Sign in with
Sign up | Sign in
Your question

The tale of old sparky.

Last response: in Systems
Share
January 30, 2012 2:09:11 PM

Hi guys,

Hoping you can help me with a few questions with an issue i'm having.

I was playing a video game about a week ago and there was a loud bang from the PSU accompanied by a white spark, I switched it all off and left it for a while, I took the psu out of the system and left it on the side, i powered it up just to see if it was dead and it powered up and the fan spun up but there was a humming sound coming from it. I figured a power rail blew and time to purchase a new PSU as the old one was about 4 years old and i've asked alot of it. I purchased a new one and when i put it in there was another big white spark and an even bigger bang. I did use a different socket and a different power lead also I thought maybe that could be the cause and decided best to change it up just in case.

I RMA'd the new psu back to the vendor after christening it Old Sparky (and as of today they've tested it and found a fault and refunded my money). I ordered a new psu (different make and model) and that arrived and that tested ok (i.e. it didn't explode when i plugged it in outside of the case like old sparky did)

I powered on the PC first off the fans spun for a second or 2 then nothing. So I took everything out and just left in one stick of ram, no video card and cleared the CMOS (unplugged it from the mains, moved the jumper and removed the battery) the machine then POSTED as though it was looking for a video card, so i put it in SLI Slot 1 and it was still posting looking for a video card also tried SLI slots 2 and 3 same as before. So i put the video card carefully to one side and put the other stick of RAM in the machine and there was no post beep what so ever. The machine would power up fans spinning no beep it would just sit there, ideally it should have post beeped as though it was screaming for a video card. I moved the ram around between all 4 slots and tried every combo i could think of until I was back to it post beeping looking for a video card with one stick of RAM in slot A1. I also switched between RAM sticks to but I'm trying not to give more info than needed.

I've tried the RAM in another PC but it doesn't recognize the memory it refuses to boot until i put the old stuff back in. I've also tried the video card in the other machine and it'll post with it in and from what i can see theres no graphics corruption, although i've not been able to test it under load because the design of the mobo i was using to test is really horrendous (the SATA connectors are to close to the SLI slot and you cant put a full card into the slot and press home because the SATA cables are in the way.

I also cant test the CPU because the only machine I have to test with is not able to take the CPU as far as i'm aware.

the specs for my machine are as follows.
1xWD 150GB Raptor
2x Maxtor 1TB HD
ASUS M4N82 - Deluxe Motherboard
4 GB RAM 2x 2GB 240-PIN UNBUFF DIMM 256MX64 DDR2 PC2-8500
1X Phenom 9850 BE
1x ASUS ENGTX560 DCII OC/2DI/1GD5 (Nvidia 560 GTX)

Test machine uses a ECS Geforce 7050MM (V.2.0) Motherboard.
a Athlon X2 2.2Ghz
2GB Ram
Nvidia 520

I had a 600W psu in there and finally replaced with a 700W that tested ok, the one that was RMA'd was also 700W

Does anyone have any idea if anything else has gone apart from the mobo? I'm thinking that has a fault on it because the video card was found and used on the test machine. Unless anyone else disagree's?

Also thanks for reading, sorry about the wall of text but I wanted to give as much info as possible.

More about : tale sparky

a c 78 B Homebuilt system
January 30, 2012 4:55:01 PM

Try the Nvidia 520 in the broken PC and see if it works.

Also, I would like to know the maker and model of the PSUs in question.

The room temperature difference that two different models of 700w PSUs can put out can literally be 700w difference (one 700w maxes at 350w and the other maxes at 1050w).

That being said, the motherboard could quite easily be the point of failure at the moment. When a PSU blows up it often sends absolutely ridiculous levels of current through the wires to anything connected to it before it shuts down, so it often takes other things down with it on the way out.

The video card often doesn't make it through this sort of thing unscathed, but if the other computer works fine with the GTH 560 in it then it may not have received any damage.

