Resolved Possible PSU issue

I currently have a network of three computers consisting of a desk top and two laptops. The hub of the network is the desktop computer which is a Dell Dimension 4700 with a 160GB Maxtor Calypso hard drive and 80547 Pentium 4 Prescott processor with Windows XP operating system. Recently the desk top has been shutting down on startup. When you hit the startup button the startup process begins and within a few seconds will just shutoff. The amber power light on the front panel will continue to blink. After the 3rd or 4th attempt the system will start fine and run satisfactorily until I shut it down. This could be several hours or even a few days later. Reviewing other post on the site leads me to believe it could possibly be the power supply unit. The computer is cleared of dust and the fan seems to work properly. I was curious if anyone is aware of a system diagnostic that when run would give a reading on the condition of the power supply unit? I hate to just through parts at the problem not knowing what the actual problem is.

Thanks

 

When it shuts down, do you ever get an error message?

 

None.

 

Arie thanks for the response.

If I understand the Extreme power supply caculator it indicateds minimum PSU wattage 38 W and recommeded PSU wattage 8.8 wattage. However I am not sure how to determine what my computers acutal wattage is. I'm a little out of my league on this one.

 

Hi

the Dell Dimension 4700 ships with a 305W power supply (documentation link)

Unlike early Dells (some of which had non-standard outputs, but unfortunately employed a normal connector which could plug into non-Dell motherboards, or vice-versa - leading to expensive mistakes!), the power supply for the Dimension 4700 appears to conform to ATX standards (scroll down) - apart from one minor detail: Dell show white wires rather than the usual yellow for +12V.

Ensure that a replacement PSU has a 24way main connector (i.e. not a 20way connector)!

I'd suggest a 350W or 450W - but if you want to be cautious and check the current ratings of a candidate replacement, you'll find Dell's figures in the documentation link above.

Also make sure that Dell haven't "done a sneaky" and used a slightly smaller PSU case size (can you borrow a spare PSU from a friend perhaps?).

==

Yes, I think that you are probably looking at a PSU issue; but also that trying to use any software to diagnose the situation might well give a misleading result, because once the computer has booted up the problem will have gone away!

If I'm correct, it's simply a function of the power supply being incapable of providing enough current until it has warmed up a bit inside; the point at which the startup process is failing is most likely the point at which the HDD attempts to spin up. The extra current to spin up the HDD is overloading the ?borderline PSU, causing the output voltage to droop, and thus causing shutdown. After a few attempts the innards of the power supply will have warmed up, and it will be capable of sourcing a little bit more current without the output voltage drooping.

[SIZE= "1"]Although we are usually troubled by computers getting too hot, within the power supply rather different considerations apply. All bipolar transistors (i.e. PNP and NPN types) tend to give (or take!) more output current as their temperature rises. As they get hotter, the voltage needed to turn them on lessens (so for the same input they turn on harder); additionally, with increasing temperature their current gain rises. These two effects combine to increase their output; the effect is actually quite substantial. I feel that you are probably looking at a power supply which has gone "borderline ", and is simply incapable of supplying enough current until it has warmed up a bit inside. Plus, it's quite common for bipolar transistors to lose gain as they get older / see a lot of use.
By contrast, field effect transistors (FET, MOSFET etc.) tend to "fade" as they get hotter - their output current decreases as they warm up, all other things being equal.[/SIZE]

Something simple which you might try to confirm all this:

- after the computer has been running for a while, rather than shutting down, choose to restart the machine. When it reboots, do you ever get this "refuse to start up" happening? If you do, then you are probably looking for another explanation.

However if the machine always restarts without problems - but when "started from cold" tends to shut itself off, particularly with orange flashing light - then I strongly suspect that you have a borderline power supply.

HTH and best wishes, HJ

 

Hugh,
Thanks for the informative post. I checked my records and the computer was purchased in 3/2005 so it has had some usage. As you recommended I performed the restart function a few times and it started without any problem. I think this gives me enough confidence to go ahead and replace the power supply unit and hope for the best. Will let you know how it works out.

 

From what you report back I think we're probably along the right lines...

Oddly, I'd just come across this same problem on a funny little thing called a "Book PC" - yes, that tiny thing under the monitor is a computer, complete with CDROM, HDD, floppy, 400MHz Celeron, early multimedia video hardware (more like TV stuff) - and a tiny PSU, all squashed into that little box. The machine was OK 10 years or so ago; but nowadays if I turn it on "from cold" it takes many attempts to get it started. Particularly after a cold night, it just doesn't like to get going in the morning I can relate to that...

When it finally consents to boot up, at the point the HDD spins up one can quite clearly hear the fan slow down as the voltage droops! And also on this computer, an orange LED flashes on the front panel while the PSU is unhappy and it's refusing to start up. Yet, after it's been on for a while and fully warmed up - no problems.

I've discovered that if I turn the mains on to it half an hour or so before I attempt to boot the thing up, it's much easier to get it to start. The small amount of heat generated in the PSU from the +5 (standby), without the cooling fan running, is enough to make quite a difference! You may find that similarly, if you apply mains power half an hour or so before you want to use the computer, it makes it easier to start up.

==

When choosing a new PSU (or borrowing one from a friend for testing), as a "sanity check" it's worth making sure that it is fitted with enough of the right kinds of connectors. Don't just go solely by the pictures in the Dell documentation though! Use the documentation as a guide - but it's much more sensible to have a quick look inside your machine to see which connectors are actually being used.

