Graphic Cards Weird overheating issue with sapphire 4850.

[Daedelus]

Ohh for gods sake is everyone here blind, illiterate, or is this forum sponsored by Corsair PSU's??

Every post says PSU, even when the guy says it works fine at 70% fan-speed. Shouldn't that take more electricity than 50% for f*cks sake.

I would recommend

1) RMA the card

2) Clean HS and reapply thermal paste

3) Monitor temps with an external solution

It could be that either your temp sensor is somehow busted or that some other component in your GFX card has become extremely sensitive to heat.

Ohh and dont listen to PPL advocating PSU upgrades. A Powersafe 400W is quite good. I have 4870 running on a Seasonic 380W.

hyeah:

I'm wondering to RMA the card..... but it happened since the first day of use itself has surprised me.....

Neways i wasn't going to go for the corsair PSU since its obviously gonna burn a hole in my pocket.

But im scared of giving it for RMA........ my experience while RMA'ing ne product is BAD @!!!!!

It goes for WEEKS..... last i sent my ASUS A8nE board for rma (forgot the issue) ... it came back to me AFTER 45 DAYS !!!!! WTF !!!!!

I called up rahul from ITwares he told me that the sapphire guys should take 4-5 WORKING DAYS for replacement.......that means 2 weeks OF DOWNTIME !!!! Shya.....

Could there be ne other problem (besides the busted temp sensor ?)..... cos i cant do without a PC !!! :no::no:

 

[asingh]

Daedelus:

This is the last time, I am going to reply to this thread. You seem to be off your rocker. You have come with a query to a problem, and you are just not listening. Please see my responses below. Hope it shows some enlightenment to you.

-- using the 4-pin to 6-pin converter that came bundled with the card.

Try to use the 2 molex to 6 PCI.E converter. Better load distribution.

-- card has a dual slot cooler (black fan-blades/silver-head)

We know this.

-- some1 pls explain me the diff b/w the 450w PSU that i have and the corsair 450W one. Advantage and Dis-Adv

The Corsair 450W is certified/tested to supply constant ripple free power at specific loads till a specific temperature. I think it is around 50C. End of the day Corsair is a much better company, and makes some of the best PSUs in the world. Its a sure shot no-worry game when using this. Try to understand. You are using a GPU which is extremely power hungry. It is just not another RAM stick or HDD in the system. IT IS THE, most power hungry component in your system. Even more than the CPU. At full load it will probably consume 120-130W. That is a lot of power. When GPUs of these types are mounted, you move to a completely different league vs. a generic workstation system.

-- my nv 6600Gt died .. so i purchased this card. W8ing till the cpu market settles with the new i5 entries. wondering what comeback AMD will respond with....

What has this got to do with your issue and/or this thread.

-- if it were my PSU issue my pc wont start OR one of my 2-hdd's/dvd writer wont get detected

No, this is an absolutely incorrect assumption. Your system 'could' have a PSU problem, and it would start each time perfect. Slowly over time, the PSU will start to deteriorate, and this problem will propagate to the rest of the system. What you get: component over heat --> component melt --> component burn out. Bye bye system. When you start to game, it is then when the system will be stressed, and the PSU will be really tested for true integrity. But how will you know, something is wrong. You will never, till it has happened. That is the problem with PSUs, there is no fail safe to tell the user that a problem is occurring. That is why, we all here are so cautious and play super safe regarding PSUs.

Even i thgt it was the psu issue... so i ran RE5 benchmark (full high settings). Kept the fan speed as 70% and monitored the GPU-load,temp,core and mem clocks.....(which were all at 100% load) for 2-3 hrs and STILL the PC DIDNT FREEZE. So i cancelled "PSU issue" from my list.

System will not freeze due to low power. It will start to shut-down, not start games, burn out. Your diagnostic methodology here is incorrect. Your issue seems like a heat issue, or a genuine bad card. I had recommended to cross reference temperatures using Riva Tuner...?

The weird this is that I HAVE TO KEEP THE FAN SPEED TO 70% !

Heating issue. You have the answer in front of you.

