PDA

View Full Version : A half second of power then crash



ThatOneGuy
12-03-2011, 07:31 PM
I think I have already found the problem, but I just want to check my reasoning, before I send back my PSU.

Build:
Asus M488TM
Phenom II x4 955
Corsair XMS3 4gb 1066
EVGA Geforce 440 1024 MB
Cooler Master exTreme 500w

Build went fine and this was my second build, so I'm familiar, but by no means an expert. The LED and computer booted up for less than a second than crashed. I then realized the PSU did not supply enough power for the video card, so I removed it.

I took it apart, checked all connections, and removed all the components until it was just the mobo and power supply. Still no boot.

I took the power supply out of my currently built computer (Thermaltake 775w), and hooked up only the mobo and PSU. It booted up, ran for about 15 secs then shut down.

This feels like a game of clue, is it the power supply?

ARchamps
12-03-2011, 10:40 PM
Sounds like the PSU. If another one worked, then that is likely the problem.

zburns
12-03-2011, 11:32 PM
I do not think it is your PSU. Neither of the PSUs worked. The smaller wattage of the two power supplies ran only a short moment. The larger wattage (775 watt) operated apx 15 seconds. Both power supplies have some form of over current protection circuits. The smaller psu likely has a lower ampere 'trip point', lets just guess and say it is 35 amps. The larger Thermaltake 775 w has a larger trip point (about 60 amps + some maybe) and it tripped out in about 15 seconds due to the higher ampere trip point.

This would indicate several possibilities. You have a component that is partially shorted, and is pulling more than enough current to trip out the smaller psu immediately. Because it takes 15 seconds to knock out the larger PSU, this would indicate the component is heating up and the initial partial short is getting worse causing the amps to increase at a rate that trips out the 775 watt psu in 15 seconds.

I am forgetting about the cooler & fan assembly on top of the cpu; what kind to you have and is it running during your 'description of what is happening'? At the moment you apply power, you do not have a complete short circuit or both psus would trip out instantly. This means that on the larger psu, some component (EDIT)on the motherboard or any other component the PSU is feeding is heating up and a partial short is getting worse and within 15 seconds trips out the larger psu.

But I need to know 'what was happening with the cpu cooler/fan combination during your test. Ie. Was one installed and running? No fan/cooler combination was installed, etc.. Need to know!!

The above is just my telling you what I am fairly certain is happening. I will look at it again in the morning first thing and give you some ideas on how to pin down the particular component. Again, what about the cooler or fan over the cpu -- is it part of your test or not?

I assume you left the cpu in place when running this test and I assume NO Ram was installed or was it? Please confirm that cpu was still installed, and Ram installed, yes OR no! And status of cooler or fan.

Regards the cooler/fan, I need to know exactly how the cpu is being cooled and was the cooling device in place and running the entire time you had the failures. Is the cooler/fan a 100% part of your 'narrative'? Sorry to be so 'pushy' with my words, but I need a 100% accurate picture.

Reason for the above edit: To clarify that whatever component is breaking down is external to the PSUs (both of them). Problem is not in the power supply -- it is external to the power supply based on the information provided.

ThatOneGuy
12-05-2011, 02:21 PM
At the moment of the boot, the back case fan spins, and the cpu fan spins.

The cpu fan is a cooler master stock fan that came with the chip, and upon boot up it spins momentarily then the computer shuts down.

I have tried booting with and without RAM, and with ram in different slots. This has caused no difference.

Other information:
The motherboard is installed with the extenders that push it off of the back of the case, and they are in the center right, bottom right, and bottom middle holes (direction right if looking at motherboard). The instructions did not specify for which holes.

You are correct about the power supply, the new power supply I just ordered does not work either.

ARchamps
12-05-2011, 02:30 PM
Try assembling the parts outside of the case to see if it is a shorting issue with the case (be careful).

Another measure worth trying is clearing the CMOS (resetting the jumper or clearing the CMOS).

ThatOneGuy
12-05-2011, 03:26 PM
ARchamps you are brilliant! It seems to be a short in the case.