Particularly, check the 6-way connector which Dell show as P6 - because I think you may well find that it's not plugged in to anything. Trying to find a PSU which comes with a 6-way connector like that will limit your choices - but with luck you'll find that you don't need it anyway.

All the other connectors employed are much more standard, and should be present on just about any ATX power supply.

The other reason for looking at what's actually on the motherboard is that occasionally later revisions of a motherboard have altered connectors, which no longer correspond exactly to the diagrams in the documentation! It only takes a minute or two to check "for real ", and you could save a whole lot of aggravation later on.

The 24-way main connector (Dell's P1) and the 4-way connector (Dell's P2) have little latch levers on the side which you push to disengage the latch - you'll soon suss it - then they come out quite easily. The one which is likely to prove most of a wrestle is the little Berg 4-way connector for the floppy disk (Dell's P7) - these are easy enough to push on, but getting them off again can be a bit tedious. Good Luck! - and keep us posted...

best wishes, HJ.

 

Thanks Hugh, You have been very helpful. I guess my next challenge will be to find a compatible replacement.

 

I had identical trouble with my HP desktop computer and it proved to be the PSU was on its way out.

A visit to my local Maplin store, showed they had several PSU of varying wattage rating. I selected a double rated wattage PSU that showed of a better quality than the one fitted in the HP computer, this was deciding factor as the original PSU was only eleven months and still under guarantee.

It should be noted as said already, to check types of connectors before buying.

 

Hugh,
Update on PSU issue. I went ahead and ordered a new power supply unit for my desk top. While I was waiting for the PSU to arrive I was doing some research on another issue and I noticed a post describing a similar PSU issue on the same model computer as mine. The following was the recommendation to cure the problem. (Remove the power cord, hold the power button in for several seconds to discharge the residue power, reconnect the power cord.) I followed these steps and the computer now seems to be working fine. Go figure. I’m going to keep the PSU I purchased for awhile just to be sure the problem doesn’t return.

 

rcflyerbob,

thanks for the update.

gosh! - and it worked?! Certainly, that would "discharge the residue power ", but then disconnecting the mains for a while does that too - like when one's fitting RAM modules, turn the mains off and (generally) wait for the LED by the RAM sockets to go out. Well, you've certainly got me thinking... could it be that discharging the power quickly is what did the trick? Unless there was some specific reason for discharging the power abruptly with the power button like that, I wouldn't have thought it a very good idea; surely it's not nice to the HDD, for example, to be "clunked" with a brief burst of power like that. So why to do it that way, rather than let the power drop nice and gently?

It would be very interesting to know, when you were having the trouble, was the mains power to the computer turned off altogether between the times you were using it - or was the computer shut down, but with the mains power left on?

The LED by the RAM modules only indicates that the +5V standby power (which Dell call +5VFP) has gone down; and there are several other places (capacitors) inside the PSU itself which could hold charge for a while after mains power is removed - but these are fitted with "bleeder" resistors, specifically for the purpose of discharging them. OK, I've only "got to grips" (intimately) with about half a dozen computer PSUs - all the ones which I've had to get to grips with at circuit level (even tacky old ones) have been fitted with these "bleeder" components (e.g arrowed blue in the picture). Which, of course, doesn't mean to say that the particular 305W PSU shipped with that model of Dell had them fitted...

Speaking of capacitors, one thing that I did turn up was that the Dell 305W power supply "Celetronix CX305N-00" (which I think is one type shipped in theDell Dimension 4700) seems to have a bit of a reputation for dodgy capacitors; and, capacitors (even good quality ones) do age with time and the stress of use. That in itself could be expected to leave you with a "borderline" PSU after a while. If your PSU has gone borderline because of capacitors, there's only one way that it would be going from this point onwards... south!

BTW - obligatory PSU warning (rcflyerbob, I probably don't need to tell you this! - but BBS pages find their way into Google so lots of people end up looking at them): If you don't have the required knowledge or facilities, do not open the case of a computer power supply as there are lethal voltages inside, sometimes in places which you wouldn't necessarily expect (e.g arrowed red in the tacky old PSU above). Similar looking components inside a computer don't generally have any voltage on the bare metal top of the can, and even if they did it would only be low voltage; but inside the power supply it's all very different - touching those cans when the volts are on would give you a very nasty surprise.

Oh heck. I hope I haven't helped persuade you to go and waste money on a PSU unnecessarily ...but I have a strong hunch that it won't turn out to be a waste in the end; I suspect the problem will come back one day, and that it will probably slowly get worse (if it's down to dodgy capacitors, then it will worsen progressively) - and that the eventual cure will be to swap the power supply. Reading up at the Dell forums, most of the people reporting "won't start / orange light flashing" seemed to find that a replacement PSU cured their woes.

best wishes, HJ

edit: lessened the strength of the wording as 6 is hardly a representative sample compared to the number of computer power supplies in the world!

 

Hugh.

Yes when I was having the problem the main power was on while the computer was shut down. As far as the PSU goes I do not have the experience to open the unit and make any adjustments. Thanks for the warning. I'm just going to see how things go over the next several days and go from there.

Thanks for all the info.

 

Update

The problem did return and I went ahead and replaced the power supply with a 350W ATX12V 2.01 Replacement Power Supply for Dell Pc. All now seems to be fine and working normally. I thank everyone who contributed to help me resolve this issue.