Question PC won't boot up and I'm stuck for ideas...

[tom_martin22]

Hi all

Having a nightmare day with my PC, grrrr 🙁

THE PC
It's an old PC based on an Intel DP35DP and Q6600. 4x2GB sticks of RAM. And an HD5450 graphics card. Windows 7. Still does me fine so no imminent plans to change unless I have to.

THE SITUATION
The PC powers up. The fans run, the optical drives light up, and I can feel the HDDs are running. But the screens stay on standby. I've got a USB wifi thing which is not lighting up. And when I move the mouse or do something on the keyboard, I get no red light on the PC to show hard drive activity. So I'm guessing from this that Windows hasn't loaded, although I wouldn't want to say 100%.

THE CLUES
CLUE1: A few months ago I was having a different issue whereby the PC would randomly turn off. I got lucky first time (for a change)!! I removed two of the four RAM sticks back then and that solved it and it's been fine since. Fast forward a few months and yesterday I finally got round to cleaning the connections on those RAM sticks, and cleaned the slots too, and then put them back in. The PC booted fine after that and showed the full 8GB. I've only used it that one time since putting them back in.

CLUE2: I have an EMU 1820M that stopped working a while ago. Because I'm not currently recording (house renovating instead!) I just unplugged the audiodock and used the onboard sound for music and internet etc. Fueled by yesterday's success with the RAM, this morning I removed the EMU PCI card and gave the terminals and slot a clean. It's since then that's I've been having the problem.

WHAT I'VE TRIED

• Removing RAM sticks, putting just one in (and trying different single sticks, in the slot that must be populated).
• Taking the EMU PCI card back out, putting it back in. Doing the hokey pokey (but not turning around cos it's all a bit stressful. Every man has his limits).
• Disconnecting optical drives.
• Disconnecting HHDs (I've got three; I disconnected two leaving just the OS drive connected).
• Removing the graphics card and giving it a clean. Setting it back in to make sure it is seated properly.
• Trying a different screen.
• Possibly other stuff that I've forgotten

If anyone has any ideas as to what else I can try to help narrow down the problem, or any hunches as to what could be the cause, then it would be great to hear from you please. The PC is overdue a clean out, which is something I do from time to time (completely take it apart, clean and re-assemble). In fact, it had been on my radar to do in the next few weeks. Perhaps a good clean and reconnecting everything might solve it. But based on some of the above, I'm not feeling all that optimistic.

Many thanks for reading and for any contributions

 

[Tac 25]

have you tried booting up the pc without the GPU?

 

[Dirtmechanic]

Did you remove the cpu from the socket at any time?

 

[tom_martin22]

have you tried booting up the pc without the GPU?

There's no onboard graphics, so I could try that but of course would have nothing else to plug the monitors into... If it's still worth trying, what should I be looking out for without the GPU connected?

Did you remove the cpu from the socket at any time?

No, that's the one thing I've never done since I built the system way back when


Thanks for the quick replies.

 

[Dirtmechanic]

I have one where I shorted the cmos by accident with the battery in and the psu turned on, and it startled me when the rig cranked up. Obviously a fault but its against good practice so I would not suggest you do it even if it is on my desk working for me. That mobo is getting replaced.

You can see the "usb wifi thing" light and such without a monitor. Those circuits on the psu have a diagram and you should verify the output if you have a failure to power. There is a trick to turn it on, there is a signal wire in the 20-24 pin that when grounded lets the rest flow. You may have a capacitor or something hot somewhere that is increasing resistance, so it may not be enough for a quick check and off. But until you verify supply it is hard to move to the tests of the next equipment. It may be a good time to bench it and clean out the dust bunnies from the psu anyway.

 

[RodroX]

There too many things going on here to be able to point out to only one component. So I would go with the basic:

Remove evertything but 1 stick of RAM and the GPU + keyboard and mouse (so out with the pci sound card, wifi card, usb sticks, hdd, ssd, everything). And see if the PC boots up and get into the BIOS.

If it does not turn On then it could be any of those components (but it will be wierd for all 4 stick of RAM to go bad at the same time so I would aim to mobo, cpu, psu and gpu instead as been culprits).

How old is the PSU non the system (Im guessing really old)?

 

[geofelt]

Sound cards seem to cause all sorts of strange issues.
Good plan to take that out for starters.
Set up with the minimum you need to boot.
CPU,cooler, perhaps 1 stick of ram(there will be a favored slot for one stick)
Graphics card. No HDD connected.
check your monitor connections.
See that your monitor input selection is what you want.
Clear the bios to default with the clear cmos procedure.

