Motherboard/PC turns on for a second then turns off


Recommended Posts

Hello,

 

Well that PC:

PNY VCQK2000-PB

Samsung 840 Pro Basic Kit 256GB

Cosair Hydro H110

Intel Core i7 4820K

ASUS DRW-24F1ST

Seasonic Platinum-760

MSI Big Bang XPower II Intel 2011, X79

NZXT H630

Arctic Silver 5

Turns on (its fans and motherboard lights turn on, not the debug display though) but then in a second turns off.

I have the RAM, PSU, and CPU installed....

So what stupid thing could I be doing?

Thanks

Link to comment
Share on other sites

Motherboards don't come with speakers (unless on the PCB itself) anyway. Cases do.

 

What was the problem, though?

 

likely forgot the 4 (or 6) pin power connector.

Link to comment
Share on other sites

Hello,

Using V points on the motherboard, only 0.8 V gets to the CPU. Strange.....why could this be?

Tested the ends of both CPU plugs both give out 12 V

Link to comment
Share on other sites

Perhaps the power delivery is borked, so it doesn't give the juice (CPUs don't run on 12V anyway, the stuff around it downsteps it to whatever voltage is requested by VID).

Perhaps you're looking at the consequence not the cause there. Try easier things first.

Link to comment
Share on other sites

Normally at this stage I start indiscriminately swapping out parts, however I realise that isn't always practical! The PSU is probably the easiest part to swap out for a test. Personally I have found the motherboard is a common point of failure. Do make sure none of the CPU/Mobo connection pins are damaged.

Link to comment
Share on other sites

Hello,

Perhaps the power delivery is borked, so it doesn't give the juice (CPUs don't run on 12V anyway, the stuff around it downsteps it to whatever voltage is requested by VID).

Perhaps you're looking at the consequence not the cause there. Try easier things first.

PSU gives me proper readout of 12 V.

Not too sure what to try here. Tried to check the CPU again and put it in place with the liquid cooler as well but that did nothing.

Link to comment
Share on other sites

powers on for a second then powers off...could be heat related (bad contact to heat sink, too much or too little thermal paste...should be a very thin amount between the two, I spread evenly using a business card as a brush). 

 

Yet another reason not to build computers for your business.

Link to comment
Share on other sites

I said CPUs don't run on 12V. VRMs convert it to proper voltage level which is way down, varies around 1V, depending on the load, power saving features and model. 0.8V is rather low for 2011, but I'm not sure what the lowest possible VID is for these. It could say the problem is with the power delivery circuit on the board, but that can't be reliably checked without swapping CPU or detecting the actual cause, if there is one.

 

Did you seat DIMMs correctly, t.i., DIMM1 and DIMM3, if using two, then DIMM5 and DIMM7, if using four? X79 is picky about that.

 

I maybe too insistent on this, but do try...

 

Also, check for bent pins, perhaps. It's all too easy. LGA2011 latch mechanism is rather tough. Try with some other cooler (even if it doesn't fit, just that it makes contact with the surface) that you don't have to mount for five minutes. It won't explode or overheat in these several seconds.

 

No, it absolutely cannot be heat related because there hasn't been any heat to begin with.

Link to comment
Share on other sites

I said CPUs don't run on 12V. VRMs convert it to proper voltage level which is way down, varies around 1V, depending on the load, power saving features and model. 0.8V is rather low for 2011, but I'm not sure what the lowest possible VID is for these. It could say the problem is with the power delivery circuit on the board, but that can't be reliably checked without swapping CPU or detecting the actual cause, if there is one.

 

 

I maybe too insistent on this, but do try...

 

Also, check for bent pins, perhaps. It's all too easy. LGA2011 latch mechanism is rather tough. Try with some other cooler (even if it doesn't fit, just that it makes contact with the surface) that you don't have to mount for five minutes. It won't explode or overheat in these several seconds.

 

No, it absolutely cannot be heat related because there hasn't been any heat to begin with.

If a mobo does not have a processor, ram, or video card, the mobo will stay on and create an error alert/beep/noise (if a case speaker is attached).  I wouldn't rule out heat as bad thermal seal could cause a shutdown pretty quickly, near that if not having a heat sink attached.

Link to comment
Share on other sites

Good then that soldering points didn't touch the conductive material. That bag is not a dielectric. It's a Faraday cage.

Yea. Interesting as to why it actually worked (I still don't know to this day) but it was suggested to me and so I tried it when I was at my wits end, luckily it worked.

Link to comment
Share on other sites

I would pull the RAM from the mobo to test also - if you get the same error, it is unrelated to the RAM. (one less possible faulty item). While most mobos would complain about this, it is worth trying because it is simple to do.

