Power Cuts Out Only when FFXIV is Used...

[Elfidan]

Since the May update my PC has been randomly losing power such as happens during a catastrophic issue ,or failure while in FFXIV:ARR. At first i believed perhaps i was overheating and so i began checking all the usual culprits. I powered back up checking cleanliness and heat of the cpu gpu heat sink and power core with no abnormality present in any of the aforementioned items. I then assumed one of the fans might be going out or a cable might be lose ,and found my search to no avail in this field as well. Next i checked all drivers and updated them as well as removing any unnecessary programs and even uninstalled FFXIV:ARR 2 times. Alas nothing changed and with the patch in June getting to the log in screen has even become impossible without an impromptu power loss. Having exhausted the fixes i can think of i decided to double check my system for flaws. I ran a full security sweep defraged the hard disc and ran a stress test in which the PC checked out perfectly. Upon reinstalling FFXIV:ARR yet again i still encounter the system power loss when pressing the play button on the launcher. And so i say to you SE nerds... HAAAAAAAAAAAAAAAAAAAALP .... that is all...

[Logo]

Since the May update my PC has been randomly losing power such as happens during a catastrophic issue ,or failure while in FFXIV:ARR. At first i believed perhaps i was overheating and so i began checking all the usual culprits. I powered back up checking cleanliness and heat of the cpu gpu heat sink and power core with no abnormality present in any of the aforementioned items. I then assumed one of the fans might be going out or a cable might be lose...

Your graphics card is probably overheating from a dead fan.

[Velox]

Could also be your power supply going bad and not being able to handle the draw from your graphics card.

[Elfidan]

Your graphics card is probably overheating from a dead fan.

That is one of the things i figured it might be Logo so i checked the fan and it functions correctly and is not loose or wobbly like a fan going bad.

I also imagined this might be the issue Velox but I've been playing on this system since beta with no hardware changes , so its not a lack of draw. It also runs any other benchmarks and stress tests with flying colors so i doubt its a the power supply going out as those would cause the same issues to occur , or at least they should.

[Raist]

If your PSU is not a certified 80+ or SLI model, than it may have been dying a slow death and is no longer up to the task--especially if it was a 650W or less. A generic PSU may only reliably provide 65-75% of it's rated capacity when under more stressful conditions (hotter temps, higher demands) while it is in good shape. As it ages and suffers more abuse, this capacity drops considerably. On the other hand, a certified model is guaranteed and tested to provide at least 80% of it's rated capacity under the same circumstances--why their free replacement warranties are so much longer, as they are designed more for the long haul.

Under moderate use, many DX10/11 level cards can draw up to around 150W on their own, plus anywhere from 45-130W for your CPU (either can go higher if you overclock), plus your drives, RAM, fans, anything else drawing power from your motherboard.... you may find your system can be pulling down around 380W just doing light work, but topping around 450W+ for FFXIV because of how it hammers everything at once (other more optimized games are less demanding because they simply manage resources better).

Pretty basic math really... 650W X 80% efficiency means ~520W with a certified model. A struggling generic trying to maintain 75% under good conditions is ~488W. If it's sloping off into the 65-70% territory like many bargain brands do after around a year or so and the hot summer heat sets in, you are down to ~425-450W. It just gets uglier if it's deteriorated even further. Should note that my ancient (by today's standards) ATI 5870 1GB calls for a minimum 500W PSU with 2 each 75W PCI-E power connectors (600W with 4 X 75W connectors for two in SLI). That card was released in 2010, and don't think it's been in production for like 2 years or more now. According to my UPS monitoring, FFXI was sucking about 380W (includes monitor and router/modem), and XIV draws about 430W on the same system.

Many PSU companies have online calculators where you can select your CPU and GPU models, RAM type, number of drives, etc. to get a recommendation on how much power you need. Some may provide a bit of overkill, but it may be something you want to look into so you have a better idea of where you stand and what you may need. Here's one from Corsair:

http://www.corsair.com/en-ca/psu-finder

I've been running one of their earlier 80+ TX650's for years now, and it's still holding it's own. The current one was actually a free replacement. The first gen model's bearing ran dry and was overheating and they replaced it with a 2nd gen in 2009 or 2010? I forget exactly when I had to swap them out. Yeah, I know they may seem pricey, but it was well worth the investment.... considering I blistered two of those ~$40 650W models from the local PC shop just after their warranty wore out (of course)! Got the Corsair on sale at BestBuy for $89 after a MIR rebate... with a 5 year warranty on it, and haven't had to buy another since. Oh, and the system is pretty power hungry.... 3GHz overclocked to 4GHz and an ATI 5870 slightly overclocked and the case is cooled by a handful of Khazhe-Yuni 110CFM (1900RPM) fans in an intake/output front/back, side/top layout for positive airflow across all critical components. Makes a great space heater for my room in the winter... heheh.

