Building a new PC - having hardware issues

[Engels]

Case:	Aluminus Ultra
Power:	ThermalTake Toughpower 1200W
Motherboard	Asus P5E3 Deluxe
CPU:	QX9650 (Core 2 Extreme)
RAM	OCZ Platinum 4096MB PC10666 DDR3 1333MHz (forgot to mention that I ran the RAM test on the Linux DVD, took an hour and a half - passed)
GPU:	EVGA 9800GX2
HD:	Western Digital SE16 500GB
Optical Drive:	Sony DRU 190A (simple drive, says DVD+-R 20X on the box)
Sound Card:	Sound Blaster X-Fi

[Trippy]

Interesting, the recommended memory slots for 2 DIMMs is opposite from the P5E3 board (that one recommends the "black" slots). Wonder why those are reversed. Don't see where it's telling you to put them in B1 and B2, though. That would disable dual channel memory.

Edit: disable

[Typhon]

Could be I'm reading the manual wrong (done a ton wrong already, so it wouldn't be a surprise).  Here's what the book says (yellow is from the book, white or blue is me, the table isn't cooperating, color-wise):

You may install 512MB, 1GB, 2GB unbuffered ECC, non-ECC DDR3 DIMMs into the DIMM sockets.

Recommended memory configuration (shows a graph, which I'll try to reproduce)

Channel A - DIMM_A1 AND DIMM_A2 - (these are the ones nearest the CPU, the "2" slots are black)
Channel B - DIMM_B1 AND DIMM_B2 - (these are the ones nearest the CPU, the "2" slots are black)

	Sockets
Mode	DIMM_A1	DIMM_B1	DIMM_A2	DIMM_B2
Single Channel	-	-	-	populated-
Single Channel	-	populated	-	-
Dual Channel (1)	-	populated	-	populated
Dual Channel (2)	populated	populated	populated	populated

[/color]

The package that the RAM came in says it's Dual-Channel.  The chart seems to say that I shouldn't be putting dual-channel RAM in as a single chip?  Am I reading that right?

How I read it last night was:  If I have a single chip, I should put it either in B1 or B2.  If I have two chips, I should be filliing the B channel first.

Filling the B-channel first causes issues (as does filling the A channel first).  Filling the "2" slots first has been the only configuration that works with two chips in - but as mentioned I was having issues with WoW.  Currently I have the DIMM_B2 slot filled, and that is working the best so far (no issues with WoW).

[schild]

Hmmmmm.

Orange and black? Shouldn't the orange match the black. As in, fill in the two orange (which I would assume is for dual-channel) and leave the blacks empty.

I could be wrong, but let's say I doubt it since I populated an ASUS board correctly 2 days ago that had 6 slots.

[schild]

Also, at least you aren't having storage issues:



What the fuck does that even mean?

[Trippy]

Could be I'm reading the manual wrong (done a ton wrong already, so it wouldn't be a surprise).  Here's what the book says (yellow is from the book, white or blue is me, the table isn't cooperating, color-wise):

Download a newer manual.

[Typhon]

Also, at least you aren't having storage issues:



What the fuck does that even mean?

Means you need to run scan disk so you can delete siht that is in that directory, but not properly in the FAT.

[Trippy]

Also, at least you aren't having storage issues:



What the fuck does that even mean?

It means something has that folder locked. Try something like this:

http://ccollomb.free.fr/unlocker/

Or just reboot your machine.

[Typhon]

Schilds Directory Issue:  Well, I recently had the same issue with the Flash drive, and it was idenpendent of what machine I had it plugged into, so I assumed what I typed above - the FAT was fucked up, something was in the directory, but was not properly in the index (or vice-versa, but if it was in the index, it would appear in a directory-viewing application, such as explorer).

P5E3:  You're right - new manual says to populate the orange first.  I'll try that and see if WoW fails.

[schild]

Enjoy WoW, I guess. It should run fine.

[Typhon]

It did, thanks again to everyone.

[Miguel]

One of your sticks of RAM is bad/flaky.  Keep in mind that the OS goes resident (including all of the memory-mapped registers, some portions of the video BIOS, etc) into fixed addresses in the addressable space of the CPU.  Hence memory testing routines do not actually cover all of the RAM located on the stick.  In other words, if you do read-write testing and you overwrite the code of the test which is in RAM, you cannot test any further and your system dies.

Anytime you run into issues like this, swap the sticks and re-run the memory test.  This will move the memory 'hole' to the other stick and allow you to test the first one (which was previously storing the OS code, parameters, etc).

[Engels]

One of your sticks of RAM is bad/flaky.  Keep in mind that the OS goes resident (including all of the memory-mapped registers, some portions of the video BIOS, etc) into fixed addresses in the addressable space of the CPU.  Hence memory testing routines do not actually cover all of the RAM located on the stick.  In other words, if you do read-write testing and you overwrite the code of the test which is in RAM, you cannot test any further and your system dies.

Anytime you run into issues like this, swap the sticks and re-run the memory test.  This will move the memory 'hole' to the other stick and allow you to test the first one (which was previously storing the OS code, parameters, etc).

