random reboots after upgrading to x2 4200+ from a64 3400+

[vexingv]

Originally posted by: myocardia
There's nothing wrong with your cpu, your power supply just doesn't have the power to run what you're trying to use it to run. That's why those Ultra psu's are so cheap; they're very, very overrated in their output capabilities.

i mentioned earlier that i had 5 IDE hard drives installed for several months (now i only have 1) and that ran without any problems. i really can't see how even a dual core cpu can pull more power than 4 hard drives. also i made a build for my brother last year using this same psu (granted it was an AM2 x2 3800+ and 2x1 gb ddr2 RAM).

i hate how this $70 upgrade is going to turn into at least $120 if the problem is indeed an underpowered PSU. any recommendations on decently priced (<$50-$60) PSU's?

 

[WildHorse]

Yeah, I'm also having trouble with my 4200+ X2 bought on the Newegg deal.

Don't know if your mobo offers these vcore settings, and NO TWO rigs work the same, but the following might at least be useful as a guidepost for you. (Again, it never works to transplant the settings from one computer to another)

After a MASSIVE amount of testing, these settings seem to sort of work for me:

vcore 1.275 volts + special 113% good for htt of 242. 242 X 11 = cpu clock 2662MHz

1.30 v + 113% good for htt of 243. 243 X 11 = cpu clock 2673MHz

1.325v X 110 % good for htt of 244. 244 X 11 = cpu clock 2684MHz, but cpu temp is alrady 63 deg C under OCCT stress test.

above that, the cpu temp goes too high.

I'm holding out some hope for a little more improvement in a couple of weeks after the ARctic SIlver 5 cures a while. We'll see.

Overall, hardly better than the 3200+ Winchester that was removed, that ran so well at 253 X 10 = 2530MHz on air forever.

CONCLUSION: So overall, if I had it to do over, I would not buy the 4200+ X2.

 

[vexingv]

Originally posted by: scott
Yeah, I'm also having trouble with my 4200+ X2 bought on the Newegg deal.

Don't know if your mobo offers these vcore settings, and NO TWO rigs work the same, but the following might at least be useful as a guidepost for you. (Again, it never works to transplant the settings from one computer to another)

After a MASSIVE amount of testing, these settings seem to sort of work for me:

vcore 1.275 volts + special 113% good for htt of 242. 242 X 11 = cpu clock 2662MHz

1.30 v + 113% good for htt of 243. 243 X 11 = cpu clock 2673MHz

1.325v X 110 % good for htt of 244. 244 X 11 = cpu clock 2684MHz, but cpu temp is alrady 63 deg C under OCCT stress test.

above that, the cpu temp goes too high.

I'm holding out some hope for a little more improvement in a couple of weeks after the ARctic SIlver 5 cures a while. We'll see.

Overall, hardly better than the 3200+ Winchester that was removed, that ran so well at 253 X 10 = 2530MHz on air forever.

CONCLUSION: So overall, if I had it to do over, I would not buy the 4200+ X2.

it seems like you're trying to OC your system. i just want to run it at default settings stably. in my initial thinking, this upgrade was worthwhile as i was still using a single core and i don't plan on making any more DIY systems (apple desktop as my next system in a few years). however, this is now turning out to be more stressful than i wanted especially if i need a new psu.

btw, scott what PSU do you use in your system?

 

[WildHorse]

it seems like you're trying to OC your system.

Yes, of course!

btw, scott what PSU do you use in your system?

FSP Group (Fortron Source) FX700-GLN ATX12V V2.2/EPS12V 700W Power Supply
Newegg Item #: N82E16817104015

 

[GeezerMan]

I would think if it was the power supply it's because it's faulty, and not a design problem. I used the exact same power supply for months in my overclocked Opteron 165 at 2.6 Ghz with 2 hard drives, 7600GT video, TV tuner, creative sound card, DVD rom ,DVD burner with no problems. If I remember correctly, Jonny Guru used the same power supply for awhile in his personal rig. It's not the best power supply you can get, but it worked fine for me and others too.

 

[vexingv]

Originally posted by: GeezerMan
I would think if it was the power supply it's because it's faulty, and not a design problem. I used the exact same power supply for months in my overclocked Opteron 165 at 2.6 Ghz with 2 hard drives, 7600GT video, TV tuner, creative sound card, DVD rom ,DVD burner with no problems. If I remember correctly, Jonny Guru used the same power supply for awhile in his personal rig. It's not the best power supply you can get, but it worked fine for me and others too.

that what i think as it's been running fine for 11 months now. but it seems like everyone keeps coming back to the PSU as the main issue. so should i get in touch with ultra support? what do i tell them, "this power supply isn't meeting the demands of an X2 4200+...and i need a new one"?

i've also been looking at newegg for a psu. would either of these fit the bill and would i be better off getting these than a replacement from ultra? OCZ StealthXStream OCZ600SXS (apparently FSP is the OEM) and FSP Blue Storm II 500.