So what do I do to correct this, should i space out my extenders more?

zburns
12-05-2011, 03:35 PM
Mounting of the motherboard. Your case is a ATX Form Factor case. The motherboard is likewise ATX Form Factor. The holes in the side of the case are an exact number of holes positioned to ATX specs that fit only a ATX motherboard. There are extra holes and they are in a pattern that will fit a Micro ATX Form Factor motherboard. There are no other holes in that area. It is impossible to install an ATX motherboard in an ATX case incorrectly. The motherboard has an 'exact' number of holes thru the board at places where there is 'on the motherboard' a deliberate 'ground plane' -- again impossible to make a mistake -- impossible to cause a short of any kind.

There is also no way to easily get a wire or wires behind the motherboard -- there are no instructions on any component that tell you to put wires behind it. Simply put, install the motherboard using the standoffs and all the holes -- no way to make a mistake.

The mother board is forced into 'one certain position' because of the rear opening in the case where the input/output terminals are.

I do not agree with setting the motherboard outside the case and hooking everything up outside the case. This only increases the possibility that you will make an error via 'static discharge' and damage the motherboard. Any thing you can do outside the case you can do inside the case. The motherboard is in one specific protected area and there is no way to cause a short by installing it in the specified holes -- and there is no way to screw up the installation of it.

Finally, you do have to use the 'standoffs' -- it is intended that you mount the mobo on the standoffs.

I am not critizing AR for suggesting this -- it is a common theme that one sees for troubleshooting a situation such as this.

ARchamps
12-05-2011, 03:37 PM
Haha it was actually zburns that gave me the idea in his post. He should get the credit :-P. Plus I had a similar issue when I switched cases.

What case do you have? Did you use all the supplied standoffs? On a normal ATX board there should be nine holes where the standoffs should align. Make sure you use the correct holes on the motherboard.

Hope this helps!

zburns
12-05-2011, 05:01 PM
A couple of comments that go with my ATX post just above. With your motherboard you received a template with holes in it to fit in the backside of case. Before installing it in the case, it is best to line the holes in the template up with the terminals and female plugs on the rear of the motherboard for two reasons: 1. To make sure you have the right template. 2. To be certain that the terminals/plugs will indeed exactly fit in the holes of the template. Because the terminals/plugs are 'soldered' to the mobo (in most cases), it is possible for some slight misalignments -- knowing this helps as you physically put the mobo into the case with the rear template already installed.

Next point. The standoffs you use are furnished by the mobo manufacturer. They are a certain height to cause the flat circuit board of the mobo to sit a fixed distance off the case side. By doing this the rear terminals/plugs are positioned heightwise off the mobo to (hopefully) go thru the template with minimum interference.

You use screws (specific length) furnished by the mobo mfg to go thru the mobo holes into the standoffs. No washers of any type are used underneath the mobo between the circuit board and the standoffs. If lockwashers are used on the topside (component side) of the mobo, the manufacturer of the mobo should say so.

When you install the mobo, do so with the case laying flat on its side so that you can gently lay the mobo in place over the standoffs and at the same time wiggle, gently position, push, etc. the rear terminals and plugs etc thru the rear template. The insert each screw into one of the mounting holes in top side of the mobo. Just get it started a few turns. Get all screws in place and started. After that run all screws fully into the standoffs but do not tighten them up yet. After you are certain that all screws went thru the holes with a little wiggle room, no forcing, etc., then go back and tighten them up. This process, sort of, assures you that there is no 'binding' at any hole and therefore no forces that will tend to cause the mobo circuit board to warp within a small area.

My point is, that the mobo 'sits' down on the standoffs in precise alignment to each standoff, and the plugs/terminals likewise fit thru the rear template without any forcing or pressure. Doing this should assure the assembler that the ATX mobo exactly fit into the ATX case in the exact position specified by the case manufacturer and the mobo manufacturer -- therefore, no possibility of any damage due to installation.

No flat washers should have been used on the top side of the mobo or the underneath side. Lock washers could have been supplied and specified by the mobo mfg for the top side (component) side of the mobo.

Given the installation being done the way I describe -- there is no way to cause a problem resulting in a short circuit. If the mobo went in 'easy' thru the template and sat directly on top of the standoffs -- hopefully a little movement back and forth -- there is no way to cause a short or the problem you are having.

