Sign in with
Sign up | Sign in
Your question

How do you detect a broken power supply

Last response: in Components
Share
July 29, 2008 3:53:26 PM

Hello

I bought a Tagan TG700-U88 BZ 700W in May. Since the beggining of July, my computer started freezing and I get blue screens on a daily basis. Because my system crashes soo much, my system files get ruined and I need to reinstall windows every other week.

I tested the ram and hard drives using diagnostic test and they check out ok. So I figured that it is either the motherbord or the power supply that is broken. I have a feeling that the motherbord or PSU has a broken capacitor.

So I would like to know what is the best way to detect a broken PSU.


P.S

As a side note, this is the third power supply I owned this year. My first PSU was the Antec SP 450, which Antec admits is faulty. They said that it had a faulty 5V line due to a bad cap. The second one was an Ultra 500 which I got discounted for $30. It stated acting up after a month. So I finaly decided to get a good durable PSU, so I figured I'll buy something from Tagan.
July 29, 2008 4:42:43 PM

That will work for you
Related resources
July 29, 2008 4:48:27 PM

Hi

Time to invest in a UPS/ powerline filter??

Not sure that an off load PSU tester will pickup a fault at this stage... But its worth a shot I guess!!

I would substitute a known working PSU to test for system stability. That hopefully pin down a motherboard or PSU issue.

The problem that since you have been using ***** PSUs with the motherboard in the past is that alone can damage components on it.

I would generally run: CPU burn (overnight), 3D Mark 2006, Gold Memory Tester, etc. to test for system stability after overclocking.

You have a good warranty on the powersupply so you should be covered in that respect!!

What motherboard, CPU, GPU and RAM are we taking about??

Bob
July 30, 2008 8:25:04 PM

Hi

Sorry I didnt respond right way. My computer crashed after writing a response and I think its going to crash again.

I'll try those programs test programs when I get the chance.

Also this is my computer

CPU: Intel E6600
Mobo: Gigabyte GA-P35-DS3L
Ram: 2 Gig Corsair 800Mhz 5-5-5-18
Video: Nividia 8800GT
PSU: Tagan TG700-U88 BZ 700W
July 30, 2008 9:13:51 PM

loafing_smurf said:

Sorry I didnt respond right way. My computer crashed after writing a response and I think its going to crash again.



Ah thats all good parts (all solid capacitors on the MB for example).

Hmmm. Interesting problem!!

I presume your temperatures (GPU and CPU) are OK? (Given the problems started in Summer - with a decent PSU)
The classic test for this is side panel off and large house fan blowing in the side...!!

What's the case you are using and is the airflow OK?

Don't bother running those tests... Your system is too unstable!!

Can you substitute any parts?? Like for example do you know anyone with a decent PSU >450W ??

RAM is going to be standard 1.8V if it is Corsair value RAM - right? Have you tried booting with one stick and then swapping over to the other?

Bob


September 5, 2008 5:56:28 PM

Hello

Sorry for replying so late. And I feel that I owe you an explaniation, since you were trying to help me. I last replied to on the end of July, which is when I had to visit my parents for a month. On the last day I had to leave, windows just wouldnt start because it crashed so many times. So I was unable to tell you I was not able to reply.

As for things now, I am trying to buy a power supply tester, but eveyone is out of stock. If the PSU is ok, I'll try to see if the memory is broken. If that is ok, it has to be the motherbord (which was once heavily used by a bad PSU)
September 5, 2008 7:51:16 PM

No worries.

I would think the MB and RAM would be OK. I reckon it's cheap gear that is vulnerable to damage from varying voltage levels from an unstable PSU (ripple or spikes/drop-outs).

Like I said before a PSU tester will generally only tell if a PSU in within the ATX specifications - static conditions - but very short variations from this in the form spikes or drop-outs will not be detected. Also excessive ripple or noise on the voltage rails may not be picked up on. The only reliable PSU testers cost a lot of money and are used by PSU reviewers.

