Does ANYONE know this PSU answer....

T

TyQuAnDo

I wish to try a number of PSU's for operation without having to put em in a
PC.

I gather there are 2 terminals to short out on the power cable loom to the
mobo, so
fooling it in to thinking it is connected to the mobo, but which terminals?
 
G

Gareth Church

TyQuAnDo said:
I wish to try a number of PSU's for operation without having to put em in a
PC.

I gather there are 2 terminals to short out on the power cable loom to the
mobo, so
fooling it in to thinking it is connected to the mobo, but which
terminals?

Connect pin 14 (green) to ground (any black wire).

Google is really good for this type of question.

Gareth
 
B

Bishoop

| I wish to try a number of PSU's for operation without having to put em in
a
| PC.
|
| I gather there are 2 terminals to short out on the power cable loom to the
| mobo, so
| fooling it in to thinking it is connected to the mobo, but which
terminals?

Make sure you put a load on the PSU you're testing.
 
T

TyQuAnDo

Thanks all
I just want to flip them on really to see if there is any life before
fitting them in a case
and then finding they are dead or the fans make a noise, so I shouldn't
think I will need 'a load' for a few seconds will I?
 
C

Conor

Thanks all
I just want to flip them on really to see if there is any life before
fitting them in a case
and then finding they are dead or the fans make a noise, so I shouldn't
think I will need 'a load' for a few seconds will I?
Nah, you'll be OK for that. Anyway, a load could be a bulb with a
couple of wires on shoved in a drive power socket.

--
________________________
Conor Turton
(e-mail address removed)
ICQ:31909763
________________________
 
J

JAD

Tom? Where are you when we need you

If you do short the terms, and if the PSU is weak or older, it can go "POOF'
in a second
 
V

V W Wall

TyQuAnDo said:
Thanks all
I just want to flip them on really to see if there is any life before
fitting them in a case
and then finding they are dead or the fans make a noise, so I shouldn't
think I will need 'a load' for a few seconds will I?
Most ATX power supplies will not start without a minimum load. They usually
need about one amp drawn, but are usually not particular whether it's the
+5V or +12V. You can use a power resister (1A @ 5V = 5ohm/5W resistor), but
an automobile tail lamp works fine, and is a lot easier to find. It also
lites up to show the supply is working. A hard drive attached to the PS
will also draw enough current to allow it to start.

As has been noted, any real test would involve the measurement of all voltages
under rated loads. Some better supplies have fans that speed up as the
temperature in the PS increases, so you can't judge the fan noise by
the unloaded, (cool), fan RPMs. Most, however have a single speed fan,
and a cheap one at that!

Virg Wall
 
V

Vanguard

TyQuAnDo said:
I wish to try a number of PSU's for operation without having to put
em in a PC.

I gather there are 2 terminals to short out on the power cable loom
to the mobo, so
fooling it in to thinking it is connected to the mobo, but which
terminals?

One of the wires in the 20-pin Molex connector going from the power
supply to the motherboard is PS-ON. When this signal is open (i.e.,
disconnected), the power supply will shut off. When PS-ON is pulled TTL
high (over 2.0 volts), the power supply shuts off. When PS-ON is pulled
TTL low (under 0.8 volts), the power supply's other outputs turn on
along with the fan. Whether turned on or off, the power supply always
supplies 5 volts to the motherboard for circuitry that must continue to
be powered even when the power supply is "off", like for "Wake on
<event>" settings in the BIOS. That is why you need to yank the power
cord from the back of the power supply when changing PCI cards.

Another user remarked that a load needs to be on the power supply, like
hooking up a hard drive. I have read that turning on a switching power
supply without a load can ruin it, that it will burn up if tested
without a load, and that's why they do not turn on unless the 20-pin
Molex connector is attached to a motherboard (i.e., PS-On pin 14 is
pulled TTL low). However, not measuring the load (amperage) capacity of
the taps means you have no idea of the quality of output capable by the
power supply; checking voltages doesn't indicate capacity. Lots of
power supplies will claim that they will put out, say, 350 watts but
actually cannot put out more than two-thirds of that. You might want
to review the Tom's Hardware articles about power supplies at:

http://www6.tomshardware.com/howto/20021021/index.html
http://www6.tomshardware.com/howto/20030609/index.html
http://www.cluboc.net/reviews/index.htm#power (and a Google search will
turn up more reviews)

You might be able to decipher how to build Tom's test rig from its
pictures and description at
http://www6.tomshardware.com/howto/20021021/powersupplies-03.html. They
really don't provide a how-to-build tutorial on all the parts you need
and a sample layout of the PCB that you'll have to etch. There's no way
to test that capacitors are defective and will start leaking since that
usually occurs after a long time. But after testing loads at the max
ratings per tap continuously for a couple days you might then want to
pop the cover off and take a look inside. Note that a multimeter (volts
& amps) won't tell you how much regulation will degrade under high load.
You'll need an scope to look at the ripple.