Endangering ..... how so ?

isn't 450W enough ? :huh:

450W is just right, but the device you are using to provide that 450W is not correct. Its not about the watt rating, its the quality. Let me put it plain and simple for you. You go to a shop, buy 2 kgs of sugar, and 2 kgs of rice. Now the shop keeper gives you the goods in: A. A paper bag, or B. A plastic bag. Both (A) and (B) will perfectly accommodate the 4 kgs of food. But which will hold up longer, which is more sturdy..? Your PSU is a paper bag, it is doing the job, but you never know, when it will rip open..!

buying that would seriously burn a hole.....

rather buy a 1tb hdd

As you wish. And have a 'bigger' hole, when your PSU burns out, so does the GPU, and the futuristic 1 TB HDD.

hyeah:hyeah:hyeah:hyeah:

I'm wondering to RMA the card..... but it happened since the first day of use itself has surprised me.....

Neways i wasn't going to go for the corsair PSU since its obviously gonna burn a hole in my pocket.

But im scared of giving it for RMA........ my experience while RMA'ing ne product is BAD @!!!!!

It goes for WEEKS..... last i sent my ASUS A8nE board for rma (forgot the issue) ... it came back to me AFTER 45 DAYS !!!!! WTF !!!!!

I called up rahul from ITwares he told me that the sapphire guys should take 4-5 WORKING DAYS for replacement.......that means 2 weeks OF DOWNTIME !!!! Shya.....

Could there be ne other problem (besides the busted temp sensor ?)..... cos i cant do without a PC !!! :no::no:

This is something, we have to live with. No use whining and crying about it here. If you got a bad component, bad luck. RMA it, and make do with what you have, and wait it out. We all do. Why the heck, you think we all are advising you PSU change, cause next time it will not be RMA for you, it will be new purchase. Check your temperature sensors, remount the card, reinstall the card. You have a long way to go before you RMA. But do get a new PSU, irrespective off RMA or not.

Note:

Try not to post using SMS language. Looks extremely juvenile and unprofessional.

Ohh for gods sake is everyone here blind, illiterate, or is this forum sponsored by Corsair PSU's??

Every post says PSU, even when the guy says it works fine at 70% fan-speed. Shouldn't that take more electricity than 50% for f*cks sake.

I would recommend

1) RMA the card

2) Clean HS and reapply thermal paste

3) Monitor temps with an external solution

It could be that either your temp sensor is somehow busted or that some other component in your GFX card has become extremely sensitive to heat.

Ohh and dont listen to PPL advocating PSU upgrades. A Powersafe 400W is quite good. I have 4870 running on a Seasonic 380W.

A. Do not use such foul language in here. Even we all know how to swear, but do not. We have civic sense, and this is not the forum for it. Nor are we your chat buddies, to tolerate such stuff.

B. We are not Blind cause we all are typing using screens and keyboard.

C. We are not illiterate cause we are all using plain and simple english.

D. This forum is totally unbiased towards computer hardware companies. We all are recommending a better PSU, cause we can 'see' the OP running towards a certain disaster.

E. If you are going to relate the fan speed of a GPU cores HSF to a PSUs power consumption, then I can relate the flutter of a birds wing in Timbaktoo to the weather in New Delhi. Please.

Yes, his temperature sensors could have and issue, or some GPU ram module, or component is sensitive to heat. He has to cross reference, and reach a conclusion. It would not be good idea to open the card, he might be able to RMA successfully still, if its a genuine card issue. Also a Powersafe <> Seasonics.

 

[dvijaydev46]

@OP: I'm again telling you, do change the PSU. Never go by the rating on sticker which is totally deceiving even in the case of VX450 (albeit positively). Your powersafe PSU might not even be able to provide 250 W with decent amount of ripple, trust me, high ripple causes components to over heat and that could also be a reason for your card's high temps. You spent on your card a lot and I hope you wouldn't like to burn it. Manufacturers don't provide warranty if the components is burnt. So do yourself a favour and get at least a CX400 for 3K. Do this first before you burn your card and other components.

Ohh for gods sake is everyone here blind, illiterate, or is this forum sponsored by Corsair PSU's??

Every post says PSU, even when the guy says it works fine at 70% fan-speed. Shouldn't that take more electricity than 50% for f*cks sake.

Check my comment above please

Ohh and dont listen to PPL advocating PSU upgrades. A Powersafe 400W is quite good. I have 4870 running on a Seasonic 380W.

So do you say that Powersafe and Seasonic are the same?

 

[comp@ddict]

ur 6600GT died? did u try roasting it in the oven? one guy managed to make his 8800ultra come back from the dead by doing this(110C temperature required)

 

[hash]

Ohh and dont listen to PPL advocating PSU upgrades. A Powersafe 400W is quite good.

