View Full Version : Power Supply Excitement

18-12-2002, 09:56 AM
On the weekend the young fellow next door asked if I could have a look at his computer, as it was completely dead. I suspected something simple like a loose power cord but whilst I was sorting through a jumble of leads he switched the power switch on the power supply at the rear of box off and back on. Well, there was a hell of a splat accompanied by a flash and followed by the electrical odour, which usually verifies all is not well. If things were not completely dead previously, they sure as hell were now.

Having regained some composure, I brought the box home and operated.

Inside I found a power supply clearly marked ďNo serviceable parts inside.Ē Canít believe this as Iím sure thereís always a fuse inside that can be replaced, so, off with the cover. Sure enough, hereís a small glass fuse shattered, with the ends firmly attached to the pc board. Hmmmm, thatís clever, a non-replaceable fuse. However, thereís also the remains of an electrolytic capacitor which had simply exploded.

Now this power supply is a Hyena (no laughing matter) model LC-235ATX (235 watt)

Personally, I think 235 watts is a bit light for a machine with CD Writer, CD Rom, Modem, HD, FDD etc. (Any comments)

Since the weekend the original supplier of the machine has couriered over a replacement power supply, identical model. Unfortunately, it has scratched on the case the word ďfaultyĒ!

Not to be deterred, I installed it, powered up, and not a sausage! Absolutely nothing. The fan in power supply doesnít turn, there is no power on any plugs, but there is 230v in.

Question Time: This is a ďsoftĒ power supply and I have nothing attached to the box. No Monitor, No Keyboard, No Mouse, nothing. I was expecting the fans to start up and the POST to give me some beeps telling me thereís no monitor. Was this a correct assumption or am I tricking myself somewhere? How can I check one of these power supplies and be confident that it is indeed faulty? Yes, I have my faithful Avo 8 and DSE Digital Multi-meter. Finally, what are the chances of the motherboard etc surviving the initial onslaught?

Thanks in advance


18-12-2002, 10:05 AM
> can I check one of these power supplies and be
> confident that it is indeed faulty? Yes, I have my
> faithful Avo 8 and DSE Digital Multi-meter. Finally,
> what are the chances of the motherboard etc surviving
> the initial onslaught?

Chances of survival - Not too good. If you can smell an odour, then somethings likely overheating or even burning/blown up. Might havejust been thast fuse however.

Use the DSE multimeter to test the power plugs, The oltages are 3 and 12volts respectively, but Im not too sure which, so I might have to get back to you, or im sure another Pf1'er could help out.

Be careful that you dont shock yourself though. I know a guy who lost feeling in his hands for a few hours coz he did it wrong, but in saying that, Ive done it twice successfully (I just cant remember off the top of my head the wire colors!).

Good luck, and try not to make that bad smell errupt from it again.. not a good sign ;) :p

18-12-2002, 10:49 AM
cheap powersuplies.....arggghhhh! ! ! !

i would get your money back for the psu and get a decent one. depending on cpu and video card either get a 250 watt or 300watt.

Graham L
18-12-2002, 03:14 PM
They say "no user-serviceable parts inside", because they don't want to be sued by the estates of people who play about inside. :D The usual reason for the fuse to blow is that either the bridge rectifier or the main electrolytic capacitor has gone short. The rectifier is connected across the mains input, and the capacitor holds the rectified 320 V. I imagine that capacitor could explode with quite a satisfying bang when it goes. That is what took out the fuse. (320V DC in 200 microfarads is ample to give a fatal shock. Use care inside power supplies.)

A power supply with "faulty" scratched on it would not inspire me with confidence. You should have got some signs of life if it is working. Is it an ATX? If so, as well as trying the power switch, it has been suggested here (a few times) that ATX often need the "reset" button (just the jumpers on the motherboard shorted) to get going from scratch. An ATX ought to have 5V on the "power switch" (and an AT will have 230VAC ;-) ).