Try to boot to bios.

If no joy,
I suspect the graphics card.

 

[tom_martin22]

You can see the "usb wifi thing" light and such without a monitor. Those circuits on the psu have a diagram and you should verify the output if you have a failure to power. There is a trick to turn it on, there is a signal wire in the 20-24 pin that when grounded lets the rest flow. You may have a capacitor or something hot somewhere that is increasing resistance, so it may not be enough for a quick check and off. But until you verify supply it is hard to move to the tests of the next equipment. It may be a good time to bench it and clean out the dust bunnies from the psu anyway.

I don't quite follow all that I'm afraid, as interesting as it sounds Apart from cleaning it all out, which sounds like a good idea and was on the to do list. If you're able to explain in another way that would be great thanks.

Sound cards seem to cause all sorts of strange issues.
Good plan to take that out for starters.
Set up with the minimum you need to boot.
CPU,cooler, perhaps 1 stick of ram(there will be a favored slot for one stick)
Graphics card. No HDD connected.
check your monitor connections.
See that your monitor input selection is what you want.
Clear the bios to default with the clear cmos procedure.

Try to boot to bios.

If no joy,
I suspect the graphics card.

I've tried booting with minimal connections. So all HDDs, optical drives and USB devices are removed. With just one stick of RAM, nothing. With no sticks of RAM, I get three long beeps. With one stick of RAM but the GPU taken out, nothing (no beeps).

I haven't tried clearing the BIOS/CMOS yet so maybe that's next although I'm not sure how tbh. But I can look into that. I have replaced the battery to be on the safe side, although it was only replaced a few months ago so I felt quite sure it wasn't that.

I haven't tried removing and reseating the CPU yet either. I'm a bit nervous to do that so am leaving it until there's nothing else to try!

Must admit that I too am suspecting the GPU as well on the one hand. But on the other, would a knackered GPU prevent Windows from loading??

Thanks so much for the replies and help

 

[geofelt]

Do not remove the cpu from the socket.
If you do, you risk bending a socket pin when putting it back.
The fact that it used to run tells me that the socket is OK.

Your motherboard manual should identify a clr cmos pair of pins.
The usual procedure is to short the pins with power disconnected for a couple of minutes.

The job of a cmos battery is to hold bios settings when the power is off and disconnected.
It likely had nothing to do with your problem.

Without a gpu, windows can still load.
But, how can you tell without a display?
Yes that is possible, and you might be able to enter keyboard commands if you can guess the navigation to do so.

Unfortunately, the only way we can diagnose pc hardware issues is to exchange components with a known good replacement.

See if you can borrow a graphics card to test with.
You could buy a used 5450 card on ebay for $15 or so.
But that is such a crummy card that I would buy something better.
It is not even as good as today's integrated graphics.
I would buy something like a EVGA GTX750ti on ebay for <$100.
If you have any interest in gaming or running a 4k monitor @60hz

 

[tom_martin22]

Without a gpu, windows can still load.
But, how can you tell without a display?

I might be missing something.... But the reason I though Windows wasn't loading is because the red HDD activity light isn't lighting. So back when everything was still connected, I waited the amount of time it would take for windows to load. I then hit a few keys on the keyboard, pressed F1 (which for me opens fooBar), did a few other things I could do blindly. And none of them got the red light to flash. I'm reasonably sure they usually would. Secondly, the WiFi dongle thing usually flashes when the PC is on, even if the browser isn't open. But there was no activity on that either. So both those made me think there was no Windows loaded. Hence why it's quite useful to know whether or not Windows would load with a knackered GPU.


Good thinking RE buying another card. But tbh it serves my purposes fine for now and I want something which uses minimal resources and is silent. So perhaps I'll buy the same for now to test with and use for the foreseeable. If/when I upgrade in the future, chances are I'll get a new machine. If not, I could just upgrade the GPU once I know what I need to get out of it, and the HD5450 could sit in a box as a spare for times like this. I've no plans to game (too addictive!), but I might one day want to plug the PC into a large TV rather than the two 19" monitors I currently run.

Thanks again for the help

 

[geofelt]

Could you have a simple solution such as the monitor not selecting the correct input, or that the connecting cable is loose or defective?
I have seen this before when I knew that windows was running.

