Power on issue - revisited

  • Thread starter Thread starter David Cook
  • Start date Start date
D

David Cook

I have a very similar question to a previous poster.

(But, I decided not to clutter up that note so I'll start this new one.)

Ok. One has a new case. One adds a power supply with a fan
inside that power-supply.

Before even a mother-board or anything else is added, one puts
a power cord into the back of the power supply and moves the
rocker switch on the power-supply to the on (aka the '1' position).

Should the fan in the power supply come on?

Should I be able to measure power in the connectors hanging off
the power supply?

(Or does logic from either the front-panel switch and/or the
motherboard's BIOS chip have to be present and fed back somewhere
before the power supply lets power be fed around?)

Dave
 
I have a very similar question to a previous poster.

(But, I decided not to clutter up that note so I'll start this new one.)

Ok. One has a new case. One adds a power supply with a fan
inside that power-supply.

Before even a mother-board or anything else is added, one puts
a power cord into the back of the power supply and moves the
rocker switch on the power-supply to the on (aka the '1' position).

Should the fan in the power supply come on?

On an ATX, no. It will not come on.
Should I be able to measure power in the connectors hanging off
the power supply?

Only 5vsb, the standby voltage. Not a good idea to run a PSU with no load
anyway. Probably won't hurt anything unless it is a very cheap PSU, but the
voltage readings don't mean much without a reasonable load. 5V no load
might be 4.5V under load. Good PSU will not show much variation, but with
no load you don't know.
(Or does logic from either the front-panel switch and/or the
motherboard's BIOS chip have to be present and fed back somewhere
before the power supply lets power be fed around?)
The motherboard pulls a pin on the connector low (do a search on
groups.google.com on testing powersupplies to find out which).
The Power switch is a soft switch that acts through the motherboard logic.

JT
 
David said:
Ok. One has a new case. One adds a power supply with a fan
inside that power-supply.

Before even a mother-board or anything else is added, one puts
a power cord into the back of the power supply and moves the
rocker switch on the power-supply to the on (aka the '1' position).

Should the fan in the power supply come on?

If ATX, no.
Should I be able to measure power in the connectors hanging off
the power supply?

If ATX, no (except the +5vsb on pin 9.)
(Or does logic from either the front-panel switch and/or the
motherboard's BIOS chip have to be present and fed back somewhere
before the power supply lets power be fed around?)

Pin 14 on the 20 pin connector must be pulled <0.6v.

To test an ATX PSU:

First, disconnect everything from the power supply (including MB.) Plug
AC cord into the rear of the PS. IF the PS has its own rocker on/off
switch, turn it ON. There should be +5v on pin 9 of the 20 pin connector
(+5vsb, usually a violet wire.) There should be >0.8v on pin 14 (PS-ON,
usually a green wire.) This is from an internal pull-up to the +5vsb.

Connect PS power leads to 1 or 2 IDE HDs (for a load). Jumper pin 14
(PS-ON) to pin 13 (ground, black wire.) At this point, fan in PS should
start spinning, drives should spin, and + 5/12v, -5/12v, +3.3v, and
+5vsb should be present at 20 pin connector.

Disconnecting pin 14-13 jumper should turn supply back off.
 
I have a very similar question to a previous poster.

(But, I decided not to clutter up that note so I'll start this new one.)

Ok. One has a new case. One adds a power supply with a fan
inside that power-supply.

Before even a mother-board or anything else is added, one puts
a power cord into the back of the power supply and moves the
rocker switch on the power-supply to the on (aka the '1' position).

Should the fan in the power supply come on?

Should I be able to measure power in the connectors hanging off
the power supply?

(Or does logic from either the front-panel switch and/or the
motherboard's BIOS chip have to be present and fed back somewhere
before the power supply lets power be fed around?)

Dave

I'm going to skip the answer since previous poster covered it, and
start with questions:

Why do you want to know, what's the problem?

Is it a name-brand power supply or unknown generic?

I would be very happy to know of even one unknown generic that's a
decent power supply, but so far, NONE are. The larger power supply
manufacturers buy parts in bulk, secure volume contracts for quality
parts, spec for workstations/servers as well as PCs, and have a
trickle-down effect such that a good power supply for a PC costs less
to make, plus they want their name only on decent hardware.


Dave
 
There are at least two reason that I want to understand this.

I saw an 'empty' case with power-supply that was rigged up
with some spooky lights inside a case that has a plexiglass
side panel, so you can see (red, blue, green) lights illuminate
the inside of this case.

So, I purchased the empty case, an ATX power-supply, and
a couple of these lights, and an instruction sheet, but quickly
realized after I brought it home that I was sorely lacking in
conceptual knowledge as to the theory behind the wiring.

(The display model in the store is plugged in and the lights
come on when the power-supply is turned on. But, there is
no mother-board in the display model at all.)

