serious electricle short! cannot locate source, may be dangerous. suggestions?

i put together a new machine, there is a lot of stuff in it(yes, i took it all out for testing).
it is not overclocked. it was locking up randomly in windows. then i turned it off for about a day. came back and it would not post. the power supply started, fans turned. no beep, no video.
now it gets intresting! i physically tapped the machine (read: slapped the side of the case) and it magically started and ran fine!
every time i have had this problem i just slap it and it works for quite some time! now, i am not stupid. i have built many, many machines. nothing is loose. no stray screws or metal objects are present. power supply tests as it should with multimeter.
loose atx connecter? i cannnot narrow down the culprit here!
maybe intermitant psu problem? i may leave the datalogger on it.
it is one of several antec 550's i got recently.
this one baffles me, as there is no apparent reason why this is happening. loose cmos battery? i'm at a loss here! ideas are appreciated.




thank you,
sendatooli.

Question. Is it in a cheapo flimsy case? Could be the mobo is grounding out at the case somewhere and the slap breaks the connection.

it is not a cheap case. not a real expensive one either. i have used this case with other mobo's for 3 years with no problems. although, this board does not have holes for standoffs in all the locations i would like to see them. i had standoffs left over after installing this board. so that may indeed be the problem. i am still surprised that if this is the problem it is still alive. i am going to look into this. indeed, the hs i am using is so heavy it may be "bottoming out" the board. thank you for suggesting this. i knew i overlooked something simple! i hope this IS the problem so i can go correct it.



thank you,
sendatooli.

If you have spare PSU, I'd try hooking that up just to eliminate that posibility. Try it inside, and out of the case. You might have gotten unlucky and gotten one of the few Antec PSU's that is defective.

it is indeed the power supply. could you please tell me more about the defective antecs? does antec know about this?

if i move the pheripheral wires while the machine is running it locks. then i jiggle them some more to get it started again.
ok, not very scientific! also, if i plug in or unplug a fan while the machine is on it locks. this should not be the case with a properly working 550 watter!

here is one thing i might point out: the psu installes "upside down" in my cases. so to install it with the fan pointed down i have to make a plate. this does not have the best metal to metal contact between the case and the power supply. but this should not matter! the psu is grounded on more than one plane. unless i am wrong about this. please let me know.

nonetheless, my voltages are very respectable while using the atx and p4 connectors. maybe i should just return it and get one of those comp-usa psu's that i say are so good. only problem is they just don't pull quite as much current as the antec for a big server. still, the comp-usa's as mentioned in my earlier post are a great value. this is like the 4th. antec i had go south on me.
unless it is the case grounding issue?



thank you,
sendatooli.

You should install the standoffs in the mobo even where they don't line up with any holes in the case because I've had to do that to keep mobos from flexing and touching my cheap cases.

Have you tried using plastic standoffs?

i am using metal AND plastic standoffs in every available hole on this board (kd7r). it still flexes. BUT, that is not my problem.
i have narrowed it down to the power supply. could anyone answer my question as to if the psu's metal case needs to fully contact the system case? all the black wires are grounds. but, the design of the antec might be dual plane. where it needs to be properly grounded to the chasis and the items being powered.
i am still intrested in what CrystalMethod knows about defective antec supplys? were they recalled or something?

power supplys are no joke. they can kill a person and torch a house! i really hope NEITHER of those things ever happens to me or anyone else! so i think i will stop using this machine untill someone that knows more than i looks at it.


thank you,
sendatooli.

Ive had problems where the computer would stop turning on but the fans would spin for a milisecond. All I did on 2 occassions was 1. I figured it was the mobo so I took it out tested it then put it back in and havent had the problem since.

2. I tested the PS inside the case using the short trick... that didnt work. Thinking the PS was fried I took it out tested it and then it worked, so I put it back in and havent had any shorting problems since.

These were 2 different setups.

So test the PS using short trick while it is in the case. If it doesnt start then take it out and put it back in. but in your situation you will have to wait until the problem happens to test it.

This will be more difficult then when it wouldnt start at all for me.

My 300W Antec SmartPower has its DC and AC grounds tied together internally with a jumper wire, but it never hurts to bolt the PSU to the chassis anyway, just in case this isn't done (as it isn't with another PSU of mine).

Are those metal standoffs on your motherboard insulated as needed? I have some mobos where the mounting holes aren't surrounded by grounded copper, and with one, the right side of the hole is at +5V and the left side is at ground, and a mounting screw or standoff will short them together if either pierces through the very thin solder mask coating.

larrymoencurly said:

I have some mobos where the mounting holes aren't surrounded by grounded copper, and with one, the right side of the hole is at +5V and the left side is at ground, and a mounting screw or standoff will short them together if either pierces through the very thin solder mask coating.

Click to expand...

Maybe its just me... but doesn't that sound like very poor planning on the mobo manufacturers part?

thanks for all the ideas. the machine has been working for a while now. it is the power supply. either i do not have it grounded to the case good enough if it needs to be or it is a bad psu. i can reproduce the problem by jiggling the whole lot of wires coming out of the psu. it will lose the 3.3v! this is why the fans spin, they are 12v. it bassicly shuts off the voltage to the cpu!
i guesss i'll have to ground it better before i complain to antec that the psu is bad.


thank you,
sendatooli.