It would be easier for you to get a parameter into the system that is almost 100% a known quantity in the form of a Corsair, Seasonic or PC Power & Cooling PSU. I would also invest in a decent UPS to protect it. Otherwise you are potentially building/testing the whole system around an unstable component at the core. Look on it as an investment to save you time, hassle, protect your system data and to protect your other system components from damage.

Like I said in previous emails the flip-side of ensuring power stability is to ensure that the DC-DC power conversion components for generating the CPU voltage from 12V, etc., etc. (mainly Mosfets) are all cooled adequately (not just the CPU temperatures). A good engineering test is literally just to feel heatsinks and chips with your finger. This is because our anatomy is generally sensitive to temperatures over 45C and so are electronic components!! :sol: 

If you want post up some pic's of your rig to see if there could be any cooling issues visible in these.

Bob
September 6, 2008 1:08:32 PM

Hello

Are you saying that I should buy a new power supply?

As a some what relevant point. I lived in a shabby apartment last year...which was probably built in the Victorian Period. The electrical outlets were not grounded. I knew this because my Belkin power bar said so. All of my components except for the Tagan 700 were used in this building. At that time I was using the $30 Ultra 500.

But now I live in a newly built apartment and the electrical outlets are grounded.

And do you think a computer shop will properly see if my Tagan 700 is faulty? What questions should I ask the computer shop to see if they properly diagnose a power supply? So I know that they dont just use a typical power supply tester?

Also my digital camera might have kicked the bucket. I had it since 2002 and it now takes pictures of purple vertical lines in a black backround. So I might have a hard time taking a picture of my case.

Thanks
September 6, 2008 2:01:53 PM

You say you install windows every other week?

Are you sure your installing the most updated drivers, MS updates, or even patches for games for your system?

When you run software diagnostic tests, that's basically more in a dos environment, rather then within windows, so I think something is overlooked in my opinion.

Might want to look in the event viewer, under systems to see if windows is complaining on drivers for the hardware, especially on a new installation of windows.
September 7, 2008 10:24:00 AM

loafing_smurf said:
Hello

Are you saying that I should buy a new power supply?


Ok so no pictures - heh heh!!

Uhmmm I had a sleep on this problem. Looking at the New Egg reviews of that PSU tester (Rexus PST-3 Digital Power Supply Tester) that would actually be quite a good way to go (especially with the LCD voltage readouts). From the level of the stability problems you are experiencing and the fact that Tagan is a good PSU make it sounds like the problems might be shown by the PSU meter. One way to test for power stability problems is to plug in something with a big load (like a delta fan) and look for drops in the voltage rails.

So I think you are right don't splash out on yet another PSU just yet. BTW is the Tagan under warranty still if you do find a fault with it??

I would still recommend to buy a UPS to try and avoid this kind of problems occurring again in the future!!

Are any components overheating in your rig or anything like that. Whenever I get instability problems I usually start poking around to feel for chips overheating or hot heatsinks!! :lol: 

You don't know anyone in the area with a stable system that could borrow your PSU to test it for a few days perhaps?

Bob
September 10, 2008 8:55:44 PM

Hello Grimmy

It might not be a drivers problem. In July, my system froze when I was instaling windows. Also, my system was runing fine for most of February of this year. Then it started crashing alot in the summer.

Hello BobWya

I think you hit the right problem. The season has been changing into fall and I've been leaving my bedroom window open. So the room temprature has been significantly lower. And I notice that my system started acting properly.

I dont get blue screens anymore or any type of system freezes. And my system now runs at performance. Where it used to take 10 minutes for Windows to load. Or very long wait times to install a video game.

I opened my case and touched some components. Among everything, the northbridge heatsink is really hot. I can touch it for a moment before my finger starts to burn. So is this normal?

I also touched the ram, and it was warm but not as hot as the northbridge heatsink.

I also checked my CPU temperature in the BIOS and it idles at 41 C.