The PSU is connected twice to the motherboard, though, so it is possible too much juice went down either of those lines and fried the board.

I doubt the CPU was destroyed in this case. Usually the damage is localized to the things the PSU is directly connected to and there is no line from the PSU to the CPU chip itself, only from the PSU to the board, then to the chip.

Try the old video card in the broken computer and let me know if anything changes.
a b B Homebuilt system
January 30, 2012 5:24:54 PM

"I've tried the RAM in another PC but it doesn't recognize the memory it refuses to boot until i put the old stuff back in."

This suggests to me the ram might be messed up. Do you have some test ram you can try in the "sparky board"?
Related resources
Can't find your answer ? Ask !
January 30, 2012 5:33:17 PM

yup, might be the RAM.
a c 78 B Homebuilt system
January 30, 2012 5:46:33 PM

RAM is a complicated beast and a very touchy thing at that. Especially if you are talking about old DDR2 level hardware.

Say you have a computer with 2 RAM slots. You shove in 1x 2GB stick in there and boot it up. Then you shut it down and stick in a 2nd 1x 2GB stick exactly like the first one and you turn it on again.

What is going to happen? Think it is guaranteed to work? Think again.

What is equally likely is that the computer will turn on for 2 seconds, recognize that the RAM amount has changed, and enter a restart loop that won't ever end. You turn it off and turn it back on, same thing.

The BIOS records configuration data for type and amounts of RAM. Especially with older hardware the computer isn't guaranteed to auto-detect this and initiate a new configuration from scratch.

To get it to work may require taking the battery out for 30 minutes and putting it back in so as to force the BIOS to reconfigure from scratch.

I am going to hazard a guess that the RAM may just work in the 2nd PC if the BIOS were forced to detect itself again.

The fact that the computer worked right when the old ram was re-inserted matches up exactly with this issue.

If you ever come across such a problem in the wild where you are trying to insert a 2nd stick of the exact same kind into a formerly empty slot and it doesn't work, but where you can put the 2nd stick in the original slot (after taking out the original stick) and it WILL work, the BIOS not scrapping the old configuration and starting over is very likely to be the problem.

This smells very much of that.

This is why I am focusing elsewhere.
January 30, 2012 6:24:47 PM

Thanks for the responses guys, i'll take each one in turn and give you guys as much info as possible.

Raiddinn

The first PSU which blew originally with a suspected rail blown I cant tell you the make and model only that it was 600W I've actually thrown that out in the bin so that ones gone. The one I've dubbed "Old Sparky" was a EZ Cool 700W Tornado PSU. The one that appears to be working now is a PowerCool GT 700w PSU 80+ Dual 12V V2.2 High Efficiency.

As i said the video card I'm unsure if it is working 100% or not, all i've basically done is attached it to the other pc I have and turned it on, see if the machine seems to boot with this ok or not. The board itself wont actually allow a full sized video card to interface with it due to poor design as the SATA connections when used prevent the card going in all the way. All i've seen with the video card is the machine booting and a screen asking for the OS. But its black and white with no random colours. But because the hard disk is SATA I've had to unplug it so I have yet to test it in windows.

Bucknutty.

I'll take the ram from the test PC and try it with the other PC that had the nukeage see what happens with that, I'm also gonna flush the CMOS on the test PC and try my RAM again with it.

Thanks again for your help guys.
a c 78 B Homebuilt system
January 30, 2012 7:21:03 PM

There are SATA cords with L shaped connectors that can be used when video cards and SATA cords need to occupy the same space.

Here is an example

http://www.eio.com/p-18002-syba-sy-cab40006-17-sata2-ca...

You might want to think about getting a few of them. I have a few laying around at my house in case the need arises because of the same sort of problem.

It is good to have cable flexibility to allow testing that isn't currently allowed. They are pretty cheap too.

In any event, regular generic PSUs tend to die in the 3 - 5 year range so the 600w made it about as long as it should have, especially if the PSU was top mounted.

