Trouble booting up at first...

November 25, 2010 at 08:39:58
Specs: Windows 7 (x64), AMD Phenom II X4/4GB DDR3
Hello, I'm having trouble booting up my computer, and I'm rather confused.

A little history: I had RMA'd my RAM twice before, and just recently, I had RMA'd my hard drive.

Now, I made a clean install of Windows 7, and noticed the errors at boot up rather quickly.
This is usually what happens: I boot up, and it BSOD's just as it starts windows. Then goes to Startup Repair on the next try. (Although sometimes, that BSOD's while even that loads)

After Startup Repair tells me it can't fix it, I restart, and it boots up fine. My attempts to recreate the error is difficult because restarting it after I get to this point goes without a hitch.

In fact, the only time I get these errors are when I've shut the computer down for an extended period (usually a number of hours).

BSOD errors (that I catch) range from 0xc0000005, 0x00000050, 0x0000001e, and here's a link to the few minidumps that were saved: http://dl.dropbox.com/u/2544832/Min...

I have tried resetting the BIOS, I have tried reinstalling drivers (probably not all of them), I have used SeaTools to check my hard drive, and that passes, and I've used Memtest86+ and that passes. This never happened on my old install of Windows 7.

So I'm stumped.

I'd like to avoid reinstalling Windows, but if I have to, I'll do it....


See More: Trouble booting up at first...

Report •


#1
November 25, 2010 at 08:46:18
Is your computer a custom built or pre-assembled. All the issues you describe could be attributed to a crappy power supply. What are the specs of your computer, including the power supply.

Also, if your RAM is not matched the RAM timings and voltages may not be correct for all modules.


Report •

#2
November 25, 2010 at 08:59:02
this is a custom built computer, more or less on a budget. RAM is indeed matched.

- CPU
AMD Phenom II X4 955

- Motherboard
ECS Elitegroup IC780M-A2
Chipsets: AMD RX780 NB, SB710 SB

- RAM
OCZ DDR3 PC3-10666 Platinum Low Voltage (2GB x2)
Timings/Voltage: 7-7-7-20, 1.60V

- PSU
APEVIA WIN-500XSPX 500W


Report •

#3
November 25, 2010 at 09:20:24
I guess I should also note the hard drives too, since I'm sure they use a bit of power.

Hard Drives:
1x 500GB SATA Seagate
1x 150GB IDE Seagate
1x 1.5TB USB Seagate External HD


Report •

Related Solutions

#4
November 25, 2010 at 09:25:52
What Video card?
High performance video systems require quite a bit of power.

Report •

#5
November 25, 2010 at 09:28:03
Doh. It's too early for my brain. haha

Graphics:
NVidia GeForce GT 220


Report •

#6
November 25, 2010 at 12:21:24
gt 220

I'm finding similar to either of these on the web, under minimum system requirements for cards with the GT 220 video chipset:

"Minimum 300Watt or greater system power supply (with a minimum 12V current rating of 18A)"

"A minimum 350W or greater system power supply (with a minimum 12V current rating of 18A)"
.....

Go into your bios Setup and find the current voltage readings. What are supposed to be +3.3v, 5.0v, and +12.0v are supposed to be within 10% of the nominal values - if any of them are not, you need to replace the power supply.
..........

Hard drives use very little power.
.......

"A little history: I had RMA'd my RAM twice before..."

Ram that is actually BAD is extremely RARE !
Even if it is BAD, it's even MORE unlikely more than one module would be BAD at the same time, and even MORE likely than that that you would have that happen to you twice !

However, you MUST use ram that is compatible with using it in your mboard !

PNY memory configurator:
http://www.pny.eu/memory_configurat...

There is NO LISTING for Elitegroup IC780M-A2

The closest there is, is IC780M-A (V1.0)

No part numbers for 1066mhz ram are listed !

PNY partnumber Info. Capacity Type Quantity
DIMM101GBN/4300/2-BX (no info) 1 GB DIMM 240-pin DIMM 1
....
DIMM101GBN/5300/2CBX (retail) 1 GB DIMM 240-pin DIMM 1
D1GBN08Q667F-SB (oem)
....
DIMM101GBN/6400/2CBX (retail) 1 GB DIMM 240-pin DIMM 1
D1GBN08Q800J-SB (oem)
....
DIMM102GBN/5300/2CBX (retail) 2 GB DIMM 240-pin DIMM 1
D2GBN16Q667F-SB (oem)
....
DIMM102GBN/6400/2CBX (retail) 2 GB DIMM 240-pin DIMM 1
D2GBN16Q800J-SB (oem)
........

If the part number of the ram you're using is NOT listed, there's NO guarantee that it will work properly when installed in you mboard model - it might work fine - it might NOT work fine !

If the ram passes memory diagnostics tests without errors, then it should work fine .

(By the way, memtest86 v 3.5 has bugs that prevent it from testing more than 4gb of total ram - you're supposed to use v 3.4 or lower if you have more than 4gb of ram installed.)

However, the memory diagnostics tests may not be using the same settings as when the operating is being used - make sure the bios settings are correct...

Some ram manufacturer's modules do not strictly adhere to the JEDEC standards that most mboards bioses use to determine ram settings.
In that case, the ram settings in the bios Setup that the bios has automatically chosen may NOT be correct.
Check the ram settings in your bios - the ram voltage, and the ram timing numbers - those should be the same as for the specs for the modules themselves. Often the ram voltage and timing numbers are printed on the label on the modules.

If the voltage setting or timings settings in the bios are different from the specs for the ram, change them in the bios. The timing numbers must be as close as you can get to the same, or slower timings (higher numbers = slower) - you won't notice the difference the slower settings make.

If you have a mix of different modules
- don't mix ram that different voltages are specified for - the bios will force the ram to use the lowest voltage, if "by spd" or similar is used (default settings) - ram that a higher voltage is specified for is more likely to not work properly in that situation.
- the bios settings must be those for the slowest timing settings of all the modules, or slower (higher numbers = slower).



Report •

#7
November 25, 2010 at 14:19:49
The only voltage readings I can find in the BIOS setup are as follows:
CPU V_core: 1.392V
VDIMM: 1.632V
V_cc5V: 5.134V

Regarding the RMA'd RAM,
Both times, only 1 out of the 2 sticks was giving me problems. The first replacement was DOA, so I had to RMA that one too. (and they come as a pair, so I have to return the good stick along with the bad)

The ones I have now, those tested okay on Memtest86+ v4.10. I did make multiple passes to be sure, and I did disable Legacy USB Support since others had mentioned that caused false positives in Memtest.

RAM settings and timings seem okay.
http://www.ocztechnology.com/produc...

CL(CAS): 7
TRCD: 7
TRP: 7
TRAS: 16

Hmm... I just realized TRAS doesn't match.

Other numbers BIOS gave me for the memory were
TRRD: 4
TRC: 27
Voltage is set at 1.60V, reads out at 1.63V, and specs say 1.65V



Report •

#8
November 25, 2010 at 16:09:43
"....I did disable Legacy USB Support since others had mentioned that caused false positives in Memtest."

Huh ? Who were "others" ?

"Hmm... I just realized TRAS doesn't match."

Timing numbers that are higher in the bios than those in the ram specs won't cause you problems.
Timing numbers that are LOWER in the bios than those in the ram specs WILL cause you problems.
If you have a mix of modules, the timing numbers in the bios must be the same as the highest of all the timing numbers, or higher.

"Voltage is set at 1.60V, reads out at 1.63V, and specs say 1.65V"

The voltage readout in the bios may be off a bit, but it shouldn't read less than the ram spec voltage - it would probably do the ram no harm to tweak the bios settings to make it read the same as the ram spec.

We've seen that sometimes for "also ran" memory module makers such as PNY, GSkil, etc., that they may rate the ram for when only one of their ram modules is installed in a mboard to make the specs appear a bit better, or they haven't actually tested their modules in all the mboard models they list their ram part numbers for - they just assume they'll work properly, based on the main chipset the mboard has (see the Mushkin web site for more info about that - www.mushkin.com ).
Sometimes you need to tweak the bios settings for "also ran" ram when more than one of their modules is installed - e.g. add .01 or .02 to the ram spec voltage, or make the timing numbers slightly larger (slower).


Report •

#9
November 26, 2010 at 00:09:32
>>Huh ? Who were "others" ?
multiple users on various forums, including a post stickied in the official Memtest86+ forums:
http://forum.canardpc.com/showthrea...

Just Ctrl+F for "USB legacy" it's there. It might not have crashed my machine, but every time I had that enabled, I had hundreds of detected errors, whether the stick was good or not.

Just to clarify: You originally mentioned memtest86, whereas I'm using memtest86+ since memtest86 does not work on my AMD64 system.

>>Timing numbers that are higher in the bios than those in the ram specs won't cause you problems.
>>Timing numbers that are LOWER in the bios than those in the ram specs WILL cause you problems.
>>If you have a mix of modules, the timing numbers in the bios must be the same as the highest of all the timing numbers, or higher.

TRAS is at 16, and specs state that it should be at 20. The other 3 timings match. (7-7-7-16 vs 7-7-7-20)
And like I said earlier, my modules match.

So, I'll be changing the TRAS to 20... and also, it seems my BIOS can only set the voltage at 1.60V and 1.70V Nothing in between.