So, based on what you posters have said, I now know that
some jumpering of wires is probably being done. (Actually,
I do NOT recall now whether the FANS were on.)
These lights probably need only 5-volts to drive them?

Another reason I wanted to know was that I had a machine go
dead (would NOT power up) last week, and since I had no
clue at all how to trouble-shoot that problem, I had to shell
out money to a fixit place just to learn whether I had a power-supply/
power-switch problem or a mother-board failure.

(They first diagnosed it as the front panel swtich, replaced that,
I took it home, it STILL would not power up, took it back, and
they now tell me I have a bad motherboard. I'm starting to understand
now why it took them 1.5 hours of labor and two tries just to
get to this point.)

Hope this satisfies the curiosity about why I wanted to understand
this stuff better.

Cheers...

Dave
 
Ric (or anyone else) -

Thanks for the very detailed explanation!

Which brings me to a related question: How do you folks
come by this knowledge? (Yeah, school of hard knocks, I know.)

What I really mean to ask is: Is there a single book (or website)
where the majority of this kind of information can be found?

(I've been building my own machines from motherboards for years,
but have never needed to venture over into hardware diagnosis/repair
until just a week or so ago.)

TIA...

Dave
 
David said:
Ric (or anyone else) -

Thanks for the very detailed explanation!

Which brings me to a related question: How do you folks
come by this knowledge? (Yeah, school of hard knocks, I know.)

Being involved in power supply R&D, design, manufacture, and test for
over 20 years? Being in charge of tech support for a company that builds
PC PSUs?
 
There are at least two reason that I want to understand this.

I saw an 'empty' case with power-supply that was rigged up
with some spooky lights inside a case that has a plexiglass
side panel, so you can see (red, blue, green) lights illuminate
the inside of this case.

What, exactly, are these lights, or at least, what is the technology
behind them, for example: LED, Cold-Cathode, Neon, plain old bulb

So, I purchased the empty case, an ATX power-supply, and
a couple of these lights, and an instruction sheet, but quickly
realized after I brought it home that I was sorely lacking in
conceptual knowledge as to the theory behind the wiring.

(The display model in the store is plugged in and the lights
come on when the power-supply is turned on. But, there is
no mother-board in the display model at all.)

Are you 100% sure there was only that much to the case/system? Often
a display model of something can have additional misc. parts to make
it do it's "thing". In this case there might've been a rigged
connector plugged into the power supply, that shorted the PS_ON line
to ground and perhaps a load on the 5V rail, dpending on what rail,
how much power the lights use. This wouldn't be needed in a system
with motherboard/etc, only to correctly wire up the lights.

So, based on what you posters have said, I now know that
some jumpering of wires is probably being done. (Actually,
I do NOT recall now whether the FANS were on.)
These lights probably need only 5-volts to drive them?

Perhaps I'm assuming you want this to be a whole system but you did
only want a case with no motherboard?

I don't know about 5V, there's not much that uses 5V BUT an LED-type
could use either 5V or 12V, depending on the size of the included
resistor, and the lighting design... if in doubt always try the lower
voltage before the higher one.
Another reason I wanted to know was that I had a machine go
dead (would NOT power up) last week, and since I had no
clue at all how to trouble-shoot that problem, I had to shell
out money to a fixit place just to learn whether I had a power-supply/
power-switch problem or a mother-board failure.

If they couldn't narrow it down more than that, you were taken for
money with no proper diagnosis... it's very easy to determine if it's
the motherboard or the switch, just remove the switch plug and jumper
or short the pins with something metal. However, only the most
severly damaged motherboard won't turn on a properly working power
supply, or of course if the board or wiring is shorted.
(They first diagnosed it as the front panel swtich, replaced that,
I took it home, it STILL would not power up, took it back, and
they now tell me I have a bad motherboard. I'm starting to understand
now why it took them 1.5 hours of labor and two tries just to
get to this point.)

In less than 1/2 hour they should've been able to swap in a different
PSU and switch for testing, have a definitive diagnosis. You were
taken by an incompetent shop. IMHO, they shouldn't have gotten a
penny if they can't even diagnose something this simple, but maybe
they didn't even have a single decent (spare) power supply to try,
only junk. Many shops do use junk PSU unless something else is
special-ordered.

Hope this satisfies the curiosity about why I wanted to understand
this stuff better.

Cheers...

Half the time a shop diagnoses a bad board, it's a junk power supply.
The rest of the time it could be a number of things, the board or a
BIOS setting or dead video card or hard drive or...

The good old swap-in-better-power-supply test is the easiest way, but
certainly NOT the same make/model, if it's junk.

If the power supply is a generic I'd go ahead and replace it anyway,
even if it turns out to be the motherboard. Also I already mentioned
how to test the switch, just unplug it and short the appropriate two
pins, are in the motherboard manual or silk-screened on the board
itself right under the appropriate pins.


Dave
 
Back
Top