The EZ Cool 700w is just not a good type and I can understand that thing blowing on day 1.

PowerCool seems to be a generic brand that is probably about equal with the EZ Cool.

PSUs that can't supply the necessary power can act in very many different ways and it is possible the PowerCool is exhibiting such problems as well. I would think it would be fine, because those parts are kinda old, but old can also mean high power consuming as well.

Did you try putting the 520 in the broken computer?
January 30, 2012 8:37:42 PM

Raiddinn said:
There are SATA cords with L shaped connectors that can be used when video cards and SATA cords need to occupy the same space.

Here is an example

http://www.eio.com/p-18002-syba-sy-cab40006-17-sata2-ca...

You might want to think about getting a few of them. I have a few laying around at my house in case the need arises because of the same sort of problem.

It is good to have cable flexibility to allow testing that isn't currently allowed. They are pretty cheap too.

In any event, regular generic PSUs tend to die in the 3 - 5 year range so the 600w made it about as long as it should have, especially if the PSU was top mounted.

The EZ Cool 700w is just not a good type and I can understand that thing blowing on day 1.

PowerCool seems to be a generic brand that is probably about equal with the EZ Cool.

PSUs that can't supply the necessary power can act in very many different ways and it is possible the PowerCool is exhibiting such problems as well. I would think it would be fine, because those parts are kinda old, but old can also mean high power consuming as well.

Did you try putting the 520 in the broken computer?


With both the 520 and the known working stick of ram in, it refuses to post so i cleared the CMOS and tried again to no avail.

I took the ram from my PC and the video card ie the 560 and took your advice regarding the cabling (honestly dunno why i didn't think of it before prolly having a blonde moment >.<) got into windows with the 560 it installed ok and ran up Star Trek online for 10 mins flew around run around a space station had no issues what so ever, so i took that out and put one of the sticks of ram in that was in my PC, (the one that went boom) and it booted fine, and i sat at the PC looking stunned and windows showed 2 GB RAM. Loaded fine no errors in the logs so i put the 2nd stick in to make it up to 4 (the 2nd of the matched pair) and it booted but it only showed 2GB RAM tried it seperately and it refused to boot, so it looks like thats gone Elvis. Replaced 520 and original RAM in test machine and here I am.

I'm guessing I got lucky with the video card but not so lucky with the stick of RAM and Mobo which I think's both gone Elvis?
a c 78 B Homebuilt system
January 30, 2012 8:52:56 PM

There are settings in Windows that can limit the max ram that a computer recognizes. Also, an amount of RAM equal to the video card RAM is usually reserved in the computer's main RAM too which could also be the reason for the drop you are seeing.

I will try to get back to you on these things in maybe 6 hrs hopefully, or if not then maybe tomorrow morning.

Anyway, I am thinking motherboard and not RAM at the moment.

If you are feeling enterprising, you can try googling about "full ram not recognized in windows" or something and looking through the links. You might be able to locate some of the things yourself that I will tell you how to do when I get more time myself if you don't do it. It would save you some time at least.
January 31, 2012 1:41:16 PM

Raiddinn said:
There are settings in Windows that can limit the max ram that a computer recognizes. Also, an amount of RAM equal to the video card RAM is usually reserved in the computer's main RAM too which could also be the reason for the drop you are seeing.

I will try to get back to you on these things in maybe 6 hrs hopefully, or if not then maybe tomorrow morning.

Anyway, I am thinking motherboard and not RAM at the moment.

If you are feeling enterprising, you can try googling about "full ram not recognized in windows" or something and looking through the links. You might be able to locate some of the things yourself that I will tell you how to do when I get more time myself if you don't do it. It would save you some time at least.