A lot of the AT supplies needed to have a certain anmount of load on one or more of the outputs before they would start up. (Some Compaqs came with a power resistor in a dummy floppy disk tray which had to be plugged into the 12V if there was only one floppy drive in the box.)

I would prefer to have an oscilloscope to work on a switchmode supply. They're not designed to be repaired, anyway.

19-12-2002, 12:13 PM
Thanks Guy's,

Donít fear for my safety Graham as Iíve managed to survive working around voltages with enough available amps to reduce the human torso to bones and buttons in milliseconds. (1500 vdc and as many amps as you like) Iím not however suggesting that other readers should be disrespectful towards these power supplies as the're perfectly capable of giving one a nasty shock, although I doubt there have been too many fatalities. :p

Anyway, Iíve crashed around the Internet and am in the process of drawing up a schematic of how I think things work. What Iím going to try is hooking up a standard automotive12v Brake/Park light to the 12v supply. That should provide a modest 26watt 2.16 amp load. Then Iíve got to figure out this "power good" line. I get conflicting explanations on this matter. Some say the motherboard tells the power supply itís waiting, whilst others suggest the power supply uses this line to tell the motherboard that power is available. Any comments anyone? Personally, I canít see how the motherboard is going to be able to send power or signal to the power supply as the only other source of power is the cmos battery and I canít see that being used. On the other hand, having a separate line to tell the motherboard that power is available is a little like leaving the porch light on so you know there should be power in the lounge. Can anyone throw any further light on this matter?

19-12-2002, 12:43 PM
The Power_Good Signal

The Power_Good signal is a +5v signal (+3.0 through +6.0 is generally considered acceptable) generated in the power supply when it has passed its internal self tests and the outputs have stabilised. This normally takes anywhere from 0.1 to 0.5 seconds after you turn on the power supply switch. This signal is sent to the motherboard, where it is received by the processor timer chip, which controls the reset line to the processor.

In the absence of Power_Good, the timer chip continuously resets the processor, which prevents the system from running under bad or unstable power conditions. When the timer chip sees Power_Good, it stops resetting the processor and the processor begins executing whatever code is at address FFFF:0000 (usually the ROM BIOS).

If the power supply cannot maintain proper outputs (such as when a brownout occurs), the Power_Good signal is withdrawn, and the processor is automatically reset. When proper output is restored, the Power_Good signal is regenerated and the system again begins operation (as if you just powered on). By withdrawing Power_Good, the system never "sees" the bad power because it is "stopped" quickly (reset) rather than allowed to operate on unstable or improper power levels, which can cause parity errors and other problems.

In most systems, the Power_Good connection is made via connector P8-1 (P8 Pin 1) from the power supply to the motherboard.

A well-designed power supply delays the arrival of the Power_Good signal until all voltages stabilise after you turn the system on. Badly designed power supplies, which are found in many low-cost compatibles, often do not delay the Power_Good signal properly and enable the processor to start too soon. The normal Power_Good delay is from 0.1 to 0.5 seconds. Improper Power_Good timing also causes CMOS memory corruption in some systems. If you find that a system does not boot up properly the first time you turn on the switch but subsequently boots up if you press the reset or <CTRL>+<ALT>+<DEL> warm boot command, you likely have a problem with Power_Good. This happens because the Power_Good signal is tied to the timer chip that generates the reset signal to the processor. What you must do in these cases is find a new high-quality power supply and see whether it solves the problem.

Many cheaper power supplies do not have proper Power_Good circuitry and often just tie any +5v line to that signal. Some motherboards are more sensitive to an improperly designed or improperly functioning Power_Good signal than others. Intermittent start-up problems are often caused by improper Power_Good signal timing. A common example occurs when somebody replaces a motherboard in a system and then finds that the system intermittently fails to start properly when the power is turned on. This ends up being very difficult to diagnose, especially for the inexperienced technician, because the problem appears to be caused by the new motherboard. Although it seems that the new motherboard might be defective, it usually turns out to be that the original power supply is poorly designed and either cannot produce stable enough power to properly operate the new board, or more likely has an improperly wired or timed Power_Good signal. In these situations, replacing the supply with a high-quality unit is the proper solution.

