Computer SHUTDOWN while Gaming (complete power off after ~20-40 mins g

[Guest Ryn]

My friend had a similar problem before with his board, I believe it's an Asus. The issue was on the PSU, not adequate enough to power his board and peripherals. I would suspect that your problem would be more on the PSU, but check your RAM too just to be sure that it is out of the equation.

[OceanSeasforMe]

Evilj, I am with the PSU train of thought. I would replace it first.

 

Might check for fluctuations on your rails, but *( I am guessing you are useing MbM5) with monitoring, (+ active logging) you would see if it was a voltage spike or dip.

 

If no mbm5... then check with a multimeter and see what voltage your rails are actually carrying from the psu.

 

Also, just for kicks, run some diagnostics on your hard drives.

[eviljelly]

Ok, thanks for the suggestions.

 

I ran memtest, looping test #5 at stock settings, with one strip of ram in the orange slot furthest from the cpu and came up with these results:

 

a) crucial ballistix ddr500 - both strips of this dual set produced multiple errors after a few minutes.

 

They also gave me a couple of bluescreens in winxp when i tried them, so i think they are faulty. RMA.

 

B) a-data vitessa ddr500 512mb strip#1 - no errors but caused the shut down problem within 10 minutes, from memtest.

 

c) a-data vitessa ddr500 512mb strip #2 - ran memtest for 30mins with no errors. I decided to try a game, and had it rendering 3d for over two hours. I left it overnight with the game still playing, but the pc had shutdown by morning.

 

d) Infineon pc3200 w/ akasa heat spreader 512mb - ran memtest for a couple of iterations - tried a 3d game and crashed within 20mins.

 

I will try memtest again and leave it on overnight this time.

 

Can anyone deduce anything from these results?

 

Does it appear more to be a PSU problem or a RAM problem?

 

I am running memtest with active logging and the 12v rail goes from 12.34 idle to 12.02 under load - with no obvious spikes or dips.

 

Thanks again for any help.

[eviljelly]

OK - I memtested the one working stick of a-data. It was fine for over an hour, so I booted back into windows and ran a 3d game for over 3 hours! The longest i've ever seen it in fact.

 

I then went back into memtest and left it looping on #5 all night. The pc had shutdown by the morning.

 

I'm guessing that if one stick of ram is a lot more stable than another - even though it itself is ultimately unstable too - means that this is more likely a bad memory issue than a psu issue?

 

I think i'll RMA both sets of ram and order some OCZ Gold PC4000 2x1gb.

 

I have seen a few threads where people haven't been able to get them running in dual mode, but hopefully they are few and far between.

 

Thanks for the help and i'll let you know if a brand new memory replacement cures my woes.

[SloppyG]

Hey Ralphie, you have the almost the same psu as me..

 

i have the 530w: http://www.tagan.com/pages/products/easyco..._tg530_u15.html

 

is it really possible that a brand new tagan psu has a fault like this?

 

 

My powerstream had some cold boot issues and seemed a bit weird after long gaming sessions in BF2. OCZ RMA'd it for me and the replacement is dead solid, BF2 ran for 8hrs straight and primed all night. There may just be something wrong with the PSU.

 

My Antec Neopower 480w won't even boot my expert and I thought that was a good unit. I'd have rma'd it but was talked into a replacement

[eviljelly]

If the memory doesn't work I'll be contacting Tagan for an RMA ASAP LOL

 

Just the excited wait for Postie now and new pressies - luscious 2gb ram sticks mmmm :-D

[eviljelly]

Ok, the memory is installed and running at stock 200mhz..

 

The early results looked positive with 3dmark06 looping it's demo for over 2hours.

 

Unfortunately Battlefield2 caused a shutdown within 10mins of running.

 

Would people concur that it is most likely the PSU then?

 

Is there any likelihood it's the CPU?

 

Thanks again.

[blinker0]

Can you list all your settings in the Bios Genie.. And Memory configuration page

[bryanberlin]

hmm I also have had the same issue but mine has only shut down 3X over 100+ hours of gaming I suspected PSU however!! check sig for commonalities

[duckman]

My shutdowns were with a ENERMAX Liberty 620 that would fail under load. Replaced with an new one and all is fine.