What if you're using a bootable ram tester, such as memtest86, off a CD or floppy? I know that it has a small 'OS' to enable the program to execute, but would it have the same effect you describe here?

[Selby]

What if you're using a bootable ram tester, such as memtest86, off a CD or floppy? I know that it has a small 'OS' to enable the program to execute, but would it have the same effect you describe here?

I've had flaky RAM that passed all of the various memtest utilities out there.  It would randomly reboot\crash with no warning every few hours, didn't necessarily need to be heavy usage or anything.  It just had an area that it didn't like using.  The utilities are useful, but they aren't the end-all of RAM functionality testing.  The only thing that worked was putting another stick in the machine.  I put that stick into 6 different machines with various motherboards for several months as an experiment back when I ran a computer lab and they all had problems with it despite passing several memory checker programs.  Ended up being a conversation piece on a desk.

[Typhon]

The problem wasn't with the ram, it was with the manual and the person reading the manual (and not getting the fixed version of the manual from asus.com).

[Miguel]

What if you're using a bootable ram tester, such as memtest86, off a CD or floppy? I know that it has a small 'OS' to enable the program to execute, but would it have the same effect you describe here?

In short, yes.  Regardless of the medium where the files are stored, the testing program will still go resident into RAM and be run from there.

It's just good practice to always swap sticks and run the test again, if you are using more than one stick.

As for the mobo design, that's just awful.  You should be able to put any stick in any slot and it should work.

[Typhon]

A new exciting problem:  system locks up hard, a squealing sound usually coming from the headphones.

History:
First time it happened it was early in the morning and I was browsing in IE 7.  I am using a USB trackball (LogiTech Trackman marble, pretty old), and I flicked the scroll wheel a bit harder then usual, making it spin fast.

Second time and every subsequent time it happend I was using my Nostromo n52 to launch my druid into the air in WoW.  When I do this, I press a couple buttons simultaneously (and quickly).  It's happened 3 more times, all with this same setup.  I play Hellgate, COD4, CoX all with the nostromo, so far no problem.

Anyone having similar lock-up problems with USB devices and Vista 32-bit?

[Trippy]

What happens if you disable the sound in-game?

[Engels]

I have one USB port -on the case- (not on the motherboard back panel) that will short out the whole machine into a lock up when I use certain periferals, wether its my digital camera or my joystick. I've checked the wire connection to the motherboard from the case, but it looks well seated, but I think that over the years, one of the cables must have become unstuck from the connector or something.

Is the USB port you are using for the device on the motherboard backpanel or one of those on the case you have to connect to the motherboard through a cable?

[Typhon]

n52 --> Belkin Hub --> Motherboard USB.
Mouse --> Motherboard USB

I just moved both (n52 is in a new slot, hub is in a new usb slot), I'll se if that changes anything.

I also try disabling WoW sound and see if I can get it to repeat.  It's confusing to me that none of the other games are having this issue.  (I also haven't had a repeat while browsing).  For reference, I'm using a Creative X-Fi Gamer card, onboard sound is disabled through the BIOS.

Lockup isn't like a normal looping lockup (were a sound-bite will loop), it's more like the last note that was playing is amplified and never stops.

For some reason I feel like I want to blame Vista, but I'm not sure why.

[Typhon]

Moving the USB connectors to other ports may have worked - not a single lock-up last night.  Seems like they are fine for harddrives, but not so fine for spastic finger-flinging.  Thanks much for the suggestion.

Edit: I got spastic while spelling spastic

[Reg]

I hade the same kind of problems when I first connected my USB headset. It caused all kinds of bizarre behaviour. Moving to a new USB port solved all of them.

[Typhon]

100% sure it was a specific set of USB ports as I locked up hard the minute I went to type on the keyboard while playing WoW last night (keyboard was only usb device left in those ports), moved them and haven't had an issue since.

Issue is kind of odd, because in all other games in that configuration the ports/devices worked just fine (which leads me to believe it's a driver issue, and not (only) a hardware issue).

[Trippy]

It's possible it's a driver issue. Some motherboards use multiple controllers to support all their USB ports. E.g. the main chipset might only support 4 USB ports while a secondary controller is used to handle another 4.

[Cylus]

I've always had USB issues with my deluxe.  Computer loves to hard lock if I use a USB device during the initial Express Gate screen.  Haven't had any issues with my memory but all 4 banks are filled.  Best of luck, in any case!

[Typhon]

Well I was wrong, it wasn't a specific port that is having issues because I locked up again last night in COD4 (although I seem to have less issues through the recently-used ports).  When I did, I notice that the Belkin hub that I was using to connect the nostromo was hot, so I switched them out (I have two of these hubs.  They are both powered, the hot one was the main, another hub and the nostromo were connected into main, with a 6-in-1 connected to the cool hub).

So I completely removed the hot hub and run only the nostromo through the cool hub (which seems like it continues to be pretty cool, only getting slightly warm).  I'd be very happy if this were the issue, cause it would mean I've found the problem, and the ports on the mobo were probably ok.