 

[Markfw]

Just to clarify, I have had many re-boot problems. Some are fixed by changing memory that seemed to test fine, others were fixed by changing PSU's that seems fine. I have problems with Corsair and Crcial memory, and SOME (not all ) Antec PSU's. Fortron Season and PC power and cooling are the three I like best. In NON-OC'ed systems, most of the parts worked fine, but in a OC'ed system, eveything has to be perfect.

My take on that, is that everyone rates their hardware (like CPU's) way lower than capable, just for this reason, padding....

 

[bryanW1995]

probably not necessary to get a multimeter, just throw in a few more mv until it's stable. I had hell getting my msi k8n plat mobo to go from an fx 55 to an opteron 180. Realistically, it is almost certainly your psu or mobo. I would lean towards the psu with that much of a drop on the 12v rail. It was very painful for me to upgrade my psu, but I've never regretted it. Earthwatts 430 or corsair vx 450 come to mind as reasonably-priced midrange psus. Throw a post over at the psu forum, john seems to know everything about those suckers.

 

[GeezerMan]

Well, my experience with software reporting on volts and temps is that it can be inaccurate. On more than one rig, I have seen .5 volt differences between what a multimeter indicated and what speedfan, or the bios showed for the 12 volt. It very well could be a board problem. Did you google the model to see if anyone had any other similar experiences? I'm sorta leaning in the direction of board or bios problem

 

[GeezerMan]

oh, how about updating the chipset drivers?

 

[vexingv]

i have the latest nforce 4 drivers. i tried to google "kn1 x2 4200 reboot" and got a few hits in forum post with similar issues, but i have yet to get any response from the 2 i have contacted. i suspected the motherboard too since ECS is a lower tier maker, but then i remembered that fry's and other places were selling x2 4200 with this kn1 board. if there were issues across the board (hehe, pun ) i would see more posts about it. i also inspected the board for leaking or bulging caps but none were found.

when i did raise cpu voltage earlier today, there was still a crash within an hour of testing. i could raise it more but i'm reluctant as i don't want to risk damaging it from the excess heat and voltage.

 

[NXIL]

Dear Vex,

that temp sure seems high to me....my Opteron 170 (sckt 939, 2.0Ghz) runs at 29C with the stock cooler....

Your new X2 has the same TDP as the 170, 89 watts.

The Venice has a TDP of 67 watts....not a huge increase, and I do not think that is enough to account for those temps.

Freezer pro 64 gets very good reviews, one reviewer notes it keeps his X2 6400 at 37c--
4800 x2 at 31-36c.....some remarks about how it is difficult to kept it clamped down properly....if the Freezer 64 is properly seated with a very thin layer of TIM with no bubbles, no impurities, etc, then it could be CPU is defective.

Hard drives: draw about 8 watts idle, 12 watts seek....so the 4 extra hard drives draw about 40W, the while X2 draws about 89W - 67W = 22W more than the single core Venice.

http://www.digit-life.com/arti.../storage/hddpower.html

Power supply review: it's OK....

http://www.ocmodshop.com/ocmodshop.aspx?a=723&p=1588

http://www.bigbruin.com/2006/xfinityapfc_7

I strongly concur with Mark, though, about PSU brands: Seasonic, PC Power & Cooling, Fortron....I especially like PCP&C, but I have not purchased one since they were bought by OCZ...hopefully OCZ will get more like PCPC rather than PCPC getting more like OCZ...but, that did not work out that way for Mercedes and Chrysler....

And as Bryan said, painful as it may be to upgrade your PSU, it is pretty much the heart of your system....

Should be something that works OK for you in here if you do decide to swap out the PSU:

http://www.newegg.com/Product/...%2c1459%2c9830&bop=And

HTH, good luck,

NXIL

 

[vexingv]

NXIL,

thanks for all the links. i did find the heat increase to be unusual, but accepted it since it's within the AMD stated 49-70 C operating temp of the cpu. the freezer pro is clamped on tightly and i followed the instructions for apply thermal paste, but i probably use a little bit to much.

i think what i may do is "buy" the antec true power 650 (seasonic) or BFG 650 psu at best buy since they have no restock fees and test everything again. if i still have issues, i can probably say it's a defective cpu and i'll have newegg RMA a replacement and test everything again. if it's the psu at fault, then i'm looking at picking up the ocz stealthxstream 600, which is actually made by FSP, or the FSP blue storm 500 wt (seasonic, PCP&C, and corsair are just a bit too expensive for me)..

 

[bryanW1995]

I think that you can get within $10-$15 of the fortron price on the cmpsu vx450. FSP has a good rep, but I think that you'd be hard pressed to get a better unit for the price than the corsair.

 

[vexingv]

Originally posted by: bryanW1995
I think that you can get within $10-$15 of the fortron price on the cmpsu vx450. FSP has a good rep, but I think that you'd be hard pressed to get a better unit for the price than the corsair.

at around $65 AR, it does price out the same with the ocz and fsp models i mentioned above. but out of curiosity, why do you suggest that one? it's only has a single +12V rail and is considered their "consumer"/budget-minded model. is this, like their HX series of psu's, also made by seasonic?

never mind, it is made by seasonic and seems to be a very good psu from the few reviews i skimmed

 

[NXIL]

i followed the instructions for apply thermal paste, but i probably use a little bit to much.