 

I think that a thermal problem here is unlikely, while still possible. I have seen some over/under/absent done thermal jobs still run for a good 15 minutes before the computer called it a day.
(also, if you have the know-how to check the voltage of components...I imagine you have done a reasonable pasting job)

 

Did you buy these parts from a local shop? If you did, they may be happy to let you use some of their old/test parts to play around with (my local shop has let me on some occasions). At the very least I would test with an alternative power supply - simply because it should be an easy test ("borrow" from another computer?) This would be the most efficient way to determine what item is at fault.

 

Alternatively if you have no way to check individual items, it comes down to randomly returning items as faulty until the problem goes away. Because you only have 3 or 4 parts, of which they are all simply connected, and not really fixable, you don't have much option but returning stuff.

Link to comment
Share on other sites

Hello,

Some of you were right :) It was a RAM issue. A odd one though.

Since this chipset supports quad channel, I tried to run it in that mode.

Layout is following:

DIMM 1

DIMM 2

DIMM 3

DIMM 4

DIMM 8

DIMM 7

DIMM 6

DIMM 5

The manual recommends for quad-channel: 1, 3, 7 and 5. All is good until I put anything in 3. Anything in 3 and it gives me that 67 error on the debug led. Also filling up odd slots, gives me a 60 but no constant reboot.

It seems 3 is faulty and would be a HUGE bitch after building the PC that ONE small part is glitching the board.

Right now it is DIMM 1, 2, 7, and 5. CPU-Z reports it is in triple channel BUT running at 800mhz when this RAM is 1600mhz. Thats something else I dont understand too well.

Link to comment
Share on other sites

Right now it is DIMM 1, 2, 7, and 5. CPU-Z reports it is in triple channel BUT running at 800mhz when this RAM is 1600mhz. Thats something else I dont understand too well.

 

Not an odd one at all. It's how multiple RAM channels have worked since they were invented, except some boards detect and run all possible combinations (for instance, in P67 and above you can put sticks in any order and slot you please) and some throw up and require to RT(f)M, to put it blunty, to seat them in those and only those slots.

 

Except in this case your board seems to be actually borked.

 

Yet it runs at the correct speed. CPU-Z reports the actual clock of 800 megahertz whereas it still does 1600 megatransfers (which is what gives *double* data rate its name). It's a technicality and one I'd like to be remedied, but well...

Link to comment
Share on other sites

Hello,

Not an odd one at all. It's how multiple RAM channels have worked since they were invented, except some boards detect and run all possible combinations (for instance, in P67 and above you can put sticks in any order and slot you please) and some throw up and require to RT(f)M, to put it blunty, to seat them in those and only those slots.

Well, chipsets back in the day were really picky but I thought modern ones really didn't care...

 

Except in this case your board seems to be actually borked.

Im worried about that :( So must frustration of putting the motherboard together just to unplug EVERYTHING now from the case and everything else to send it back to MSI, get a replacement and not be able to test it until the 7th of January....

 

Yet it runs at the correct speed. CPU-Z reports the actual clock of 800 megahertz whereas it still does 1600 megatransfers (which is what gives *double* data rate its name). It's a technicality and one I'd like to be remedied, but well...

CPU-Z says its running in triple data rate mode, not double. I would understand if it was running in double your logic but...
Link to comment
Share on other sites

Im worried about that :( So must frustration of putting the motherboard together just to unplug EVERYTHING now from the case and everything else to send it back to MSI, get a replacement and not be able to test it until the 7th of January....

 

CPU-Z says its running in triple data rate mode, not double. I would understand if it was running in double your logic but...

 

It's DOA, right? Perhaps you can press on the retailer? Or choose a different board instead... after all, you did buy a thing that never even worked (correctly) to begin with. If distance sales agreement is in order (that is, anything you didn't buy in a brick and mortar store, but ordered online and got in a package), retailer in question might even be obligated to immediatey exchange or refund (depends on the country; it's EU consumer protection law I'm basing it on).

 

 

CPU-Z says its running in triple data rate mode, not double. I would understand if it was running in double your logic but...

 

You're confusing things now. There's no such thing as triple data rate. It's still DDR. And neither does it become QDR or quad data rate just because there's now four of them. Each module in and of itself works at this specific rate, being able to (in theory) send data *twice* every time the internal clock ticks. You don't sum them together.

 

Consider the following example I very much like, if that helps:

SDR SDRAM on a single channel is sort of a car on a single-lane country road, speed limit 90 mph.

DDR on a single channel is the same car, but towing a trailer, still on that pitiful country road. However, more stuff can be delivered now, because of the trailer.

DDR on a quad channel is the same trailer-towing car on a four-lane highway. Now, four cars can go along each other and carry four times more stuff, but the speed limit is still 90 mph.

Link to comment
Share on other sites

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.