Computer overheating with BSOD's - please help!

HBgamer

Honorable
Dec 29, 2014
78
0
10,630
Hi all. I'm a brand new member and this is my first post so sorry if description / terminology is a little vague. I have used Tom's Hardware for support for as long as I can remember so I'm sure you guys can assist me here.

For the last 2 years I had been running an I5-3450, GTX660 and 8GB RAM with the stock cooler on the 3450 and one main fan on my Freshtech Galaxy 3 case. Never had an issue with temperatures and the whole thing ran like a charm.

I recently decided to jump to an MSI GTX 970. When I went to install I found that my current 750w PSU didn't have the 2 PCIE connections I needed. I was short of disposable income at the time so I went on ebay and, after trawling through hundreds of positive reviews for it, I went with a £30 Artic 750w PSU which had what I needed. I've been told that going with a cheaper PSU is not a bright idea but with my budget and the reviews stating how reliable it was I jumped in anyway.

Once I had both the PSU/GPU installed everything seemed to be going OK. What I then noticed was, immediately after starting up, it would sit on the Gigabyte mobo screen for far longer than usual. Between 1-2 minutes as opposed to the usual 3-5 seconds.

Past the slow startup I used my PC as usual, but during the first 12 hours I experienced a couple BSODs which I had never had before. I wasn't doing anything stressful on the rig it just appeared to be random. I also wasn't quick enough / didn't have the knowledge to log what the warnings showed so sorry I can't provide here. Once the PC restarted it again sat on the Gigabyte screen, started up slower than usual then carried on like nothing happened.

Anyway that was 3 days ago and I haven't suffered any since. However, and I believe this to be linked somehow, I now have a serious overheating problem during intense gaming. Each time I play Assassin's Creed Unity, my fan speed begins to ramp up like crazy. After about an hour of play I then experience a complete PC shutdown with no warnings. The screen goes black, system powers down, it restarts then I'm greeted by a small WIndows message saying I didn't shutdown properly etc.

Is this a result of my CPU or GPU's temperature reaching it's limit then shutting down to prevent further damage? No other games seem to have the problem and I understand Ubisoft's newer games (such as Far Cry and particularly Watch Dogs) have caused many a headache with their poorly optimised engines. It's extremely worrying as I happen to have just installed the new components. From what I've read previously on TH the PSU cannot cause the overheating and if it was faulty would not have provided the ample power to begin with?

I then turned my attention to the CPU. While nothing has changed I am concerned from my fiddling I may have created an issue somewhere. I do recall the 24-pin mobo PSU connector giving me grief when I tried to remove the initial supply. I followed instructions from TH not to put pressure on anything but the connector and after 20mins of struggling it finally came loose. When I then went to put the new connector back in, one end was being slightly difficult but with a slightly harder push it finally fell into place. Could I have caused any damage to the mobo here?

Lastly, I ran Prime25 to give my CPU a good run and monitored the temps via SpeedFan. After 10 minutes I had 3 readings. 'Temp 3' (could someone explain why this appears under GPU?) was at 71C. Then the two core readings were 74C (Core0) and 78C (Core1). Are these abnormal or OK for the CPU running at 100%? I am going to try and see how the temps are when playing AC Unity and will report back when I can.

Apologies for the essay. I wouldn't be surprised if my limited knowledge is making me miss something obvious but hopefully there's a quick / simple / cheap solution! If you require further info / readings please let me know and I'll get them on here ASAP.

Thanks for your time. HB
 
Solution
I can't fully answer the question about heat, but I presume you have a top mounted PSU and the new one draws less air from the case than the old one did. My comment is valid only if you noticed a temperature increase while using the same GTX 660 GPU. If you noticed it with the new GPU, then the difference could be the new GPU exhausts more hot air inside the case. Since we don't know the exact model number of the old GPU, the new GPU (MSI sell several GTX 970) or the old PSU, this is speculation.

If your old PSU is a quality unit, then you could use an adapter like this one: http://www.amazon.co.uk/Molex-PCI-E-Express-Converter-Adapter/dp/B00FHLQURO/ref=pd_sim_computers_4?ie=UTF8&refRID=1P3EEJQZC1V1CSWWZ58Q

HBgamer

Honorable
Dec 29, 2014
78
0
10,630
Thanks for the quick reply! Out of interest how does the PSU affect the temperatures?

If my previous PSU only has 1 PCIE how do I go about adapting it so I can have 8 and 6-pin plugs for the GTX 970?

Sorry should have added - it only has the single 6-pin connector!
 
I can't fully answer the question about heat, but I presume you have a top mounted PSU and the new one draws less air from the case than the old one did. My comment is valid only if you noticed a temperature increase while using the same GTX 660 GPU. If you noticed it with the new GPU, then the difference could be the new GPU exhausts more hot air inside the case. Since we don't know the exact model number of the old GPU, the new GPU (MSI sell several GTX 970) or the old PSU, this is speculation.

If your old PSU is a quality unit, then you could use an adapter like this one: http://www.amazon.co.uk/Molex-PCI-E-Express-Converter-Adapter/dp/B00FHLQURO/ref=pd_sim_computers_4?ie=UTF8&refRID=1P3EEJQZC1V1CSWWZ58Q
 
Solution

melonhead

Distinguished
Sep 21, 2010
329
0
18,860
temps can definately cause bsods. however, the temps that you are at, i dont think is the cause of this. to really determine the cause, the list of memory dumps would definately help to see the cause of it.
i would say that my guess is the GPU (possibly in conjunction to the PSU). i would guess that there is a driver conflict between the old drivers, and the new ones. of course, this is just a guess without knowing more.
on the other hand, it could be the cheap PSU. i would not suggest using adapters though as you run the risk of damaging components by doing this. can you do it? yes. but it is not the best practice.
you are better off buying a quality PSU that has the proper connections. in fact, it could be the new (current) PSU, solely causing your problems right now.

