Question Previously working PC switched to new case, won't start up

[cpm18]

Brief background: Had a case from an old PC build that I reused to build a new system last year. The power button on the old case had given me problems with the old system and would sometimes take multiple pushes to start up. The new build was built in this case and everything worked fine, except the power button was still a bit troublesome. System has been running fine for last year.

Recently, I've had a few power outages, including a brown out. My system was connected to a UPS so I don't know how these power issues may have impacted it. But after the latest power issue, I was finding myself unable to power on. Being fed up with this old case, I ordered a new case and planned to migrate everything into the new case.

I finished moving everything into the new case and powered it on. Everything started coming up but after 5ish seconds, fans, lights, etc turned off. Oddly, the RAM which lights up was still lit up even though the rest of the system was down. I opened the case back up and started checking connections, couldn't find anything wrong. Subsequent attempts to power up result in lights coming on for a split second, fans starting to turn, then immediately everything shuts down. I am getting a red LED for my CPU on my motherboards debug lights, right before everything turns off. I've continued testing, tried pulling RAM, graphics card, drives, etc. Went down to the bare minimum, just to see if I could get it to start up but continue to get the same brief start up, then shutdown, with the CPU LED error.

At this point, I believe either the CPU or motherboard may need replacing. While the error light indicates CPU, the CPU was never disconnected during the case move and had been working fine before. It's possible the power issues could have damaged it, but I do not think it is improperly seated. I'm also not sure if the motherboard is damaged and incorrectly reporting the CPU problem.

Does this sound like a CPU or motherboard issue?

 

[Ralston18]

Update your post to include full system hardware specs and OS information.

PSU: make, model, wattage, age, condition (original to build, new, refurbished, used)?

Disk drive(s): make, model, capacity, how full?

Make and model case(s)?

Do you hear any beeps? Some pattern of long and/or short beeps at boot time. Does the motherboard have its' own speaker?

 

[Paperdoc]

Here is one common cause of your problem - may not be YOUR cause, but you can check.

Every case comes with Stand-offs, often pre-installed, sometimes in a bag. These are short metal items about ¼" long with a threaded shaft out one end, and a threaded hole in the other. They MUST be placed in the CORRECT spots under the mobo. Most mobos have 9 mounting holes in three rows of three, but the exact locations of them vary from one mobo to another. Ideally, you should have one Stand-off under each mounting hole in the mobo for support AND to create a space away from the case's back plate. BUT you must NOT have any under the mobo where there is NOT a mounting hole! If you look closely at those nine mounting holes you'll see little metal "fingers" around each. The mobo is supposed to be grounded to the case back panel at these points (by the screw and Stand-off) ONLY. If you have a Stand-off under the mobo at the wrong place, it can short out a mobo trace on the back side.

So, IF your case did NOT have any Stand-offs already installed in its back panel, FIND them. If they WERE installed. you need to check their positions. Study the MOBO mounting holes. Carefully unscrew it from the back plate and remove. Compare closely the Stand-off locations (OR install if needed) to the mounting hole locations and ensure EVERY one matches up. The case's back plate has many threaded holes so that you can re-position any Stand-off to a mobo mounting hole location. Carefully re-position the mobo in the case and verify that there are Stand-offs ONLY where they match the mounting holes of the mobo, then re-install the screws.

 

[Karadjgne]

How was the motherboard moved over? Was the cpu cooler removed and cleaned/repasted or did you just move it directly over in one piece?

Brownouts/blackouts, sudden shutdowns don't hurt hardware, they only ever affect software, data loss. And with an actual UPS, none of that should have affected the pc at all.

But a move can affect hardware, especially the cpu if it moved because the pressure of the cooler changed.

I'd start out with inspecting the socket, make sure there's no bent pins, make sure the cpu is placed back where it should be, repaste.

 

[cpm18]

How was the motherboard moved over? Was the cpu cooler removed and cleaned/repasted or did you just move it directly over in one piece?

The CPU was not removed from the motherboard, nor was the fan cooling it, everything moved together. I would assume it is still seated properly without bent pins but I can't say for certain as I wanted to avoid tampering with it, unless necessary.

 

[cpm18]

So, IF your case did NOT have any Stand-offs already installed in its back panel, FIND them. If they WERE installed. you need to check their positions. Study the MOBO mounting holes. Carefully unscrew it from the back plate and remove. Compare closely the Stand-off locations (OR install if needed) to the mounting hole locations and ensure EVERY one matches up. The case's back plate has many threaded holes so that you can re-position any Stand-off to a mobo mounting hole location. Carefully re-position the mobo in the case and verify that there are Stand-offs ONLY where they match the mounting holes of the mobo, then re-install the screws.

I had considered that something could be shorting. I tried testing this by removing the motherboard from the case and laying it on a piece of cardboard next to it, then reconnecting and powering up. This should have prevented anything contacting, but the result was the same as when the board was in the case. It's possible something did cause shorting initially and the damage is already done, hard to tell at this point.

 

[Paperdoc]

It's entirely possible the problem is not a short. You have done the "breadboard" basic process and it would not start outside the case. If you have not done the following yet, try it.

