[Okta Webinar] Learn how to a build a cloud-first strategyRegister Now

x
?
Solved

-5v rail keeps failing in the same system.

Posted on 2009-12-16
25
Medium Priority
?
1,455 Views
Last Modified: 2013-12-11
Hi guys,
I have a system thats a few years old, its a P4 3.2ghz prescott, 2gb ddr1 RAM, Radeon x1950 Pro, and its been through about 4 Power Supplies's. each time the -5v rail dies. Ive used Generic 450Watts, a 550Watt ThermalTake, and a 750Watt Arctic PSU for it and they all had the same problem.
I am wondering what might have caused this to happen 4 times and whether or not any of my broken PSU's can be repaired?


Cheers
0
Comment
Question by:true_blue
  • 11
  • 11
  • 2
  • +1
25 Comments
 
LVL 12

Expert Comment

by:hathehariken
ID: 26066991
the -5v rail is only present in the ATX connector.
to the best of my knowledge, the -5v rail gets used somewhere in the memory controller and the AGP/PCIe slot(s)

do you have any graphics card?
if not, are the slots free of dirt and dust?

how clean is the cabinet and motherboard?

can you visually check for conducting debris like pins, staples that may be lodged on the motherboard?
0
 

Author Comment

by:true_blue
ID: 26067370
Hi yes I have a ATI Radeon x1950 pro 512MB (AGP)  graphics card like I said.
There is no clutter inside the case as I clean it every few months along with the fans. The contacts on the AGP port look fine to me, and the system ran great for months at a time but then all of a sudden the PSU would blow. I have gone through 4 of them now and no matter what wattage or brand they are they always seem to fail after 6-12 months.
Could it be a faulty motherboard?
Also when I had the thermaltake PSU installed I noticed after removing it that the 4pin port on the motherboard had burns on the top 2 prongs. and also on the PSU's connector. I gussed it would be down to lack of power from PSU so went for a 750Watt the next time and then now the same thing happened but without the burns this time.

0
 
LVL 2

Expert Comment

by:curtisla
ID: 26068879
true_blue

I do venture the advise that MB is defective. There is a outside possibility that your CPU is damaged as well.

The thermaltake PSU 4 pin connector being discolored is of no aid as this is a +12 Volt supply to the CPU power supply onboard the MB. The -5volt supply comes to the motherboard on a white wire in the 20 pin molex connector from the PSU.

According to the site formfactor.org the -5 volt is not being used as of 2004.

So why your unit is frying the PSU's is a indication of a defective MB. The power supply buses are heavily bypassed with filter capacitors. It is usually one of these that causes the problems.

Sorry
0
Reclaim your office - Try the MB 660 headset now!

High level of background noise often makes it difficult for employees to concentrate fully on their jobs – or to communicate clearly on calls. The MB 660 headset helps you create a disruption free workspace.  

 
LVL 26

Expert Comment

by:PCBONEZ
ID: 26068886
-5v was only used for ISA slots.

It's been optional in power supplies since ATX Spec ver:1.3 [year 2003].
Few PSU's even have it anymore.

Considering you are running a P4 Prescott you don't likely have any ISA slots and so there is nothing to send -5v to on the motherboard.
- So why do you want it?

.
0
 

Author Comment

by:true_blue
ID: 26069161
ok then im very confused.. why and how would a modern motherboard blow out a -5v rail if its not being used, and why would that prevent the PSU from supplying power properly?
I.E. the machine turns on and goes off again once the OS starts to load. Also after plugging in a few neon lights and fans to the PSU and turning it on manually it cuts out after a few secs.. but if I leave it running with nothing plugged in it runs constant, but I can also notice the fans arent spinning as fast as they did before the PSU stopped working.
Yet when I use my Dr Power Thermal Take PSU tester on the PSU's all lights are solid green except the -5v.
What does this mean and why arent the PSU's working anymore if the tester says all needed rails are fine?

Thanks for everyones help by the way..