Powersafe adviser ? Suggesting a Powersafe 400 watt to someone with a 4850. Get outa here man.

 

[atiamd]

Dude, you should understand that a good quality PSU goes a long way. There is not only wattage that you have to take into consideration, but how much amperage is there on the +12V line(s) etc. Also, the efficiency of the PSUs are to be observed. Typically we find 80%+ PSU in the league of quality ones. The manufacturers that make best PSU are: Corsair, thermaltake, seasonic, Cooler Master REAL POWER, PC P&C, OCZ, Antec etc.

You should actually read upon this. Go to a site called jonnyguru.com. for more information. I forgot where there was an excellent article written by The Sorceror is....... Maybe he can answer this. Or send him a PM asking for it.

 

[The Sorcerer]

^^ Hmm? Ignore nitant and other troll posters who are simply hellbent in spoiling the forums,I am sure many will agree. He/she/it is simply trying to derail the thread for their own fun and games, whereas some will advise to bake cards and eat them while its hot! One can't make horse out donkeys. Besides its troll mating season at TE. Unless one is planning on a mating ritual with such samples, why bother even reading them? We do have an option on the forums to ignore such posters afterall, so use it.

=

Try the following in the same ascending order:

#Uninstall the drivers, google, download and install driver sweeper, select ati and run it. #Restart the pc and then install the latest drivers+ the hotfix. See if it makes a difference

#Change the psu.

Don't be under the impression that changing the psu would fix the issue. Possibility exists that the psu might have done some damage. Replacement is inevitable. In any case, don't risk your system with a substandard psu or else be ready to make rounds with the replacement centres and if they find any physical damage/burn marks, warranty is naturally voided.

 

[comp@ddict]

One can't make horse out donkeys.

or make sheep bark, they only blah blah

agreed with sorcy here

@ OP, my recommendation wud be:-

Corsair CX400W @ 3.1k

Corsair VX450W @ 4.1k

Make ur pick..

 

[nitant]

My only point was that people rather than advsiing him on how to prevent his card from overheating, assume that his PSU has some how killed his system.

I mean you have absolutely no evidence of him PSU acting up, no proof, no nothing but still advocate PSU upgrades.

It doesnt make sense.

In the year 2002, when India had only 230W and 300W PSUs from crappy companies, I ran my brand new system with a GeForce 4 Ti 4600 and AMD Athlon XP 2100.

The final result of an underpowered PSU. NADA. NOTHING. NILCH. After a week of running, the PSU would just shut off as it couldnot handle the load. That is all. I got my PSU RMA'ed 3-4 times and finally got an Antec 400W, which worked perfect for years.

Also you do realize that the power circuitry on most high-end components is also very high-end. A simple Philips Power unit on the iPod ensures charging from 5V USB as well as the 12-14V Firewire interface.

My mobo today can regulate CPU voltages to the 5th decimal. I seriously think 3.1K for a PSU and 2K for a Heatsink is a waste of money.

I doubt GPU manufacturers assume that everyone who buys a 4870 will also buy a 3.1K PSU. Warranties are not offered on assumptions, they are offered on facts. In order to ensure a product is not RMA'ed, every manufacturer jumps through N-number of hoops to take care of every situation. I am sure that they will also invest a few $ in a couple of power regulators.

In any case, most GPU's have a 3yr warranty, get em replaced after 2 yrs and you might even get upgraded, like my friend who got a 9800GT instead of 8800GT for Rs 500 more.

Not much more tha I can add on this topic, my only last suggestion is help the guy with his current problem. Not future ones, its his life, his money, his future, not yours.

Its like a guy goes to a hospital because he had a minor heart attack, and the doctors suggest a full blown heart transplant only.

 

[Superbad]

My only point was that people rather than advsiing him on how to prevent his card from overheating, assume that his PSU has some how killed his system.

I mean you have absolutely no evidence of him PSU acting up, no proof, no nothing but still advocate PSU upgrades.

It doesnt make sense.

In the year 2002, when India had only 230W and 300W PSUs from crappy companies, I ran my brand new system with a GeForce 4 Ti 4600 and AMD Athlon XP 2100.

The final result of an underpowered PSU. NADA. NOTHING. NILCH. After a week of running, the PSU would just shut off as it couldnot handle the load. That is all. I got my PSU RMA'ed 3-4 times and finally got an Antec 400W, which worked perfect for years.

