Sign in with
Sign up | Sign in
Your question
Closed

RAM Unstable at stock speeds

Last response: in Memory
Share
November 16, 2010 3:02:29 PM

Hello again guys!

So, recently I got a new mobo, ram, and gfx card.

The RAM I got was some nice Mushkin DDR3-1600 2x2GB Silverline Stuff.
http://www.advantionline.com/Mushkin-4-GB-2-X-2GB-DDR3-...

Now the problem I have is that the RAM is rated at 1600 Mhz, but at first I noticed in BIOS Post that it runs at only 1333 Mhz. So I ask the guys at the store how to solve this, and they tell me to set the RAM speed at 1600Mhz manually in the BIOS. So I do this, and it works. However, the programs I use most frequently, (Firefox and Just Cause 2) at that moment, seemed to be crashing roundabout every 5 minutes. I had been running in this mode all day, and it was annoying me quite a bit. Now this morning, (it is evening now), I thought, wait a minute, when I first got the RAM, I left it at 1333 for a couple hours, and never had a problem. So you can guess what I did, I turned it back to the AUTO setting, and surprise surprise, it has been running flawlessly all day.

Now I know that you might think, your overclocking it, you need to up voltage etc, but I have checked all bios settings, voltage, timings and frequency, and they are all at the specified stock settings for the RAM dimms. What confuses me is that I am not overclocking, im restoring stock speeds, and then it crashes. I am not getting BSOD's, the programs just freeze up, and I have to use TaskMan to kill the process every 5 minutes.... Once it has happened that I needed to use the reset button.

I am so confused :cry:  help please...

Thanks in advance already!
a b } Memory
November 16, 2010 3:34:09 PM

Quote:
when I first got the RAM, I left it at 1333 for a couple hours, and never had a problem


AMD Phenom 2 965 @4Ghz
Cooler Master V8
4GB DDR3-1600 ram


Download and install CPUZ. Check the RAM speed when BIOS is overclocked to 4Ghz. and RAM transfer rate set in BIOS at 1600MHz. If the CPU is overclocked to 4GHz the RAM speed is taken along with it. So if the RAM is set at 1600MHz it is being taken up to increased speeds from there. The likelihood of speeds far above 1600Mhz are occuring and the system crashes occur.

http://www.cpuid.com/softwares/cpu-z.html
Score
0
November 16, 2010 3:44:49 PM

I will have a look once I get back to my main rig. Actually, right now the CPU is not overclocked to 4Ghz, since I got a new board, and haven't bothered upping the clocks yet, and neither bothered to update that part of my signature... probably should. Either way though, I overclock it solely by upping the multiplier and voltage. For the RAM, I have a setting that lets me choose between a bunch of speeds: 1066, 1333 and 1600 MHz.

I actually never overclock through the base clock, because with my old system I never cared about overclocking my RAM and didn't want to affect it's speeds.

Ill have a CPU-Z when I get back to main rig though
Score
0
Related resources

Best solution

a c 347 } Memory
November 16, 2010 4:26:27 PM

First, boot from Memtest let run for 2 passes - http://www.memtest.org/

BIOS, Load Defaults, CPU OverClocking {Auto} -> {DOCP}, and DRAM O.C. Profile -> {DDR3-1600MHz}.

Installation, I tend to recommend A2/B2 vs A1/B1.

Next, I would recommend the 996744 for your ASUS M4A88TD-M/USB3 http://www.mushkin.com/Memory/Blackline/996744.aspx or 996657 http://www.mushkin.com/Memory/Blackline/996657.aspx

Otherwise, exchange for 996805 http://www.advantionline.com/Mushkin-4-GB-2-X-2GB-DDR3-...
Share
November 17, 2010 1:41:00 PM

http://valid.canardpc.com/show_oc.php?id=1488077

Theres a CPU-Z for stock CPU, with RAM set to 1600 MHz in BIOS.

@jaquith, what do you mean by "Installation, I tend to recommend A2/B1 vs A1/B1. "

For as far as I am concerned, this all looks fine, however, I could be wrong..
Score
0
November 17, 2010 1:44:21 PM

Oh and I let the memtest run for 2 passes, no problem at all.

Also, I noticed that Firefox is no longer crashing so much today. I looked up on it, it seems like it might be because the version I have is known to be unstable with flash on certain machines....

Furthermore, the only thing I still have issues with are games.

One session I can run a game for hours without a problem, another moment it'll crash every 5 minutes. Not a BSOD, just a game freeze up, and depending on the game I can either fix it somehow without killing the process, or I have to kill the process. Once I've had to use the reset button for the first time since 4 years...