Edit/P.S - It got hot this weekend, mostly on Saturday. It caused and I notice that my system crashed alot. So I had to rebuild my OS.
September 11, 2008 12:43:54 AM

Yeh

I've got a watercooled rig but all my expansion cards slots are full and keep overheating due to me leaving out enough airflow overthem!! I may have to replace the case with one the newer models (120mm exhaust vs 2x 80mm exhaust at the rear).

What case have you got there? Maybe a case upgrade with a model with better airflow? Stagnenet air around the northbridge allows that heat buildup. You want the heatsink to get hot (good = TIM contact is OK) but then you want airflow through the case to take the heat out.

Something like:
http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=162&products_id=24017
is the daddy!!

Its worth throwing money at good airflow it makes life a lot easier day to day... See I am still not taking my own advice and paying for it!!

Bob
September 12, 2008 10:31:20 PM

At the moment I left the case side panel open for cooling and it hasnt crashed all day.

Although, it crashed yeasterday and I got this code in the blue screen.

STOP 0x000000 F4 (0x00000003,0x8A3C74C8, 0x8A3c763C, 0x805D1142)

Looking it up its seems like a hard drive or RAM problem. I'm actualy thinking that my hard drive could have broke from formating soo many times.

But my computer has been running non stop since last night and it hasent crahed and is running at performance. And this could be because I left the side panel open.


Anyways I'm probably going to buy this Antec 300 case. It looks like I cant go wrong buying it.

http://www.ncix.com/products/index.php?sku=29812&vpn=TH...
September 12, 2008 10:50:53 PM

An idea hit me!

STOP 0x000000 F4 is a RAM or hard drive problem. But cooler temperature seem to solve the problem.

When I touch the ram heat sink it is warm. Perhaps the heat sinks are not doing their job properly.
September 13, 2008 6:54:32 AM

Believe it or not.....I actually had a ram heatsink contribute to ram in my system burning up. Seriously. I started having blue screens when my system was only a couple of months old. Mind you I've got a 250 mm fan in the side of the case too, so not like cooling is an issue. Anyway, started getting blue screens and all, long story short, I'd gotten a lot of gear for Christmas. And dad had bought one stick of ram without heatsinks, one with. He just found a good deal on the other one, so picked it up. Got everything assembled, worked for a while, tested the ram after reformatting and what not when blue screens started. Got errors, the stick with the heatsink actually had the errors. Pulled it, and like yours, warm to the touch, and had a funky smell to it to. Replaced with another stick with no heatsink, and not one issue in that respect since. Oh well, live and learn.
September 13, 2008 7:10:36 AM

Or it could be the Manufacturer process of applying the adhesive. I brought Crucial Ballistx memory that had a bunch of bad reviews. I even thought it was attributed to heat as well.

But since I was OC'ing, I pretty much pushed it voltage wise, although I replaced the HS that were currently on them. I spent a good hour cleaning off the adhesive that kept the HS on the Ballistix memory.

Long story short, I used Mushkin memory on my current system. So far the Ballstix I was having problems with, work fine at stock settings, with mild OC on the E4400 system that I use it on. Heh, and on top of that, I'm using the Muskin HS on the Ballistix, since the HS were allot easier to take off and put back on the Ballistix. :lol: 
September 13, 2008 3:12:49 PM

This is the ram I have.

Corsair XMS2 TWIN2X2048-6400 2GB

http://www.ncix.com/products/index.php?sku=18581&vpn=TW...

I'm either thinking about investing in cooling or new ram all together.

RAM cooler

http://www.ncix.com/products/index.php?sku=21240&vpn=CM...

RAM heatsinks (problem is removing the existing heatsinks)

http://www.ncix.com/products/index.php?sku=31434&vpn=AR...

RAM I am looking at

Mushkin HP PC2-6400 HP2-6400 2GB

http://www.ncix.com/products/index.php?sku=21095&vpn=99...
September 13, 2008 3:26:35 PM

I use these on my Mushkin, which were on my Ballistix:

Thermaltake CL-R0026 Aluminum Heatsinks only

I used a hand held IR Thermometer to measure external temps on the HS. Normally both ram kits would register 40-50C which would depend on the load on the system. With those replacement HS by thermaltake, the temps dropped to 36-43C.
September 13, 2008 3:37:43 PM

Thoes looks good but I need to figure out how to replace the heatsinks.