If you are not looking to qualify the claimed capacity or quality of the
outputs for the PSU, and all you want is an OK-notOK check, I suppose
you could get something simplistic like the Antec ATX Power Supply
Tester for around $10 to $15 USD. CompUSA has one (Powmax;
http://www.powmax.com/Power Tester.htm) for $15. I believe the Antec
tester only checks the +5V output using its 20-pin Molex connector (but
says it has leads to test +12V and +3.3V) while the Powmax tests Full
test on +3.3V, +5V, +12V, -5V, -12V, and +5Vsb outputs concurrently.
However, with either of these units, you still need to get a good
multimeter. Of course, you go industrial and get something like the
http://www.prodigit.com/e3600.htm for lots and lots of money, so
spending the money on something like Tom's tester rig would be a lot
cheaper but requires you be handy with a soldering iron, PCB etching,
parts selection, and basic electronic repair and testing, and buy a
couple other test equipment like a digimeter that can measure high
currents.
 
K

kony

Thanks all
I just want to flip them on really to see if there is any life before
fitting them in a case
and then finding they are dead or the fans make a noise, so I shouldn't
think I will need 'a load' for a few seconds will I?

Writing in generalizations only, a cheap/junk power supply "might" be
destroyed by power-on without a load. It's not likely these days, but
possible.

A moderately better, consumer/PC grade power supply won't damage
itself, but may not run without a load attached to the +5V rail. You
should always assume you need a load on the 5V rail. A hard drive is
less than the typical 2A load the manufacturer spec requires, but it
generally enough.

A better, high-end power supply may or may not need a load on the 5V
rail, it can vary based on the target system/market. It WILL monitor
3V rail. Some power supplies monitor the 3V rail internally, but
sometimes externally instead, rather through the 3V sense line on the
motherbaord connector/leads. The 3V sense line might be paired with a
3V supply line on the same pin, so there's 3 3V conductors instead of
only 2, total, going to the motherboard via the main ATX connector.
By doing that you already have the sense line connected to the supply
line.

This 3V sense line is pin 11. The manufacturer can choose to put ONLY
the sense line on pin 11, or place BOTH the sense line and a 3V rail
line on pin 11, which means two physical wires at pin 11. All too
often I see this point neglected, and as a result a perfectly good
power supply might be declared dead by conventional (power-on shorting
and 5V load) methods of testing.

In order to keep one of these PSU (with only the sense used on pin 11)
powered on, even for more than a fraction of a second, you MUST also
connect the 3V sense line to the 3V rail, one of the other 3V pins.

When a power supply is connected to the motherboard this sense pin is
connected to the rail, but not when the connector is removed. A load
on the 3V rail can also be required but it's even less common, and
placing a load on the 3V rail wouldn't (by itself) accomplish this
requirement to have the 3V sense line connected.

So in summary, yes, you need a load. The far easiest way to test a
power supply isn't to fool with all these external loads and shorting
power-on-to-ground, etc.. the easiest way is just to keep an old/junk
motherboard lying around with a power switch connected. I actaully
soldered a switch onto an old board I have here. If you don't care
about seeing the motherboard hardware monitor/BIOS reported voltages
then this old motherboard doesn't even need to have ANY other parts
plugged into it, no cpu or memory, video card, etc... a bare
motherboard with a power switch attached will turn on a motherboard.
Even motherboards that die from "most" causes, except something
shorting out, are still functional to turn on a power supply and keep
it running.


Dave
 
R

ric

TyQuAnDo said:
I just want to flip them on really to see if there is any life before
fitting them in a case
and then finding they are dead or the fans make a noise, so I shouldn't
think I will need 'a load' for a few seconds will I?

Often, yes. The PSU will shut down immediately without it.
 
G

Guest

TyQuAnDo said:
I wish to try a number of PSU's for operation without having
to put em in a PC.

I gather there are 2 terminals to short out on the power cable
loom to the mobo, so fooling it in to thinking it is connected
to the mobo, but which terminals?

Connect the Power_On* signal of pin 14 (green wire) to any ground
(black wire, such as pin 13 or 15 next to it. If you're concerned
about causing damage by connecting the wrong pins together, use a
500-1000 ohm resistor to make the connect. Most AT power supplies and
some ATX supplies will not start unless a load is applied to the +5V
line, and the automotive 12V lightbulbs mentioned by some work fine,
or you can use a resistor rated for approximatley 10 ohms and at least
10 watts. Connect either device between any pair of red and black
wires. A few ATX supplies also require a similar load for the +3.3V
line (any pair of orange and black wires). All ATX supplies will turn
on at such loads, but not all will provide good voltage regulation
without higher loads (this is not an indication of good or bad
quality).