I ran furmark on my new card for a good 30 minutes, but everything seems fine, no artifacts. Could it be that the Catalyst 10.10 is just bad?

Graphics card is a new Sapphire HD6870: http://www.advantionline.com/SAPPHIRE-Radeon-HD-6870-1G...

Score
0
a c 347 } Memory
November 17, 2010 2:13:18 PM

N.Broekhuijsen said:
http://valid.canardpc.com/show_oc.php?id=1488077

Theres a CPU-Z for stock CPU, with RAM set to 1600 MHz in BIOS.

@jaquith, what do you mean by "Installation, I tend to recommend A2/B12 vs A1/B1. "

For as far as I am concerned, this all looks fine, however, I could be wrong..

It was a typo I meant A2/B2 DIMM slots to install the memory.
Score
0
a b } Memory
November 17, 2010 3:28:23 PM

Also what do you have the voltage for the RAM set at in BIOS -- the modules you have are rated at 1.55v and the default spec for DDR3 is 1.50v so it might be the voltage is not getting set correctly when running at 1600 so it is not stable at 1600 when gaming but the modules may still be stable at 1333 @ 1.50v --- try manually setting the voltage to 1.55v if you have not done so already.
Score
0
a c 347 } Memory
November 17, 2010 4:29:12 PM

^ That would be scary if 0.05V causes a stability issue @ stock speeds; RAM and I know Mushkin is good {stable} ±0.3V~±0.2V or ~1.2V~1.8V.

The RAM passed -> 2 passes. OP = "...memtest run for 2 passes, no problem at all"
Score
0
November 17, 2010 5:53:19 PM

Ok, I don't mean to be asking the obvious question here, but I gotta feeling that it might be my PSU... however, I never had any problems with my GTX460 before, which I had OCed to as far as I could get it, as well as an OC on my CPU to 4GHz when running Just Cause 2.

PSU is also sucking in air like a beast now, very noisy, whilst before it never was..
Score
0
November 17, 2010 6:17:13 PM

Alright guys, so im just having a bit of fun now.... Just decided to run Prime95 and Furmark simoultaniously, and seemingly as soon as I ran prime95 whilst furmark was already running, I noticed drops in the 3D performance... and CCC indicated that the GPU wasn't funning at full activity.

As soon as I OC'ed the CPU to 4Ghz, which always went just fine, I got hardware errors, and after about 2 seconds, I got a bunch of fun yippy lines and pretty shapes running through the screen, followed by a BSOD, and a lovely crash....

and now, as im writing this, I am running prime95 alone, with all the OCs off, running everything stock, and after lets see: about 8 minutes and I get hardware failures on each core... :pfff: 

So,
RAM - Check
GPU - Check
CPU - Maybe faulty
PSU - Maybe faulty, or affecting CPU.

I think honestly though that it is the CPU being close to death, since adding the GPU load on top of it does not affect the time until the system breaks sweat.....

lovely guys.. and here in Bahrain, this is gonna be one stupidly expensive repair... :cry: 
Score
0
a c 347 } Memory
November 17, 2010 7:32:07 PM

RAM - Check
GPU - Check
CPU - Maybe faulty <-- 1/10
PSU - Maybe faulty <-- 6/10
OC - Maybe faulty <-8/10

Suggestion, go to a place with a GOOD return Policy and 'try' a new PSU; I recommend Corsair {Japanese capacitor, Single Rail, over/under & fault protection}. IF you have a faulty PSU that's throwing 'dirty power' at your components you do NOT want to be replacing EVERYTHING... Go now!

Calculate the PSU 'proper' size -> http://www.corsair.com/psufinder/default.aspx
Score
0
November 17, 2010 8:05:36 PM

Ok. So I tested the CPU twice at 3.4 Ghz, with everything at factory settings, and everytime after about 10 minutes Prime95 tells me I get hardware errors.

HWmonitor is telling me my voltages are all screwed up, but I have always been getting horrible voltage readings with MWmon, don't really know why...

-12v = 6.08 v
VCore = 3.84 V

amongst the few that are off
???

Anyways CPU-Z gives me a nice 1.386 volts for Vcore, so that seems better.