You have to get past the lack of the ability to display the bios first.
I suspect you are correct that windows is not loading.
Only then can you move along to windows loading or not.
Once you can see the bios, you will know that the pc hardware is sound.

Windows may not load because of a faulty driver or start up app.
The way to diagnose that is to start windows in safe mode(f8)
That loads windows with only essential safe drivers.

Next to the graphics card, you may have a bad motherboard.
It is most unusual for an intel processor to go bad.

I would not go so far as to buy a replacement motherboard.
$250 will buy you a much stronger current gen processor like the i3-10100 with integrated graphics, a lga1200 motherboard and 16gb of ddr4 ram.

 

[Dirtmechanic]

I don't quite follow all that I'm afraid, as interesting as it sounds Apart from cleaning it all out, which sounds like a good idea and was on the to do list. If you're able to explain in another way that would be great thanks.



I've tried booting with minimal connections. So all HDDs, optical drives and USB devices are removed. With just one stick of RAM, nothing. With no sticks of RAM, I get three long beeps. With one stick of RAM but the GPU taken out, nothing (no beeps).

I haven't tried clearing the BIOS/CMOS yet so maybe that's next although I'm not sure how tbh. But I can look into that. I have replaced the battery to be on the safe side, although it was only replaced a few months ago so I felt quite sure it wasn't that.

I haven't tried removing and reseating the CPU yet either. I'm a bit nervous to do that so am leaving it until there's nothing else to try!

Must admit that I too am suspecting the GPU as well on the one hand. But on the other, would a knackered GPU prevent Windows from loading??

Thanks so much for the replies and help

Sure! Not everybody has a multimeter, or a multimeter with frequency. They should start at the wall and verify 60 hertz but cell phones are bad at that kinda thing. Here is a picture of what one would test a psu for and you see a meter is needed.
https://www.google.com/search?q=atx...biw=412&bih=764&dpr=2.63#imgrc=IVW7qoacJp47_M

 

[tom_martin22]

A couple of updates ... I forgot that I have a spare PC I picked up a while back for parts. It's onboard graphics so no joy there, but I took out the PSU and temporarily fixed it up to my main PC. I only connected enough for testing purposes, namely the large motherboard connector, the CPU connector and the system fans. With a stick of RAM in and the GPU in (but everything else still disconnected), I get the same result. IE nothing!

Does that mean the finger is pointing to the GPU?

The other update is a non starter. I took the GPU back out and gave it another visual inspection. I noticed two of the gold teeth were shorter than the others and had a semi EUREKA! But just been googling that and turns out it's meant to be like it. As in this thread:

https://forums.evga.com/GTX-690-Missing-half-part-of-gold-pinwarranty-voided-m1932991.aspx

Could you have a simple solution such as the monitor not selecting the correct input, or that the connecting cable is loose or defective?

Good thinking but I've checked all that - I've tried different monitors, a TV, VGA only, DVi only, HDMI only, etc.

Sure! Not everybody has a multimeter, or a multimeter with frequency. They should start at the wall and verify 60 hertz but cell phones are bad at that kinda thing. Here is a picture of what one would test a psu for and you see a meter is needed.

Thanks. I've got a multimeter. I've had a read up and think I can test the PSU, but best to do it when I'm not as stressed and a bit clearer of mind so that I don't do anything stupid. That's actually what led me to try the other PSU in the meantime. Guess the fact I'm getting the same results with that is pointing the finger away from the issue being with the PSU??

Big thanks

 

[geofelt]

Psu testers are junk. A tester costing less than $5000 can confirm a dead psu, but not confirm proper operation.
It does not sound like the psu is the issue since another psu showed the same results.

 

[Dirtmechanic]

A couple of updates ... I forgot that I have a spare PC I picked up a while back for parts. It's onboard graphics so no joy there, but I took out the PSU and temporarily fixed it up to my main PC. I only connected enough for testing purposes, namely the large motherboard connector, the CPU connector and the system fans. With a stick of RAM in and the GPU in (but everything else still disconnected), I get the same result. IE nothing!

Does that mean the finger is pointing to the GPU?

The other update is a non starter. I took the GPU back out and gave it another visual inspection. I noticed two of the gold teeth were shorter than the others and had a semi EUREKA! But just been googling that and turns out it's meant to be like it. As in this thread:

https://forums.evga.com/GTX-690-Missing-half-part-of-gold-pinwarranty-voided-m1932991.aspx



Good thinking but I've checked all that - I've tried different monitors, a TV, VGA only, DVi only, HDMI only, etc.