0
 
LVL 26

Expert Comment

by:PCBONEZ
ID: 26070260
The testers are built to accommodate both new and old PSU specs so they have the -5v LED.
-
The -5v doesn't come on with your PSU tester because the power supply doesn't have a -5v rail.
It didn't go away. - You never had it to start with.
-
There is SOME problem obviously. But missing -5v isn't it.
.
0
 
LVL 26

Expert Comment

by:PCBONEZ
ID: 26070271
Look at the label on the side of these power supplies and tell me how many amps they have on -5v.
If it's not there, the unit doesn't have a -5v rail.
.
0
 
LVL 2

Expert Comment

by:curtisla
ID: 26073069
One small point of interest is that I can not locate any information on the brand of MB being used except for the family group of  P4 3.2ghz prescott CPU. What is the manufactor of the MB and what is the part number. Then we could locate the internal power usage list and tell you if the MB does use the -5 volt supply at all. Also, some mfgs. have been know to stray from the published specifications to limit the exterior hardware theat will work with their equipemnt. I am not saying that this is the issue but you never know. Please update the information on the MB. Thanks
0
 
LVL 26

Expert Comment

by:PCBONEZ
ID: 26073887
Um,
He asked what the motherboard make/model is.
I would like to know that too.
-
I would also like to know the same for those PSUs.
.
0
 

Author Comment

by:true_blue
ID: 26075082
Ok guys I will have this info for you within a couple of hours.. I am not home right now though.
And yeah you are right, the 4 optional ports that come with the PSU's werent used on this motherboard but for some reason even when connecting the -5v to the tester it still shows as dead. Which is kinda odd IMO.

Will post back with more info soon,

Cheers!
0
 

Author Comment

by:true_blue
ID: 26075645
Hiya guys,
The motherboard is an MSI (MS-7060)
Link;
http://eu.msi.com/index.php?func=proddesc&maincat_no=1&prod_no=158
Picture;
http://s2.imgimg.de/uploads/IMG2950ec2f5fb6jpg.jpg

The 2 most valuable PSU's I used are

Thermaltake TR2-500 (w0093)
http://www.newegg.com/product/product.aspx?Item=N82E16817153028

Arctic Power 700
http://www.compucase-hec.co.uk/psu_arctic_power_500-600-700.htm

My graphics card was;
Sapphire Radeon x1950 Pro 512mb (AGP)
http://www1.sapphiretech.com/us/products/products_overview.php?gpid=168



If you guys need anymore info please let me know!


Help Always appreciated.


0
 
LVL 26

Expert Comment

by:PCBONEZ
ID: 26076481
Thermaltake W0093RU 500W - has no -5v rail.
Arctic Power 700 - has no -5v rail.

That's what I was saying before.
-5v didn't 'go away'. - You never had it in the first place.

What exactly happened to cause you to suspect a PSU problem?
- Did it not boot? - Not come out of standby? - Crash somehow?

MSI boards of that age are fairly well known for having capacitor problems.
I'll see if I can find anything specific to that model.
.
0
 

Author Comment

by:true_blue
ID: 26076774
Ok 3 years ago I built this PC using a new motherboard, new graphics, new RAM, new PSU and a used CPU.
The PSU was 450w cheap/light-weight one that came with the PC case.
After 6 months the PC switched itself off suddenly, and would not turn back on.
I watched inside the case as I pressed the power button and noticed the CPU fan was jumping slightly I did this but not actually spinning.
I assumed it was a dead PSU, and replaced it with the Thermaltake 500w I spoke of earlier.
This one lasted me about 1 year, and then the same thing happened, PC switched itself off and would not turn back on.
After inspecting inside the PC this time I noticed the 12v port on the motherboard and the connector on the PSU were both burnt.
I then used a 350w generic PSU from my old PC until the Arctic Power 700 arrived.
The 350W PSU lasted about 3 days and then the same thing happened, PC switched off and would not turn back on.
Again after inspecting I noticed the CPU fan would jump slightly when pressing the power button but not actually fire up.
I put the Arctic Power 700w in and it lasted me 14 months until last week when I was playing a PC game and noticed smoke coming out of the PC case, the PC was still functioning properly but I immediately flicked the power switch to prevent damage.
I then inspected inside the PC, and looked for any sign of burnt out caps etc.
I did not find anything that looked like it could have been burnt. So I guessed that the graphics card may have overheated and began to fry while playing the game.
So I removed the card, took off the heatsink and fan housing, cleaned the chip and inspected it. But it actually looked OK and so did the rest of the card. So I re-applied arctic silver 5 and attached the heatsink/fan back together.
I then did the same with the CPU and found no sign of burns or anything so applied more thermal paste and put it back together.
Then I disconnected the PSU from all components and manually started the PSU by crossing the on/off pins on the 20pin connector.
The PSU started OK and seemed to run properly so I then inspected the memory, sound card and found nothing that looked abnormal.
After using compressed air to clean the PC a bit, I reconnected the PSU and started it up.
The PC got past post boot and then switched itself off!
I waited 30 secs and tried to turn it back on but it wouldnt work this time.
So I replaced the PSU with another spare I had, and it still wouldn't work.
I thought it may be a faulty on/off switch so I crossed the 2 pins on the motherboard to manually switch the PC on, but still the CPU fan jumps a little and nothing.
After all that I decided the motherboard must have fried when/if the PSU blew so I tested the memory, hard drive in another PC and both were damaged.
I havent had a chance to test the graphics or CPU yet but I am not holding my hopes up.