Anyways faulty OC im sure it isn't, since I ran the tests only once with OC, but now without OC it crashes.

as depressing as it seems, im actually hoping the PSU is dead, since that is the most optimistic situation... :pfff: 

Anyways when I get a chance I'll run by the local store and have a go with another PSU there (I am friends with the guys, so they let me use the hardware at the store without paying :bounce:  ).

so your saying, do not at all use the rig until I go to the store or have another PSU?
(im on my lappy now)
Score
0
November 17, 2010 8:06:52 PM

just remembered

one of my 12v rails seems to be putting out 14.08 volts.... :pfff:  Isn't that outside of 10% ATX specification for margins? (whatever it's called)
Score
0
November 17, 2010 8:13:10 PM

14.08 according to HWmon, of which I question it's accuracy
Score
0
a c 347 } Memory
November 17, 2010 9:16:58 PM

Are you using ANY of the crappy ASUS Utilities to OC? If yes, DELETE them and Manually set everything...
Score
0
November 18, 2010 6:35:50 AM

I am not OCing. I have my CPU since I got it on the stock speeds for 95% of the time. It is just very occasionally that I OC it, and when I do, I use AMD overdrive, or do it in the BIOS. The only OCing tools I ever use are MSI Afterburner, AMD Overdrive, or just plain BIOS. Never any crap utilities from the vendor

But, should I leave my machine untouched until I get to test another PSU?
Score
0
a c 347 } Memory
November 18, 2010 12:59:21 PM

Load Defaults until it is sorted out - too many variables otherwise. Upping Voltage is a common mistake and only slight differences w/OC are actually needed -> causes instability. Voltage = Heat

Same goes with RAM OC'ing.

'I' always look for the LEAST Voltage = Stable. Less = More
Score
0
November 18, 2010 1:25:03 PM

I have loaded defaults on everything. there is no OC, and NO voltage changes at all.

Everything is at stock, as if it just came out of the factory,

now, normally my voltage is at 1.392, whilst when I stress it it plummets to 1.321 volts. (CPU-Z, so reliable), and it still crashes within 10 minutes. My other voltage readings are also screwed up.

Now I've done plenty of OCing before, and I know that 0.08 volts can make the difference.

On saturday im off to the one and only computer store in this country, and gonna test with another PSU, if that doesnt solve, CPU.

I had a look at the receipt of this PSU, and it is about 2 years, 11 months old... and I've put it through absolute hell over the last 3 years by using it in a lot of different machines, for testing and crap...
Score
0
a c 347 } Memory
November 18, 2010 2:43:43 PM

I agree [RE: "0.08 volts can make the difference'] only to the extent of outside normal ranges {example DRAM V ±0.3V~±0.2V rarely cause issues with 'good' RAM} depending upon 'which' voltage you're adjusting. Voltage outside of operational parameters is an entirely different matter.

Clearly, 'faulty' or 'dirty' PSU can easily cause either or both voltage spike and/or drops. I am very picky about PSUs and prefer PSUs as I've already stated.

Good Luck on Saturday!

Score
0
November 18, 2010 2:50:34 PM

Just one question about saturday then: If with another PSU the problem is resolved, would this be a good PSU to go for?

Cooler Master Silent Pro M700
http://www.advantionline.com/COOLER-MASTER-Silent-Pro-7...

Or just a list of all the PSU's they have
http://www.advantionline.com/Power-Supplies/

I know that the prices will shock you, (see in dollars below the Dinars), but this is unfortunately the only store in this country where I can get this sort of stuff.

Anyways thanks a lot for helping me out with this. Don't know where I'd be without your help, or tomshardware for that matter.
Score
0
a c 347 } Memory
November 18, 2010 3:11:34 PM

What a LIMITED List!!! Then yes the Cooler Master Silent Pro M700. You need 650W with your CPU/(1)GPU + OC. The CM Pro M700 lack fault protection.

My preference is the Corsair lines, but...I cannot justify the Pro 850.

Specs CM Pro M700: {Japanese capacitor, Single Rail, over/under & fault protection} - http://new.coolermaster.com/product.php?product_id=4202
Score
0
November 18, 2010 5:56:04 PM

Alright well I'll let you know how it goes when I finish,

in the meantime, thanks a lot.

Moderators: Please change the thread type from a discussion to a question with a best answer, so that I can choose one later!

Hope this solves the problem :ange: 
Score
0
a c 347 } Memory
November 18, 2010 6:46:31 PM

No big deal - I don't look for BA option topics.
Score
0
a c 163 } Memory
November 18, 2010 6:58:44 PM

@OP: Now you can select the best answer.
Score
0
a c 347 } Memory
November 18, 2010 7:23:25 PM

^after the PM - thanks...
Score
0
a c 163 } Memory
November 18, 2010 7:40:25 PM

Your welcome ;) 
Score
0
November 18, 2010 8:37:12 PM

jaquith said:
No big deal - I don't look for BA option topics.