[Reg]

Your adventures have convinced me not to attempt to build a PC from scratch. This kind of stuff would have driven me wild with frustration long ago. It's worth spending money not to have to go through it.

[rattran]

His case was sort of an outlier. Normally, everything goes together smoothly, the system works, and you save a pile of money.

Other than laptops, I can't see myself ever buying a pre-made.

[Typhon]

I continue to be plagued by the USB prompted lockup.  This morning while going through all things Vista I found that I hadn't actually completed the sp1 install like I thought I had, so I completed that.

Googling for others with the same issue it seems like Vista Ultimate comes up fairly often (which is what I have, cause I'm a moron who doesn't like to keep his money).  I do notice that the date of the last posts I seem to see ends in January of this year.  Again my fingers are crossed that maybe this (Vista SP1 will resolve the issue).

@Reg - I'd have to agree with rattran, with the caveat that you can avoid this stuff by doing a good amount of research and reading - I experienced many issues because I was ignorant (ordering wrong stuff and not checking and reading the latest mobo manual), some issues due to hardware (motherboard fucking up proper during bios flashing) and some due to crappy choise of operating system (just stick with XP until people stop bitching about Vista).

That said, if you can't be bothered to do all that research, it's definitely worth your while to buy a pre-built system.  I'd also advise that you don't get a high-end system, they tend to have just-release hardware and drivers that often need a bit more baking in the oven then are worth the bit of extra performance they provide.

[Typhon]

NECRO: ARISE FROM YE SHALLOW DEEP GRAVE!!!

Thought I'd update this in case someone else ever has as many issues as I did when building a new machine and this can in someway help, or is just looking for schadenfreudder (fodder for schadenfreude).

Two updates to this:

1) My instability problem was a combination of motherboard + specific set of RAM.  Both were fine when playing with other partners, the combination were less then fine.  Solution was to increase the RAM voltage and decrease the RAM clockspeed.  Finally after a year and a month I had a stable system (found this information on the Asus site, with my motherboard as a topic). 

So, if you are having random system lockups (complete lockup, not blue screen of death, just sudden and complete lockup of the system that requires a cycling of the power.  frequently accompanied by looping sound) - try increasing RAM voltage and dropping the RAM clockspeed.

2) Yesterday evening I turned on the machine and heard a long beep, followed by three short beeps.  That motherboardese for, "your video card has shit the bed".  Moved it to the other slot (motherboard supports SLI) and no luck - seems definite that that video card died an early death.  I was positive that I had registered the card with EVGA, so I went to the site to see what type of warranty they had (although I was afraid it was just 1 year.  It was 1 year, unless you had registered the product within one month, in which case it was a limited lifetime warranty.

Turns out I didn't actually register the product, I had just registered on the website.  So I got to receive EVGA email for a year... but I now have a dead card 2 months after the end of warranty.  Sure have had crappy luck with this go round of computers.  I used to be pretty religious about filling out warranties but ended up feeling like it was just a trick to get contact info for advertising.

[rattran]

Evga once declined to honor a warranty, telling me that I hadn't registered the motherboard properly on their site. No explanations as to what was improper. Which is why I won't buy evga anymore.

[Miguel]

My instability problem was a combination of motherboard + specific set of RAM

...

One of your sticks of RAM is bad/flaky.

I R WIN?? 

You wouldn't believe the shit that gets soldered onto PCB's and sold as a DIMMs in the retail space nowadays.  High quality RAM is worth the extra $$ IMHO.  This trick is finding out who sells it (almost nobody)!

Glad you got it ironed out!

[Typhon]

RAM is OCZ, which I thought was decent/high quality (it sure was expensive).  And since I've upped the voltage and dropped (1333 to 1000MHz) RAM is fine, no specific issues or defects with the RAM.  The problem that people were reporting seemed common to this motherboard and the OCZ DDR RAM.  Guessing - this RAM pulls a bit more juice then everyone else's, so it's necessary to jack the voltage up a bit to keep it stable.

EVGA is trying to give me a "one time replacement" as we speak, so I can't really complain about them either (I really didn't fill out the warranty info, I just registered on the website).

[Numtini]

My OCZ 1333 ram won't even boot at 1066 whatever I set the voltage too. Just gave up and reverted to the default 800.

[Strazos]

FYI, OCZ is trash. I have seen problems with their ram in the past. A company trying to be a big player is really frustrating sometimes, for the consumer.

I only buy Corsair, personally, but I would not argue with someone buying Crucial either. I'm sure others here can think of other reputable brands.


Also, I still don't like your PSU; ThermalTake is trash. Unfortunately, beyond SeaSonic (or Corsair units utilizing their design) or Antec, I'm at a loss to think of other decent brands.


EDIT: What I mean to say is that in my experience (note: I'm not a pro at this), PC building is one activity that really exemplifies the maxim that you get what you pay for; it is ALWAYS worth the extra money to buy the better brands. Sure, sometimes you'll be ok buying some off-brand, but I don't think it's worth the risk. I'm sure I'm not the only person who almost wanted to vomit or cry at some point when faced with a mysterious major PC problem.