I'm no longer at my computer at the moment (Thanksgiving weekend stuffs) but this will require further study to see if I can set it to a higher degree of accuracy. I'll post back in a couple days with results of my fiddling.


Report •

#10
November 26, 2010 at 05:16:24
I have found that many times RAM can't be run at the advertised specs. Try setting to Auto and see what timings are used then.

DDR3 specs call for 1.5V. Manufacturers that juice their RAM do so to get higher clock speeds or timings out of them. I only buy RAM that is made to run at spec for the voltage. If you want to overclock that RAM then have at it.


Report •

#11
November 26, 2010 at 05:35:09
"Just to clarify: You originally mentioned memtest86, whereas I'm using memtest86+ since memtest86 does not work on my AMD64 system."

I know from experience that memtest86 produces false errors in up to two of it's tests in it's set of tests on systems with some main chipsets with the original Athlon cpus too - Test 5, and another one that produces infinite errors in a loop. However, you can set memtest86 to not use the specific tests that produce the false errors.

I did notice the 4.x version you quoted was higher than the highest version of memtest86 - V 3.5 - that was on the memtest86 web site when I looked not long ago. If memtest86+ is a different program then I don't know anything about it - false errors when legacy USB is enabled is a bug that shouldn't be tolerated.

I think your ram error problems have a lot more to do with the bios having the one timing number too low.

Microsoft's Windows Memory Diagnostic has no problem testing ram on AMD cpu systems, but it can't test more than 4gb:
http://technet.microsoft.com/en-us/...


Report •

#12
November 26, 2010 at 12:44:11
>>I have found that many times RAM can't be run at the advertised specs. Try setting to Auto and see what timings are used then.

I have never messed with the RAM settings in the BIOS before, so the Auto settings gave me 7-7-7-16 instead of 7-7-7-20 as stated in the ram specs.

>>If memtest86+ is a different program then I don't know anything about it - false errors when legacy USB is enabled is a bug that shouldn't be tolerated.

It definitely is a different program (although still based off the original memtest, just with added compatibility with more chipsets and allowing for the two tests on the original memtest86 to work on AMD64 systems). http://www.memtest.org/
A few people I know personally and it seems many people online, have used it with success, so I don't see a problem. I am not aware of any new updates to the original memtest86 program since i began using it.

And if you don't trust that, I did run Windows Memory Diagnostic recently and that came up with no errors. So in summary, there are no detected errors on my current memory sticks with both Windows Memory Diagnostic and memtest86+.

And I only have 4GB of ram, so having more than that won't be a problem.


As stated previously I'll change the timings as soon as I get back home, and I'll let you know what happens then.


Report •

#13
November 26, 2010 at 13:02:53
Memtest86+ is a legitimate open source memory test based on the original.

sleepy, have you ever had the rig run well with that brand/ model of RAM?

"I have never messed with the RAM settings in the BIOS before, so the Auto settings gave me 7-7-7-16 instead of 7-7-7-20 as stated in the ram specs".

The spec I am most concerned about is the voltage. JEDEC specs call for 1.5V default settings. Many folks use RAM that is not built to JEDEC specs successfully. That said, I see no reason to use non standard RAM. Look at the link below for more on DDR3 specs.

Try running with only one stick of RAM installed to see what happens.

One other thing to cover. Is the processor properly identified in the BIOS? Are the settings in use correct. Sometimes an older version of a BIOS won't properly configure a processor.

Your processor requires a motherboard capable of supporting a 125W processor. According to the ECS site (link below) your board does support 125W processor.

Do you have the additional power connector inserted directly into the board near the processor ( 1 x 4-pin 12V Connector)?

http://www.ecs.com.tw/ECSWebSite/Pr...


Report •

#14
November 27, 2010 at 21:06:10
So I fixed the timings, and I still had blue screen error as Windows was starting. There was no dump, and I was unable to see what kind of blue screen it was. (It just flashed for a fraction of a second before rebooting)

And like before, Startup Repair comes up, and says it cannot fix the problem. It shuts down after it's done; I turn it back on, and Windows starts up just fine.

>>sleepy, have you ever had the rig run well with that brand/ model of RAM?
when I first built it, it worked fine. no problems until about 2-3 months in when i had random freezes and crashes. After a long while, I ultimately found the memory to be the problem. Only one stick was bad, and since I was really busy with school and could not afford to be without a running computer, I ran with only one stick of memory successfully without problems. Once I had time to be without a computer, I returned both sticks (as per RMA instructions) for replacement.

>>Try running with only one stick of RAM installed to see what happens.
I will remove one stick tonight, boot it up tomorrow and report here with the results.

>>Is the processor properly identified in the BIOS? Are the settings in use correct.
The processor is indeed properly identified. As for settings, I'm not sure what you're looking for, but it's running with a CPU V_core voltage of 1.40V, which seems to be within spec, from what I've gathered online.

>>Do you have the additional power connector inserted directly into the board near the processor ( 1 x 4-pin 12V Connector)?
Yep, the 1x4 pin connecter near the processor is firmly in place.


Report •

#15
November 27, 2010 at 21:21:27
"OCZ DDR3 PC3-10666 Platinum Low Voltage (2GB x2)
Timings/Voltage: 7-7-7-20, 1.60V"

Interesting that the RAM is called "Low Voltage" but requires 1.6v rather than the JEDEC standard voltage which is 1.5v.

Which HDD has Win7 on it? Hopefully the 500GB Seagate. What do you need the 150GB IDE drive for? I suggest you eliminate it, at least while troubleshooting. And unplug the external drive as well.

Your power supply isn't a good one. I doubt it's the problem but you never know. It's a little light in the +12v amperage.

http://www.newegg.com/Product/Produ...

You didn't get the GT 220 for gaming, did you?


Report •

#16
November 27, 2010 at 23:51:50
>>Interesting that the RAM is called "Low Voltage" but requires 1.6v rather than the JEDEC standard voltage which is 1.5v.
Yes... I found that odd as well. That is, when you all told me about it.
And actually, I should have been more specific. The RAM actually requires 1.65V, and the BIOS automatically set it to 1.60V.

>>Which HDD has Win7 on it? Hopefully the 500GB Seagate. What do you need the 150GB IDE drive for?
Indeed, the C drive is on the 500GB sata. The IDE contains an old Win2k install and I have been too lazy to go through and find the files I want to keep.

I'll disconnect those when I start this computer up tomorrow morning as well.

>>You didn't get the GT 220 for gaming, did you?
haha, yeah, I did not get the GT 220 for gaming. And the few games I play on here don't have very taxing requriements at all.


Report •

#17
November 28, 2010 at 09:06:09
At this point I would try to increase the RAM voltage to 1.70V and see if that makes things stable.

I suggest you go about it this way. First set the RAM to auto and let the voltage and timings be set by the board. Then manually change ONLY the voltage to 1.65V or 1.70V.


Report •

#18
November 29, 2010 at 12:08:28
>>I suggest you go about it this way. First set the RAM to auto and let the voltage and timings be set by the board. Then manually change ONLY the voltage to 1.65V or 1.70V.

Okay, I've placed the timings to auto and set the voltage to 1.65V. Started it this morning, and got a blue screen. Then got another blue screen as Startup Repair loaded. Was a PAGE_FAULT_IN_NONPAGED_AREA (0x00000050) blue screen, naming rdbss.sys to be at fault. I've had blue screens as Startup Repair loaded before, so it's not surprising. Usually, by the second time around, Startup Repair will load just fine. It did a disk check today, instead of nothing, in case that matters.

I will now set the voltage to 1.70V, shut it down for a number of hours and report back here.


Report •

#19
November 29, 2010 at 12:20:12
Read the link below from MSoft on possible causes of that error.

http://technet.microsoft.com/en-us/...


Report •

#20
November 29, 2010 at 15:50:29
>>Read the link below from MSoft on possible causes of that error.
So I figured I could do a proper disk check before I shut the computer down, out of the many methods listed in that link.

Before I had booted up the computer, I changed the voltage to 1.70V. (Of course, since I already had it running earlier today, there were no blue screens this time)

I scheduled a disk check and proceeded to shut the computer down. Before it did, a MEMORY_MANAGEMENT blue screen flashed for a bit.

I booted it up again to run the disk check. I left it alone, checking every so often. I can't say with 100% certainty, but it seems that disk check ran twice... my guess is that it crashed once and at reboot, it started another disk check.

In any case, when it successfully completed, I did not notice any blue screen, but I did notice there was a minidump around the time when the disk check finished/when the computer booted.

Minidumps here:
http://dl.dropbox.com/u/2544832/Min...

112910-23197-01.dmp is the MEMORY_MANAGEMENT dump at shutdown
112910-23010-01.dmp is the NTFS_FILE_SYSTEM dump that I didn't see at disk check completion/reboot


Report •

#21
November 29, 2010 at 23:34:42
okay, first time booting after a long period of being off, with 1.70V going to the RAM. I'm still running with one stick.

Got an bsod error I hadn't seen before as windows started:
GS_FALSE_POSITIVE_MISSING_GSFRAME
Dump here: http://dl.dropbox.com/u/2544832/112...

It shut down, and started up Startup Repair, telling me it couldn't do anything, and then shut down.