Who says you are getting the BA? JK, but just to be sure I'll wait till after saturday, just in case anything comes up or it doesn't solve the problem... don't want mousemonkey to close the thread just yet. (no offence to you mousemonkey, but we all know you do this :sol:  ).
Score
0
a c 163 } Memory
November 18, 2010 10:17:58 PM

Change the last time to 27 and set to 1.65V. 9-9-9-27-2T@1.65V.
Score
0
November 22, 2010 8:39:59 AM

Alright guys, dropped off my rig at the store 2 days ago, and tested there on the spot with a new more powerful PSU. Didn't solve the problem.

Now it's either the new Mobo being a bad bake, possibly bad VRMs since my vcore drops significantly under stress, or a CPU gone bad.

these issues started exactly at the moment that I got a new mobo (using same old CPU), so my money is on the mobo.

I left it with the guys to test it though, I CBA to do it myself and they only charge about 12 bucks to do it. If the mobo appears dead, then no charge since its warranty covered.

they should give me a call any moment now to tell me what the problem is.
Score
0
a c 347 } Memory
November 22, 2010 11:59:50 AM

A 'bad' CPU will certainly effect the RAM, and 'bad' chocks will certainly effect voltage. If it is indeed the case of voltage drops with a new PSU then I would lean towards MOBO.

We'll see.
Score
0
a c 163 } Memory
November 22, 2010 12:38:57 PM

I think that is the mobo, a CPU fail isn't very commune and since this problems only start with the new mobo, is a bad point for it.
Score
0
November 22, 2010 1:04:40 PM

Alright, well I just got a call from the guys, and they said that they replaced the RAM and now it is working.

Strange though, I didn't get any errors, and they did... must be because weve been testing it at the store and RAM deteriorates like that if it is broken.

Anyways im gonna chuck a heavier PSU in their regardless, since when I run both prime95 and furmark at the same time (yes I know it is overkill) 3D performance comes to a crawl. Now Prime95 uses all unused CPU cycles, so you can have it running in the background and you should not notice any performance impact, and furmark gets all the CPU power it needs.

When I do this it seems that there is not enough power in the system, and the gfx card throttles back I think.

Anyways they are replacing the RAM free of charge since I only bought it about 8 days ago.

Strange though im not used to this from mushkin.
Score
0
a c 347 } Memory
November 22, 2010 1:16:42 PM

Just goes to show never rule-out anything. Better now then later and I still like Corsair & Mushkin.

It's very ODD that Memtest didn't throw ANY errors on a 'bad' stick?! :pt1cable: 
Score
0
a b } Memory
November 22, 2010 1:25:16 PM

How much do you trust these guys ?? --- I'd definitely ask them to give me back the old ram in any case !! -- Might be they swapped out the MOBO (and will file a warranty repair) but then also want to make some $ from you so are saying it was the ram to sell you some new modules ! (do you have the Serial # from the MOBO written down somewhere so you can check if they replaced it ? )
Score
0
a c 347 } Memory
November 22, 2010 2:16:53 PM

^True possibility - first 'good' tested RAM I've seen as 'bad,' but the OP's RAM was not on 'my' list of recommended. Duh?! Who knows...
Score
0
November 22, 2010 3:08:32 PM

No I got the ram at the same time as I got the mobo and new gfx card. I got all three at the same time and started getting issues.

They are just replacing the ram with identical new sticks, covered by warranty so no big deal there.

I did keep all the receipts and yes I will check to be sure if the did not replace anything I did not "authorize".

I trust these guys in that aspect, it is just very hard for me to trust someone else with servicing my own computer, but I'll live. Im sure you know the feeling.

anyways I'll keep you updated.

Now time to return my brand new SGS to the store I bought it from... and they have one stupid return policy, refund only in gift coupon form...

WTF am I gonna do with a 730 US Dollar gift voucher that is not exchangeable for cash?
Score
0
a c 347 } Memory
November 22, 2010 3:37:19 PM

Duh what's SGS?

I can think of a bunch of ways to spend $730: 3 Monitors, GPUs, SSD, Printer - Xmas Gifts... How long before the 'Gift Card' expires? <- keep that in mind.
Score
0
November 22, 2010 4:07:26 PM

SGS= Samsung Galaxy S

Yes I know I paid too much for it but welcome to bahrain!

