Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I was forced to move my computer yesterday for texturing work to be done on the ceiling. Took the opportunity to blow it out really well with a can of compressed air.

It worked fine for quite a while after that, but it's now having major problems. It will stay up for maybe 10 mins and then just keep rebooting over and over. This prompted me to poke around.

During one of the reboot loops, I decided to go into BIOS. CPU temp was at 47C, so that is clearly not the issue. Didnt have enough time to look at anything else.

Anyways, I went looking inside the case and all the fans were working, but then i felt something rather hot even though I hadnt even grabbed onto it. Simply being near my hand was enough to feel the heat. On AMD computers (3800x2, I think), there's a 4-pin connector that attaches near the CPU. (I forget what it's called, but no, it's not the large 24-pin connector that powers the motherboard. Unfortunately it would be hard to google and find the name of it, as I am currently stuck with using my phone.) If i had to put a number on it, I'd say the Molex connector that feeds the 4-pin connection to the PSU is 125-130 degrees Fahrenheit on the outside of the plastic, which means probably 140+ degrees on the inside where the connection is. This seems to me to be extremely hot, as I've never felt any Molex connector get hot before, much less 125-130 degrees. Furthermore, one of the pins of the Molex connector is discolored (yellow/brown plastic on the female connector.)

Any input on whether or not this could be the problem? Is this a relatively normal temp, or out of the ordinary?

Edit: Picture of the hot Molex connector. Note the discoloration.

enter image description here

share|improve this question
"here's a 4-pin connector that attaches near the CPU. (I forget what it's called," 4 pin MOLEX 39-29-9042 connector at the motherboard. – Aki Jul 25 '11 at 10:55
Yes... that's it. Took pics with my phone but cant upload them to :( It's not that connection that I'm worried about but rather what's at the other end, where it connects to the power-supply's connector. – Michael Jul 25 '11 at 11:03
Are there inductive loads like an induction motor on a fridge? If so then a surge protector might solve the problem of spontaneous rebooting. – Aki Jul 25 '11 at 11:21
up vote 2 down vote accepted

Clearly this is not ordinary, and something probably went wrong with the PSU. I would not turn it on again due to a fire hazard and that it sounds like it could damage your mother board and other components.

I would change the PSU, and just hope that the current one has not fried the electronics.

share|improve this answer
Picked up a new 550W Antec PSU. It's been working for 30 minutes or so. Keeping my fingers crossed, saying prayers while waiting for web pages to load, giving my computer puppy dog eyes in an attempt to convince it to keep working, and yet rather optimistic! If this turns into one of those "spend 3 weeks figuring out what's wrong with my computer" deals, I am going to cry! – Michael Jul 25 '11 at 23:03
@Michael It sounds like you are going to be fine so far. – KCotreau Jul 25 '11 at 23:12
Hope you're right. Edited the OP to show a picture of the suspect connector, btw. – Michael Jul 25 '11 at 23:43
By any chance, on the back of the original power supply - is there a switch to change voltage (120/240) and could you have switched it to the wrong voltage while you were cleaning the system? – Blackbeagle Jul 26 '11 at 0:42
@Blackbeagle Yes, there is a switch on the old one to switch in between 115 and 230, but it was set on 115. The PSU is roughly 5 years old, and it's gotten heavy use (computer is on roughly 15 hours per day.) I probably shouldn't be surprised it went bad, and it was probably something that I did when I was cleaning it anyways, but nope... it wasn't the 115/230 voltage switch. I'm elated that the problem only cost me a few hours and $70, though. Every other time my computer has gone down in the past, I've poured in dozens of hours of diagnostics time. – Michael Jul 27 '11 at 5:15

I would say that the PSU is the only problem only if your computer would shut down, not restart. I've seen random restarts several times and it always seems to be a CPU issue. Could you boot into a linux live cd and open a terminal. Type sensors which will give you:

  • Your CPU temp
  • Your motherboard temp
  • The voltage of all your PSU rails
  • Other useful information

I have a sneaky feeling that sensors will point you (and us) in the right direction. Keep running it every couple seconds and see how you computer changes. Hopefully you'll get to see the values right before your computer crashes.

If you have access to another PSU and sensors isn't telling you anything, I would try switching them up and seeing if the problem persists.

One more idea, might be a long shot but worth mentioning. I was once cleaning my computer with air to get the dust out. When rebooting I had intermittent crashes which were caused by small dust particles being stuck in the CPU fan and blowing into the RAM slots (even with the RAM inside, gets in the cracks). Cleaning the memory and running a memtest fixed it.

share|improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .