Power consumption/draw question

Hi,

I have an "older" desktop I built several years ago, and I'm having issues with it. I'd like to mitigate the issues if possible without replacing parts yet.

My issue is... first of all, it restarts seemingly randomly. I believe this is a power issue. The build includes the following:

Antec 1200 case
Intel i7 980X CPU
Asus P6X58D Premium mobo
1 Intel 120GB SSD
2 2TB WD Black HDDs
1 Nvidia GTX 470 factory OC (I forgot the make/model)
HTO Claro sound card (though I don't use it much anymore, I mostly use the HDMI sound via the video card)
Multi-card reader (I forget the make/model)
BD burner (I forget the make/model)
12GB DDR3 RAM (Corsair)
1000W I believe Antec PSU (it's one of the top brands, if it isn't Antec)

It's connected to a 24" Dell monitor and a 55" Samsung TV. I also have an external HDD (powered separately) connected via USB2.

Other than a wired keyboard and wireless mouse, that's it. I had 2 other external HDDs connected before, but those are disconnected right now. It's running Windows 7x64 Ultimate.

Anyway, my first question... I know this CPU can draw up to 40+ amps... it's powered (along with other things plugged into the same breaker) via a 20A breaker circuit right now. I don't have another option. Could the actual wall power be an issue?

My second question: I know power supplies degrade over time, but I got this power supply in mind... I shouldn't need 1000W to power this machine... BUT, could the power supply be insufficient now?

It's definitely not a software issue... I've reinstalled Windows several times with the same effects. The case/components are clean, and I've checked temps... all the temperatures are nominal/normal.

Unfortunately, I think I need a new case, because I believe there's a short in the top of the tower now where the power/reset buttons are... barely tapping the top of the case causes the machine to reboot. However, even when it's just sitting still with no external vibrations or anything, it will reboot as well. I did trip this dang breaker a few times getting all my stuff plugged in, so I think I'm near the limit on this breaker (and yes, I unplugged a few things as well afterward to lessen the power load on the breaker).

Suggestions? I have a multimeter (2 in fact), but I'm not sure what exactly I should be checking at this point. When running, the machine performs flawlessly... I don't think any of the electronic hardware is malfunctioning.

Last but not least, would a UPS help with the amperage draw between the wall and the machine at all? I know the PSU does a bit of its own management regarding that, but would a UPS aid it in a useful way in my situation?

 

Greetings, mcsmc.

My advice would be to try to make 100% sure that the random restarts are caused by hardware and not by Windows. Not talking about the "tapping the top of the case" - those are most certainly a short. However, if the machine is rebooting otherwise then you need to make sure that you don't have a software problem in conjunction with hardware issues - make Windows show you the Stop Error Message (the BSoD).

I'm a huge UPS booster, but I doubt that adding a power conditioner will help your symptoms. Looks like your problems are all internal.

If this were my machine I'd tear it apart to fix those shorts, monitor Windows minidumps for any BSoD information, and grab a replacement PSU to troubleshoot. I've never known a single computer to trip a 20A circuit breaker unless something was wrong with the power supply.

 

For info: here's an online PSU calculator - comes in real handy sometimes.

 

Amps are only comparable at the same voltage, your cpu may draw 40amps but only at 3V through a DC to DC converter so this example cpu will draw 3VX40= 120watts.

40 amps at the wall socket @ 120V is 4800watts:-D In total your computer will only draw around the 3-400watt mark maximum so you have no problem in the wall socket since it can provide about 3KW and the PSU can run 1KW.

If you suspect an AC problem and have problems with other electronic devices in the house or suffer frequent blackouts or grey outs a UPS will help.

 

Alright, case in point... I let this thing run ALL night last night playing episodes of Star Trek: TNG and it hasn't restarted all night. However, before I made my initial post in this thread, it'd restarted while not running anything multiple times (literally just logged onto Windows, no extra programs running). The system log doesn't show anything (this was also happening at another location... I moved a few months ago).