I turned it on again, and while windows started, all the applications that are run at startup (anti-virus, dropbox, etc) did not run. I got Application Error messages telling me "This application was unable to start correctly. (0xc0000005)"
I try to run Internet Explorer and Google Chrome and neither started up.

So I shut down, start it again, and everything's fine.

Also to note, I do use the computer for a while after, and when everything's fine, it stays that way.


Report •

#22
November 30, 2010 at 04:45:56
Upon rereading all your responses I am wondering if the issue is the motherboard. The fact that after booting once the computer will boot fine without errors the second time could indicate a poor connection or cold solder joint. One other possibility is an errant case standoff or loose screw in the case.

First step in troubleshooting would be to remove the board from the case and verify you have standoffs under each screw hole and ONLY under each screw hole. You do use standoffs in the case? Some cheap cases have raised bosses instead.

Are you always using the first RAM slot? Maybe you should try using some other slot with one stick. If it won't boot when using the second slot on the first bank then maybe try the fist slot in the second bank. Understand, I am groping for explanations here. The possible causes of the errors you are getting include the board, RAM, CPU, PSU, drivers, etc.

Googling the latest error yielded some interesting hits. I am not versed enough to decipher the meaning of most of it though.

Going back to the hardware/RAM issue it seems telling that upping the RAM voltage changes the type of error you get.

ECS has recently released a BIOS update that is related to memory voltage. You may want to consider flashing it. Link below.

http://www.ecsusa.com/ECSWebSite/Do...


Report •

#23
November 30, 2010 at 17:40:15
>>First step in troubleshooting would be to remove the board from the case and verify you have standoffs under each screw hole and ONLY under each screw hole. You do use standoffs in the case?
My case does use standoffs, yes. I have verified their existence and have verified that they are all under every screw hole.

After doing so, I tried booting again, crashed at startup again, and during startup repair (as opposed to when Startup Repair was loading), I got a REFERENCE_BY_POINTER bsod error, 0x00000018. Another new one, yay.

I am still using the same RAM slot, but I think I will reduce the voltage back to the auto selection at 1.60V before changing the slot and trying again.

Or do you recommend that I flash the BIOS with the updated version before continuing?

>>Understand, I am groping for explanations here. The possible causes of the errors you are getting include the board, RAM, CPU, PSU, drivers, etc.

I understand. I knew early on that this would be hard to diagnose.

It does require a bit of patience, since I can't really go through a checklist of things to do very quickly, since the errors only occur after the computer has been shut down for a number of hours.

Just so you know, I did purchase a new PSU from Cooler Master at Fry's Electronics over the weekend. Still in its plastic though. To be honest, I just chose the one that had the least number of returned/refurbished items on the shelves.
http://www.newegg.com/Product/Produ...
Reviews are mostly good on newegg, but there does seem to be a good number of 1-egg reviews. Then again, it's still better than my current PSU.


Report •

#24
November 30, 2010 at 19:41:34
"My case does use standoffs, yes. I have verified their existence and have verified that they are all under every screw hole"

The possible problem is never with the ones that you can see are in the right places - it's the ones that may be in the wrong places that may be shorting something out.
You can't see metal standoffs that are in the wrong places from the top side of the mboard. Depending on the design of your case, you may be able to see the bottoms of the metal standoffs in the screw holes that are under the back side of the mboard if you take off the case side panel on the back side of the mboard (that side is riveted on on some cases), or you may need to remove the mboard to make sure,
Loose screws you have lost track of can cause the same problem if there is enough room for them to get behind the mboard.

"I think I will reduce the voltage back to the auto selection at 1.60V "

That's a bonehead thing to do, after what we have told you, It does no harm to the ram to use a bit more that the stated voltage, as long as you don't go too far over the specified voltage for the ram, e.g. if you increased it to .1 over the voltage specified on the module, you might have problems but it probably wouldn't hurt it; ,2 over might damage it. The detailed specs for the ram, if they exist (they do for major brand name ram) will tell you what the limits are.
.
"...do you recommend that I flash the BIOS with the updated version before continuing?..."

NO!
NEVER flash your bios when you're not sure whether your system is working properly !


Report •

#25
November 30, 2010 at 20:34:58
Standoffs:
As per OtheHill's instructions, I did go in and remove the motherboard, verifying that all standoffs were under screw holes on the mobo.

RAM:
Okay, but having new errors, and these errors being outside of the times when I normally see it, does worry me. At 1.65V and 1.60V, I only got BSODs when loading or starting Windows or Startup Repair. At 1.70V, I indicated BSODs at shutdown and while Startup Repair was running, in addition to the errors at previous voltages.

RAM Specs here: http://www.ocztechnology.com/produc...
Nothing on the packaging, either. Just that site. All it says is 1.65V. Nothing more.

BIOS:
I'm not the type to go in and do things without reason. I know that flashing the BIOS can potentially cause me trouble. That's why I asked in the first place, since OtheHill suggested it.


Report •

#26
November 30, 2010 at 21:04:16
Tubes

Did you see this in my #22? "ECS has recently released a BIOS update that is related to memory voltage. You may want to consider flashing it. Link below".

http://www.ecsusa.com/ECSWebSite/Do...


Report •

#27
November 30, 2010 at 21:23:19
OtheHill

"Tubes

Did you see this in my #22? "

Yes.

I still insist on.....

NEVER flash your bios when you're not sure whether your system is working properly !

....unless specific info can be found that describes symptoms the same or very similar to what he is getting qand that updating the bios version cures the problem, E.g. on the ECS web site in the bios update info, or on an ECS mboard forum, or when searching on the web with his mboard model and his BSOD messages, etc.

You don't update the bios version when a cpu is installed that the bios doesn't recognize properly - you install a cpu the present bios version does recognize fine, then update the bios, then install the cpu you were trying to install in the first place.- , so why would you update the bios version when you're not sure if the mboard is working properly with this ram in it ?.
If he can install other ram that does work fine in this mboard, then in that case, flashing the bios isn't any more risky that it normally is,
......

sleepyd

"As per OtheHill's instructions, I did go in and remove the motherboard, verifying that all standoffs were under screw holes on the mobo."

You didn't say so in your post before that.

The ram timings are listed under Specifications on that page you pointed to.
As for more specific info, I wouldn't be at all surprised if that doesn't exist on their web site.


Have you tried any ram that the Jedec standard 1.5v for DDR3 is specified for ?

Kingston ram listed for your model.

Go here:
http://www.ec.kingston.com/ecom/con...

Under Model, enter IC780M-A2, click on Search.

You click on one of the module part numbersw to see more info, then click on Data sheet in that info to see specific specs.

E.g.
Data sheet for the first one listed:
http://www.valueram.com/datasheets/...

It uses the JEDEC standard voltage for DDR3 ram - 1.5v
Voltage range - 1.5v + or - .075v

You're probably doing your ram no harm at all by selecting 1.7v - .05 over the 1.65 specified for the module.
Is that one timing number still tweaked in the bios so it's the same or higher than that same number in the timing numbers specified for the module?


Report •

#28
November 30, 2010 at 21:27:38
>>Have you tried any ram that the Jedec standard 1.5v for DDR3 is specified for?
The ram I have is the only DDR3 ram in the building and the only DDR3 ram I've purchased, so no.

>>Is that one timing number still tweaked in the bios so it's the same or higher than that same number in the timing numbers specified for the module?
OtheHill had earlier instructed me to reset all other timings back to auto and only adjust the voltage, so that timing number should be back to 16.

Edit: specs state 20, so it is lower than that specified on the site. I did notice a review in the tab stating that 16 was its stock setting... "To gauge its full potential, the OCZ Platinum was overclocked from stock settings of 7-7-7-16 at 533 MHz to stable settings of 10-10-10-23 at 746 MHz."

Does the RAM itself indicate what its timings should be, or does the BIOS just sorta figure something out?


Report •

#29
November 30, 2010 at 21:34:13
"OtheHill had earlier instructed me to reset all other timings back to auto and only adjust the voltage, so that timing number should be back to 16."

As I said above...

"Timing numbers that are higher in the bios than those in the ram specs won't cause you problems.
Timing numbers that are LOWER in the bios than those in the ram specs WILL cause you problems.
If you have a mix of modules, the timing numbers in the bios must be the same as the highest of all the timing numbers, or higher. "

You WILL have problems if ANY of the timiing numbers are LESS than specified for the module - 7-7-7-20 !


Report •

#30
November 30, 2010 at 21:51:23
So, voltage back to 1.70V and T_RAS back to 20 then?

Report •

#31
November 30, 2010 at 22:07:35
Yes, try that. Whatever you try, that timing number should be 20, or higher.

Report •

#32
November 30, 2010 at 22:42:06
So Windows started up all right, but since it hasn't been too long since it was last on, I can't say for sure if this solved it. Not until tomorrow, anyway.

However, not long after it started, I did run into a MEMORY_MANAGEMENT bsod.
Dump here: http://dl.dropbox.com/u/2544832/113...

To my knowledge, nothing else was going on other than programs running at startup, like the anti-virus.

I'm posting from this computer now, so currently no problems.


Report •

#33
November 30, 2010 at 22:47:50
By the way, OCZ has made the typo 10666 over and over again for this module - there's no such thing - it's supposeed to be 1066.

""To gauge its full potential, the OCZ Platinum was overclocked from stock settings of 7-7-7-16 at 533 MHz to stable settings of 10-10-10-23 at 746 MHz."

