Mysterious System Hangup at POST

Status
Not open for further replies.
Archean

Archean

Posts: 5,650   +102
Yesterday night, I shut down my system (i.e. desktop pc) normally, and went to do whatever I needed to do. Also, I temporarily overclocked (just for testing purpose, as usually I do not OC) HD 5770 (GPU @ 900 Mhz, and Memory @ 1300Mhz), it worked perfectly without any issue beside these not that much of aggressive OCing, I ran couple of games without any issue for an hour I think, then restored the clocks to original settings (i.e. 850 MHz for GPU, and 1200 MHz for memory); and continued using the PC for a while without any issue at all.

The system Specs for this machine are:

1. Core 2 Quad 8400
2. Intel DP35DP Motherboard (with BIOS updated to last released version)
2. 6GB DDR2 (Kingston)
3. 500GB WD Caviar HDD + 200GB Seagate Barracuda HDD
4. HP DVD/RW+ & LG DVD Reader
5. Radeon HD 5770 1GB
6. 19" Viewsonic LCD Monitor (@1680x1050)

When I tried to turn on the system few hours later; it would turn on and a cursor would just blink on the left edge of the screen. No beeping, nothing except all three keyboard lights were on (stuck as well couldn't turn them off).

Anyway, I first disconnected HDDs + DVDs, reduced the RAM to only one module of 2GB, nothing showed up at all.

Changed the graphic card (have an old 7600GT lying around) nothing at all, and I know the card is fine as it works fine in the other machine.

I disassembled the pc, brought out the motherboard to see whether something may be shorting out, again nothing.

Checked the PSU, it works fine on the other pc with E6300 + 9600GT, similar 2 HDDs and DVDs, no issue it was fine.

The long of and the short of it, as last resort I pulled out the motherboard battery, left the motherboard alone for the night, this morning i placed the original graphic card back on after putting the battery back in, and viola, everything worked just fine.

Now, I have reassembled the pc, everything is back where it was, and everything is working just fine, infact I am writing this thread from this same machine.

My question is, what could have caused this? and how? and why there wasn't any beep?

Edit:
I just went through Event Log of the motherboard's BIOS, and it has two recent errors:

i. Intel ME Interface Non-Functional
ii. Intel ME Interface Initialization Failed

As far as I know, Management Engine is the interface between the firmware and the host. The same is backed up by Intel as Drivers and applications on the host that wish to interact with Intel ME through host interface would use the ME Interface host windows driver.

So the scope has somewhat narrowed, and the thing is what caused it to become 'Non-functional' & its 'Initialization Failed', and how it ended up fully functional again, considering the scenario I described above.
 
Cosmic Alignment.......

Well, I think the BIOS reset is what did the trick, and not any of the other diagnostic attempts. Red says that sometime s*** happens, and then your BIOS resets, for better or worse. (loose quote).
I turned on my E7300 machine one day, and while watching the crawl, I'm told that the CPU is running @ 2.2 Ghz. I was of course dismayed, and reset the CPU multiplier to 10.0, from 8.5. How it got to be 8.5 remains a mystery.

That's the what of it, but my lack of in depth tech knowledge prevents me from saying exactly why.

One time I read an article saying that "cosmic rays" might affect computer operations as the gates became progressively smaller.

I read a lot of crap about Y2K being the end of the world also, but as we all know, that didn't pan out.

I also tried to install a copy of Ubuntu one time, goofed up taking the disc out, rebooted at the wrong time or something, and I got an error message saying, "the BIOS is being reloaded from backup". (Gigabyte dual BIOS board). Wow, dodged a bullet there too.

In any event, it seems an OS can alter , damage, or modify the BIOS HAL firmware. (note 1)

Note 1; Since "alter" and "modify" are close to synonymous in this context, the sentence even has a "built in backup". But at no point am I claiming that this makes it more informative.
 
Fair enough; although with all due probabilities ...... may be one in a million at that, it happened. Anyway, Thanks for the insight Captain :)
 
Well by stroke of luck or perhaps the other way round, it happened again about an hour ago.

Same symptoms, all three keyboard lights on / stuck ......as nothing works, no beeps etc.

However, I believe I know what is causing this though, and that is, my 16GB Corsair USB Flashdisk was attached and I rebooted the system, and I think probably last time around similar thing may have happened. However, I am unable to get to the reason why this would cause these hangings.

Right now, I have pulled out the battery and cut off the power to system, and will check in few hours if it returns to normal behavior and boot. As right now, I am out of ideas about getting around this issue except waiting.
 
I think the issue is resolved for good (hopefully ;)).

After spending couple of hours trying to figure out what may be the cause and disassembling the whole system again, it turned out that the casing's front panel USB connector's actually was the issue, as soon as i disconnected them every thing worked out as it should be. Its a minor functionality loss, as I rarely used them any way. Regards

Edit:

Well unfortunately I was wrong, and it isn't solved.

Anyway, the latest is, I've contacted Intel about the issue, and they said the issue is with PSU, so I've replaced it and bought another Cooler Master 500W PSU (replacing the older one), but it didn't helped either.