Anyway I have new parts on the way to build a new PC within the next few weeks, but the main purpose of my topic is to find out what went wrong with all those PSU's and whether or not I can salvage any of them?

I used my Thermaltake DR power PSU tester to test the PSU's and they all show green lights as ive said earlier.. but they wont stay on long if I draw any power from them.
http://www.thermaltake.com/product_info.aspx?PARENT_CID=C_00001036&id=C_00001037&name=dr.power&ov=n&ovid=n



Cheers again guys, hope you can help me understand what went wrong.
0
 
LVL 26

Expert Comment

by:PCBONEZ
ID: 26076785
[Not a lot of time to look.] I didn't find any actual complaints but I did find some close-up photos and it appears to have Chemicon KZG for the 16v VRM caps and a few misc OST brand caps scattered about the board.
Both of those are known to fail with no bloating.
The VRM output [aka Vcore] had Rubycon MBZ which are great caps.

Vid card had Chemicon KZJ which are suspected of the same problem and KZG [same base electrolyte] but they are so uncommon it's rare to see then dead or alive.
There are two un-sleeved aluminum canned caps on the vid card I couldn't make out.
What does the writing on those say?

Should note:
It's not at all unusual for a manufacturer to change make/model of caps midstream in production over availability issues. Three or four same make/model/revision motherboards might all have completely different compliments of caps. [So what I saw in the photo may not be what you have.]

.
0
 

Author Comment

by:true_blue
ID: 26076887
hmm im in a different world with the electrics side of things but from what I see one of them has the number F1 10 16 but they both have been marked with a blue pen and cant make the other one out..
I can take a Hi res pic of the graphics card for you without the fan housing on if you like?
0
 

Author Comment

by:true_blue
ID: 26076917
theres another that says NKG25.00:G09 (could be 609)
Im assuming you meant the radeon card and not the onboard graphics?
0
 
LVL 26

Expert Comment

by:PCBONEZ
ID: 26077036
Thank you.
I'm not sure what those are but they are not the problem caps I was worried about.
The problematic caps would have an FZ somewhere in the writing.
They nearly always pop their top and sometimes they outright explode.
Fairly common on video cards and occasionally seen on motherboards.
.
I'll walk through your earlier failure history post in a bit.
I'm on the wrong PC right now.
.

0
 

Author Comment

by:true_blue
ID: 26077099
Ah ok, thanks mate, its getting a bit late here but will be up for the next few hours! In the meantime I will have a more thorough inspection of the motherboard and see if I can find anything I missed.

Cheers
0
 

Author Comment

by:true_blue
ID: 26077540
Hi, I checked the motherboard again very carefully and all I could see is one of the caps between the memory modules and the CPU socket was slightly bloated at the top. I could clearly see the silver part was raised like a lump.
The rest of them seem flat at the top.
The writing on the bloated one was like this:
OST, 0519, RLX, 105c, 6.3v, 1000*F
Asterisk represents a symbol I could not recognise but kinda looked like an "A".
And the number right next to the bloated cap on the motherboard was CE2.

Hope this helps!
0
 
LVL 26

Accepted Solution

by:
PCBONEZ earned 2000 total points
ID: 26077557
Okay, had time to read through the history in detail.
-
Looks to me like an intermittent or partial short in +12v in/on the motherboard in the CPU power distribution.
- Here's why.
Your CPU uses about 90 watts. Adjusting for typical VRM efficiency it's actually 'pulling' ~117 watts.
All of that is from +12v and passes through the 4-pin connector that burned.
117w @ 12v --> 9.75 amps.
Those connectors are rated for [something like] 7-8 amps each, since there are two 9.75 amps should never be a problem.
-> Something is pulling more amps through there..... at least occasionally.

Most likely suspects:
A ground to the case [back-side of mobo].
A bad MOSFET [power transistor/regulator] in the VRM.
Bad MOSFET controller chip that's leaving them on too long. [Bad or it's receiving a bad feedback signal.]
A cap that intermittently shorts. [Least likely, unusual to fail that way.]