Have you been overclocking the mhz of this ram all along ? If so, knock it off ! Set the mhz to the stock setting ! If you want to overclock, you can do that later - find a site that has tested overclocking settings with the same modules on the same mboard model, otherwise there's no guarantee the settings that are stable with another mboard model will be stable with this one.

"Does the RAM itself indicate what its timings should be, or does the BIOS just sorta figure something out?"

Most bioses use JEDEC standards - your ram doesn't, at least, 1.65v is not the JEDEC standard for DDR3 ram - 1.5v is. !,65v would be low voltage it bit were DDR2 ram, but it's high voltage for DDR3 ram. . The ram module has a tiny chip on one end that contains SPD (Serial Presence Detect) data that the manufacturer has entered - some manufacturers enter all the proper data about what the module is specified to do, but some enter only a minimal amount of data, or some data may be improperly entered. If the bios is set to detect the ram "by SPD" or similar, which is usually the default, if there is enough accurate data on that tiny chip, usually the bios gets the voltage and timing settings for the ram right, or the ram timing numbers can be higher, but if there isn't enough data or if it's been improperly entered, the bios is a lot more likely to set settings wrong for the ram,
E.g. Kingston probably enters all the SPD data correctly, OCZ may not.
So - you need to check the settings the bios has chosen, especially when your ram doesn't adhere to Jedec standards, to make sure they're not improper. In the case of your mboard, it sounds like you or the bios can't set voltages in .05 incraments, so the bios could have choosen either a lower or higher voltage - lower isn't a good thing - but if the data had been entered correctly, the bios should have gotten all the timing numbers the same, or higher if it can't do that..


Report •

#34
November 30, 2010 at 22:59:18
Thanks for the info. Good to know.

>> it sounds like you or the bios can't set voltages in .05 incraments
Since that post, I did figure out how I can set voltages in my BIOS, which does so in increments of 10mV. Or 0.01V, as it were.


Report •

#35
December 1, 2010 at 13:19:23
>>Have you been overclocking the mhz of this ram all along ? If so, knock it off !
No no, I stated that I saw that on one of the review tabs in the OCZ site. They mentioned 16 as stock setting instead of 20, so I thought that odd, and I thought it was worth mentioning.


Also, results of booting up this morning:
Blue screen while windows was starting. Flashed for a bit, didn't catch it. Oddly enough, instead of rebooting into Startup Repair like it normally does, it tried to start Windows normally. Again, however, it bsod'ed on me during the "Starting Windows" screen. Only in this second reboot did it boot into Startup Repair. Once there, it attempted to repair disk errors, and completed it's repair. Reboot, and Windows starts fine.

I don't know how much of that is relevant. The only difference is that it tried to boot into windows twice before going to Startup Repair. Usually Startup Repair does nothing, but occasionally, it will "repair disk errors".

Voltage is still at 1.70V, TRAS still at 20, still using the same slot.

So, next steps? Shall I move the RAM to a different slot now?


Report •

#36
December 1, 2010 at 15:39:41
I would try a different slot. I would also email tech support at ECS and ask them about the BIOS update I mentioned. With all the errors you are having something is going on.

Clarify something for me. You stated after you performed a clean install of Windows 7 you noticed the errors you describe right away. Does that mean they weren't happening prior to the clean install?


Report •

#37
December 1, 2010 at 17:58:33
"Blue screen while windows was starting. "

Quote the STOP 0Xx00000xx error .

Is it the same error each time ?

STOP errors can be but aren't necessarily caused by hardware problems - you probably did previously have errors caused by hardware or settings for it - you having the one timing number too low - and because of that data may be damaged on the Windows installation such that you can't fix the problem unless you install Windows from scratch again.


Report •

#38
December 1, 2010 at 17:58:46
>>I would also email tech support at ECS and ask them about the BIOS update I mentioned. With all the errors you are having something is going on.
Okay, I'll go ahead and do that.

>>You stated after you performed a clean install of Windows 7 you noticed the errors you describe right away. Does that mean they weren't happening prior to the clean install?
Prior to the clean install, no. These errors were not happening.

Prior to the clean install, I had a dying hard drive. Got S.M.A.R.T. warnings and everything. I now know that I should backup my files. I just copied all the files I wanted to keep onto my external (I was lucky it died slowly). I have yet to migrate all those files to the new install and new (actually, refurbished) hard drive until I get the current problem solved.

And still prior to the failing hard drive, I had memory errors. Random freezes and BSODs when doing absolutely nothing. Using memtest86+, found one of the sticks to be faulty. When I received new ones from OCZ, I immediately tested them and again found one of the two sticks to be faulty. Got new ones again, and memory checked out okay.
(but two weeks later I got the SMART warning from the hard drive)

So it's been rough. At least the previous two problems were consistent with the errors they were spitting out.


Report •

#39
December 1, 2010 at 18:42:41
>>Quote the STOP 0Xx00000xx error . Is it the same error each time ?
I can't tell you since it flashed for a fraction of a second before rebooting. I did some digging just now and found that I can make it so that it doesn't reboot when faced with system failure. Hopefully that helps in the future. I didn't realize this before, since sometimes when I get a blue screen it didn't reboot, and sometimes it did.

>>Data may be damaged on the Windows installation such that you can't fix the problem unless you install Windows from scratch again.
Since I haven't quite "settled in" to this install of windows with all my programs and things, it wouldn't be too much of a task to gather what little I need from this computer and perform another windows install.


Report •

#40
December 1, 2010 at 18:43:36
"Using memtest86+, found one of the sticks to be faulty. When I received new ones from OCZ, I immediately tested them and again found one of the two sticks to be faulty."

Were those memory tests done before or after you discovered you needed to disable Legacy USB in your bios to avoid false errors because of bugs in memtest86+ ??

If they were identical to the modules you have now and the bios had that one timing number set to 16 instead of the minimum 20, and if your voltage in the bios was set to less than the voltage spec for the modules, then it's not surprising you got ram errors, which could very well have been FALSE !.

The odds of new major brand ram module being bad may be, say, 5%, that percentage I suspect would be higher for OCZ (and GSkil) ram, but still, the odds of you receiving two modules that are bad , if the bios was set to proper settings, is extremely remote.


Report •

#41
December 1, 2010 at 19:10:35
Or the BIOS could be wacked causing the problems. I don't know what that update is about but it states it is about memory issues.



Report •

#42
December 1, 2010 at 19:44:21
>>Were those memory tests done before or after you discovered you needed to disable Legacy USB in your bios to avoid false errors because of bugs in memtest86+ ??
After. When I had that enabled, I'd get something like a 1000 errors or something. Disabled, there were only about 5-10 on the bad sticks, more or less.

>>The odds of you receiving two modules that are bad , if the bios was set to proper settings, is extremely remote.
Maybe, maybe not. I got this ram as a package deal with the AMD processor I'm now using from newegg. To be fair, it seems since that time, a good number of people have been RMA'ing their sticks according to feedback on newegg. And at least one person more than once. http://www.newegg.com/Product/Produ...

In any case, that's stuff's all over with, and the sticks I have now check out okay on memtest86+, but I can test them again since they were tested before all this troubleshooting in this thread.


Report •

#43
December 1, 2010 at 21:45:42
"To be fair, it seems since that time, a good number of people have been RMA'ing their sticks according to feedback on newegg."

You've always got to take into account when you read those reviews that
- people who have problems are A LOT MORE LIKELY to post a review than those who don't. The number of them complaining is likely to be way out of proportion.
- they may not have their bios ram settings set a compatible way with their mboard either !


Report •

#44
December 2, 2010 at 11:19:34
>>- people who have problems are A LOT MORE LIKELY to post a review than those who don't.
- they may not have their bios ram settings set a compatible way with their mboard either !

Of course.
I read the individual reviews from those that seem knowledgable instead of just looking at the number of bad reviews. Like you say, it's not reliable, but it gives me a feeling about the product at least. In any case, my RMA issues are over and done with (hopefully) so I'd rather move on.

=======================

Okay, so this is what happened last night: I decided to play World of Tanks, an online game that allows me to play a quick 15 minute game.

The program had been run successfully before without problem. the program (not windows) crashed, giving me a dialog box stating some uinexpected error with Windows Visual C++. Now I know I installed Visual C++ when I installed Windows Debugger so I could read the minidump files. I uninstalled every program associated with Windows Debugger including Visual C++. I also uninstalled the game for good measure.

And later the computer crashed on me while typing in IE to post on here. Left it off for the night, turned it on this morning. I get the screen telling me if I wanted to boot into safe mode or start windows normally. I left it alone to start normally. Surprisingly, it didn't crash at startup.

Get the "Windows recovered from an unexpected shutdown" message as Windows opens. Error was 0x0000001E. I noticed the timestamp on the most recent dump file was right when I booted instead of last night, so I'm not sure how that works. The dump file just before that one was 2 days ago.

While looking at the timestamp, it crashed with a BSOD, error was 0x00000050. Back on the computer, no bsods this time.

The two aforementioned minidumps here:
http://dl.dropbox.com/u/2544832/Min...

So... this is different. With the random bsods, it's feels like I'm back to having RAM problems again.

Oh, also, due to all that stuff I did last night, I left the ram stick in the same slot. Everything's still the same. Nothing in the BIOS has been touched. Still 1.70V, TRAS 20.