My ram heatsinks are not held on by clips. I think they attached with thermal glue.

Perhaps I can freeze the ram and heatsinks will snap off?
September 13, 2008 3:48:51 PM

Heh... I first used those on my Crucial Ballistix, they were glued to the HS. I ended up just pulling them off. Then a spent a good hour or so cleaning all the adhesive off them before putting those thermaltake HS on, mubbling why did they do that?? :lol: 

I also wondered if I damaged it while pulling or cleaning them, later on. :sweat: 

I basically used rubbing alcohol to get as much off the memory chips. But I still had blue screens after that, so I ended up with Mushkin. Then I decided to take the Mushkin HS off which was way easier, and put those clip HS on my Ballistix for my E4400 machine which it still works fine. Its just weird that I have Crucial memory with Mushkin HS. :lol: 

Edit:

So mainly my problem with the Ballistix was with my OC. The Mushkin didn't to much better since I was on a P6N Plat with a 650i chipset. That is where your thread is not simliar to what happened to me, but is kinda linked to over heating ram.
September 13, 2008 4:19:32 PM

Thats a bit akward. But whatever solves the problem. And shouldnt the Thermaltake HS be better than the Mushkin HS?

Looking at RAM prices and heatsink prices. I feel tempted to just buy new RAM. New 2 GB of ram can cost $10 more than new heatsinks.

But I just might go for new RAM heatsinks.
September 13, 2008 4:24:52 PM

Heh.. the problem was never solved. My MB (P6N Plat) broke so now I'm on a different MB (DFI P35, which OC better), but the thermaltake HS are on the Mushkins.

Since my Ballistix no longer had HS, I used the Mushkin HS on the Ballistix, and used the on a different system, which is paired with my E4400 system. It's basically a long story, hence why I was trying to cut it short.
September 13, 2008 5:16:21 PM

I wouldn't recommend using AS5 on the memory. It does contain silver particles, and using it on memory is different then on a CPU IHS or without.

Usually RAM HS have a pad that keeps it in contact for the memory chips to the HS. If want to use a thermal grease, use the ceramique stuff. Problem there is it going to be messy (putting it on each chip) and it prolly won't fill the gap like a thermal pad would, unless the clip actually can get an even pressure to assure that contact for all the chips is flush.
September 13, 2008 6:04:43 PM

I see

So there is no need for AS 5 or aftermarket thermal paste/adhesive?
September 13, 2008 6:11:11 PM

Not really. As far as the adhesive, that would be more for permanent application, so you would really need to be sure you would never want to take it apart ever again.

Here's a quote on the caution:

Quote:
CAUTION! Arctic Silver Thermal Adhesive is a permanent adhesive. If you do not follow the instructions for diluting the adhesive with Arctic Silver Thermal Compound, any components you attach together with full-strength Arctic Silver Thermal Adhesive will stay attached forever.


All I can say is be careful if you get that stuff. The thermal pad I use for my Thermaltake HS work fine as is, and would not get as hot as for say a CPU or GPU, which would matter more then ram chips in using thermal grease.
September 14, 2008 4:32:37 PM

loafing_smurf said:
I see

So there is no need for AS 5 or aftermarket thermal paste/adhesive?


I have successfully fitted watercooled RAM sinks to some old Geil DDR RAM I had. It was a biattch to get the stock RAM sinks off (necessitating soaking them in some thermal compound remover solution). It was a bit touch and go to get them with a screwdriver! Thankfully the compound on the old thermal pads with greatly weakened so I didn't end up pulling parts of the RAM chips off with heatsinks!!

I used a tiny bit of Silca gel to insulate visible circuit traces beside the RAM chips and then MX2 on the surface of the chips. MX2 is practically non-conductive and as such is a better alternative to AS5!! Amazingly I got the board to POST in a recent test!!

