconds, and indeed initially my BIOS had all default values, but, strangely, the system time and date when I got into BIOS were in fact accurate! I guess that BIOS has some sort of SRAM that holds the current date/time without the need of an external charge, and/or the 10+ seconds I had 'shorted' the CMOS circuitry was not enough to affect the date/time field for some reason.

Thank you both again!

RL

RayLopez99 wrote:
On Friday, December 13, 2013 9:50:31 PM UTC+8, Flasherly wrote:


I wouldn't care even to get into a portion of the hardware

complexities you've pursued without avail with that ASUS MB.



ASUS, however and of course isn't inherently all bad.

Flasherly, Paul:

You won't believe this, but I am 'tickled pink' that I took Paul's advice to reseat the memory--and now things are back to normal! So after all of this it was indeed the memory sticks somehow not being seated correctly--amazing!

What is interesting is that normally I am very tender with handling memory sticks. I make sure my feet are not too dry, that I don't have static electricity (sometimes I even wet my feet a tiny bit), I touch the refrigerator (a kind of ground, though in this country in southeast Asia they don't use a true ground), then I very gingerly seat the memory sticks by gently snapping them into place with the supporting clips. This time however, I figured since the mobo was ruined anyway, that I would pull out the sticks completely, then, holding my hand in back of the mobo so not to bend it too much, roughly 'snapped' the memory sticks in place by pushing them firmly into the slot, and the white handles on each side of the memory stick sockets audibly 'snapped' into place without me even touching them--a departure from my normal gentle protocol. I figured since the mobo was ruined, what's the harm? I was actually inspired by Paul saying 'by the time you read this you'll already
have a new board' and figuring why not prove Paul wrong by trying to test the board one last time? I did not expect even a 1% chance of succeeding, since to me it looked like the memory sticks were firmly seated. But low and behold--I got a POST. Then, since ACHI mode for the SATA drives was not set (since I had drained the CMOS battery by shorting the jumpers earlier) Windows would not boot, but I remembered it was because IDE was set, so I changed it to ACHI and was able to log into Windows upon reboot. Problem solved! I saved about a half day to a day of work (not to mention having to mess with thermal grease, set off screws for the mobo, cables, etc etc etc and about $65 for a new ASUS or Mega mb).

Now I intend to test using Prime 95...hold on while I do this... and double check that my memory and uP are not double clocked (I wish ASUS had a setting for doing this easily, but it appears to be buried somewhere)... as I type this... all is well with Prime95! All four cores are working fine... "Lucas-Lehmer iterations"...Pass1, Pass2.....Test 1- 14 + all OK, excellent!

You guys, especially Paul, saved me money. And not in 10 years would I have guessed it was the memory, since the sticks looked so firmly seated.

BTW a curious artifact: I had reset the CMOS by draining the charge, shorting the jumper, and waiting at least 10 seconds, and indeed initially my BIOS had all default values, but, strangely, the system time and date when I got into BIOS were in fact accurate! I guess that BIOS has some sort of SRAM that holds the current date/time without the need of an external charge, and/or the 10+ seconds I had 'shorted' the CMOS circuitry was not enough to affect the date/time field for some reason.

Thank you both again!

RL

If it was a problem like that, it's a wonder
your computer could run at all (with memory
not seated fully or making contact).

You'd think the computer would crash during POST.

*******

If the date and time are correct, after clearing
the CMOS, then immediately entering the BIOS to
verify the time, then it probably did not get reset.

If you boot the computer, have an Internet connection,
and NTP polls an Internet server, then the time can
be corrected for you. But if you're careful
to enter the BIOS first, right after clearing CMOS, then
the time should return to an earlier time.

The "lowest" value of time, varies from generation to
generation of computer. I have computers here that
default to "1970", as well as computers that default
to a later date, in the 2000+ range. But the
time should be sufficiently different, for you
to notice. The CMOS well in the Southbridge, the
CMOS RAM and the RTC (real time clock), should
be sharing the same 3VSB-like power source. So
if one gets cleared, so should the other.