 

HBgamer

Honorable
Dec 29, 2014
78
0
10,630




Thanks for your input melonhead. I have purchased a dual molex to 8 pin adaptor which I was going to try with my previous PSU (which was a far better brand). As I said in my opening post I don't have the disposable income right now to buy a £70-100 PSU but will definitely look to do so in near-future. Is it possible I can survive with the adaptor for a couple months before upgrading? And will the dual molex provide the same 150w that an 8pin would?
 
To clarify my suggestion about using an adapter, in a private message I told the OP to use it for testing only, not to run benchmarks or anything that could possibly exceed the old PSU's available power. If the GPU doesn't get enough power, then it could cause a shutdown, but uit shouldn't be damaged. Also note that MSI provide a 6-pin to 8-pin adapter; therefore one has to presume the GPU doesn't exceed 150W.
 

HBgamer

Honorable
Dec 29, 2014
78
0
10,630


Sorry to cause confusion with the private message. I'm new to the ins and outs of how these forums work and the best way to contact other users!

With regards to the adaptor, I'm still slightly confused if I'm able to use it for more than testing i.e. general day-to-day processes and gaming?

 
If it's http://www.amazon.co.uk/CiT-750W-Power-Supply-Rails/dp/B002Q8HFEI/ref=sr_1_1?s=electronics&ie=UTF8&qid=1419878912&sr=1-1&keywords=CIT+ATX+12V+Model%3A+750UB then I'd use it for testing, but it isn't a 750W PSU; the 12V rails correspond to a 500W PSU (40A).

A PSU like this EVGA 600 B is not a high quality PSU, but it's a lot better and more powerful (49A on the 12V rail) than the 750W PSUs you are using: http://www.amazon.co.uk/EVGA-600W-Bronze-Power-Supply/dp/B00EN3FD20/ref=sr_1_3?s=electronics&ie=UTF8&qid=1419879345&sr=1-3&keywords=600w+psu
 

HBgamer

Honorable
Dec 29, 2014
78
0
10,630


Thanks for the suggestion. I will save the link for when I can afford the expense!

I'm still no closer to solving the immediate issue though. If I can use the molex adaptor with my current supply can I get away with it for a month or 2 until I buy the one you suggest?

 
You'll most likely be fine, but the reason why the card has a 6-pin connector plus an 8-pin connector is that it can draw a bit more than 150W at full load (approximately 164W according to http://www.guru3d.com/articles-pages/geforce-gtx-970-sli-review,4.html). Full load doesn't happen too often while gaming, but that GPU could draw up to 14A from the 12V rails. In other words a quality 500W PSU would be fine, but a 600W is better. I'm not sure your old 750W PSU really delivers 20A on each 12V rail; but the new PSU with 32A on each rail should have no problems at all. If the old PSU works, then you'll know for sure the new one doesn't deliver more than half its rating or voltages and/or ripple are out of specs.
 

HBgamer

Honorable
Dec 29, 2014
78
0
10,630


Hello again. Apologies for delayed feedback, I've been a bit caught up with work and had to wait for delivery of my new PSU. I came into some money (god bless Gumtree) so I went hunting for an EVGA 600B as prescribed. I couldn't get hold of one due to stock / slow delivery etc but then I found a semi-modular Corsair CXM600W for a few £ cheaper and went for that (I read tons of reviews and although 56a rather than the EVGA's 59a they seemed to match up).

I have now installed the PSU and I don't seem to be suffering any issues with Assassin's Creed Unity at present. I do, however, still have the slow start-up which sticks on the mobo screen for a good 2-3 minutes.

I've done a fair bit of reading the last few days and a few people have mentioned that an outdated Gigabyte BIOS are known to have issues with Nvidia's 900 series? I've just looked at my BIOS version and I'm still rocking an 06.04.05 from March 2012! I think there has been at least a dozen updates since then but I literally know nothing about this area at all. Is it worth attempting myself and should I post this as a new subject in the relevant category?

Thanks in advance for your help.

HB
 
Updating the BIOS is not difficult, but we could provide better advice if we knew what Gigabyte motherboard you have. Best is to download the BIOS file to a USB stick and then flash from the BIOS; don't do it from the Windows utility as that's more risky. Your motherboard should have dual BIOS and upgrade instructions should be in the manual.
 

HBgamer

Honorable
Dec 29, 2014
78
0
10,630


Thanks for the speedy reply. I have a Gigabyte Z77-D3H motherboard.
 
Make sure you get the latest BIOS for your motherboard (there are several versions and Gigabyte have a BIOS version like F18, F21, etc.) Use Q-Flash as found in section 4-2 of the manual. Before flashing the BIOS record your settings if you want to use the same configuration again once the procedure is complete.
 

HBgamer

Honorable
Dec 29, 2014
78
0
10,630


Good evening guys. I've just attempted to access my BIOS and I am having an issue. Due to the lingering POST screen there isn't any feedback at all when I press Del or End to access BIOS/Qflash. The first time I pressed End once and waited to see if it would go through once the POST screen sorted itself - but no dice. The second time I simply kept pressing the button periodically to try and catch the 'window' when the POST screen responds before loading Windows. But again no luck.

What other options do I have here?
 

HBgamer

Honorable
Dec 29, 2014
78
0
10,630


I have tried pressing continuously, sporadically, different speeds. Just won't register. The keyboard is not fully-powered at this stage as the CAPS etc won't light up.

It is a Microsoft Sidewinder (X4 I think) and has been connected to each of my 2.0 ports during tests with no luck