I wouldn't have attempted this 2-3 years ago so definitely go with good quality thermal pads if you are just starting out!! It was all a bit alarming (especially if you know the price of 2Gb of Geil PC3700 RAM some 3-4 years ago)!! :hello: 

Bob


September 14, 2008 4:34:39 PM

MX-2 is a great product. What I really like about it, no cure time. :D 
September 15, 2008 1:09:53 PM

Here is what I just ordered.

- Scythe Kama Wing Adjustable Aluminum Memory Heatsinks

- Corsair Dominator Airflow Memory Fan

- Arctic Silver ArctiClean 1 & 2 Thermal Surface Cleaner & Purifier 60ML Kit (2) 30ML Bottles

- Mg Chemicals Label & Adhesive Remover

- Arctic Cooling MX-2 Thermal Compound Paste NON-ELECTRICAL Conductive 4 Gram


First I'm probably going to soak a retractable knife with the adhesive remover. Then carefuly pry off the memory heatsinks. Then finish off any left over thermal adhesive.

After that I'll use the thermal pads that come with the memory heatsink to attach the new heatsinks.

Then I'll attach the memory and Dominator cooler.

And hopefuly this will be the end of all the crashing.

September 15, 2008 9:33:16 PM

loafing_smurf said:

First I'm probably going to soak a retractable knife with the adhesive remover. Then carefuly pry off the memory heatsinks. Then finish off any left over thermal adhesive.


I actually soaked my Geil RAM sticks in Arctic Thermal Surface Cleaner fluid. The sinks still needed quite a bit of force to pry off. Just be careful you don't rip chunks of the RAM chips off with the sinks...!! :whistle: 

Bob
September 16, 2008 12:58:31 PM

^dat was my fear after I got the HS off my Ballistix, as well as the serial number that were on them. The Lifetime warranty was then screwed for me, but it seems the 4gb kit is working fine on my E4400, which still has a 2.8ghz OC on them. :wahoo: 
September 17, 2008 6:11:41 PM

bobwya said:
I actually soaked my Geil RAM sticks in Arctic Thermal Surface Cleaner fluid. The sinks still needed quite a bit of force to pry off.


How much can you "soak" a stick of memory? Do you wet it a little bit or can you submerge it in adhesive remover?

I'm starting to regret buying memory heatsinks and should have got new ram...
September 17, 2008 9:00:57 PM

loafing_smurf said:
How much can you "soak" a stick of memory? Do you wet it a little bit or can you submerge it in adhesive remover?

I'm starting to regret buying memory heatsinks and should have got new ram...


No I litterally soaked the RAM sticks (i.e. submerged) for 24 hours in the Artictic Thermal Material Remover (1). Mind you I wanted to fit a MIPS watercooled heatsink to the RAM (so I had to get the sinks off). THIS WHOLE PROCEDURE IS NOT FOR THE FAINT HEARTED!! :non:  :non:  :non: 

Personally if I was you I would that fancy Corsair RAM fan to you existing sticks and not bother taking off the sinks. If the sinks get hot then they are transferring heat outwards from the RAM already. Don't always believe that someone elses problems apply to you. A bit of active cooling can usually overcome these minor cooling problems anyway.

It does sound like you need a better case rather than new RAM anyway. The Antec 900 has very good cooling by all accounts. Modern systems need really good airflow to remain stable.

Good case + good PSU = the bedrock of any good system.
:sol: 


Bob
September 23, 2008 5:37:55 PM

My order of computer parts came in the mail. And everything is now installed.

It seems that some problems disappeard and new problems appeared.

Problems solved

- I no longer get "STOP 0x000000 F4" blue screens.

- I got an external temperature gauge and the RAM surface idles at about 30 C. And this is with the ram cooling attached.

- My north bridge idles at 48 C.

New problem

-I now get the STOP 0x00000077 (0xC000000E, parameter 2, parameter 3). Which means a broken hard drive, loose connection or broken cable.

- There seems to be no loose connections. So perhaps my hard drive took a beating form all the formats and is now broken?