Thanks. I've got a multimeter. I've had a read up and think I can test the PSU, but best to do it when I'm not as stressed and a bit clearer of mind so that I don't do anything stupid. That's actually what led me to try the other PSU in the meantime. Guess the fact I'm getting the same results with that is pointing the finger away from the issue being with the PSU??

Big thanks

Lol Absolutely! I had an extra case and swapped mobos because that is a lot of work and work has an hourly rate I did not care to pay!

 

[tom_martin22]

An update... Some bright spark pointed out that with another PC I could try the GPU in that. No idea why I didn't figure that out for myself. So I've just gone and done that and lo and behold it worked. So the GPU is off the hook.

So all fingers starting to point to the motherboard now? I'm definitely going to give the PC a strip down and clean up and see if something has come loose, unseated, or if a fluff ball of dust is causing havoc... But now know it ain't the GPU, is unlikely to be RAM ('I've tried different sticks, and get no bleeps with RAM installed), and is unlikely to be the PSU as I tried the one from the other system and got the same result. That being said, the borrowed PSU is only 250W (the existing is 430W). But seeing as I tried the PC with only the bare amount of stuff connected I didn't think that would matter. Please correct me if I'm wrong though!

Thanks for all the help. I'll'get stripping

 

[Dirtmechanic]

An update... Some bright spark pointed out that with another PC I could try the GPU in that. No idea why I didn't figure that out for myself. So I've just gone and done that and lo and behold it worked. So the GPU is off the hook.

So all fingers starting to point to the motherboard now? I'm definitely going to give the PC a strip down and clean up and see if something has come loose, unseated, or if a fluff ball of dust is causing havoc... But now know it ain't the GPU, is unlikely to be RAM ('I've tried different sticks, and get no bleeps with RAM installed), and is unlikely to be the PSU as I tried the one from the other system and got the same result. That being said, the borrowed PSU is only 250W (the existing is 430W). But seeing as I tried the PC with only the bare amount of stuff connected I didn't think that would matter. Please correct me if I'm wrong though!

Thanks for all the help. I'll'get stripping

Newegg and others have wattage calculators for the purpose of psu sizing by component. They are pretty cool.

 

[tom_martin22]

I've had a breakthrough!!! I'm just not sure how and it's left me completely puzzled. With the spare PC to play with, I started swapping components between that and the broken PC.

Order of events...

- I put two sticks of RAM from the broken PC into the spare PC. There's no OS on that, but during boot and in BIOS it showed 4096MB so if nothing else I knew it was registering it.

- I did the same with the other two sticks of RAM from the broken PC. Same result.

- I wasted two hours trying to put an old copy of XP on the spare PC just so that I could test the RAM further, but I couldn't get it installed. Grrr

- Based on a suggestion from someone else, I tried the broken PC with the GPU back in, but no RAM. Just to see if I could get to BIOS. I got the 3 beeps but nothing else. No graphics and monitors stayed off.

- Seeing as the RAM slots were empty, I thought what the heck and stuck the RAM from the spare PC into the broken one. I checked whatever specs I could to ensure compatibility with the mobo and most seemed ok (one or two I could not get an answer for). It worked! The screen came on and I had graphics. Unbelievable and very surprised to see that screen come on.

- I removed that "spare" RAM and stuck in two of the RAM modules that belong in that PC. They worked as well. Graphics/screen came on.

- I removed those sticks and put in the other two. The worked as well.

It goes totally against the grain of every test and trial I've done to date. But hey, things are looking up for now.

I guess the finger is pointing towards either the RAM or the motherboard. Way back in the first post, I said:

CLUE1: A few months ago I was having a different issue whereby the PC would randomly turn off. I got lucky first time (for a change)!! I removed two of the four RAM sticks back then and that solved it and it's been fine since. Fast forward a few months and yesterday I finally got round to cleaning the connections on those RAM sticks, and cleaned the slots too, and then put them back in. The PC booted fine after that and showed the full 8GB. I've only used it that one time since putting them back in.

Could it be that the RAM is causing an intermittent fault? But if so, would that prevent the graphics from loading / the screens from coming on? Based on the marathon of posts and scenarios above, if my PC continues to work what tests can anyone recommend to help determine what the problem is?

Sorry for the long post!

Cheers