Dummy Load:

Power Supplies need load on the +5v AND the +12v in most cases

19-12-2002, 01:42 PM
Top stuff ďgfĒ, an excellent explanation.

Better still, I can understand it. (I think) :D

So, if I have one of these ATX power supplies sitting all alone on my bench and I provide a load to the +12v & +5v outlets, I should have all the appropriate voltages at the various other outlets including power_good. Now, this doesnít mean that the power supply is necessarily 100%, but failure to provide the appropriate voltages would surely mean the supply is faulty.
Is that a fair assumption?

Incidentally, should the power supply fan at least attempt to run, even if the power supply is not prepared to go power_good?

Thanks again


19-12-2002, 02:01 PM
The fan is usually run from a +12 v rail, so yes, if the supply is actually operating (into a load) then the Power Good will not stop the fan from running (normally), unless the voltage is just too low. These supplies are more go/no go really, unusual to have low volts.

The supplies are reasonably simple, just a rectifier to turn the mains to DC, and a switchmode gated chopper to supply the secondary voltages.

You really do need an oscilloscope to be able to fault find. Its very easy to blow switching semiconductors when repairing (expensive then)

Graham L
19-12-2002, 03:12 PM
I liked being able to repair linear supplies with just a multimeter and a soldering iron, with the supply turned off. Switches are bitches. :D

Graham L
19-12-2002, 04:14 PM
Also ... the 5VSB (standby) line is the only active supply until you pulse the "power switch" line.

20-12-2002, 09:44 AM
Ahhhhhhhh! Graham,

Just when I thought I had my head around this, we need a pulse on the 5v Standby Line.

Whatís got me licked is where does this pulse come from if thereís no power until its arrival?

Talk about confused.

Someone must have spent a lifetime making a power supply so complicated! :

20-12-2002, 09:48 AM
Actually BM, Graham did tell you.

>the 5VSB (standby) line is the only active supply until you pulse >the "power switch" line.

The 5vSB line supplies the power for the pulse.....

Momentary connection of the power on line to this will power it up.

Graham L
20-12-2002, 03:29 PM
Ahem, I've had a look at my book by Scott Mueller (caught you GF :D), and noted a few things down for you.

The 5VSB is on pin 9 (purple wire). That should be on whenever the power is connected.

PS_On, on pin 14, with a green wire, is grounded to turn on the main supply. If it is at "5V", the supply is "off", except for the 5VSB. This is done by the motherboard logic , on the basis of pulses (in up or down directions) provided by "wake-on" devices, and the "Power Switch" ... all of which use the SB supply. You can turn on the supply by keeping pin 14 at ground.

pins 15,16,17 (with black wires) are ground. 19 and 20, with red wires are 5V.

20-12-2002, 03:33 PM
Ok, so I was ONLY 180 degrees out? close?

It is a few years since I repaired one, admittedly. Its just not worth it commercially.

Its Xmas, so I beg forgiveness...I am only getting a lump of coal anyway :(

20-12-2002, 09:10 PM
Isnít it spooky how coincidences fall sometimes?

Whilst we are dealing with the peculiarities of these ATX power supplies a mate calls in with two flat boat batteries. No problem, pull out the old 30amp Battery Charger, hook everything up and nothing! The rectifier bridge is shot. (Probably stuffed by the last person who borrowed it). So, off to DSE and an hour later and $10 poorer all is repaired and the batteries on charge.

Talk about a contrast in designs. Wondering if thereíll be a third and what sort it will turn out to be.

Anyway, thanks for the help with these beastly little ATX supplies.