The point of that last paragraph is that it's worth investing in the higher quality if you haven't already... it will pay off in the long run.

[Elfidan]

Again it isn't a heat issue there's no heat in the box or even on the singular components. This set up has been running this game for 6+ months ,and still runs benchmarks and stress tests no problem. It only started after the last update in may and only for this game. I've seen similar complaints but i really cant seem to nail this down to anything tangible.


My Rig:

Power: thermaltake TR2 600w 80+ rated
GPU: EVGA NVIDEA 560 TI
CPU: AMD Hexcore
Cooling Unit: Master Cool Hyper 212
RAM: 8gb Corsair Vengeance low pro
MOBO: ASUS P8Z77-VLX
HDD: Seagate 1TB
SSD: Samsung

[Raist]

um.. sure your motherboard properly supports your CPU? The one you listed looks like a board for Intel and not AMD. It could be marginally supporting that CPU until you try to run it full bore in XIV and it's buckling under the draw--could be that something along one of the busses suffers a dip in power just enough to trip a reset. Heck... capacitors could be drying out or something else in voltage regulation could be causing it to skew too much and it throws something out of whack in the north/south bridges. Seen that happen with Gigabyte many times---bad caps around the CPU socket.

Don't think you can completely rule out your PSU yet either. That card can potentially draw 170W on it's own---210W if it's the limited edition (the former calls for 500W PSU, the latter a 550W). So you could possibly be hitting 300W+ just on raw computing power. Have to remember, XIV is slamming your CPU, GPU, RAM, NIC, and drives in a big way. You may not be getting the same effect with other applications. Are you sure that your stress test is hammering things in kind? If you have a smart UPS on your system, you may be able to log or somehow view the power draw of the system during the tests and compare it to what goes on with XIV running. Alternatively, you may be able to do something similar with tools like GPU-z and CPUID's HWMonitor to try to determine just how much you are actually pulling.

Just don't feel entirely confident in a 600W PSU for that level of hardware... even though it carries an 80+ rating. The vidcard recommendation is 500W for a standard card in a standard system... you've got two drives in there and an advanced cooler on the CPU. That makes me wonder if you throttling the system somehow (CPU, or if you have a special edition GPU that is overclocked by default, or you've tweaked it yourself). If so, that 500W recommendation can go out the window. It may sound impractical, but if you know anyone with a higher end PSU that you can test against, it may be worth the effort just to rule it out.

Then there's the issue of hot pockets that can come into play as well.... IDK how many times I've seen devices crap out because they got a little too hot in one spot. Not just Vid Cards either. Creative Labs cards... NIC's.... RAM chips getting too hot in the channels between them... one or more cores of a multi-core CPU getting a little too hot. It's so easy for these little gremlins to creep up on us... and they usually rear their ugly heads suddenly in the summer. I'm guessing it's generally been cooler the last 6 months?

[Menae]

um.. sure your motherboard properly supports your CPU? The one you listed looks like a board for Intel and not AMD. It could be marginally supporting that CPU until you try to run it full bore in XIV and it's buckling under the draw--could be that something along one of the busses suffers a dip in power just enough to trip a reset. Heck... capacitors could be drying out or something else in voltage regulation could be causing it to skew too much and it throws something out of whack in the north/south bridges. Seen that happen with Gigabyte many times---bad caps around the CPU socket.

He must have said his specs wrong. AMD CPUs are not even remotely compatable with Intel boards. They don't physically fit in the socket.

Elf, do you get any sort of logdump when you crash? Posting that in a tech forum (this one might not be quite the right option, though go ahead and try if you like) is a great first step. If you AREN'T getting a logdump...

Otherwise, I'd have to guess it's your PSU as well. They are usually the first thing to fail on a computer, and that model in particular has a number of complaints on Newegg about shutting down under load. It's possible yours wasn't initially defective, but either has been gradually dying or spontaneously died right around the update you mentioned. If you know a techie with a multimeter, you can have them test it for you. Don't test it yourself unless you know what you are doing, though, as they are VERY dangerous to mess around with.

[Kerberon]

It really does sound like every bad PSU I've ever had.

[Squa11_Leonhart]

bad psu, TR2's are known for failing.