Some of the Intel chipsets (the modern ones like
yours), there is actually a RESET pin for CMOS,
so it's not actually a "draining" action. But
not all the manufacturers tap into that pin,
instead choosing to use the old draining method.
The nice thing about the RESET pin just for the
CMOS well, is the action is instantaneous. The
draining thing, users may notice a time dependence.

Always unplug the computer, before clearing CMOS,
unless instructed to do otherwise by the motherboard
manual. I think some Intel motherboards, their
instructions can be slightly different.

I've had instances, where I cleared CMOS, and the
time did not reset. And I then had to do it again.
Annoying.

Paul

On Saturday, December 14, 2013 2:47:29 AM UTC+8, RayLopez99 wrote:

I took some time to go through the settings, and discovered Turboboost is not the same as Overclocking, so I kept it on. Likewise for Speedstep.

I will run Prime95 for longer but so far no errors.

What is interesting is that Memtest86+, as I indicated in my original post, was run for over 8 hours and did not detect any bad memory, despite the sticks at the time not seated correctly. So to a degree Memtest86+ gives a false negative reading and a false sense of security.

RL

On Saturday, December 14, 2013 3:46:57 AM UTC+8, Paul wrote:
If it was a problem like that, it's a wonder

your computer could run at all (with memory

not seated fully or making contact).

Indeed. And how did the problem get worse over time? That is, for a few weeks I was getting an occasional BSOD, then more frequent, then I would get the "USB device over current status detected will shut down in 15 seconds" error? the only thing I can think of is that the memory stick somehow was gradually getting loose, getting unseated, over time.




You'd think the computer would crash during POST.

Right.


If the date and time are correct, after clearing

the CMOS, then immediately entering the BIOS to

verify the time, then it probably did not get reset.



Possible, but some other settings were reset, such as ACHI vs IDE for the hard drive.


If you boot the computer, have an Internet connection,

and NTP polls an Internet server, then the time can

be corrected for you.

Well maybe this is the key, since I did connect to internet--if done through BIOS--since the internet Ethernet cable was plugged in, even though the hard drive was inaccessible the first boot (since IDE was set, not ACHI). So if it's possible for BIOS to access the internet, then indeed the clock was reset from the internet. But I think only the OS can access the internet and then it updates the BIOS maybe.



Some of the Intel chipsets (the modern ones like

yours), there is actually a RESET pin for CMOS,

so it's not actually a "draining" action. But

not all the manufacturers tap into that pin,

instead choosing to use the old draining method.

Yes maybe the reset only affects parts of the BIOS, but why is unclear.




I've had instances, where I cleared CMOS, and the

time did not reset. And I then had to do it again.

Annoying.

Yes, that is what happened to me I believe.

All's well that ends well...

RL

On Fri, 13 Dec 2013 10:47:29 -0800 (PST), RayLopez99
wrote:

What
is interesting is that normally I am very tender with handling memory
sticks. I make sure my feet are not too dry, that I don't have static
electricity (sometimes I even wet my feet a tiny bit), I touch the
refrigerator (a kind of ground, though in this country in southeast
Asia they don't use a true ground), then I very gingerly seat the
memory sticks by gently snapping them into place with the supporting
clips. This time however, I figured since the mobo was ruined anyway,
that I would pull out the sticks completely, then, holding my hand in
back of the mobo so not to bend it too much, roughly 'snapped' the
memory sticks in place by pushing them firmly into the slot, and the
white handles on each side of the memory stick sockets audibly
'snapped' into place without me even touching them--a departure from
my normal gentle protocol.

-
Uh-huh. . .