Additinal points

- I replace the CPU fan and thermal paste with Artic cooling MX-2. I noticed a decrease in temperature by 1 degree.

- I also replaced the thermal paste from my video card. It was a nightmare when the artic clean dissolved the orginal thermal paste. The liquid remover got under the VGA memory. I thought that was it and my video card was broken, but I guess it still works because I'm using it right now without problem.

- The video card now idles at 63 degrees C instead of 67.
September 24, 2008 3:21:49 PM

loafing_smurf said:

New problem

-I now get the STOP 0x00000077 (0xC000000E, parameter 2, parameter 3). Which means a broken hard drive, loose connection or broken cable.

- There seems to be no loose connections. So perhaps my hard drive took a beating form all the formats and is now broken?



When does this problem occur?? When Windows is starting up or while Windows is running?

The stop code actually is described as meaning:
"the drive went unavailable, possibly a bad hard drive, disk array, and/or controller card."

So you basically need to rule out the disc controller (MB I presume??), cable (non-latching SATA cables are evil - worst design in history) or disc drive.

Can you boot into a live CD Linux distro like Ubuntu/Kubuntu or Konoppix? Is this more stable than booting from the HD?

Have you updated your Gigabyte MB BIOS to the latest?

You say you are reinstalling Windows regularly... A fresh install - correct? Are you updating the drivers to the newest versions?

What model HD are you using? What kind of interface (SATA/PATA)? Is the cable shielded?, latching? ??

Spinright 6.0 is a very useful tool for diagnosing HD faults:
http://www.grc.com/sr/spinrite.htm

You just have to work along the chain till you have found the weak link... Prey that it is only one link that is broken!! :whistle: 

Bob
September 28, 2008 11:55:29 PM

Good news, it appears that all problems are gone.


But, first bobwya

I'd like to mention my gratitude for helping me. This is a problem that started in July, and you persisted to help me to the end of September. If you are in Canada, I'd owe you a beer. :) 

In terms of solving the problem, I blew compressed air in the system again. Then checked all the connections and closed my system case slowly. Just to make sure nothing came loose while closing.

I reinstalled windows and updated the drivers. I must admit that I havent been installing the drivers for the past week, because I'm afriad that my system would crash again. And I can only download 25 GB/month.

My system has been running for 2 days and nothing has gone wrong. I'm instaling all my windows programs and video games. Because I am confident that everything is fixed.

Additional Notes

The average speed of the RAM cooler is 2400 RPM and it keeps the ram at 34 degrees C. If the ram cooler is slowed down to 800 RPM, the ram temperature hits 50 degrees. I immidately increased the cooler speed, out of fear that my system would crash.

At first I had problems viewing websites with flash player. Like youtube and websites full of advertisemnts (toms hardware for one). But updating windows fixed the problem.

Thats all, and it seems like my problems are solved.
September 29, 2008 12:13:16 AM

Hi mate,

If you are ever in Cambridge, UK then you owe me a beer... Well not really since you did all the work yourself!!

It is not **** about dusty machines being unreliable (well OK your old P2 system might not have cared)... Especially when the PSU gets clogged up with crap... Oh dear he says looking at current rig covered in dust... :sol: 

Glad that every thing appears to be working now but I would still run stability tests overnight to check you have it nailed... That might sound harsh but if I build a system that can't survice an x264 encode/ CPU burn process (1/core) overnight then I consider to be "broken"!! Also you might be disappointed when the weather gets warmer next year and the machine starts to fail again.

25Gb/month ouch!! I usually download about 4x that much!! I hate all these petty download caps... :sol: 

Bob
September 29, 2008 3:21:56 AM

Just to add some insight...
Which windows do you have?
If Vista, check to see if your ram was running at the manufacturer's suggested voltage ratings, if it needed more than 1.8 volts, that was your problem right there.
That and slightly slowing down timings does the trick more so than not for vista bsod's and phantom restarts.

Edit: and you think 25gigs a month cap is bad? mMy cap is 20, and i pay 55 bucks a month ffs, F#$k videotron
!