[eviljelly]

ok here are my bios settings - they are basically stock:

 

FSB Bus Frequency............................. 200

LDT/FSB Frequency Ratio....................... Auto

CPU/FSB Frequency Ratio....................... Auto

PCI eXpress Frequency......................... 100

 

CPU Voltage .................................. Startup

LDT Voltage .................................. 1.20v

ChipSet (NF4) Voltage ........................ 1.50V

DRAM Voltage ................................. 2.60v

+0.03 if not 3.2V ............................ disable

 

Memclock (DRAM Frequency) .................... By SPD

1T/2T Timing (Command Per Clock).............. Auto

CAS Latency (Tcl)............................. Auto

RAS# to CAS# delay (Trcd)..................... Auto

Min RAS# active time (Tras)................... Auto

Row precharge time (Trp)...................... Auto

Row Cycle time (Trc).......................... 12T

Row refresh cyc time (Trfc)................... 24T

Row to Row delay (Trrd)....................... 3T

Write recovery time (Twr)..................... 3T

Write to Read delay (Twtr).................... 2T

Read to Write delay (Trwt).................... 3T

Refresh Period (Tref)......................... 3120

DRAM Bank Interleave.......................... Enabled

 

DQS Skew Control.............................. Auto

DQS Skew Value................................ 0

DRAM Drive Strength........................... Auto

DRAM Data Drive Strength...................... Auto

Max Async Latency............................. Auto

DRAM Response Time............................ Normal

Read Preamble Time............................ Auto

IdleCycle Limit............................... 256

Dynamic Counter............................... Disable

R/W Queue Bypass.............................. 16x

Bypass Max.................................... 4x

32 Byte Granularity........................... Disable (4 Bursts)

 

Am going to set it to OCZ recommended settings from bleeding edge forums found here:

http://www.bleedinedge.com/forum/showpost....973&postcount=4

 

and will let you know results. Thanks.

[eviljelly]

Ok so i set it to this: (taken from BleedingEdge)

 

FSB Bus Frequency............................. 250 MHz

LDT/FSB Frequency Ratio....................... Auto

CPU/FSB Frequency Ratio....................... 8

PCI eXpress Frequency......................... 100

 

CPU Voltage .................................. 1.400v

LDT Voltage .................................. 1.30v

ChipSet (NF4) Voltage ........................ 1.60V

DRAM Voltage ................................. 2.60v

+0.03 if not 3.2V ............................ enable

 

Memclock (DRAM Frequency) .................... 200

1T/2T Timing (Command Per Clock).............. 1T

CAS Latency (Tcl)............................. 3T

RAS# to CAS# delay (Trcd)..................... 4T

Min RAS# active time (Tras)................... 8T

Row precharge time (Trp)...................... 3T

Row Cycle time (Trc).......................... 11T

Row refresh cyc time (Trfc)................... 16T

Row to Row delay (Trrd)....................... 3T

Write recovery time (Twr)..................... 3T

Write to Read delay (Twtr).................... 2T

Read to Write delay (Trwt).................... 3T

Refresh Period (Tref)......................... 3120

DRAM Bank Interleave.......................... Enabled

 

DQS Skew Control.............................. Auto

DQS Skew Value................................ 0

DRAM Drive Strength........................... Normal 3 or 4 (called Level 6 or 8 in other bios')

DRAM Data Drive Strength...................... Level 2

Max Async Latency............................. 8ns

DRAM Response Time............................ Normal

Read Preamble Time............................ 6ns

IdleCycle Limit............................... 256

Dynamic Counter............................... Disable

R/W Queue Bypass.............................. 16x

Bypass Max.................................... 7x

32 Byte Granularity........................... Disable (4 Bursts)

 

Of course this was running a San Diego 3700+ which should be 2.2ghz stock instead at 2ghz.

 

The result was it did feel a lot more stable. I was actually able to game for over an hour or so - so I was encouraged.

 

However when i ran 3dMark06 overnight it had still shut down by morning.

 

I also got a couple of bluescreens - which made a change from total shutdown, an improvement even?- and a further shutdown an hour into gaming the following day.

 

So taking heart in seeing the relationship between fiddling and stability I continued tweaking.

 

Since then I have clocked my cpu to 2.5ghz @ 1.44v - which is stable in superPi, prime95 and occt - and I have also increased voltages across the board including - (LDT @ 1.4v, DRAM @ 2.8v, chipset @ 1.7v).

 

The problem is that it's crashing within 10mins of gaming again, so I've obviously reduced stability with a setting or two. Maybe the cpu i guess. I will reduce the settings later and try some more tests.

 

So it certainly appears that changing memory and fiddling with bios settings does affect the stability - but ultimately the shut-down issue persists, it just changes from 5mins to 5 hours.

 

The weird thing is that it's rock solid in 2d and windows. I can work on Flash, Photoshop and Dreamweaver with Winamp and Azureus running in the background all day long with not so much as a flicker of memory errors or whatnot.

 

I have emailed Tagan twice to see what they say and try and get an RMA but as yet no reply.

 

Can any uber-talented pc guru maybe point me to an obvious error in bios settings or offer further suggestions?

 

Should stability really be *so* difficult to achieve if one is leaving everything on stock?

 

Does this further point toward the PSU as being the culprit?

 

Thanks again for any help.