micky said:
Amazing, at least to me. Thanks.
As to my not connecting anything but the mother board, not the other
outputs too, surely the floppies and the CD drives use almost nothing
when they are not in use. The hard drive would use somet. Do
computers with cheap power supplies have a hard time displaying the
entire POST or the Setup/BIOS screens if there is no harddrive?
Or are the fans enough?
You might want something around 12V @ 1A load, and fans might draw
around 0.5A. Your motherboard doesn't power the processor from +12V
(most likely), so your +12V rail might only have fans on it.
Hard drives are a reasonably light load. The motor uses +12V, and
the current could be between 0.3A to 0.6A or so (idling). A couple hard
drives, with data cable disconnected, might serve as a load. Hard
drives can draw up to 2.5 amps during spinup, for the first ten
seconds, so the loading isn't constant as such. The power supply
has to supply more current, right after you turn on the circuit.
After ten seconds, the spindles are up to speed on the drives,
and the current flow level drops back to 0.3A to 0.6A on
each drive.
When testing power supplies of unknown quality, you don't
use your "good" hard drives
That's the problem with
using hard drives, is what happens if the power supply 12V
rail is running at 15V.
If your ATX supply had a 12V @ 15A rating, and you connected
two hard drives, the hard drives will draw up to 5A for the
first ten seconds, settling down to 1.2A or less, once up to
speed. If you use too many hard drives, you could violate
the rating, due to the spinup current draw. In this example,
six hard drives would be "max" (as then, you'd be drawing
15A for the first ten seconds).
Hard drives are convenient as a load, since the connector
is ready to go, and easily mates with the ATX supply. As long
as you're aware of the spinup current, and don't use
too many, it makes a fine source of loading.
*******
I use power resistors from an electronics store, for test loads,
and those can be used for test purposes. This is an example
of what you can find for resistors. These particular ones have
a metal body, and can handle 25W or 50W. The 50W ones are
a bit bigger. And power resistors get good and hot. On my
load box, I use an 80mm fan for cooling, across the resistors.
There is material between the metal body and the wire wound resistor
inside, so the metal should not be electrified.
http://www.galco.com/techdoc/nte/25wm220_cp.pdf
The 25WM012 is a 12 ohm resistor. If connected across the 12V
rail, the current draw is 12V / 12 ohms = 1 amp. The power
dissipated, is 12V * 1A = 12 watts. And 12 watts, in such a
small package, should have some cool air across it. If you
were only dumping a couple watts, it might be sufficient to
convection cool it.
You adjust the ohms of the one you purchase, to get the
loading you desire. If you used a 3 ohm resistor,
12V / 3ohm = 4 amps. And 12V * 4A = 48W, and that 50W resistor
is cooking. The 25W resistor is probably in serious trouble.
You can use combinations of resistors, in series or in
parallel, to help spread out the heating. I don't generally
set up my load box resistors, to get near their name plate
rating in watts. I try to keep the watts within reason,
using a fraction of the watt rating so the surface
temperature won't be too high.
Some ceramic resistors, can operate at 300C, in which
case, you don't need the fan, but don't touch it
The 25WM012 above can take 275C, but I probably wouldn't
push that one that hard. I have some wirewound ceramic
power resistors, and you can tell by looking at them,
they'll take a lot of heat.
Generally speaking, light bulbs are not recommended as test
loads, at least when power is tightly constrained. To
give an example, I bought a DC adapter (wall wart) at the store,
something like 12V @ 2A, connected an automotive light bulb,
perhaps 12V @ 1A, and it causes the DC adapter to shut off.
The bulb would not light. So I can't test that adapter with the bulb
as a load. The reason for this, is when the automotive light
bulb is cold, the initial current draw can be double or more,
the "running" current flow level. When the filament is hot,
that is when it draws 1A. My DC adapter, being a switcher,
detects the overload instantly, and won't allow the cold
filament overload to last for more than a fraction of a second.
More crude power technologies, don't react immediately like that,
which is why they might allow the bulb to warm enough not to continue
to be an overload. Switching supplies can be set up, to shut off
very quickly, before the bulb even gets warm.
So in some ways, the "bad habits" of the disk drive
(the 2.5A spinup current), are mirrored in the operation
of the automotive light bulb (fraction of a second, cold
filament current flow). As long as you're aware of
the electrical characteristics of the initial inrush,
you can figure out a safe number of loads to use.
If the ATX supply had a 12V @ 15A rating, and you connect a
12V @ 1A automotive light bulb, no problem. The initial
cold filament surge current will be handled with ease.
But if the total loading of light bulbs, say eight light bulbs
in parallel, is slightly over half the supply ampere rating,
there is a chance the supply would "turn off on overload",
due to the initial cold filament surge. This is one reason,
I use real power resistors, because while they do have a
temperature coefficient, it isn't nearly as non-linear as a
light bulb.
There are metals with a pretty low temperature coefficient of
electrical resistance, such as manganin and related alloys.
The resistor used inside a multimeter, when making current
flow measurements, will have a low temperature coefficient,
and the resistance value won't change much at all when it
gets warm. By comparison, my load box would be inferior to
one of those. But using the automotive light bulb, as a
current shunt inside the multimeter, while cheap, would
throw any measurements out the window (50% error plus).
You use this stuff, for utmost accuracy, when the resistance
must remain stable. For best results, this stuff has to
be handled with care, annealed and the like.
http://en.wikipedia.org/wiki/Manganin
http://www.o-digital.com/uploads/2179/2192-1/Manganin_Shunt_42.jpg
You wouldn't waste that stuff for making load resistors, as
it would be too expensive.
So those are examples of some loads and resistor types.
Since you have hard drives on hand, those are the
closest thing within reach, to draw a bit of current.
Paul
