I think it's the power supply...

[mrcaptainbob]

I have a Dell Dimension 2350. Yeah, it's old, but it works fine. Except after the lightening storm a few weeks ago. Although it was plugged into a spike protector power strip, one of those for computer use(!) the darned thing won't fire up. Despite the fact the strip was shut off at the time! There is power from the supply cord, so I'm assuming it's the power supply that's been compromised. But....How can I verify it? I'm sure there is (are) power outs from it, but how can I check for voltage? And what should I expect? DC? 12V? 48v? ....?
I do have a 'spare' tower with a known good power supply. Can I swap them?

[HayZee518]

pwer supplies

the power out of the four pin molex plug is 5 volts dc.
if you have a sata system then the old power supply won't work even though it may look the same.
the old system used a parallel interface which is the flat cable assembly coming off your hard drive and may be piggy-backed to your cd drive.
a sata system uses a whole different cable.
check the wires and look for a very small molex connector marked P1

[mrcaptainbob]

But is the power supply??

Got to thinkin' about the power supply. The cabinet is layin' in front of me with the side off and the power supply box out, but still connected. What I'm thinkin' about is that I should have stated my problem rather than maki9ng assumptions and misleading. The issue is the power on button on the front of the machine will not turn on the machine. So....it could be the button, the power supply, or something in between. I see the button switch has six light gauge wires going to what I think is the motherboard. It could be a bumbed motherboard, too. So....how can I tell what to do? Pow2er at the switch first? I presume the power supply would be always feeding the motherboard to make the switch work. Or at least that circuit of the motherboard. If I probe for voltage at the motherboard/on/off button connector do I check to cabinet ground with the black meter probe and the red meter probe goes to the pins? Or would that maybe fry a component?

[HayZee518]

motherboard

look on your motherboard for an 8 pin connector
these are the pin-outs:
1: HD LED + 2: Power LED +
3: HD LED - 4: Power LED -
5: Reset 6: Power
7: Reset 8: Power
for the front panel connections
atx power supplies lists your computer [dimension 2350]
as shown above, reading with a multimeter on low dc volts will give you what you are looking for
all the power goes to the power supply and out the individual wire harnesses.
atx2024, P4 connector, 8 pin cpu connector, 4 pin molex which is power, a smaller version of the four pin to your floppy drive, 2SATA connector and a PCI Xpress connection.

[adnadeau]

Go to the Dell web site and find /support/Manuals/us/en/19/Product/dimension-2350 and download the Dell Dimension 2350 Series manual. On pg. 52 of the manual is a pinout of the power supply connector to the motherboard, P1. This shows an 20-pin ATX-type computer power supply connector. If you pull all the connectors off the motherboard and all the peripherals, Pin 14 is the 'PS On' circuit and has to be connected to a COM pin to fire up the power supply. A load of some sort should be connected between one of the voltage outputs and COM for the supply to work. I use a 100 ohm 10 watt power resistor between +12V and COM, but a 12V auto lamp will work okay. You can Google 'ATX power supply test' to come up with web sites that will give you a detailed test procedure.

[mrcaptainbob]

Power supply testing...

DANG! You guys are good! I will take some time and digest what you told me and study that manual. Thank you for sharing your knowledge...and your help.

[mrcaptainbob]

Power supply check.

Checked the power supply according to directions and all is good with that. I will check that button switch on the front for continuity. But....if the power supply is good...and that button on/off switch on the front is good, does that mean something in the motherboard is fried?

[HayZee518]

mb connections

if you can find a schematic for your motherboard, you can trace the front panel connections to the motherboard.
instead of a conventional relay, chances it would go to a gated triac. three terminals are on the triac.
the two ends are power in and power out, the middle is the gated on-off signal.

[adnadeau]

further tests

Visually check your motherboard for "bad caps". This was an issue for Taiwanese-made boards a number of years ago. I don't know if Dell's were affected by this. These are the electrolytic capacitors that are scattered all over the motherboard. The tops should be flat, not bulged up, and the bottom rubber plug where the leads come out should not be visible, and there should not be any goo.
If that's okay, try plugging the power supply back on the mb and checking the voltages on the plug. With AC supplied to the power supply but the computer off, you should see +5V between any black wire and the purple +5V Standby wire. You can stick your meter leads down into the back of the plug to measure. You should also see +5V between black and the green Power On wire. If you can manage to keep the meter leads in place here, push the computer power on button--the voltage between the green & black wires should go to zero. The pwr button is momentary, but the mother board latches the green lead low if everything is working.

[mrcaptainbob]

Issue update....

Took a while, but I finally got back into it. The power supply checks okay. After reconnecting it and plugging it in there is a small green light near the main mother board plug that does light up. I verified 5v from the purple to the black and do get 5v there, at this point. However, there is no voltage between the green and black wire from the power on switch. All the little 'cans' look like each other and do not show any abberations as you described, Adnadeau.
Would it prove the MB bad if I were to shunt from one of the unused power supply black wires to the black of the power on switch? I'm thinking that if the power on switch got power (although just temporary at this point..) that it would sort of 'back flush' the voltage to the green and back through the MB to the point where the 'compromise' is at?

[adnadeau]

There needs to be a voltage on the green wire for the power supply to turn on. I told you 5 volts, but I checked on an old supply I had and it's actually 3 volts, measured to any black wire. You can try using another black wire to run to the switch; all the black wires on the power supply connector are tied together and to the case ground. I'm reasonably sure the power from the power supply green wire goes into the motherboard at the PS connector and then runs to the Control Panel Connector J24 and then to the On-Off switch on the front panel. The black ground wire should take a similar path to the switch. If one of the connectors or connector wires or printed circuit board paths is open, then the switch won't turn on the supply. See if there is +3V on the green wire where it enters the back of the 20-pin connector from the power supply. Measure to any black wire in the same connector.

[mrcaptainbob]

There is 5v at green on the 20 pin connector. Would it mess things up if I jumped from the green 20 pin directly to the on/off green? Also, it sounds like I can check for continuity of all blacks as they are ground? One more thing...should there be continuity from the 20 pin green to the on/off green?

[mrcaptainbob]

Hmmm...I re-read what you wrote and now have another question, Adnadeau. You said there should be 5v at the switch wire to turn on the power supply? Yet, the power supply has the 5v at it's pin, but the switch wire had none. Does the 5v go to the switch green first, then through the MB, then to power supply green to turn it on? Or in the reverse order? Could it be a bad switch? Maybe stuck in an open position? You describe the switch green wire momentarily going to zero when depressed. It must shunt it to one of the other wires, I presume. What voltages should be read on the others at that point?

[adnadeau]

I think the fact that the power plug green wire has 5 volts and the switch green wire has 0 volts is the heart of the problem. Why isn't the 5 V getting there? When you put your meter test lead on the green wire at the switch, where are you putting the other lead? If your put it on the black wire at the switch, you can't know which wire might be bad. Try using your ohmmeter to check out just the cabling from the switch to it's plug on the motherboard. If the ohmmeter leads are stuck in the proper holes on the plug going to the switch and the switch is pressed, you should see 0 ohms. Then the problem would be in one of the circuit paths in the motherboard.

[mrcaptainbob]

Removed the switch from the unit along with it's connector. The switch (orange and blue) show continuity when the switch is depressed. All other wires show continuity when individually checked. Would/should the led's show continuity? One is marked 'power' and uses the green and black wires. The other is marked 'HD ACT' and uses white and red. The pin connector also has a brown shunt wire at two pins on the end. There is no continuity from green to black, nor from white to red when read from the pin connector.