PC shuts off like pulling a cord from wall socket. It happens mostly on idle after BeamNG session, but it happened after adding 14th police car on WCUSA. No warnings, no errors, just typical power loss message in logs. It seems to happen after heavy power demands, but it can be delayed by 1 or 2 hours, so that makes it bit weird. Has anyone has similar with some other components being at fault? Seasonic fanless PSU, very expensive so kinda would not like to end up getting MB if it is not in PSU.
More detailed specs of your computer could help. Is your Processor overclocked? What model power supply are you running? What Video card? Monitor your temps too, make sure nothing is overheating.
Specs: i7-6700 with stock clocks Gigabyte Z170-UD3 16GB Kingston Hyper-X 2666Mhz Gigabyte 1050Ti Windforce 4GB OC stock clocks Samsung SM951 256GB M.2 NVME PSU Fanless Seasonic S-12 460W CPU cooler is Thermaltake Macho Rev B, I doubt they make them much bigger than that, with dual fan setup. Voltages, from some reason I have always got those split second zero readings, I think those are just flukes, but if none other gets those, then those might be valid data: Temps, nothing odd here, bit hard to make out individual cores, but they are in perfect range: I made bit of stress testing above, with this and again no issues: I think my cooling is quite ok, tomorrow is dust removal day, but as issue has been pretty much as long as I have had machine together, I doubt that little dust does anything. Only happens when using BeamNG, well usually hour or two after and quite rarely. PSU did make weird loud noises when rather new, but those did stop when moved to city so electricity is now better.
I think that 460 watt power supply is going to be getting pushed pretty hard by that system. Power supply might be overheating and crowbarring to protect itself and the system from damage, have you got any other power supplies to test with?
Sadly I don't have other PSU's, but I'm not sure about PSU being pushed very hard as CPU is NON K version, only 65W TDP and 1050Ti is also something around 65W. Problem is that issue should happen on load quite constantly if PSU would be overheating, but it has happened now only once on load, other times always 1-2 hours from the load, so there has been time to cool off, in theory. In practice anything really is possible, maybe overload is such minimal that PSU gets into some weird fault mode and freaks out after some time. Total system draw from wall socket, including monitor and audio amplifier has been around 360W at most, but how that turns out to PSU watts some with electricity math fresh in memory might be able to compute. Anyway if no cases with MB or GPU causing such shutoffs is known, then I would be quite confident of issue being fixed by new PSU and I would have no problem buying one, but it has to be very quiet, like breathing quiet, that is why I have overkill CPU cooler to keep rpm of fans down when I'm in "noise hurts" -mode, rest of times I just play death metal louder Update: This is normal desktop usage, Excel, browser etc. ~40 minutes, data polling rate is 250ms as in previous charts:
Sorry I'm being a tool, blame my cold. I missed that you mentioned it happens at idle too. Yeah, sounds like a Power supply fault to me, also don't forget there's more than just the processor and video card drawing current in your system, fans, drives, the motherboard etc. You'd be suprised how quickly it all adds up. Also TDP doesn't equal power draw, that relates to how much power the item dissipates as heat.
I wonder if HwInfo's CPU package power is power or thermal power? Gigabyte's software shows same, that is 52.6W for CPU maximum I could get it to and for GPU power that is 64W maximum I could get that to. Currently I don't have drives connected (sorry I lied, DVD burner seems to be connected, but without disc that is not drawing much of current), only that M.2 SSD and it draws quite little power, 4 fans of course add something, also 4x DDR4 sticks. I did run this system with 430W PSU and during that I had no issues of shutting off, but man that PSU were screaming, it was lot over 10 years old and I left it to countryside as it was really too old to be used anymore. Surely it could be that if I could of tested that PSU longer there would of been shut off issue too, but these annoying problems are hard to figure out what is exact condition to make issue happen. Maybe I could find some relatively ok and cheap PSU to test out, any tips of such?
Intermittent problems are a PAIN to troubleshoot, just when you think you have it figured out and everythings been fine for a few hours, boom, the problem happens again... Best I could suggest is you RMA the old power supply for testing. If you have the cash I'd suggest you upgrade to something a bit more powerful, maybe silver rated or higher. You could also try a cheap brand name higher wattage power supply. Edit: HWmonitor shows my overclocked I7 2700k as drawing the same power as stock, sooo, take that as you will.
Silver would be downgrade though, my current one is Platinum rated, If I don't remember wrong this cost something like 160 euros and now that you reminded me, thing has 7 years warranty. I guess I buy some cheap PSU first though, I can't be without computer too long. What do you think of this https://outervision.com/b/ADKUzp ? They recommend 280W PSU
My bad my initial search was showing me results for bronze rated supplies. I'd say atleast get another power supply rated over 400w, maybe I'm just a little paranoid over an accident I had a long time ago when hardware drew a lot more power then todays hardware and had a power supply literally explode while I was gaming, I believe it was a Cooler master.