Report •

#45
December 2, 2010 at 12:32:43
High end games are well known to be a lot more likely have bugs in them than most programs, and they will not necessarily work properly on every system that meets at least the game's minimum requirements. If you ONLY have problems with a specific game or some games, then the problem is probably caused by bugs in the game(s) .

If an online game that worked fine before no longer works fine, it may be you are now playing a different version of it that has bugs that only now produce errors on your system. Loading an older version, if available, may work fine on your system.
.......

And then there's the very real possibility your existing Windows installation was damaged when the bios had the ram voltage set too low and the one timing setting set too low. It may very well be that the only way to fix your problems now that your bios ram settings are okay is to re-install Windows from scratch.
.......

0x0000001E - KMODE_EXCEPTION_NOT_HANDLED

0x00000050 - PAGE_FAULT_IN_NON-PAGED_AREA

Apparently, both of those errors are usually caused by software problems, not hardware problems, and you can get those errors in many different situations.

You haven't provided enough info for me or us to possibly pin down what causes those two errors.

In most cases, the stuff in brackets after the STOP: 0Xx00000xx is specific to your own system and useless regarding you or us searching with that on the internet, although sometimes the first one or ones can tell you something.
You will find a lot more "hits" when you search on the web with the Stop error if you don't include the stuff in brackets when you search for possible causes.

There is often a problem file named on the same BSOD screen, often at the end of the text.
Or
There is a link to "More info" or similar - if you click on that, a problem file may be named.

If the screen text indicates it made a minidump, there can very useful info in that, but it's text is not all readable unless you use specific programs to "read" or "translate" it, and I don't read other people's minidumps.
There is info on the web about what programs you need to use to "read" or "translate" a minidump, and how you interpret the results to possibly find what the problem is actually caused by.
Or there are sites on the web (not this one), where you can send them a minidump and they will interpret it for you.


Once you have noted more info about you error, search the web using: 0x0000001E or 0x00000050 and any file name stated or possible cause found by interpreting your minidump(s) to find "hits" that are a lot more likely to help you fix your problem.

Any minidump with a date and time older than when you last corrected the bios ram voltage and the one timing setting and left it that way may not be relevant at all now.


Report •

#46
December 2, 2010 at 13:52:41
It's not really that high end, but it demands more than most of what I do on this computer.
World of Tanks Minimum Requirements:
OS: Windows 2000 / XP /Vista / 7
CPU: Pentium 4 2.0 Ghz
RAM: 1 GB RAM
HDD: 3 GB Free
Graphics Card: nVIDIA GeForce 6800 series

World of Tanks Recommended Specifications:
OS: Windows 2000 / XP /Vista / 7
CPU: Pentium 4 3.0 Ghz or faster processor
RAM: 2 GB RAM
HDD: 3 GB Free
Graphics Card: nVIDIA GeForce 7600 GT series / ATI Radeon X1600 series or better

By the way, neither of these blue screens appeared when the game was running. The program crashed, not windows itself.

I did link to a zip file containing the dump files in my previous post. I also had other dump files in previous posts, but like you said, those aren't relevant anymore.

In any case, I uninstalled the program that let me read the dump files in the first place (Debugging Tools for Windows) as I mentioned in my previous post, so I'm now using the same program on my old laptop to read it.

dump files are as follows:

KMODE_EXCEPTION_NOT_HANDLED (1e)
This is a very common bugcheck. Usually the exception address pinpoints
the driver/function that caused the problem. Always note this address
as well as the link date of the driver/image that contains this address.
Arguments:
Arg1: ffffffffc0000005, The exception code that was not handled
Arg2: fffff88000c04fa9, The address that the exception occurred at
Arg3: 0000000000000000, Parameter 0 of the exception
Arg4: 0000000000000078, Parameter 1 of the exception

Debugging Details:
------------------


EXCEPTION_CODE: (NTSTATUS) 0xc0000005 - The instruction at "0x%08lx" referenced memory at "0x%08lx". The memory could not be "%s".

FAULTING_IP:
fltmgr!GetContextFromStreamList+99
fffff880`00c04fa9 488b4820 mov rcx,qword ptr [rax+20h]

EXCEPTION_PARAMETER1: 0000000000000000

EXCEPTION_PARAMETER2: 0000000000000078

READ_ADDRESS: GetPointerFromAddress: unable to read from fffff80002ead0e0
0000000000000078

CUSTOMER_CRASH_COUNT: 1

DEFAULT_BUCKET_ID: VISTA_DRIVER_FAULT

BUGCHECK_STR: 0x1E

PROCESS_NAME: svchost.exe

CURRENT_IRQL: 0

EXCEPTION_RECORD: fffff8800892cd88 -- (.exr 0xfffff8800892cd88)
ExceptionAddress: fffff88000c04fa9 (fltmgr!GetContextFromStreamList+0x0000000000000099)
ExceptionCode: c0000005 (Access violation)
ExceptionFlags: 00000000
NumberParameters: 2
Parameter[0]: 0000000000000000
Parameter[1]: 0000000000000078
Attempt to read from address 0000000000000078

TRAP_FRAME: fffff8800892ce30 -- (.trap 0xfffff8800892ce30)
NOTE: The trap frame does not contain all registers.
Some register values may be zeroed or incorrect.
rax=0000000000000058 rbx=0000000000000000 rcx=0000000000000058
rdx=fffffa8004a96f20 rsi=0000000000000000 rdi=0000000000000000
rip=fffff88000c04fa9 rsp=fffff8800892cfc0 rbp=0000000000000000
r8=fffffa80045c2a90 r9=fffff8800892d0b8 r10=fffffa800553d8a8
r11=0000000000000011 r12=0000000000000000 r13=0000000000000000
r14=0000000000000000 r15=0000000000000000
iopl=0 nv up ei pl nz na pe nc
fltmgr!GetContextFromStreamList+0x99:
fffff880`00c04fa9 488b4820 mov rcx,qword ptr [rax+20h] ds:d030:00000000`00000078=????????????????
Resetting default scope

LAST_CONTROL_TRANSFER: from fffff80002cafa39 to fffff80002c75740

STACK_TEXT:
fffff880`0892c5b8 fffff800`02cafa39 : 00000000`0000001e ffffffff`c0000005 fffff880`00c04fa9 00000000`00000000 : nt!KeBugCheckEx
fffff880`0892c5c0 fffff800`02c74d82 : fffff880`0892cd88 00000000`00000000 fffff880`0892ce30 fffffa80`0553d8f8 : nt!KiDispatchException+0x1b9
fffff880`0892cc50 fffff800`02c738fa : 00000000`00000000 00000000`00000000 00000000`00000000 fffffa80`00000030 : nt!KiExceptionDispatch+0xc2
fffff880`0892ce30 fffff880`00c04fa9 : fffffa80`0402a430 00000000`00000000 fffff880`0892d070 fffff880`00c0432e : nt!KiPageFault+0x23a
fffff880`0892cfc0 fffff880`00c04533 : fffffa80`0553d8a0 fffffa80`0553d8a0 ffff0000`0b69825b fffffa80`045c2a90 : fltmgr!GetContextFromStreamList+0x99
fffff880`0892d040 fffff880`03fe0426 : fffffa80`0553d8a0 00000000`00000000 fffff880`0892d1d8 fffff880`0892d1d8 : fltmgr!FltGetStreamHandleContext+0x43
fffff880`0892d070 fffffa80`0553d8a0 : 00000000`00000000 fffff880`0892d1d8 fffff880`0892d1d8 00000000`00000040 : aswMonFlt+0x35426
fffff880`0892d078 00000000`00000000 : fffff880`0892d1d8 fffff880`0892d1d8 00000000`00000040 fffff880`0892d120 : 0xfffffa80`0553d8a0


STACK_COMMAND: kb

FOLLOWUP_IP:
aswMonFlt+35426
fffff880`03fe0426 ?? ???

SYMBOL_STACK_INDEX: 6

SYMBOL_NAME: aswMonFlt+35426

FOLLOWUP_NAME: MachineOwner

MODULE_NAME: aswMonFlt

IMAGE_NAME: aswMonFlt.sys

DEBUG_FLR_IMAGE_TIMESTAMP: 4c865084

FAILURE_BUCKET_ID: X64_0x1E_aswMonFlt+35426

BUCKET_ID: X64_0x1E_aswMonFlt+35426

Followup: MachineOwner

===========================================================
===========================================================
PAGE_FAULT_IN_NONPAGED_AREA (50)
Invalid system memory was referenced. This cannot be protected by try-except,
it must be protected by a Probe. Typically the address is just plain bad or it
is pointing at freed memory.
Arguments:
Arg1: fffffa8006dfba10, memory referenced.
Arg2: 0000000000000000, value 0 = read operation, 1 = write operation.
Arg3: fffff880049780b8, If non-zero, the instruction address which referenced the bad memory
address.
Arg4: 0000000000000002, (reserved)

Debugging Details:
------------------


Could not read faulting driver name

READ_ADDRESS: GetPointerFromAddress: unable to read from fffff80002ef30e0
fffffa8006dfba10

FAULTING_IP:
nvlddmkm+36a0b8
fffff880`049780b8 8b8f101a0000 mov ecx,dword ptr [rdi+1A10h]

MM_INTERNAL_CODE: 2

CUSTOMER_CRASH_COUNT: 1

DEFAULT_BUCKET_ID: VISTA_DRIVER_FAULT

BUGCHECK_STR: 0x50

PROCESS_NAME: System

CURRENT_IRQL: 0

TRAP_FRAME: fffff88002f8e8f0 -- (.trap 0xfffff88002f8e8f0)
NOTE: The trap frame does not contain all registers.
Some register values may be zeroed or incorrect.
rax=fffffa8002dda7e0 rbx=0000000000000000 rcx=0000000000000000
rdx=000000000000000d rsi=0000000000000000 rdi=0000000000000000
rip=fffff880049780b8 rsp=fffff88002f8ea80 rbp=0000000000000000
r8=0000000000000000 r9=0000000000000000 r10=0000000000000001
r11=fffff88002f8ea50 r12=0000000000000000 r13=0000000000000000
r14=0000000000000000 r15=0000000000000000
iopl=0 nv up ei pl zr na po nc
nvlddmkm+0x36a0b8:
fffff880`049780b8 8b8f101a0000 mov ecx,dword ptr [rdi+1A10h] ds:00000000`00001a10=????????
Resetting default scope

LAST_CONTROL_TRANSFER: from fffff80002d3a8f2 to fffff80002cbb740

STACK_TEXT:
fffff880`02f8e788 fffff800`02d3a8f2 : 00000000`00000050 fffffa80`06dfba10 00000000`00000000 fffff880`02f8e8f0 : nt!KeBugCheckEx
fffff880`02f8e790 fffff800`02cb982e : 00000000`00000000 00000000`00000000 00000000`00088600 fffff880`047da687 : nt! ?? ::FNODOBFM::`string'+0x40ec0
fffff880`02f8e8f0 fffff880`049780b8 : 00000000`00000000 00000000`00000000 00000000`00000000 fffffa80`02db7000 : nt!KiPageFault+0x16e
fffff880`02f8ea80 00000000`00000000 : 00000000`00000000 00000000`00000000 fffffa80`02db7000 00000011`00000001 : nvlddmkm+0x36a0b8


STACK_COMMAND: kb

FOLLOWUP_IP:
nvlddmkm+36a0b8
fffff880`049780b8 8b8f101a0000 mov ecx,dword ptr [rdi+1A10h]

SYMBOL_STACK_INDEX: 3

SYMBOL_NAME: nvlddmkm+36a0b8

FOLLOWUP_NAME: MachineOwner

MODULE_NAME: nvlddmkm

IMAGE_NAME: nvlddmkm.sys

DEBUG_FLR_IMAGE_TIMESTAMP: 4cb9dd0e

FAILURE_BUCKET_ID: X64_0x50_nvlddmkm+36a0b8

BUCKET_ID: X64_0x50_nvlddmkm+36a0b8

Followup: MachineOwner
---------


Report •

#47
December 2, 2010 at 14:14:31
What I meant by high end games is anything more sophisticated that the simple games that come with Windows or the relatively simple puzzle or kids games you can play on the web.

Try searching on the web for info about interpreting those "read" minidumps.

E.g.
Info here says the Windows Debugger may be of help:
http://apcmag.com/beat_those_bluesc...

I haven't found that "reading" them properly helps much myself, I've almost always figured out what the problem was some other way.


Report •

#48
December 2, 2010 at 15:56:17
>>What I meant by high end games...
in that case, yes, it is a high end game.

>>Info here says the Windows Debugger may be of help:
Yep, that's how I read the minidumps in the first place. But it doesn't do much in the way of interpretation.

I'm no expert at this at all, but from what I've gathered online, both errors appear to be very common and can be caused by memory, drivers, or programs installed on the computer.

Your link tells me to look at the module name part of the dump.
The first error involves aswMonFlt.sys, which is an Avast AntiVirus driver. The second error mentions nvlddmkm.sys, which involves the NVIDIA graphics driver.

I'll uninstall avast, then I think I will run memtest86+ and rule them out first.

If that doesn't work, I'll uninstall and reinstall drivers (chipset drivers, graphics drivers, etc) in safe mode. Should there be a particular order in how i uninstall and reinstall them?

If you have any other suggestions, I'd be happy to hear them.


Report •

#49
December 2, 2010 at 16:53:00
At this point I think you might save time by reinstalling ONLY windows with updates and drivers. See if the system is rock solid at that point. Then add one item at a time and test.

If it were me I would also consider the BIOS update. I don't usually recommend flashing the BIOS. However in this case the update purpose clearly indicates it deals with memory settings.


Report •

#50
December 2, 2010 at 17:01:57
You could try searching the web with their file names and matching STOP error.

I don't see any indication in those memory dump "reads" that you have a ram problem.

It's a VERY good idea to DISABLE the resident module(s) - the part that runs all the time looking for suspicious activity - from running of your anti-malware software or third party firewall software, BEFORE you install software, especially more sophisticated software like video and printer "drivers" , and large software packages, because the resident module(s) MAY interfere with the software installing properly, and if it does you usually get NO indication the software didn't install properly while installing it. In anti-malware progams, you usually must disable the resident module in the program itself - disabling the entries found for it from loading in msconfig - startup usually does not do that.

Unless the instructions for installing a device tell you otherwise.......
You DO NOT install drivers for a device while booting into Windows, if the software for the device has not been installed yet - when Windows detects a generic device or New Hardware while booting, you allow it to search for drivers, it doesn't find any, and it wants you to show it the location of the drivers - CANCEL that, continue on to the desktop, and install the software for the device using the proper installation from a CD or the proper installation file that you downloaded from the web.

If you DID install drivers that way, go to Control Panel - Classic View - Programs and Features and Un-install the software you installed, reboot, DO NOT install drivers while booting, and install the software the right way !

The same applies no matter when Windows finds New Hardware !


Report •

#51
December 3, 2010 at 01:48:44
I agree, I think reinstalling Windows will help us narrow down what's wrong with my computer.

Well, I wanted to get a few files off of there. In doing so, I ran into a couple more stop errors. First one occurred after the Starting Windows screen, but Windows wanted to do a disk check since it did not shut down successfully, apparently. It did not run diskcheck and it errored out, giving me the following:

SYSTEM_SERVICE_EXCEPTION (3b)
An exception happened while executing a system service routine.
Arguments:
Arg1: 00000000c0000005, Exception code that caused the bugcheck
Arg2: fffff880049f1027, Address of the exception record for the exception that caused the bugcheck
Arg3: fffff88003dcf0f0, Address of the context record for the exception that caused the bugcheck
Arg4: 0000000000000000, zero.

Debugging Details:
------------------


EXCEPTION_CODE: (NTSTATUS) 0xc0000005 - The instruction at "0x%08lx" referenced memory at "0x%08lx". The memory could not be "%s".

FAULTING_IP:
nvlddmkm+1e5027
fffff880`049f1027 493b5008 cmp rdx,qword ptr [r8+8]

CONTEXT: fffff88003dcf0f0 -- (.cxr 0xfffff88003dcf0f0)
rax=0000040000000000 rbx=fffffa80030020f8 rcx=fffffa8002c48768
rdx=00000000c1d003f0 rsi=0000000000000050 rdi=00000000c1d003f0
rip=fffff880049f1027 rsp=fffff88003dcfac0 rbp=00000000c1d003f0
r8=0000040000000000 r9=fffffa80030021f8 r10=00000000c1d003f0
r11=fffffa8002c48750 r12=0000000000000001 r13=00000000c1d003f0
r14=0000000000000001 r15=0000000000000030
iopl=0 nv up ei pl nz na po nc
cs=0010 ss=0018 ds=002b es=002b fs=0053 gs=002b efl=00010206
nvlddmkm+0x1e5027:
fffff880`049f1027 493b5008 cmp rdx,qword ptr [r8+8] ds:002b:00000400`00000008=????????????????
Resetting default scope

CUSTOMER_CRASH_COUNT: 1

DEFAULT_BUCKET_ID: VISTA_DRIVER_FAULT

BUGCHECK_STR: 0x3B

PROCESS_NAME: csrss.exe

CURRENT_IRQL: 0

LAST_CONTROL_TRANSFER: from 0000000000000050 to fffff880049f1027

STACK_TEXT:
fffff880`03dcfac0 00000000`00000050 : fffff880`03dcfaf0 fffffa80`030d9000 00000000`c000009a fffffa80`030020f8 : nvlddmkm+0x1e5027
fffff880`03dcfac8 fffff880`03dcfaf0 : fffffa80`030d9000 00000000`c000009a fffffa80`030020f8 fffff880`049c5ce4 : 0x50
fffff880`03dcfad0 fffffa80`030d9000 : 00000000`c000009a fffffa80`030020f8 fffff880`049c5ce4 fffff880`03dcfb30 : 0xfffff880`03dcfaf0
fffff880`03dcfad8 00000000`c000009a : fffffa80`030020f8 fffff880`049c5ce4 fffff880`03dcfb30 00000000`00000050 : 0xfffffa80`030d9000
fffff880`03dcfae0 fffffa80`030020f8 : fffff880`049c5ce4 fffff880`03dcfb30 00000000`00000050 00000000`00000001 : 0xc000009a
fffff880`03dcfae8 fffff880`049c5ce4 : fffff880`03dcfb30 00000000`00000050 00000000`00000001 00000000`00000000 : 0xfffffa80`030020f8
fffff880`03dcfaf0 fffff880`03dcfb30 : 00000000`00000050 00000000`00000001 00000000`00000000 00000000`00000000 : nvlddmkm+0x1b9ce4
fffff880`03dcfaf8 00000000`00000050 : 00000000`00000001 00000000`00000000 00000000`00000000 fffff880`049c7b72 : 0xfffff880`03dcfb30
fffff880`03dcfb00 00000000`00000001 : 00000000`00000000 00000000`00000000 fffff880`049c7b72 fffffa80`00000000 : 0x50
fffff880`03dcfb08 00000000`00000000 : 00000000`00000000 fffff880`049c7b72 fffffa80`00000000 fffff880`04a124e0 : 0x1