Hey Vex,

the instructions for Arctic Alumina run 7 pages:

http://www.arcticsilver.com/pd...s_aa_amd_dual_wcap.pdf

Looks like a little goes a very long way....just a tiny dab of it in the center of the CPU heat spreader--too much and it acts like a thermal insulator.

60C for a non-over clocked X2 still seems high to me....

HTH

NXIL

PS: congrats, man of a thousand posts....

 

[fetuchin]

Try graphic card drivers. Believe it or not it happens from time to time and, as a matter of fact it happened to me.
So, do uninstall nvidia video drivers, re-start and run your system with standard vga windows drivers. Do as many tests as possible and let the system run. If after around 24 to 36 hours you're still alive and running,... voilà! Then restart your computer and re-install nvidia drivers from scrathc.
If this doesn't make the trick, i'm sorry but have no other clue. (anyway believe me and do try it before saying it's nonesense)
Good luck

 

[vexingv]

i was in the middle of composing an update when the system rebooted (while being stress tested in orthos)...

so i picked up the seasonic-made antec truepower trio 650 wtt psu from best buy on the way home. i installed the new psu and x2 4200+ again. i was able to run orthos for a measly 15 minutes before the system rebooted just now. so i definitely don't think it's the psu that's at fault.

i'm thinking of maybe stress testing it one more time using an ubuntu livecd before i RMA the cpu. is there anything similar to orthos/stress prime for ubuntu/linux?

 

[CTho9305]

Originally posted by: myocardia
Set your vcore manually in the BIOS to 1.35v, the minimum for any X2 @ stock speed. If it's still restarting, it's because of your psu, which seems to be undervolting your +12v rail by a fairly large amount.

Are you an idiot? There are a bunch of X2s that run on less than 1.35V, and running CPUs at higher voltages than required is a fantastic way to wear them out early.

vexingv, if a CPU doesn't work at it's rated voltage, it's defective or you're not cooling it. It may be that your board under-volts it (a crappy voltage regulator?), but if it's not, don't over-volt your CPU. Do you have any other motherboards (or friends' motherboards) you could use to test your CPU?

 

[vexingv]

Originally posted by: CTho9305
vexingv, if a CPU doesn't work at it's rated voltage, it's defective or you're not cooling it. It may be that your board under-volts it (a crappy voltage regulator?), but if it's not, don't over-volt your CPU. Do you have any other motherboards (or friends' motherboards) you could use to test your CPU?

unfortunately, i don't have access to another 939 board. it could be possible that the board is undervolting, but it has been able to supply power for my 3400+ with no issues and that cpu has a requirement of 1.4 V. the board will be a year old in about 3 weeks and i purchased it from newegg who doesn't have anymore in stock. the only other option would be to RMA to ECS, but i'm not sure what they could do about it (board is operating fine with the 3400+ after all...) or if they even have any instock. also, i don't want to be without a pc for however long it can take for a mobo rma (took me a month to get my epox back after bad caps).

as some have pointed out, the cpu seems unusually hot (39/40's while idle; 60C+ during load) considering that it's not overclocked and i'm using a freezer pro 64. my single core 3400+ which has the same clock speed tops out at 44C during orthos testing. i think this may be a defective cpu (like how there are always some in a batch/platter that don't seem to cut it during certain stresses or clock speeds).

update: i was typing this post on my powerbook and when i looked up, my system had just rebooted itself! this time it wasn't even being stress tested as it had just reboot itself earlier and i left it idling. something's not right...

 

[imported_seferio]

Your temps are off probably due to bad mounting of the heatsink or a bad thermal paste application. I would do two things, redo the heatsink + thermal paste thing and try running on a single stick of ram. The X2 has a different/updated memory controller compared to your 3400+ and that could be causing the reboot problems.

 

[vexingv]

Originally posted by: seferio
Your temps are off probably due to bad mounting of the heatsink or a bad thermal paste application. I would do two things, redo the heatsink + thermal paste thing and try running on a single stick of ram. The X2 has a different/updated memory controller compared to your 3400+ and that could be causing the reboot problems.

b/t the 3400+ and x2 4200+, i've probably installed/reinstalled maybe 5 or 6 times now. so i think i've got the technique down to a T. though memtest showed no errors, i'll try testing w/ one stick.

btw, the system also rebooted within minutes of using an app called cpu stability test

 

[imported_seferio]

Memtest is not a guarentee that everything is fine. I've had situations where memtest doesn't fail for like 10+ hours and then I assume everything is fine and the comp still crashes sometimes. I see memtest more as a way to make sure the sticks themselves aren't bad.

 

[vexingv]

i just tried stress testing with orthos using only 1 memory dimm installed. did this for each dimm and both times the system rebooted. any other ideas?

 

[Markfw]

Do you have other memory to test with ?