So basically, I am almost out of ideas.

By the way the system has booted again after a dozen tries, and is working without any issue atm. But as soon as I shut it down, usually it returns to above detailed state.

Also, I just started running unigine heaven benchmark with all settings as high/extreme; and so far the system is stable and working totally stable under load as usual.
 
I would definitely check your VGA BIOS version against known hang issues.
Having said that, it could also be insufficient power delivery through the PCIe x16 slot (version 1.0 ?), since you can run a 9600GT OK - a definite possibility if it's the 59watt version rather than the 96watt.
From a troubleshooting perspective, you can probably rule out the PSU if both the original and it's replacement is in the 40A (12v) range. A PCIe 6 pin only provides 75W/6.25A which is not a big power draw.
Next stop would be the board- if the HD 5770, PSU and harddrives all test good through diagnostics or by swapping into a working system. Intel boards are usually no-frills, over-engineered parts that are durable if not overly exciting. I think the DP35DP is an old-school electrolytic capacitor affair- but if no bulging/leaking caps are evident then thats about as far as I think you can go. Any PCB trace damage for instance is going to be minute if the problem is occasional/intermittant.
 
Thanks for the reply DBZ, I really appreciate that.

Alright now to the points you raised:

1. If VGA bios would have been issue, the system would display when I place my older 7600GT in the system which didn't need any external power connector, and also works alright in the other PC, infact that card was in this same system few years ago.

2. I have been unable to confirm whether the slot is PCI-E 2.x, however, if I am wrong please correct me, I believe 9600GT draws more power than 5770.

3. Well according to documentation of the PSU it deliver 18Ax2 (on two different 12V rails), that is 36A in total (I am not so good at the electrical side of stuff, so again please correct me if I am wrong). And 9600GT needed something like 26A on 12V, I think.

4. There isn't any leaking / bulging caps visible on the motherboard.

5. I haven't tested HDD's but I don't think the issue is with them, as since this issue has started I tried to boot system countless times with nothing connected at all, not even keyboard; and the problem still persists. Also I've disconnected both optical drivers, and result is the same.

6. I have tried swapping the memory and took at perfectly working DDR2 800 sticks from the other PC, and it doesn't help either.

I am now tilting more towards something is up with the motherboard, some short etc. which I am unable to find at the moment.

Edit:
I just remembered a peculiar piece of info, i.e. according to Intel's IDU (which provide all sorts of info about temperatures as well) tells that CPU temperature is 26C, now that seems bit odd to me, as I remember Core 0, 1 and 3 would hover around in early 30s, and Core 2 in late 30s or early 40s.

One more thing I'd add to this is, usually the GPU temperature hover around late 40s to 52C, however, If i jack up the fan's stock speed from 35% to say 60% the temperature immediately drops to 42-43C. During full load I haven't checked the temps but as system never crashed, no untoward incidents etc. I don't think that is an issue. As its early summer now the ambient temperature is around mid 20s (C) so I don't see any real untoward reading in this as well.
 
Sorry for double posting, but I don't want this to be lost in a bigger post.

Do you guys think if somehow usb port(s) is/are shorted the system can stuck on POST with all three keyboard lights on?
 
I just remembered a peculiar piece of info, i.e. according to Intel's IDU (which provide all sorts of info about temperatures as well) tells that CPU temperature is 26C, now that seems bit odd to me, as I remember Core 0, 1 and 3 would hover around in early 30s, and Core 2 in late 30s or early 40s.
Click to expand...

I think that the Core temperatures are fine. That should not be an issue.
 
Yes, I think that as well, I will try to get I/O Hub and Memory Controller HUB temps if and when I can get the system boot again.

Usually, I have to disassemble the whole system, let it sit there for a while, then reassemble and after putting in bit more elbow grease it would agree to boot normally.

Also I mentioned Event Logs of motherboard with regard to ME in 1st post, no such error has been recorded later on, that I checked last night.

Rit: what you think of my last line of thinking shorted out usb port(s)?
 
So, if I understand your post correctly- the Intel board has POST problems with other cards (7600GT / 9600GT ) ? If that's the case then it looks as though the board is the culprit- whether you can track down the cause might be problematic at best. I presume this a "cold boot" problem - i.e. when it happens it will invariably be the first time it's started each day, and not if the system is powered down then restarted a short time later ?
If this is the case then the problem could be a dry (or cracked) solder joint. A burnt out or broken PCB trace should be able to be viewed with reasonably keen eyesight, but I would think this is unlikely unless the board has shorted or been stressed/warped.

I don't think the GPU is at fault but just for general info; the HD5770 max power draw is slightly higher than the 9600GT ( 108w versus 95w- assuming the 9600GT has a 6pin PCIe connector, if not then the 9600GT is 59w)
 
I think you are probably right about board; I have checked it with third power supply, tried it with old 9600GT which was on this board few weeks ago, and same issue nothing at all.