The "breadboard" process is a sequence to identify the source of the problem. It starts with the system outside the case on a non-conductive surface like a dry board; hence the name. Remove most mobo components, leaving only the essentials: mobo, PSU connected to it, ONE stick of RAM, CPU AND its cooler in place and plugged in, no drives, no graphics card. Use a screwdriver to short out briefly the two pins of the Front Panel header where the case On / Off button power leads normally are connected. If it will start up and keep going, that's a good start. Shut down by turning off the PSU. However, if it will not start and keep running, you now the problem is with those basic components. Among those, I'd try first substituting a different PSU - does not need to have high power capability. Maybe try a different RAM stick, or none.

Once you do get a basic start-up, on the next try add ONE item - the graphics card and monitor OR, if the video is generated on-board, plug the monitor into that rear port. See if that runs, and this time you can see on-screen errors like no keyboard and no boot drive. Shut down.

Next add the keyboard and try. If that still works, try adding one more item each time - maybe more RAM. Eventually if it keeps starting up you can add the boot drive and the system should boot up completely. Then you can try other things like USB devices, etc. At some point you will get failure and you know the last item added is the likely culprit.

 

[cpm18]

It's entirely possible the problem is not a short. You have done the "breadboard" basic process and it would not start outside the case. If you have not done the following yet, try it.

The "breadboard" process is a sequence to identify the source of the problem. It starts with the system outside the case on a non-conductive surface like a dry board; hence the name. Remove most mobo components, leaving only the essentials: mobo, PSU connected to it, ONE stick of RAM, CPU AND its cooler in place and plugged in, no drives, no graphics card. Use a screwdriver to short out briefly the two pins of the Front Panel header where the case On / Off button power leads normally are connected. If it will start up and keep going, that's a good start. Shut down by turning off the PSU. However, if it will not start and keep running, you now the problem is with those basic components. Among those, I'd try first substituting a different PSU - does not need to have high power capability. Maybe try a different RAM stick, or none.

I had been trying something similar to this. I pulled the motherboard out and laid it on a piece of cardboard. I have only the CPU hooked up to it, and I am powering the motherboard and CPU. But I had not tried the method to short out the 2 pins for the power button. I tried that but my result appears to be unchanged. The board powers up for a split second, I see the fan make a turn and the red LED for CPU error is lit up, before everything shuts off again. So unfortunately, I cannot test any additional components since I am unable to get even the basic pieces to run.

I didn't have a spare power supply other than one I use in a PC I am currently using for work, so I am reluctant to remove it from there just for testing purposes. I had hoped I might have stumbled across some clues to narrow down which is the faulty component (motherboard or CPU) but its still hard to say. At this point, I might just have to make my best guess and order a replacement part. I will probably try a new motherboard first, as the CPU was never disconnected from the motherboard during my initial case migration, so it seems unlikely that the CPU would have been damaged.

Appreciate all the suggestions.

 

[Paperdoc]

Seems a good starting point form what you already know. The symptoms suggest falure is in the PSU, the mobo, or the CPU chip.

By the way it IS important that, when you do this testing, the CPU cooling FAN is connected to the CPU_FAN header and DOES start up immediately. Many mobos will NOT let you start up if the CPU_FAN header does NOT receive immediately a valid speed signal from a fan plugged into it. IF that fans is suspect and does not start up right away, you can fake it for testing by plugging into that header any fan you KNOW is good. If THAT gets it to start up you will know the FAN may be the problem.

 

[cpm18]

By the way it IS important that, when you do this testing, the CPU cooling FAN is connected to the CPU_FAN header and DOES start up immediately. Many mobos will NOT let you start up if the CPU_FAN header does NOT receive immediately a valid speed signal from a fan plugged into it. IF that fans is suspect and does not start up right away, you can fake it for testing by plugging into that header any fan you KNOW is good. If THAT gets it to start up you will know the FAN may be the problem.

I did have an old fan which had worked the last time it was used (~2 years ago). I tried plugging that one into CPU_FAN to test, but results appear to be the same. Since I am not 100% sure the fan still worked, I can't eliminate the fan as an issue, but does seem more likely that it is CPU, motherboard, or PSU.

 

[Paperdoc]

Actually, you probably have at least one other fan already plugged into the mobo for a case vent function. Try using that plugged into the CPU_FAN header instead. The mobo does NOT prevent booting when it has no fan on a SYS_FAN header.

 

[Karadjgne]

The basic theory is that aircoolers are 'stock', so the cpu has a built in sensor attached to cpu_fan header. If it sees 0rpm or non-populated cpu_fan header it's assumed the cpu has no cooler installed so protects itself from burning up in seconds by stopping any boot process and shutting down instantly.

Default is Protected, but that can be disabled in bios. (not recommended without reason)

 

[cpm18]

I ended up taking a gamble that it was a bad motherboard (even though the motherboard showed a CPU error light). Turns out it was correct as I received the new board, switched everything over and I am back up and running.

Thanks to everyone who offered suggestions and ideas to troubleshoot.