FOLLOWUP_IP:
nvlddmkm+1e5027
fffff880`049f1027 493b5008 cmp rdx,qword ptr [r8+8]

SYMBOL_STACK_INDEX: 0

SYMBOL_NAME: nvlddmkm+1e5027

FOLLOWUP_NAME: MachineOwner

MODULE_NAME: nvlddmkm

IMAGE_NAME: nvlddmkm.sys

DEBUG_FLR_IMAGE_TIMESTAMP: 4cb9dd0e

STACK_COMMAND: .cxr 0xfffff88003dcf0f0 ; kb

FAILURE_BUCKET_ID: X64_0x3B_nvlddmkm+1e5027

BUCKET_ID: X64_0x3B_nvlddmkm+1e5027

Followup: MachineOwner
=======================================================

After I shut it down and booted windows again, I realized that Windows Aero was no longer running. Furthermore, my mouse cursor was invisible. A couple reboots did not solve this situation.

I also ended up with a second stop error while trying to navigate my way without a visible cursor:

MEMORY_MANAGEMENT (1a)
# Any other values for parameter 1 must be individually examined.
Arguments:
Arg1: 0000000000003452, The subtype of the bugcheck.
Arg2: 000000013fecd000
Arg3: fffff70001082dd0
Arg4: 0000000000000000

Debugging Details:
------------------


BUGCHECK_STR: 0x1a_3452

CUSTOMER_CRASH_COUNT: 1

DEFAULT_BUCKET_ID: VISTA_DRIVER_FAULT

PROCESS_NAME: itype.exe

CURRENT_IRQL: 0

LAST_CONTROL_TRANSFER: from fffff80002cf1073 to fffff80002c7d740

STACK_TEXT:
fffff880`027b9068 fffff800`02cf1073 : 00000000`0000001a 00000000`00003452 00000001`3fecd000 fffff700`01082dd0 : nt!KeBugCheckEx
fffff880`027b9070 fffff800`02c5068a : fffffa80`0479e6a0 fffffa80`00000000 fffff8a0`000000fc fffff880`00000000 : nt! ?? ::FNODOBFM::`string'+0x339e3
fffff880`027b9920 fffff800`02f62dcf : fffff8a0`01a53590 fffff880`027b9c20 00000000`00000000 fffffa80`047a3060 : nt!MmCleanProcessAddressSpace+0x96
fffff880`027b9970 fffff800`02f3b635 : 00000000`40010004 fffff8a0`01a78d01 000007ff`fffde000 fffffa80`04808b60 : nt!PspExitThread+0x92f
fffff880`027b9a30 fffff800`02c5a1db : 00000000`0021b2a0 00000000`00000000 00000000`00000001 00000000`00020019 : nt!PsExitSpecialApc+0x1d
fffff880`027b9a60 fffff800`02c5a620 : 00000000`0021b5b0 fffff880`027b9ae0 fffff800`02f3b74c 00000000`00000001 : nt!KiDeliverApc+0x2eb
fffff880`027b9ae0 fffff800`02c7ca37 : fffffa80`047a3060 fffff880`027b9ca0 00000000`0021dfc8 00000000`00000004 : nt!KiInitiateUserApc+0x70
fffff880`027b9c20 00000000`77920caa : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : nt!KiSystemServiceExit+0x9c
00000000`0021b238 00000000`00000000 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : 0x77920caa


STACK_COMMAND: kb

FOLLOWUP_IP:
nt! ?? ::FNODOBFM::`string'+339e3
fffff800`02cf1073 cc int 3

SYMBOL_STACK_INDEX: 1

SYMBOL_NAME: nt! ?? ::FNODOBFM::`string'+339e3

FOLLOWUP_NAME: MachineOwner

MODULE_NAME: nt

IMAGE_NAME: ntkrnlmp.exe

DEBUG_FLR_IMAGE_TIMESTAMP: 4c1c44a9

FAILURE_BUCKET_ID: X64_0x1a_3452_nt!_??_::FNODOBFM::_string_+339e3

BUCKET_ID: X64_0x1a_3452_nt!_??_::FNODOBFM::_string_+339e3

Followup: MachineOwner
---------

===========================================
First one blames NVIDIA, the second, the NT Kernel.

Since I feel that it will remove a lot of variables with all programs and things I've installed, I will proceed with reinstalling windows and install drivers as per Tubesandwires' instructions.


Report •

#52
December 3, 2010 at 11:45:10
When trying to install Windows, as the Starting Windows screen appeared, I ran into a IRQL_NOT_LESS_OR_EQUAL (0x0000000A)

Quick internet search tells me this error may be caused by an overheating CPU, bad memory, or bad device drivers.

Did it again. Same error. Will test memory.


Report •

#53
December 3, 2010 at 12:33:36
Flash the BIOS.

Also try borrowing or begging a stick of different RAM to try.

Boot directly into the BIOS screens and see what the processor temps are. I doubt the processor temps would be an issue on a cold boot.


Report •

#54
December 3, 2010 at 13:47:25
DON'T flash the bios ! That's a last resort !

That error is not necessary cause by "bad" ram - it has many causes - if your ram still tests okay with memtest86+, that's probably NOT the cause.

If new ram is "bad", assuming the settings for it are are correct - the settings the bios is using for it by default, or the settings it has been set to by the user ithe bios is using, are compatible with it, it's immediately "bad", when you first install it.
Other than that, "bad" ram is extremely RARE !
When people get ram errors when they test it for ram that previously worked fine in the same mboard, almost always the ram has a poor connection in it's slot(s), and there's actually nothing wrong with it.
If the ram is not 100% comptible with using it in the mboard, or if the settings the bios is using for ram that is compatible with it are not compatible, the ram is likely to fail a ram test, even when there's nothing wrong with it.

The ram passing memory tests doesn't necessary indicate the ram is working properly with the operating system - e.g. your incorrect ram settings in the bios didn't cause errors with memtest86+ - but usually if the ram tests fine, it's working fine in the operating system too.


Report •

#55
December 3, 2010 at 15:06:03
The issues you describe, and I have read over every bit of this post, seem awfully familiar. Look here.


I have used a bios flash many times and never had it cause an issue. I understand the problems it can cause but given your issues I recomend it.

Likely


Report •

#56
December 3, 2010 at 15:11:25
May also try running speedfan and check your voltages.

Likely


Report •

#57
December 3, 2010 at 16:24:31
Thanks for weighing in Likely.

I haven't booted into Windows today yet, so I'll check that speedfan program later.
CPU temp in between memory tests: 40C
There's a memtest in progress right now, but I do recall when I was checking the CPU temp that the CPU voltage was around 1.4V and the memory was around 1.72V or 1.73V

So results of memory test: The stick I had been using, fails consistently with about 25-40 errors on test 5 at the same range of addresses. I did 2 passes on all 4 slots, same result each pass.

The other stick that I removed earlier in this thread has currently passed the first pass. I will test it again in another slot to make sure.

So, opinions on what's going on with the RAM? And where to go from here?

I figure I'll try to reinstall windows again with the 'good' stick.


Report •

#58
December 3, 2010 at 20:27:35
So I reinstalled Windows without trouble, this time using the other memory stick.
I installed the chipset, LAN, RAID, Sound, and graphics drivers, one at a time, shutting down and restarting each time for good measure.

I then installed SpeedFan. I checked the +3.3V, +5.0V, and +12V readings off SpeedFan, since I never did see those values in the BIOS.

I don't know how SpeedFan works, but the voltages seem very, very off.

Vcore1: 1.02V
Vcore2: 3.09V
+3.3V: 1.73V
+5V: 6.85V
+12V: 4.74V
=============
-12V: -4.21V
-5V: -0.43V
+5V: 5.43V
Vbat: 3.18V

I don't think these are correct. I'd probably have to go in with a multimeter.

So you don't have to go look for it, my PSU is:
APEVIA WIN-500XSPX 500W

Currently, I have a NVidia GeForce GT 220 graphics card, a 500GB Seagate SATA HD, and a LG SATA CD/DVD drive hooked up.

And no blue screens so far.

EDIT:
Went in with a multimeter.
Read voltages of 3.39V, 5.16V, and 12.08V for my +3.3, +5, and +12V wires.


Report •

#59
December 4, 2010 at 06:35:43
You usually don't need to use Speedfan or similar.

Most bioses will show you the current voltage, fan rpm, and temp readings in it somewhere - e.g. hardware monitor.
Take a look at those - they are likely to be more accurate than in SpeedFan, if there's nothing wrong with the mboard.