I think your point about cracked/dry solder joint make sense, as Memory I/O & Controller etc. ran very hot on this motherboard from day one, something like 60-90C range (according to Intel local rep. these were not alarming temperatures, beside it ran for years perhaps I guess reaching the limit of soldering's tolerance).

Anyway I've disassembled the motherboard again, and I am going to send it along with processor & GPU to the vendor I bought it from, for warranty. I am sending Graphic card as well because it wasn't giving display on the other machine, although I think it probably is fine.

Just out of curiosity, although 5700 series relatively cooler, this 5770's backside of PCB felt really hot ..... is that normal? (for reference I've explained its temperatures etc. in previous post).
 
Yes, I have disconnected front mounted USBs, and also have half a dozen on these on the backside panel. Also I just noticed 5770 is causing the system to hang on another system board, I am not sure, perhaps the motherboard has also fried it or something?

Edit:
@DBZ & Rit
I have also confirmed at last that the PCI Express x16 on this board is indeed an 2.0 slot.

Anyway guys, Thanks alot for your precious time, and I will update whatever happens with regard to replacements/warranties :)

Just incase there are issues with availability of same board as replacement, which one you guys would suggest I should get (I'm saying this because obviously this a 3 years or so old motherboard, and not sure they will have any more in stock)
 
Just out of curiosity, although 5700 series relatively cooler, this 5770's backside of PCB felt really hot ..... is that normal? (for reference I've explained its temperatures etc. in previous post).
Click to expand...

The temperature of the card will depend on the ambient temperature in the room as well. In summer my HD 4890 hovers around 52-55C idle without the AC on (@ 55% fan speed).

Since we can't exactly determine what the issue is you did the right thing by sending it back to the vendor. It's sometimes the only way of solving such unpredictable issues. :)

Hope all is well once you get your stuff replaced.
 
Yes indeed, and also do you have any suggestion on the replacement motherboard? As I think given the time span this board is at the end of its warranty period (fortunately for me it hasn't yet expired ;)); so I am pretty sure I won't get the same board, they will offer probably something else.
 
No OCing, higher end-midrange motherboard should do it; as right now I had no plans on upgrading to i7, infact I planned on skipping this generation, and wait for sandy bridge to showup.

Just for clarification, I am setting no budget as I pointed out earlier due EOL for this product, most likely I would be offered some different option as replacement (probably from Intel's line as well), however, I may be able to do a bit of bargaining, so I am in no position to set budget (being purchasing from this same vendor since what 1992 i think; and been in similar position once, hmm in the late last century ;)) though, this make me feel bit older.
 
I would suggest something like the ASUS P5Q Pro. It's a pretty solid board for about $130.

EDIT: Here it's available for $104.99 after a $20 mail-in rebate.
 
I don't know if it's still available (EPA and all that), but there used to be cans of refrigerant available at electronics stores, that you could use to diagnose heat related problems. You know, give suspected ICs a little squirt, when the big chill hit, it was better than Santa Claus for figuring out "who's been naughty, and who's been nice".
 
Thankyou Rit.

And Captain, I haven't seen those in a while either, but as I said after spending almost a week trying to figure out what is wrong; I am just sending it back for warranty. But thankyou nonetheless for your valuable input.

I'll let you guys know what happens next with regard to progress on this issue, Best Regards to everyone.
 
Archean said:
....as Memory I/O & Controller etc. ran very hot on this motherboard from day one, something like 60-90C range (according to Intel local rep. these were not alarming temperatures, beside it ran for years perhaps I guess reaching the limit of soldering's tolerance)
Click to expand...

Using the vernacular of the day....WTF!
60-90C for a P35 northbridge is insane. The lower-mid end of that scale is usually reserved for nVidia's SPP's (northbridge/chipset). The hottest boards I have ever owned were both 680i SLI mobo's (EVGA and Asus) and both topped out at around 75C (and that's with overclocked/overvolted RAM) -active overboard cooling is a must at that temp. A P35 chipset working hard with overvolted RAM sitting above a hot GPU would likely max out at 50-55C, maybe 60C under a full load stress test. If the boards system sensor is not out of whack then something is amiss.
Technically the Intel guy is right- most Intel chips are rated 100-110C (P35 is 106C - beware link is a pdf)...but still
 
That link gives error when I try to open, but your words are sufficient for me. I really appreciate your time and help DBZ.
 
Well guys as it turned out the store from where I purchase my IT stuff has told me that Intel do not have any similar motherboard to replace my damaged one (only have DG43NB in stock, which is inferior in some specs, except chipset I guess). Hence, came the offer to pay the difference and get Asus P5Q Deluxe motherboard, and I've gone for it, it should be here today sometime,then the painful process of reinstalling everything ........ again.
 
Status
Not open for further replies.

Similar threads

M
Looking for advice on new PC build
Replies
4
Views
660
mbk34
M
S
Dell Latitude E5470 Video Issues
Replies
1
Views
695
Mark Fuller
M
bruno167573
What should i do?
Replies
3
Views
97
bruno167573
bruno167573

Latest posts

Back