If you're getting drops to 0v it could be a bad solder connection. Sniff wall socket for burning smell where PC is plugged in to (the nose knows all), and also listen for any zaps... bad plugs would not be a surprise, though surely you'd have smelled it by now. Also check other sockets in the room and switches. It would be obvious if it's been occurring for a while now. Seasonic usually has a good 5 years warranty on most of their products, mine in fact has 7 years of warranty on it, though the cheapest models may have 3 years only. If the machine is shutting off for no reason, first thing I would do is test with a known good power supply. EVGA makes some good cheap supplies to test with, even a 430w 25~30$ (USD) power supply should work for testing purposes. Could be something as simple as a surge-protection sensor on the BIOS. I always keep a cheap one around to test with. If it cures your woes, buy a better one when you can afford it. Modern gaming machines rarely take over 300~350w of power, to power during gaming sessions, unless you have LOTS of fans, lots of hard drives, and mega-over-clock and a REALLY BIG video card with or without an overclock. While I do have a 750 watt PSU here, it's way overkill. I bought it because it should be able to stand up to anything for a lot of years, even when it gets older and puts out less power. A power supply unit 3 years old might only put out 70~75% of it's rated power, while other supplies might go 5 years and still put out 85% max rated power. Other cheap units might only put out 60% power after 5 years. Sometimes it's luck of the draw, sometimes it's a cheap component substitution, temporary or permanent by the manufacturer due sometimes to shortages at assembly time, or just an older model with a revision in it to drop manufacturing costs. It happens in a lot of sectors of the technology business - for example, Gigabyte motherboards does this. Power supply manufacturers do this too. Cheaper components doesn't always mean a worse product though, often times it's just merely to compete and still meet market demands - after-all no one's going to buy a 200$ overclocking motherboard for a Core 2 Quad machine when the machine is 8~10 years old today, but they might pay 40~60$ to replace the motherboard and just make it work again. So now off-topic, I'd surely replace the power supply, and the surge strip, you never know. That'd be my first guess. If that does not solve it, RMA, and get a different brand same-chipset motherboard (if you do not do a cross-ship RMA - credit card needed here). A bad VRM on the motherboard or gimmicky solder connection can do this. That could be anywhere. A goofy pin on the CPU socket could do it, too. Once you use a surge strip for 2~3 years you MUST replace it as it's lost most of it's surge protection. Though SEASONIC power supplies have pretty good built-in supression. Though, I will say this much, if you have other equipment on the circut your PC is on, and sometimes even if you're pulling close to 100 amps with all appliances in your home combined, and your house only has 100 amp service (basic, for American homes, is 100 amps, your area may differ, substitute as necessary), you could have a 'brown out' noticeable when the lights flicker and some appliance (furnace or center air-conditioning etc) kicks on. At once point 10 years ago if I turned on both my computers, it would blow a 20-amp circut breaker and I would have a fun trip in the dark to the fuse panel. To recap if you are lost with my BOOK of a post: Give a cursory check over your power source (surge strip, plugs and wall switches on same circut for shorts, a quick arc will cause exactly what you're seeing, but not long enough power loss to shut the PC off). Also, wiggle ATX power to motherboard plug, and 12v cpu plug gently while machine is on in BIOS, to see if there's a loose connection, also check the GPU 12v plug the same way. If it resets, turn off machine and try to crimp the female connectors on the PSU to board plugs needed or tape in place. Replace power supply. Buy a cheap EVGA model 430w~500w if money is tight, I've used them they work well enough and can be trusted. Don't jam the paper-clip into the plug hard to test though or it'll spread!!! If this solves your issue, keep running it or replace with 500w Seasonic unit later on. Amazon sells power supplies pretty cheaply in the USA not sure about rest of world. Replace motherboard. MSI uses cheap VRM's as does Gigabyte on their <150$ (USD) motherboards and some may be analog and prone to acting up - or causing FIRE conditions, so if you have that brand MSI or possibly cheaper GIGABYTE Z-170 or H-170 chipset board, get another motherboard through RMA or buy an ASUS motherboard with that chipset. Make sure it has the same ports you need. You may have to call Microsoft to get your Windows re-activated though. You should NOT need to buy a new OS. Seeing these issues, you have either an over-current or over-power condition detected by the supply, or a bad solder connection (most likely). Bad solder connections happen with shipper's mis-handling of delicate products, more often than you'd like to think. You just have to find it! --Good luck! --- Post updated --- Cooler master makes CHEAP power supplies, worse than the EVGA models I mentioned. Seasonic is the best, up there with PC P&C, Sparkle's better ones for the server market were always reliable too, their cheapies not so much though. There was another name or two out there, but if you had one POP real loud that could have been it's inside-fuse blowing out from over-power condition to prevent fire. If a cap explodes spectacularly in a BOOM type of fashion, it will short and blow the internal fuse also. Bad cheap power supplies will just explode and take your components with them, vs trip a fuse, though :x EDIT: To clarify, a lot of companies like Cooler Master don't MAKE power supplies but use an OEM like, PC P&C, SeaSonic, FSP make their own (among other companies who are less reputable I won't even mention).