The readings for +3.3v, +5.0v, and +12.0v are supposed to be within 10% of the nominal values.
If those values for what is supposed to be +3.3v, +5.0v, and +12.0v are similar to the readings from SpeedFan , in that case, going by your multimeter readings ......Your mboard is damaged !!.

You could try connecting another PS to see if that changes anything, but I suspect it IS your mboard that's damaged.


Report •

#60
December 4, 2010 at 10:31:56
I have checked every menu and submenu item on this BIOS, AMI BIOS v2.61. I can only see the +5V reading (5.16V in this case). I see fan rpm, I see CPU and Memory Voltages, I see CPU temp and System temp, but none of the other voltages in question.

Since the +5V reading seems okay, I think SpeedFan is just reading some other value unrelated to the aforementioned voltages.


Report •

#61
December 4, 2010 at 11:13:12
Try using CPUID instead. I have gotten bad readings with speedfan in the past. May be that the sensor type needs to be set in speedfan, not sure.


Report •

#62
December 4, 2010 at 11:37:01
"The readings for +3.3v, +5.0v, and +12.0v are supposed to be within 10% of the nominal values. "
Maybe I misread something here but it seems the multimeter readings all show within 10%. I have had issues with speedfan as well depending on sensor settings. Admittedly I was reaching a bit with that but seeing how a meter was used rather than just a program I am inclined to follow the meter. With speedfan was I was looking for was the voltage the memory was set at. I have found that the bios is not always dead on. I am not saying that speedfan is I just wanted a second opinion over the bios.

Likely


Report •

#63
December 4, 2010 at 12:18:50
The current voltage readings in the bios Setup have almost always been tweaked to be quite accurate.

SpeedFan makes assumptions depending on the voltage monitoring chip or the same built into one of the main chips - it can be a little off, but not that far off.

The bios and Speedfan are interpreting the same information. SpeedFan can't find the info if the hardware monitoring capability is not built into the mboard. If a sensor isn't hooked up to a particular monitoring chip input, SpeedFan's reading will be incorrect but it never change in that case.

However, in this case, maybe that's what that voltage related fix bios update OtheHill keeps mentioning is about. There may be more info in the release notes for that fix where you get the manufacturer's bios updates for your mboard.
Newer bios updates usually have all the fixes of previous versions, unless a previous fix had problems.
If there's something wrong with that in the bios, then it's likely SpeedFan will report similar.

Otherwise, the only times I have seen where the +3.3v, +5v, and +12v readings in the bios are more than 10% out and the PS is NOT defective was when the mboard was damaged.


Report •

#64
December 4, 2010 at 13:17:59
"I have checked every menu and submenu item on this BIOS, AMI BIOS v2.61. I can only see the +5V reading (5.16V in this case). I see fan rpm, I see CPU and Memory Voltages, I see CPU temp and System temp, but none of the other voltages in question."

"AMI BIOS v2.61" is the "overall" bios software version. It's a software "framework" that is modified by or for the mboard manufacturer (or brand name system builder) to suit the specific mboard model. Lots of different makes models have that same AMI BIOS v2.61 overall version. Whoever makes the SPECIFIC bios versions for the mboard model chooses what to show you the the bios Setup.

I took a look at the Bios Setup settings in the manual for your mboard model. You're right, the +3.3v, +5.0V, and +12.0v readings are not there.
That's BOGUS of ECS to do that ! I have seen mboards as old as made in the mid 90's that have those voltage readings, and almost all mboard manufacturer model's bioses for mboards that have monitoring capability show those voltages in the bios ! (Brand name system bios versions often show you less info than the mboard manufacturer's bios versions, especially on laptops and netbooks.)

"Since the +5V reading seems okay, I think SpeedFan is just reading some other value unrelated to the aforementioned voltages."

I think what's going on here is , as I said in my last post....

"If a sensor isn't hooked up to a particular monitoring chip input, SpeedFan's reading will be incorrect, but it never change in that case."

The specific bios versions will not show readings for hardware monitoring inputs that are not hooked up to a sensor, but SpeedFan doesn't "know" which inputs have no sensor hooked up to them.
.......

I looked at the bios updates for this model.

All it says for the latest version is:

"1.Updated Adjust memory voltage rule."

Not enough info.



Report •

#65
December 4, 2010 at 14:03:35
>>Try using CPUID instead.
Yeah, same voltages show up there too.

>>With speedfan was I was looking for was the voltage the memory was set at. I have found that the bios is not always dead on.
Ah. Well, if you want the values for memory from the BIOS, the BIOS reads out at 1.728V, since we set the RAM voltage to 1.70V earlier in this thread.

>>Whoever makes the SPECIFIC bios versions for the mboard model chooses what to show you the the bios Setup.
mmm I see. Makes sense.

>>SpeedFan can't find the info if the hardware monitoring capability is not built into the mboard.
Seeing that there is no BIOS readout of the +12V and +3.3V, I agree. The functionality doesn't exist and SpeedFan and CPUID are spitting out erroneous values. +5.0V does exist in the BIOS, known as Vcc5V. It matches up with my meter reading, so I do believe that's what it is.

Current status:
So far, running on the single 'good' stick. Drivers installed, installed the browsers, installed anti-virus. Even installed the World of Tanks game and gave it a go for a few hours. Had a cold boot this morning, no problems. So far so good.

I purchased a pair of Kingston RAM I saw listed in the motherboard manual as certified to be compatible with the motherboard. 1.5V, CAS Latency of 9.

It's beginning to look like this was a memory problem. Not sure what the exact cause is, however. Poor RAM, bad mobo, incompatibility, i dunno.


Report •

#66
December 4, 2010 at 14:52:09
Refer back to #10 above for my opinion on that.

I might mention at this point that ECS built a reputation in the enthusiasts community by building boards that got the job done for not the highest cost. That said, they have always had more than their share of problems with quality control. Back in the Slot A AMD days ECS had a string of boards that would not hold CMOS settings. The issue was never really solved completely. For those reasons, I stay away from that brand. Along with Jetway, and a couple of others.

I am surprised your board is not monitoring all the voltages.

Did you email ECS about what to expect with that BIOS update?


Report •

#67
December 4, 2010 at 14:58:27
The accuracy of multimeters varies considerably. The cheaper it was to buy, the more likely it's not as accurate, especially for resistance related (e.g. ohms, diode tests, transistor tests, etc.) and fractional voltage measurements.
A really cheap multimeter may not be as accurate as the current reading in the bios Setup.

"I purchased a pair of Kingston RAM I saw listed in the motherboard manual as certified to be compatible with the motherboard. 1.5V, CAS Latency of 9."

You WILL have problems if you mix ram modules that are specified to run at different voltages. If "by SPD" or similar is used in the bios, which is usually the default, the bios will set the voltage for ALL the ram modules to the LOWEST voltage of the installed modules.
Sometimes you can custom set the ram voltage in the bios to that of a higher voltage, but that's not always do-able.
E.g.
The Kingston ram can tolerate 1.50v + or - .075v, but custom setting the bios ram voltage to 1.65v to accommodate the OCZ ram will fry the Kingston ram.

In this case, you could custom set the voltage to 1.65, the Kingston ram can tolerate 1.50v + or - .15, but the Kingston ram would get hotter.


Report •

#68
December 9, 2010 at 19:17:00
>>That said, they have always had more than their share of problems with quality control.

When I get a proper job, I should be able to build more than just a budget rig, but for now, I'll be sticking with this.

>>Did you email ECS about what to expect with that BIOS update?
Here's their not-so-useful response:

ECS Support(USA) Posted : GMT 2010/12/09 17:53:41
Dear Valued Customer,

Unfortunately we do not have any further information about this BIOS release.
However I do not see this model of memory listed for this model motherboard. As you can see on the following link:

http://download.ecs.com.tw/dlfileec...

So please try to use one of the model listed on the QVL list for this board or you can check with the Memory manufacture to se which modeles of their memory they recomend for this model motherboard.


Thank you


>>You WILL have problems if you mix ram modules that are specified to run at different voltages.

I don't plan on mixing them. I've had enough troubles with memory that I'd rather not risk giving myself MORE troubles.


That said, I have just now installed 4GB of RAM that was listed in the manual, as they linked me. I also switched out the PSU to cover my bases. The other (really cheap) PSU was getting on about 2 years old, anyway.

Well, barring any unforeseen future problems, I think I'm good.
Thank you all for helping me out. And I think I came out of this a bit more informed about the innards of my computer.


Report •

#69
December 9, 2010 at 21:42:18
Hope it runs steady now.

Report •

#70
December 10, 2010 at 02:42:39
sleepyd

Quoting from #68 above:
"When I get a proper job, I should be able to build more than just a budget rig, but for now, I'll be sticking with this. "

I've been wanting to know for a long time what a "proper job" means.

I enjoyed your thread and learned much from it. Thanks. Good luck.

And while I discovered last night through a diagnostic test that I may have to have a little funeral for my hard drive - just waiting for confirmation from Tubes and OtheHill - (ref. thread "Help me resurrect a throw-away MS-6714, #51), I'll also believe in my dream of building "more than just a budget rig" one happy day, while sticking with what I have, and while wondering if ever I'll get a "proper job" (or want one).


Report •

Ask Question