Also you do realize that the power circuitry on most high-end components is also very high-end. A simple Philips Power unit on the iPod ensures charging from 5V USB as well as the 12-14V Firewire interface.

My mobo today can regulate CPU voltages to the 5th decimal. I seriously think 3.1K for a PSU and 2K for a Heatsink is a waste of money.

I doubt GPU manufacturers assume that everyone who buys a 4870 will also buy a 3.1K PSU. Warranties are not offered on assumptions, they are offered on facts. In order to ensure a product is not RMA'ed, every manufacturer jumps through N-number of hoops to take care of every situation. I am sure that they will also invest a few $ in a couple of power regulators.

In any case, most GPU's have a 3yr warranty, get em replaced after 2 yrs and you might even get upgraded, like my friend who got a 9800GT instead of 8800GT for Rs 500 more.

Not much more tha I can add on this topic, my only last suggestion is help the guy with his current problem. Not future ones, its his life, his money, his future, not yours.

Its like a guy goes to a hospital because he had a minor heart attack, and the doctors suggest a full blown heart transplant only.

Seems you have given the answers.
Even when a person suffers a minor heart attack the causes are mainly due to increase in cholestrol in the blood stream. So though doctors may not advise heart transplant they do suggest to reduce the cholestrol in the blood .this requires change in lifestyle and eating habits because he is future candidate of many attacks to come if does not bring changes in lifestyle.
Regarding others suggesting better power supplies it is with better efficiency which means that the power supplied to the components is pure dc or almost dc. The thing with cheap psu is that they may provide the required wattage but will also provide the components with ac power which does the most damage . These power supplies when stressed to the limit gives out more ac power which
Is fatal causing burns to other components. I hope i was clear.

 

[asingh]

My only point was that people rather than advsiing him on how to prevent his card from overheating, assume that his PSU has some how killed his system.

I mean you have absolutely no evidence of him PSU acting up, no proof, no nothing but still advocate PSU upgrades.

It doesnt make sense.

Yes, it makes absolute sense. Most people told him to change his PSU, just so that he is safe. No one said, that it would exactly solve his problems. Its like you see a child with a stomach ache and open shoelaces. What do you do...? Am absolutely sure, asking him to tie his shoelaces will not cure the stomach ache, but will make him a bit safer...right..?.right..! And regarding over heating being the root cause -- even that is debatable. I told the OP to do a couple of things, but he has not. He has done jack-squat apart from speculating about holes in his pocket, and the discomforts of RMA.

In the year 2002, when India had only 230W and 300W PSUs from crappy companies, I ran my brand new system with a GeForce 4 Ti 4600 and AMD Athlon XP 2100.

No, you cannot make this comparison. Technology which is 8 years past cannot be compared to today. No. That is unrealistic, illogical, and just trying to make a point. Or you have been out of the hardware scenario for past 7 years, and suddenly realized things. That GeForce 4 series was an AGP card.

It had no external power connectors. It was pulling all power from the motherboard dock. The max a motherboard dock today can supply is 75W, and that to PCI.Ex16 2.0 standard. So I doubt your card was ever drawing so much power. Where as the HD4850 has 3 connectors. Each 'can' supply upto 75W. 75 x 3 = ?????. Am sure you can see the differential between cards. Its like comparing a peanut to an apple. Pure nonsense. See the difference below.

*courtesy Video Card Reviews and Specifications - Welcome To GPUReview.com! - GPUReview.com

The final result of an underpowered PSU. NADA. NOTHING. NILCH. After a week of running, the PSU would just shut off as it couldnot handle the load. That is all. I got my PSU RMA'ed 3-4 times and finally got an Antec 400W, which worked perfect for years.

Well congrats, for three successful RMA's due to you damaging the equipment your self. Within a week, your so called PSU was hiccuping, and coughing blood all over the PSU. Nice. You want the OP to do the same with such an expensive card. This is no sound method to use hardware, and test it to the limit, burn it out, and get a replacement.

Also you do realize that the power circuitry on most high-end components is also very high-end. A simple Philips Power unit on the iPod ensures charging from 5V USB as well as the 12-14V Firewire interface.

Auto voltage detection is different to clean, ripple free power, which can be supplied constantly at certain loads, at high temperatures. Which is what a good PSU does.

My mobo today can regulate CPU voltages to the 5th decimal. I seriously think 3.1K for a PSU and 2K for a Heatsink is a waste of money.