I would love to rip this baby apart and fix the shorts, but the top part of the case where I think the short is, is riveted shut. I don't have a rivet gun. Maybe I'll try to double check all of the wires leading out of there (to the mobo), but I'm pretty sure they are okay.

Thanks Caliban, yes I've used that tool before... thanks for the link, though, I need to make it a prominent bookmark!

I knew there had to be something I wasn't figuring quite right... it was kind of hard for me to imagine a little computer chip sucking more power than an entire room's load. :-D

Thank you for that explanation... yes, I know the AxV=W calculation but I definitely wasn't thinking about that last night haha.

I have a receiver, TV, this desktop, and my laptop (ok... portable desktop -- Asus gaming computer), and several chargers connected on the same breaker. The breaker did trip a few times when I tried to plug the desktop in... but from what it sounds like, it's not a power draw issue, but likely some short somewhere in the desktop. I honestly don't know the reliability of the power here, but it seems pretty reliable... I haven't noticed any issues other than with this desktop, regarding the electricity here.

I'm guessing the short would make a UPS complain as well, huh.


Also, I'm not POSITIVE that the short is within the riveted part, but that is where literally a light tap will make the computer reboot... even plugging something into the front USB port (same area) will trigger it. Anyway, thank you ALL for your responses and I will post any actual results I figure out.

 

I did some very simplistic troubleshooting and disconnected (one at a time) the various things connected from that top part of the case... and wouldn't you know it, it's the reset switch. Since the reset switch isn't all that useful anyway and I don't use it much, I don't think I'm going to bother trying to fix it... simply disconnecting it from the motherboard has fixed the "tapping" problem. I just have to wait and see if that was the only issue, or if there's something else causing any other rebooting business. I'm pretty sure that it was so sensitive, that even vibrations from walking on the floor, etc. could have been tripping that. Of course, I have a good sound system with a pretty bangin' sub... so certain low frequencies could have been tripping it as well.

If unplugging the reset button is the only thing I have to worry about, then I'm happy.

 

Good job! Hope that's all it was - thanks for the feedback.

 

I apologise for asking this here (maybe I should have PM'ed Rikky or started a new thread), but I didn't know a better place. I have some questions about Rikky's statement above, a discussion of which I hope will clear my confusion.

1. I had believed that CPUs of this age operated at a voltage of 1.5V ? (I understand that the value is irrelevant to Rikky's point)

2. I had thought that CPU's were powered by a 12V line through a separate PSU connector (known as the "P4"). So with 100% DC - DC efficiency that would make 10A (120W/12V) draw on the PSU (in Rikky's example), [which is not unreasonable]. Is my thinking correct ?

3. Is the DC - DC converter of the switched mode type, or is it a linear regulator ? The large coils near the CPU make me think it's s switched mode, but I see no evidence of the complex circuitry which I assume is needed to implement a switched mode DC - DC converter (eg, control chip, transformer).

Dumb_Question
21.January.2014

 

You don't need complicated circuitry, you need one reference oscillator to switch the DC on and off very fast,a power mosfet to switch the power, a capacitor to smooth power,a small 'coil , choke, inductor' to limit the current and a bridge rectifier diode.

Once this circuit produces the new DC voltage close to what you need i.e 12V to 1.5V, it will do this quite efficiently producing little heat if the circuit is well designed.

The voltage will then be reduced further by a linear voltage regulator for example taken down from 1.5V to 1.3V, this is inefficient because any power blocked by the voltage regulator will be dumped into the case as heat, so should only really be used in lower power situations or very small voltage reductions.

You can see which solution is the best, a real switching DC to DC converter but it's also more expensive requiring more components and design.

The number and design of each circuit in a motherboard and psu is entirely dependant on the manufacturers design, they're all different. Intel will provide a reference design with reference voltages and leave it up to the motherboard and PSU manufacturer to implement a power solution.

As motherboards have advanced they have taken more control of the DC to DC conversion converting the psu 12V line into usable voltages which 90% of the power will be drawn from.

My example was to show the difference in wattage at the cpu and wall socket.