Always, handle the mobo by the edges of circuit board -- keep hands off all components, always!

zburns
12-05-2011, 06:45 PM
ThatOneGuy: If you take my two posts, the 5:35pm and the one just above this one, they are intended to tell you 'how it is impossible to make a mistake' when installing the motherboard in the case. Therefore, if it works outside the case, but does not work inside the case, the conclusion has to be 'something' you are inadvertly doing when you put it into the case -- you just do not recognize it, some how it is bypassing you.

In my post just above, look at the para that begins like this: When you install the mobo . . . This para and the two shorter ones that follow it tell you exactly how the motherboard should 'feel' when installed before tightening the screws and after tightening the screws. If the motherboard feels 'jammed' hard in place before the screws are tightened down, somethings is wrong. With all the screws just halfway screwed in, the motherboard hopefully has some 'play' (very slight movement); if there is no play, you can tell the same thing, if all the screws can be tightened down with nothing more than finger pressure. Use two fingers lightly on the screwdriver (flat or philips); should be able to turn it with little effort -- tells you the motherboard hole is not pushing hard against the screw.

The motherboard cannot be 'forced' into place.

ARchamps
12-05-2011, 07:10 PM
I agree that building outside the case (also known as "bread-boarding") is a debated topic. In my opinion, if you take the same precautions that you would do installing the case, you shouldn't have any problems. This includes always making sure you are grounded and handling your components with care.

Keep us posted!

ThatOneGuy
12-05-2011, 07:54 PM
The case I have is Xion AXP 100 case. The case supplied 3 stand-offs which I had originally put in 3 bottom holes.

After moving the standoffs to one top left, one middle, one bottom right, i hooked up everything and it fired right up. I just left my little brother's house (was building for him). Everything was smooth as could be from there.

I sincerely want thank all of the posters in the thread for not only your help, but your quick response. Since this build was for my 12 year old little brother he was eagerly awaiting it's completion, and I was reaching my boiling point.

Once again, thanks for your help, because without it, I may have never determined the problem.

zburns
12-05-2011, 08:53 PM
First, I am glad you got it working. But only three standoffs may be a problem.

I think you have to fill up all the motherboard mounting holes. If you only got three standoffs, call the manufacturer, get them to send you what you need. Better idea if you bought the case locally and the seller has another one, go there and get the full compliment from a fresh box. (You cannot go buy them because they have to be exactly the same height, and they are probably metric)

Lets assume the case requires 9 standoffs for nine holes in the motherboard. The motherboard has or MAY have a groundplane that is dependant on all or just some of the standoffs being in place. If you only used three standoffs to begin with, you may have used mobo circuit board holes or left holes unused and 'that process' then 'interfered with the groundplane of the motherboard' and thereby causing your shut down problem -- there is no way to know (1) this caused the problem (2)the current location is such that the absense of the others does not cause the problem.

However, you have to assume that the motherboard manufacturer requires all mounting hole screws to be in place fully screwed into all the standoffs in all the ATX holes in order for the motherboard to operate correctly and more importantly operate in a SAFE MANNER. From a safety view point, I refer to anything that relates to the possibility of fire and / or an electrical shock hazard.

I am not trying to alarm or scare you. Your computer case and the components, within, meet certain standards for fire and electrical safety. When you fail to use all the standoffs, you do not know what potential problem (s) are possible.

ThatOneGuy
12-06-2011, 09:08 AM
After doing some research I found that some cheaper cases are built with machined in "standoffs". This appear as raised screw holes, hence only being supplied 3 standoffs.

It's worth noting for other first time builders that the standoffs should be placed in the flat screw holes and the raised ones should fill in the other gaps.

ARchamps
12-06-2011, 09:27 AM
Yes some cases are like that. My case (Silverstone PS06) had 3 or 4 raised screwholes. You just have to ensure that the other holes have the standoffs placed correctly.