Ouch, Cooler Master is one of my do not touch brands, they do make surprisingly short lived stuff at times, but somehow they manage to do good stuff on other times. This setup of mine draws less power actually than my Core 2 setup did, while computing power is slightly improved, compared to your CPU there hasn't been much other improvement than power needs are a lot less. At idle power taken from wall socket is around 50 Watts, that is without display. With i3 that was 35 Watts, old laptops were something like that. @bob.blunderton thanks from your book, always appreciated! This is Gigabyte board indeed and Ultra Durable 3 variety, I think it was something like 130 euros, that might been close to $150 when I bought it some 2 years ago I believe. Luckily here power is really stable, I don't even need UPS here as there are no issues with a power, all power cables run under ground, much better than countryside where slightest wind caused issues with power. Ok, so I look around for EVGA PSU then, it would be enough to get this PSU replaced.
Yeah, found that one out the hard way... But hey learning with sparks is fun right? She definitely let go with sparks, and holes blown in voltage regulators, no blown caps thankfully so no fishy smell, unlike the time a amplifier I had, had a Filter cap go boom...
My friend, have you considered staying bit farther away from electrical devices? Things seem to like to blow around you :-\ Back when I did exams for computer tech papers, behind me was some girl who somehow managed to blow PSU while installing graphics card, this was before time of ATX standard, so it is still beyond me how that was possible, but that was smelly white smoke cloud for sure! Not much of such during my career, but at electrician line of studying we did manage to cause few fires as only instruction we got was to test it, so we tested, quickly learned those wire colors too, after smoke had dissolved
Lol, in my defense the amplifier was 50 years old and I've only shocked myself at mains voltage 5 times now, I swear!
It's okay working with PC's so long as you don't touch the AC prongs while plugging something in, and you don't open up & touch the guts (or heatsinks!!!) on a computer power supply. I've touched those heatsinks, and I don't know what voltage or amperages run through them (in this case a 10 year old Antec 350w atx PSU - core 2 era), but it will work a little better than morning coffee does. Thank goodness I didn't have to use the loo at the time (bathroom). AC power won't usually KILL you it will wake you up, now 220v~240v from europe or Asia will possibly harm you or make you bounce around the room if not prepared. That's a lot more than our current here, but it's the amps that get you. AC is easy to let go of, because it alternates, allowing your body's nerve signals to get your hands to release between pulses. DC is the scary one though, it won't let you let go as it's steady, direct current. Though 12v you might even not be able to feel. Just avoid running it through your chest/head and you should be just fine (e.g. don't touch with both hands). @fufsgfen You're quite welcome, when someone needs a detailed lecture/theory on things I am definitely the man for it.
I've re-capped a few PSU's mostly out of old Compaq's and HP's the dreaded Bestec power supplies! Thankfully I've never shocked myself with one yet, but there's always a first time for everything!... as far as voltage at the heatsinks, I've measured about 1,000v with my multimeter on them. So yeah, kinda high voltage lol...
Let me tell you this, European 240V is blue and leaves bad taste to mouth at least by my experience Who at Evga has decided that Supernova is good name for PSU a device which is know to blow up if it is poor quality? That must be one of the most ridiculous branding efforts I have seen. 550W Evga is 100, 550W Seasonic is 100, 600W Be Quiet is 75 in sale, normally 95, 550W Corsair is 60, quality brand 500W Bronze models are not available, but would be around 60-70. This was cheapest while not appearing completely horrible, however will take long time until that would be here. https://www.computeruniverse.net/en/products/90627632/evga-netzteil-atx.asp I could of course swap i3 back in, lowering power requirements should change something too.
Lowering power requirements may help if it's tripping an over-power or over-current condition, or there's issues with the motherboard VRM (it's questionable, but obviously the i3 wouldn't make them work so hard). This is not a fool-proof test by any rate though, so I wouldn't really bother much. I would just get the Seasonic at that rate if your warranty is out already on yours. If it doesn't help, return it - even if you pay shipping and a restocking fee, it's still some money back in your hands vs not having it at all, and still having a problem. If that doesn't fix it send the motherboard back.
You are using under well 200W. https://www.techspot.com/review/1269-nvidia-geforce-gtx-1050/page8.html ~150W that it drew from the wall which means due to effiency that the system it self is using around 135W. This is what I assume to be a i7. I can guarantee you that you have plenty of wattage. 460W will get you a LOT farther than you think. You could even run a GTX 1080 Ti and be under 400W. https://www.anandtech.com/show/10325/the-nvidia-geforce-gtx-1080-and-1070-founders-edition-review/30 Not recommending doing this but two GTX 1080's can draw around 440W. Two GTX 1080 Ti's with a i7 6850K uses under 600W https://overclock3d.net/reviews/gpu_displays/asus_gtx1080_ti_strix_oc_sli_overkill3d_review/2 Something else is going on that is causing your issues.