Not to mention,I bough this phone at one of those "noob" stores, which doesn't sell loose parts. Could go for monitors though, have considered eyefinity before, scrapped it because of a broken 5870. With my knew 6870 though, could do so.

this is a nice way to be forced into eyefinity! :lol: 
Im just thinking of what to buy.... ill need to buy like 3 24 inch screens in order to spend that much here and still have left over.

screens are dirt cheap here, phones and loose pc parts are not.....
Score
0
a b } Memory
November 22, 2010 4:18:26 PM

IF you do go the Eyefinity route just remember - To enable support for more than 2 monitors, “active” DisplayPort™ adapters/dongles are required (or monitors with direct DisplayPort™ connectivity must be used). And the Active adapters/dongles run about $100 US. so if buying monitors you might want to get at least one that already has direct displayPort connectivity (while those monitors are more expensive they are usually better quality and the price difference is made up by not needing the active adapter !)
Score
0
a c 163 } Memory
November 22, 2010 4:21:48 PM

With that gift, I'd build another rig and re-sell it... :lol: 
Score
0
a c 347 } Memory
November 22, 2010 5:05:56 PM

Hmm...I would recommend CF with 2-HD 6870 with 3 screens, otherwise on many games the frame rates drop-off too slow. I have both 2 & 3 screens; 3 for gaming and 2 screens for work. I tried 3 screens and 'to me' is was too distracting. Right now 2 screens: Work @ center + Forum @ right {typically Work screen2}. Tried-> | Email | Work | Server | and drove me nuts. ...Some of my staff likes 3 screens.
Score
0
November 22, 2010 5:24:56 PM

ahahaha yeah I know the whole story behind eyefinity. I was actually in the process of setting a set up, but I received a DOA and I was leaving the country to go back to bahrain, so I needed another card quick, got a 460 at the store I was working.
Score
0
November 22, 2010 5:31:44 PM

I'd love to xfire 2 6870's, but im running a uATX system, in a uATX case. I know tis possible, just need to find the right mobo for it, which I can't find here.

A friend of mine is running it on a single 5850, and he says its fine, but his expectations aren't very high... I need buttersmooth always.

Probably won't do eyefinity though, its just a matter of what happens at the store tomorrow and how much of a bitch the store is gonna be about cash return.
Score
0
a c 347 } Memory
November 22, 2010 6:05:06 PM

It's all about the dough ($); otherwise we'd all have 40+" monitors and obscene rigs. Maybe in PC heaven. :) 
Score
0
November 25, 2010 5:09:16 AM

Alright guys,

Just got my rig back yesterday, been installing windows on it.

They tested the RAM and found it to be faulty, so somebody deserves a BA.
Furthermore after they replaced the RAM the system failed to boot properly from a cold boot, a warm boot worked though. They also reinstalled windows to see if the install went bad, but the problem remained. At that point they decided that it was either the SSD that had gone toast, or the Mobo, and basically they could not be bothered to test both, so they just replaced both with new ones. :D 

I checked out all the paper work and it all seems to be up to scratch and legit.

Furthermore just to be sure I bought the Silent Pro M700, and replaced that at home. I did this just to be sure, however, my voltage readings on HWmonitor are still screwed up. My guess is that I just have a 2yr old version of HWmon, so that doesn't really help.... :pfff: 

Anyways the whole system is running perfectly stable now, however, they screwed up my sound card. I've now got a really annoying hiss in the background (also in digital mode which I use ALL the time) and it keeps on cutting up the sound, suddenly for .5 second I don't get any sound and then it comes back. My amp also shows no digital input at that time. After half a second it will come back, then go away and on and well its really quite random and very agitating....
What do I do? Bring it back to them? I never bought the card from them...
Score
0
a c 163 } Memory
November 25, 2010 11:40:21 AM

@jaquith: You get it :) 
Score
0
a b } Memory
November 25, 2010 12:58:50 PM

Since you say they reinstalled windows - the sound problem night be either a codec or driver issue (try updating the driver for the sound card) and then also try downloading and installing the K-lite Codec pack (it is a set of codecs that will playback pretty much any format audio\video file and all have been tested to work together) -- Also while at it check n your sound manager and mute any channels you are not using (ie. if not using a mic and that channel is open it can pick up distortions that get played back as background noise)
Score
0
November 26, 2010 6:15:35 AM

Ok. Well I don't know what has been happening lately, but this machine is finally working perfectly again.

First the soundcard issue, then the next day I got the strange "cold boot" issue the guys in the store had. I force turned the machine off, and started it up again and it all worked, including sound card no problems.... strange.

This morning I started up my machine without any problems, and everything works fine. I don't know if I will get any new problems or if these problems will come back, but for now I think it's time to put an end to this thread.
I'll BA on Memory issues!


Hey thanks a lot for the help guys, I really appreciate it!

:sol: 
Cheerz
Score
0
!