Great that your motherboard has an ALU for CPU voltage calculation. Which is good as useless. Do you even understand what voltage means, what power draw is, what resistance of equipment does to power draw,how important the amp (the main factor) current is. Voltage is just a potential difference between two points. It is at the start and end of the wire, its the current which flows in it which makes sense. The potential difference insures, that a current will flow, and when it has a device attached to it with resistance you get power draw. This is critical. So do not give us moon shine about 5 decimal place accuracy.

I doubt GPU manufacturers assume that everyone who buys a 4870 will also buy a 3.1K PSU. Warranties are not offered on assumptions, they are offered on facts. In order to ensure a product is not RMA'ed, every manufacturer jumps through N-number of hoops to take care of every situation. I am sure that they will also invest a few $ in a couple of power regulators.

Your doubt is false. They do assume, and list it all over the reviews, and official specifications. I researched your GeForce 4 series card, and no where did I find PSU requirements, or power requirements. Why...? There was no need to specify, cause it pulled all power safely from the AGP slot, which would suffice just perfect. A GPU does not have power regulators, which will smoothen out spikes, and improper supply from the source.

In any case, most GPU's have a 3yr warranty, get em replaced after 2 yrs and you might even get upgraded, like my friend who got a 9800GT instead of 8800GT for Rs 500 more.

If your GPU is burnt out due to a bad PSU, NO, the RMA will be slammed back in your face, with no change.

Not much more tha I can add on this topic, my only last suggestion is help the guy with his current problem. Not future ones, its his life, his money, his future, not yours.

Well, we worry about the future and money of posters here. If you do not, sad. What a lame and pathetic attitude. And we are helping him, by giving suggestions. We are not shop keepers selling hardware, get the job done, close the deal, and close shutters.

Its like a guy goes to a hospital because he had a minor heart attack, and the doctors suggest a full blown heart transplant only.

Nice analogy, sure. If the minor heart attack shows complete heart failure, I guess what you suggested, full blown transplant is a way to go.

 

[hash]

@ nitant and the other fool

Why do you try to make sense of bullsh1t. Running high end stuff needs decent PSU's has been tried and tested.

 

[Daedelus]

OK guys...... 3 days of my horror is over...

I completely formatted my pc (taking backups of 500Gb of data is a b****)and installed the DEFAULT drivers that came with my mobo/graphic card CD (not the latest)

and VOLIA !!!!! everything is working FINE.....

no issues.... till now....... kept my pc up for benching some games and for downloads(the entire night) .... so far no problems.

I installed the asus probe utility and monitored the voltage levels... no major problems...

Took an image of my C: (with 9.8 catalyst) using acronis for backup and then installed 9.12 catalyst. and the screen went GREEN on 2 occasions and my pc froze while watching an high-def video in the span of 4 hrs. So i guess the new ati drivers are giving an issue with my nforce board...... so i reloaded the saved backup image of my c: and the entire sunday went smoothly...

Afraid to install the latest catalyst... for now///

One more stange thing is that prototype didnt work at 1920x1080 reso on my samsung 2233sw monitor.... i had to change the reso to 1600x1200(highest possible)..... similar issue with ghostbusters..... while RE5 worked at 1920x1080.

Will be posting the voltage readings at night... tell me if somethings wrong.

 

[Gautham]

Well, congrats, I guess. Another note on drivers. I get the latest drivers from my card manufacturer's site rather than nVidia's site. My card is a Sparkle 9800GT, so I go there to get the latest drivers for my exact card model. I can't assure anything, but imho it's the best way to avoid problems with official drivers.

 

[asingh]

Drivers from OEM official site = Driver from nVidia site. They have to be the same, unless you are using modded or hacked version. Same goes with ATI series.

 

[satyanjoy]

ATI graphics card with nvidia based mobo is not the best combination

 

[The Sorcerer]

ATI graphics card with nvidia based mobo is not the best combination

With great people with such comments, who needs techboi to give us a laugh hyeah: OR unethical dealers to spread lies and damage a great forum. Here you go!!!

 

[Godfather]

I thought i logged into TE. :S

This thread felt like i am in TDF/Tetti/... :S

 

[asingh]

ATI graphics card with nvidia based mobo is not the best combination

Do you really know what a chipset means...? It is an enabler, not a limitation.

 

[satyanjoy]

Do you really know what a chipset means...? It is an enabler, not a limitation.

yes dude.. i know it enough to use it as a user..if just do some googling you will find lots of problem with these issue and without any solution from tech geeks like u..i m out of here..enjoy