There's no need to buy a power supply tester, and since none will give
you an accurate indicatio of the voltages, money is better spent on a
digital voltage meter, which can also be used for testing batteries
and almost anything electrical.

Most ATX supplies include at least one extra wire on the 20-pin plug
to sense the voltage at the plug and compensate for any losses in the
cable, especially for the +3.3V. But supplies are designed to work
even without those extra wires because a break in one of them could
cause make the power supply overcompensate and put out too much
voltage on that line.
 
G

Guest

V W Wall said:
Most ATX power supplies will not start without a minimum load.
Untrue.

They usually need about one amp drawn, but are usually not
particular whether it's the +5V or +12V.

Most will start without a load, but if a load is needed it usually has
to be on the +5V. Also some supplies need a load to meet voltage
tolerances.
 
G

Guest

kony said:
A better, high-end power supply may or may not need a load on the
5V rail, it can vary based on the target system/market. It WILL
monitor 3V rail
This 3V sense line is pin 11. The manufacturer can choose to put
ONLY the sense line on pin 11, or place BOTH the sense line and a
3V rail line on pin 11, which means two physical wires at pin 11.
All too often I see this point neglected, and as a result a perfectly
good power supply might be declared dead by conventional (power-on
shorting and 5V load) methods of testing.

In order to keep one of these PSU (with only the sense used on pin 11)
powered on, even for more than a fraction of a second, you MUST also
connect the 3V sense line to the 3V rail, one of the other 3V pins.

When a power supply is connected to the motherboard this sense pin is
connected to the rail, but not when the connector is removed

I don't know the official ATX spec, but I've never seen an ATX supply
without a power wire on pin 11, and I doubt that 2 wires could handle
well the full +3.3V current rating of some supplies (and I'm referring
to older supplies without a 6-pin connector and extra +3.3V wires).
Also every sense wire I've seen was connected internally to its output
voltage through a low-value resistor, 100s of ohms or less, as a
safeguard in case the sense line broke. So the supply can still
regulate the +3.3V even when pin 11 isn't connected to any external.
 
K

kony

I don't know the official ATX spec, but I've never seen an ATX supply
without a power wire on pin 11, and I doubt that 2 wires could handle
well the full +3.3V current rating of some supplies (and I'm referring
to older supplies without a 6-pin connector and extra +3.3V wires).
Also every sense wire I've seen was connected internally to its output
voltage through a low-value resistor, 100s of ohms or less, as a
safeguard in case the sense line broke. So the supply can still
regulate the +3.3V even when pin 11 isn't connected to any external.

Who knows, maybe it was just a fluke, manufacturing defect... I do
know the 3V sense circuit wasn't internally connected to the 3V rail,
or perhaps it was but the resistance was too great to keep the unit
running.


Dave
 
G

Guest

TyQuAnDo said:
I think the best thing is to use an old mobo.
Another reason I was trying to be cautious is I bought a brand new
case + PSU, fitted everything, turned it on, smoke poured out the
pc after about 3 seconds and anything that the dc cables touched
inside was burnt out. This happened to be the data cables for the
HD and floppy drive, sending the current up to the drives
knackering both new items. Of course the retailer would only
replace the PSU as the other items were not supplied by them.

I'm cheap and don't even want to risk even my old equipment to a bad
power supply, so I use dummy load resistors, which let me apply
maximum loads, whereas a motherboard is unlikely to draw more than
100-150W. Also with the supply outside of the case, compare its
temperature with that of a known good one driving the same load
because a faulty supply will often heat up at a much lower power
level. This isn't usually caused by poor air flow but by faulty
design or components.

What are the name and address of that dealer, and what was the brand
of power supply?
 
G

Guest

kony said:
This 3V sense line is pin 11.
When a power supply is connected to the motherboard this sense pin is
connected to the rail, but not when the connector is removed external.

Who knows, maybe it was just a fluke, manufacturing defect... I do
know the 3V sense circuit wasn't internally connected to the 3V rail,
or perhaps it was but the resistance was too great to keep the unit
running.

I've looked inside a few power supplies, including some cheap ones
with several parts missing, and the sense wire was always connected to
the output through a resistor. Without such a resistor, a broken
sense wire connection could make the regulator think that the voltage
had gone to zero and overcompenste until it reached the upper limit.
 
T

TyQuAnDo

EBUYER; need I say more?
The problem is I don't think they know which cases (inc PSU) to put in their
'Mid price' range,
some of the cases in the low end range are better.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top