I think either long term higher [than should be] current and/or an occasional ground has been frying the PSU's.
- and
When the last PSU blew out it took those other parts and the mobo chipset with it. [Sent out a power surge.]

The catch-all phrase is "bad motherboard".

~~
There are some other things that could do it too. The ones I mentioned just seem most likely.

Have a couple pics but I'll post separately.

.
0
 
LVL 26

Expert Comment

by:PCBONEZ
ID: 26077570
Yes - You have bad OST caps then.
Mentioned that you might earlier on.
OST sometimes don't even bloat when they go bad.
~ Might not be the only problem though.
.
0
 

Author Comment

by:true_blue
ID: 26077605
Ok thanks a lot for explaining all that, so the motherboard is trash... what about the PSU's? Is there a chance I can fix them reletively easily or is it a waste of time? I mean the tester DOES say they are working but how can i stop the cutting out when under medium - heavy power load?
I am sure there were no parts of the case/mounting screws touching any part of the board that they shouldnt have been.. so guess the board was just made with a fault if thats the reason?

Thanks!
0
 
LVL 26

Expert Comment

by:PCBONEZ
ID: 26077666
This is basically how a VRM works.

VID is for "CPU Voltage ID"
It's an IC that senses what the CPU is and tells the MOSFET Controller IC [labeled IC] what voltage to regulate Vcore to.

IC [as I just said] is the MOSFET Controller IC.
It turns the two MOSFETs off/on to adjust the Vcore voltage.
Gets a feedback from Vcore and compares to what VID says Vcore should be.

The MOSFETs are very fast electronic switches.
If the upper switch is shut then Vcore goes up [toward +12v].
If the lower one is shut then Vcore goes down [toward ground].
These switches operate VERY VERY fast and part of the "Out Caps" job is to smooth the voltage 'bumps' from the switching action.

The "In Caps". These smooth the 'bumps' from the power supply which as 'Switch Mode Power Supply' suggests also has some switching action going on inside. [Though not as simple as a VRM.]

~~
The VRM regulates the CPU main power.
There is another CPU regulator called Vtt which adjusts the Signaling Voltage.
RAM also has a Vtt and a Main Power regulator.
Slots, the Chipset, and misc other IC's have POL Regulators [Point Of Load] which are usually one or two ICs and a cap or two.
The Chipset [primarily for the Memory Controller section] may have it's own Vtt regulator or it may 'steal it' from the RAM or CPU Vtt regulator.
.
VRM-Simp-01.jpg
0
 
LVL 26

Expert Comment

by:PCBONEZ
ID: 26077705
The motherboard might be repairable [-might-] if you are into that sort of thing.
Same with the PSU's.
- As you haven't done it before I would attempt it only as a learning project and don't expect it to actually work when you are done. Fair chance it will but don't get you hopes up. That last crash sounded pretty bad.
You'll need soldering equipment and a few other tools. [Some investment.]
You should go here for that: www.badcaps.net. [to the forums]
There are Mobo, PSU, LCD screen & tv, Audio, and even some MAC experts there that do component level work for a hobby or a living.
0
 

Author Comment

by:true_blue
ID: 26084977
Thanks for all your help PCBONEZ, you have been very informative and ive learnt a lot from your posts!
I will try and fix the PSU's and see what happens.. I don't think I will bother with the motherboard because if I screw something up then I could end up damaging other parts in the process.
I wish I knew 3 years ago that the motherboard was faulty, I could have saved a lot of money on parts it kept killing, although I did have my suspicions after changing almost every part except the mobo and still having the same trouble!

Cheers again! :)
0

Featured Post

Independent Software Vendors: We Want Your Opinion

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

This paper addresses the security of Sennheiser DECT Contact Center and Office (CC&O) headsets. It describes the DECT security chain comprised of “Pairing”, “Per Call Authentication” and “Encryption”, which are all part of the standard DECT protocol.
This article shows how to use a free utility called 'Parkdale' to easily test the performance and benchmark any Hard Drive(s) installed in your computer. We also look at RAM Disks and their speed comparisons.
Please read the paragraph below before following the instructions in the video — there are important caveats in the paragraph that I did not mention in the video. If your PaperPort 12 or PaperPort 14 is failing to start, or crashing, or hanging, …
With just a little bit of  SQL and VBA, many doors open to cool things like synchronize a list box to display data relevant to other information on a form.  If you have never written code or looked at an SQL statement before, no problem! ...  give i…
Suggested Courses

834 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question