I have run across the 3 GB limit on the 32 bit version of Windows . Although the test machine is running the 64 bit version of windows 7. Also to make sure I wasn't tripping i took the memory from my machine and took out the original memory from the test system, and replaced it with the "bad" memory and the system refused to boot, just posted as though it was looking for RAM replaced it with its "twin" and the machine happily loaded up into windows.
a b B Homebuilt system
January 31, 2012 1:48:58 PM

I think its safe to conclude that "old Sparky" killed that 1 ram stick. Thats not bad, as it could have been much worse and a 2x2gb kit of ddr 2 800 is only around $50.
January 31, 2012 2:04:21 PM

bucknutty said:
I think its safe to conclude that "old Sparky" killed that 1 ram stick. Thats not bad, as it could have been much worse and a 2x2gb kit of ddr 2 800 is only around $50.


I totally agree mate, I contacted the vendor who i bought that crappy psu from and had a chat with them for 30 mins or so, from what they've said if i can get an engineer's report that states that there psu nuked those components then they'll look at replacing them for me.

As far as i'm aware it voids the warranty on pretty much everything. The problem i have now is trying to find a mate or family member who would write one up for me as cash is pretty tight atm to say the least.
a c 78 B Homebuilt system
January 31, 2012 2:10:29 PM

Perhaps the PSU did kill the one stick of RAM, but that should be determined by using MemTest86+.

If I read it right, the test computer was able to boot with both sticks from the broken PC and it just looked like 1 stick wasn't contributing to the amount of RAM available.

If so, the above MemTest86+ program can easily point out whether the RAM is completely broken.

It is also possible that there is a windows setting that is limiting the maximum amount available to the OS and the system problems are just due to the fact that RAM is touchy in general.

To see whether there is a setting limiting the RAM available to Windows, hit start, type in msconfig, go to the boot tab, advanced options, and see if the maximum memory box is checked and if so what the limit is that is typed into the box.

Sometimes Windows checks this box and writes in a number by itself with no user involvement for whatever reason and it is worth checking since it takes only a few moments to do.

If this setting is checked, Windows can report a wide variety of maximum memory, and it is completely separate from the amount of memory that a 32 bit OS is capable of recognizing.
January 31, 2012 2:37:37 PM

Raiddinn said:
Perhaps the PSU did kill the one stick of RAM, but that should be determined by using MemTest86+.

If I read it right, the test computer was able to boot with both sticks from the broken PC and it just looked like 1 stick wasn't contributing to the amount of RAM available.

If so, the above MemTest86+ program can easily point out whether the RAM is completely broken.

It is also possible that there is a windows setting that is limiting the maximum amount available to the OS and the system problems are just due to the fact that RAM is touchy in general.

To see whether there is a setting limiting the RAM available to Windows, hit start, type in msconfig, go to the boot tab, advanced options, and see if the maximum memory box is checked and if so what the limit is that is typed into the box.

Sometimes Windows checks this box and writes in a number by itself with no user involvement for whatever reason and it is worth checking since it takes only a few moments to do.

If this setting is checked, Windows can report a wide variety of maximum memory, and it is completely separate from the amount of memory that a 32 bit OS is capable of recognizing.



Well you learn something new every day, thanks Raiddinn never knew that setting was there, although I've not had to mess around with msconfig to much, just disabling stuff in the startup sequence and messing around with the boot menu and stuff. I just checked it on the test system and the box is unchecked although I cant go switching memory in and out just yet because my bro's on it and in the middle of a raid on WoW, but i'll try it prolly after dinner and post back some findings.
January 31, 2012 8:20:33 PM

Okiday, I've run the test system back up with both sticks of ram from the system with the assploding psu the system only detected 2Gb ram and so did memtest when i booted it up using a USB stick. It didn't even find the 2nd stick and i changed the settings etc to try and get it to look for more than 2gb RAM but it didn't. I think its safe to conclude that the 2nd stick of RAM is dead.

Thanks for your help guys with this, once again you rock :) 
a c 78 B Homebuilt system
January 31, 2012 9:00:54 PM

I guess that does kinda narrow it down.

Hope that is all that got fried other than the PSU.
!