Glad your system is up and running!

zburns
12-06-2011, 10:05 AM
Good comments and a partial answer relating to the cause of the problem, probably. The 'standoffs' are not 'machined in', they are punched in via a 'tool' and 'die' in an automated process. It is an effective way to get the 'standoffs' for free, probably, takes all of 2 seconds to 'punch' them in on the automated punch press machine -- it is a turret machine with 32 and up machine tools that 'do the machine software's bidding'.

Your motherboard is made up of multiple layers of printed circuit board fiberglass. One or two of these layers may be 'full board size' pieces of copper called 'groundplanes'. Here is one Wikipedia URL address on the topic: http://en.wikipedia.org/wiki/Ground_plane Here is the subject to look for: Printed circuit boards. Here is the statement that matters: Ground planes can also be placed on adjacent layers to power planes creating a large parallel plate capacitor that helps filter the power supply.

Each mobo manufacturer is going to 'mess' with the groundplane based on his 'engineers' knowledge of groundplanes and how to use them effectively. Please read the para 'Printed Circuit Boards' on Wikipedia -- it is very informative. The para is short and tells you that a full ground plane or part of one (I think) can be used as a capacitor as a 'circuit design aide' regards the power distribution over the motherboard.

The last paragraph simply means that the 'groundplane' is important to the mobo operation; therefore, it is likewise important that all of the mobo ATX specified mounting holes be filled with screws that screw into the mobo area of the case side -- whether standoffs are used or the 'cheap' metal punched standoffs are used -- does not matter. The groundplane of the mobo is connected to the metal side of the case via the 'screws that go thru the ATX specified pattern of holes in the mobo'.

It is entirely possible that 'ThatOneGuy's mobo shut down both psus he initially tried because of the absence of one or more screws at one of the ATX locations.
---------------
This entire series of 'post' or 'answers' is full of a lot of disscussion and 'what was done' and 'what to do' but nothing absolutely definitive about the cause of the problem. But there has to be a cause. The lower watt psu shut down instantly; the higher wattage shut down in 15 seconds -- this is a 'senario that electrically involves a 'capacitor charging up' because of the 15 second delay in the 775 watt psu shutting down. There are multiple groundplanes in the mobo board makeup; one maybe used as a 'capacitor' -- circuit effectively altered because of a missing screw(s) -- Overcurrent (OCP) or Overvoltage (OVP) protection circuits built into the psu shut down the psu in both cases.
-------------------

zburns
12-06-2011, 10:34 AM
Here is the URL for the ATX spec: http://joule.bu.edu/~hazen/LinuxCluster/atx_201.pdf

See page 8 & 9 -- page 9 is most helpful, I think. From what I see, the Standard ATX full size spec calls for (EDIT) ten holes. In looking at the spec, I see two places where the number holes for ATX are specified. One is a table on page 9 which clearly states (EDIT) nine holes. The other is on page 13 and is a mechanical drawing that identifies all the holes for 2 or 3 mobo specs including the ATX and the micro ATX. The drawing seems to specify 10 holes for ATX -- so there is a discrepancy I will have to chase down within this document and correct this post.

Some holes may be punched and raised, as mentioned above; several holes have to be flat (not raised) to accomodate other non ATX versions. Standoffs are used here.

Viewer note: There is a problem reading the ATX specification and determining whether nine or ten holes -- has to do with the older AT spec. Will correct this later today. Counting the holes in a ATX mobo may or may not work, since one hole may be optional for ATX. -- looks like a govt doc.

zburns
12-06-2011, 02:15 PM
Message to 'ThatOneGuy',

'After moving the standoffs to one top left, one middle, one bottom right, i hooked up everything and it fired right up.' The above are your comments.

Your case has what I call 'punched' or 'formed' standoffs in place of add on 'standoffs' for approx 6 or 7 of the holes in the mobo. So I am assuming that you anchored the mobo to these six or seven 'formed' standoffs as well as used 3 screw in standoffs -- giving you a total of nine or ten screws going thru your mobo into an equal total number of 'formed' standoffs plus the three added standoffs.

I got the wrong impression. I assumed, wrongly I hope, that you only had three screws holding the mobo in place. I assume, all the time, you had nine or ten screws holding the mobo in. The thought of only three screws being used caused me concern about potential groundplane problems and electrical/fire safety considerations.