Poorest excuse I've ever heard (just kidding, also been there,
somewhere, I'm sure). Do me a favor, though, ease up on that feet
stuff while the refrigerator is running. I learned "rough-and-tumble"
electronics, computer assembly, from an old television repairman apt
to laugh at "getting bit" by CRT flyback voltages (while jumpering off
leads to a PCB solely as if by intuition for a fixed product). Sure,
there's something of a touch (tolerances - plastics, delicate screws,
all that crap) to handling an assembly, but, damn sam, it's for sure
as hell need not be that delicate or fragile with proper observance
and care . A proper memory seating is simply a positive expectation
and knowledge that it's done right, even after, at times, a little
requisite finagling, filing - whatever it takes to make it fly - to
fitting the pieces fitted correctly together;- sort of like a word
processoror and programming: do either long enough, and they all start
merging into one big cludge that looks the same.

Hm...surprised you got a refrigerator. I didn't when I lived two
years in Thailand;- just wasn't a "native" sort of thing (roundabount
Bankok and Udapah), as I prefered it, adapting close to the fringe
with the green, green jungle nearby. Miss those gorgeous reds in
sunsets, (even when the night heat remains for a couple hours exactly
as if the sun were beating down on one's back like a demon from hell.
Heh - I worked in "open areas," often on concrete, so it was
definately hot. So hot that not everybody could do it... just nowhere
near what those tough old Thais were capable, though. Food, too,
maybe near half-potency strength for bellying-up on the hot-assed
peppers in just about everything they eat.)

On Saturday, December 14, 2013 6:54:37 AM UTC+8, Flasherly
Uh-huh. . .



Poorest excuse I've ever heard (just kidding, also been I learned "rough-and-tumble"

electronics, computer assembly, from an old television repairman apt

to laugh at "getting bit" by CRT flyback voltages (while jumpering off
Hm...surprised you got a refrigerator. I didn't when I lived two

years in Thailand;- just wasn't a "native" sort of thing (roundabount

Bankok and Udapah), as I prefered it, adapting close to the fringe

Yes, that makes sense. A master repairman is much stronger than an amateur like me for stuff like repair.

As for hot, it's very true what you say. I'm even closer to the equator than Thailand now, and it's always something like 90 degrees except when it's raining, and the sun always rises and sets at 5:30 to 6:30 AM/PM year round. That's why the people are brown. And in the tropics they'd be black. If you are white it's a death sentence, literally, as albino skin animals for example (including the fabled albino elephants sometimes fortuitously found for the King of Thailand) end up getting skin cancer and dying. Revenge is taken up north however, since any of us brown skinned or black skinned people would develop rickets from lack of sunshine Vitamin D and expire.

RL

On Fri, 13 Dec 2013 19:21:47 -0800 (PST), RayLopez99
wrote:

As
for hot, it's very true what you say. I'm even closer to the equator
than Thailand now, and it's always something like 90 degrees except
when it's raining, and the sun always rises and sets at 5:30 to 6:30
AM/PM year round. That's why the people are brown. And in the tropics
they'd be black. If you are white it's a death sentence, literally,
as albino skin animals for example (including the fabled albino
elephants sometimes fortuitously found for the King of Thailand) end
up getting skin cancer and dying. Revenge is taken up north however,
since any of us brown skinned or black skinned people would develop
rickets from lack of sunshine Vitamin D and expire.

-
Watching a televised Nature series, subject: Why are fair-skinned
people considered the most beautiful among all Northern people...

Something some time ago like that, anyway, but funny how that works,
the blue-eyed, blond-haired and fair-skinned complexion, as for a
brand Westerners most covet among statistical measures typical of
beauty. I never really broke through any Thai cultural barriers,
though;- enough to ask for food, get off a bus, in a general but
limited and inarticulate way of getting around when buying sapphires,
which I liked to do for such times as I had idle;. . .Leisurely,
always, knowing exactly how that all-day 90-degs feels (at 100 percent
humidity), when never appearing rushed, slow and easy when dealing
with Thais on their terms;- but of course patience does tend to go a
long way for a foreigner in any country.